IHSAA Boys Basketball Championships to air live on FOX Sports

Volume 357 — August 9, 2007 — Number 6,pp.525-628
Article Summaries
Immigrants and Health Care — At the Intersection of Two Broken Systems
S. Okie
Becoming a Physician: Terra Firma — A Journey from Migrant Farm
Labor to Neurosurgery
A. QuiГ±ones-Hinojosa
Pay for Performance, Version 2.0?
T. H. Lee
Prophylaxis versus Episodic Treatment to Prevent Joint Disease in Boys
with Severe Hemophilia
M. J. Manco-Johnson and Others
A Single Cycle of Rituximab for the Treatment of Severe Pemphigus
P. Joly and Others
Complement C3 Variant and the Risk of Age-Related Macular
J. R.W. Yates and Others
NXY-059 for the Treatment of Acute Ischemic Stroke
A. Shuaib and Others
Acute Ischemic Stroke
H. B. van der Worp and J. van Gijn
Current Concepts: Drug-Induced Immune Thrombocytopenia
R. H. Aster and D. W. Bougie
Intravesical Foreign Body
N. Mondaini and R. Bartoletti
Asymptomatic Bladder Stones
Y. Daneshbod
Case 24-2007 — A 20-Year-Old Pregnant Woman with Altered Mental
A. J. Cole, J. W. Henson, M. H.A. Roehrl, and M. P. Frosch
Safer Drugs for the American People
G. D. Curfman, S. Morrissey, and J. M. Drazen
Prophylactic Treatment for Prevention of Joint Disease in Hemophilia —
Cost versus Benefit
G. Roosendaal and F. Lafeber
Rituximab and Pemphigus — A Therapeutic Advance
L. A. Diaz
The Tension between Needing to Improve Care and Knowing How to Do It
A. D. Auerbach, C. S. Landefeld, and K. G. Shojania
Adjunctive Antidepressant Treatment for Bipolar Depression
Glucose Regulation in Young Adults with Very Low Birth Weight
Use of Physicians' Services for Medicare Beneficiaries
Video on Orotracheal Intubation
Case 14-2007: A Man with Pain and Swelling of Both Eyes and the Right
Pregnancy in a Patient with Congenital Erythropoietic Porphyria
Preventing Medication Errors
Medication Errors
Medical Management of Vulnerable and Underserved Patients: Principles,
Practice, and Populations
New and Evolving Infections of the 21st Century
Vaccine: The Controversial Story of Medicine's Greatest Lifesaver
n e w e ng l a n d j o u r na l
m e dic i n e
images in clinical medicine
Asymptomatic Bladder Stones
healthy 26-year-old driver had
routine laboratory tests performed to renew his driver’s license. Hematuria (3+)
was incidentally noted. He reported having had
no urinary tract symptoms. Abdominal radiography of the kidneys, ureters, and bladder showed
large bladder stones (Panel A, with a Foley catheter tube visible). During cystotomy, two large
stones, measuring 8 cm and 3 cm in diameter,
were removed (Panel B). These stones were found
to contain both calcium phosphate and uric acid
and were infected with Proteus mirabilis. No structural bladder-outlet obstruction or other anatomical abnormalities were identified. The patient
recovered quickly from the surgery and was discharged 2 days after the procedure. The infection
was treated with 500 mg of ciprofloxacin twice
daily. Whereas a small stone (only millimeters in
size) often causes significant pain as it travels
through a narrow ureter, bladder calculi may become large and yet not cause symptoms.
Copyright В© 2007 Massachusetts Medical Society.
Yahya Daneshbod, M.D.
Dr. Daneshbod Pathology Laboratory
Shiraz 71347, Iran
[email protected]
n engl j med 357;6  www.nejm.org  august 9, 2007
AUTHOR Daneshbod
august 9, 2007
Immigrants and Health Care — At the Intersection
of Two Broken Systems
Susan Okie, M.D.
t a primary care clinic in Montgomery County,
Maryland, where I volunteer, the patients are
uninsured immigrants from Latin America or West
Africa. Many are day laborers, house cleaners, or
construction workers; most do
not speak English. Several months
ago, I saw a middle-aged Hispanic
baker with profound weakness,
fatigue, limb swelling, and severe muscle pain, who had to be
hospitalized for myxedema. Fortunately, a local charity agreed
to pay most of her hospital costs,
and she’s now receiving thyroid
hormone–replacement therapy
— but with regular care, her hypothyroidism could have been
diagnosed earlier and hospitalization averted. Another day, I
tried to persuade a reticent West
African man who had been tortured in prison that psychological counseling might help his
chronic pain. However, mental
health services for uninsured
immigrants are sparse, and the
man was reluctant to venture to
a distant part of Washington, D.C.,
to a program for torture survivors. A third patient, a man in
his 40s, came in with a nearly
empty bottle of eyedrops, which
he had brought from Ghana to
take for glaucoma. The disease
had already blinded him in one
eye, and the vision in his other
eye had been fluctuating. He
needed a complete eye exam and
visual-field testing, but arranging timely referrals to specialists is often difficult for caregivers treating the uninsured. I
wrote him a prescription, and
we managed to set up an appointment at a hospital-based
ophthalmology clinic that ac-
cepts a limited number of uninsured patients.
For recent immigrants — especially the estimated 12 million
who are here illegally — seeking
health care often involves daunting encounters with a fragmented, bewildering, and hostile system. The reason most immigrants
come here is to work and earn
money; on average, they are younger and healthier than native-born
Americans, and they tend to
avoid going to the doctor. Many
work for employers who don’t
offer health insurance, and they
can’t afford insurance premiums
or medical care. They face language and cultural barriers, and
many illegal immigrants fear that
visiting a hospital or clinic may
draw the attention of immigration
officials. Although anti-immigrant
sentiment is fueled by the belief
that immigrants can obtain federal benefits, 1996 welfare-reform
legislation greatly restricted im-
n engl j med 357;6  www.nejm.org  august 9, 2007
Immigrants and Health Care — At the Intersection of Two Broken Systems
migrants’ access to programs
such as Medicaid, shifting most
health care responsibility to
state and local governments.
The law requires that immigrants wait 5 years after obtaining lawful permanent residency
(a “green card”) to apply for federal benefits. In response, some
states and localities — for instance, Illinois, New York, the
District of Columbia, and certain California counties — have
used their own funds to expand
health insurance coverage even
for undocumented immigrant
children and pregnant women
with low incomes. Other states,
however, such as Arizona, Colorado, Georgia, and Virginia,
have passed laws making it even
more difficult for noncitizens to
gain access to health services.
Whether or not they have
health insurance, immigrants
overall have much lower per capita health care expenditures than
native-born Americans,1 and recent analyses indicate that they
contribute more to the economy
in taxes than they receive in public benefits. In a study from the
RAND Corporation, researchers
estimated that undocumented
adult immigrants, who make up
about 3.2% of the population, account for only about 1.5% of U.S.
medical costs.2 Many immigrants
do not seek medical treatment
unless they are injured or acutely
ill; at our clinic, patients with
type 2 diabetes often have florid
symptoms and even incipient renal damage by the time their disease is diagnosed.
One study found that annual
per capita expenses for health
care were 86% lower for uninsured immigrant children than
for uninsured U.S.-born children
— but emergency department
expenditures were more than
three times as high.1 Although
U.S. hospitals must provide
emergency care without first
asking about income, insurance,
or citizenship, early diagnosis
and treatment in a primary care
setting are both medically preferable and a better use of resources. “If people keep postponing medical care because
they’re so concerned about being sent back over the border,”
noted Elizabeth Benson Forer,
executive director of the Venice
Family Clinic, a venerable free
clinic in Los Angeles that serves
many immigrants, “then you
can end up with some pretty
horrendous health situations.”
Immigrants live, work, and
attend school in communities
throughout the country; laws
and bureaucratic barriers that
reduce their use of key preventive health services, such as immunizations and screenings for
infectious disease, make for bad
public health policy, and denying immigrants primary care ultimately increases health care
costs for everyone. For example,
labor and delivery costs for undocumented immigrant women
are covered under the federal
and state emergency Medicaid
program, but most states do not
cover prenatal care, and there is
no coverage for family planning.
Some of my patients say they
would like to use oral contraceptives or an intrauterine device or undergo a tubal ligation,
but they can’t afford it. And immigrants, like native-born Americans, are vulnerable to chronic
diseases; as my colleague, nurse
practitioner Lois Wessel, notes,
“Even the 25-year-old day laborers are eventually going to become 45-year-olds, probably still
undocumented, with hypertension and diabetes.  .  .  .  Life in
America is going to make them
become not so healthy.”
Recently, a bipartisan group
of U.S. senators, with White
House support, introduced an
immigration bill that offered the
best chance in years of achieving
substantial reform of a dysfunctional system. However, the bill
met with opposition from both
conservatives and liberals and
was killed in the Senate this past
June, quashing all hope of immigration reform during the current administration. State legislatures this year are considering a
record number of anti-immigrant
measures, and the Senate bill’s
demise heightens their chances
of passage. “You will see the
states and cities scrambling to
pass their own laws and regulations, and you’re going to get a
completely contradictory set of
policies,” Senator John McCain
(R-AZ) predicted in a Washington
Post article on July 8. In many
areas of the country, one consequence is likely to be reduced access to health care for immigrants.
Noncitizens make up about
20% of the 46 million uninsured
people in the United States.
Hospitals generally do not collect information on patient immigration status, and there are
no reliable national figures on
hospital costs for undocumented
immigrants. Nevertheless, the
soaring cost of uncompensated
care (see graph A) has made the
problem of providing care for
uninsured immigrants a hot political issue, particularly in border states and those (such as the
southeastern states) whose immigrant populations have grown
rapidly in recent years. Some
uninsured immigrants needing
emergency treatment (including
pregnant women, children, adults
n engl j med 357;6  www.nejm.org  august 9, 2007
Immigrants and Health Care — At the Intersection of Two Broken Systems
Cost of Uncompensated Hospital Care
(billions of $)
with dependent children, and elderly, blind, or disabled patients
with incomes below Medicaid
thresholds) qualify for emergency Medicaid coverage. In many
other cases, hospitals receive no
payment for their care, although
in 2003 Congress appropriated
$250 million per year for 4 years
(starting in 2005) to partially
compensate hospitals for treating undocumented immigrants.
A recent study found that although emergency Medicaid
spending for immigrants in
North Carolina grew by 28% between 2001 and 2004, it still
represented less than 1% of the
state’s Medicaid budget.3 More
than 80% of that spending was
for childbirth and complications
of pregnancy, and major injuries
accounted for nearly one third
of the rest. In California, emergency Medicaid spending for
uninsured immigrants for fiscal
year 2007 exceeded $941 million, according to Kim BelshГ©,
secretary of the California Health
and Human Services Agency.
“Clearly, there are medical needs
faced by this population,” said
Belshé, “and the emergency room
is not the most cost-effective
Estimated No. of Unauthorized Immigrants
Living in the U.S. (millions)
Cost of Uncompensated Care (Panel A), Number
of Unauthorized Immigrants (Panel B), and Number
of Uninsured People (Panel C) in the United States.
The annual cost to U.S. hospitals of uncompensated
care (charity care plus bad debt) has been rising,
although the fraction of total hospital expenses represented by such care has remained relatively constant
at about 5 to 6% since 1980. The number of unauthorized immigrants present in the United States has also
been increasing, although the estimates are uncertain. Treatment of unauthorized immigrants contributes to uncompensated care costs, but the main reason such costs are increasing is the rise in the
number of people who lack health insurance. Immigrants represent only about 20% of the uninsured.
Data on uncompensated care are from the American
Hospital Association; data on unauthorized immigrants are from the Pew Hispanic Center; data on the
uninsured are from the U.S. Census Bureau.
No. of People Not Covered by Health
Insurance (millions)
place for [addressing] them.” In
addition, undocumented immiAUTHOR: Okie
grants may account
for as much
as $750 million annually of the
cost of uncompensated care in
California hospitals — about
10% of theRETAKE
total — since
they represent about
3rd 10% of the
n engl j med 357;6  www.nejm.org  august 9, 2007
Figure has been redrawn and type has been reset.
Please check carefully.
JOB: 35706
ISSUE: 08-09-07
Immigrants and Health Care — At the Intersection of Two Broken Systems
state’s emergency department
patients, according to Jan Emerson, vice president of external
affairs for the California Hospital Association. “Almost half of
the hospitals in California are
currently operating in the red,”
she said. “It would not be fair to
place the blame solely on undocumented immigrants, but
certainly, they are a contributing
The chief sources of outpatient care for uninsured immigrants are public clinics and
community health centers. Such
clinics are often sparse in suburban and rural areas that have recently faced an influx of immigrants. Even in cities with strong
community-clinic networks and
a long history of serving immigrants, access to care is uneven.
For example, at the Venice Family
Clinic, a bilingual nurse educator
runs health and exercise classes
in Spanish and English for patients with diabetes, pregnant
women receive free state-subsidized prenatal care, and there
are regularly scheduled clinic
sessions for victims of torture
and human trafficking. Yet arranging specialty referrals is a
constant challenge — it usually
entails sending patients to outpatient clinics at county hospitals, where some have to wait as
long as a year for an appointment. In the Washington, D.C.,
area uninsured women, including undocumented immigrants,
can get free annual mammograms and Pap smears through
subsidized cancer-screening programs, but follow-up treatment
for abnormal findings other than
cancer is usually not included,
and many clients have no source
of primary care, as noted by
nurse practitioner Wessel, who
works monthly at one such pro528
gram. “Patients come in year after year” for Pap smears — “but
they’ve never had their blood
pressure checked,” she said. “We
don’t check it, because all we’re
financed to do is cervical- and
breast-cancer screening.”
In states seeking to expand
insurance coverage, the question
of including undocumented immigrants is a thorny one. About
1 million of California’s 4.8 million uninsured residents are undocumented adults, and about
136,000 are undocumented children.4 As part of a proposal for
comprehensive health care reform, Governor Arnold Schwarzenegger is seeking to provide
(through Medicaid and the State
Children’s Health Insurance Program, or SCHIP) to all children
with family incomes at or below
300% of the federal poverty level, regardless of immigration
status. Although there is considerable public support for insuring undocumented immigrant
children, Republican state legislators “do not believe that state
general fund revenues should be
invested in people who are here
illegally,” said health secretary
Belshé — “and that extends to
The federal Medicaid program has always been restricted
to U.S. citizens and legal residents, but recent federal and
state laws designed to strengthen enforcement of eligibility
rules have created new barriers,
even for infants and children
who are citizens, and have had a
chilling effect on other programs providing health services
for immigrants. The 2005 Deficit Reduction Act requires all
persons applying for or renewing Medicaid coverage to provide
proof of identity and U.S. citi-
zenship. Since that law went
into effect, at least eight states
have reported dramatic declines
in Medicaid enrollment, and
some Medicaid-eligible infants
and children have gone without
immunizations and needed
medical care because of delays
in coverage.5
In Georgia, which last year
passed a law requiring immigrants to show proof of legal
residency in many situations,
“we’ve started seeing a lot of
kids not going to the doctor,”
said Flavia Mercado, a pediatrician who runs the International
Medical Center at Atlanta’s
Grady Memorial Hospital. “A lot
of my clients are leaving and going to other states, and a couple
are even going back to their
country. Everyone is very fearful.” She said that Atlanta organizations are scaling back health
services for Hispanics and have
stopped sponsoring Hispanic
health fairs, fearing that they
will be raided by police or immigration officials. Meanwhile,
faced with rising health care
costs and increasing numbers of
uninsured persons, the state’s
Medicaid program has sharply
reduced benefits: it recently
stopped paying for prenatal care
for high-risk women and for
nonemergency hemodialysis. Although immigrants make up a
minority of the uninsured, Mercado said media reports regularly blame illegal immigrants
for the worsening problems of
the state’s health care system.
Anger over high medical costs
and reduced access to care no
doubt contributes to anti-immigrant sentiment; the remedy,
however, is not immigrant bashing, but health care reform.
“As an American citizen, I
understand that you want to
n engl j med 357;6  www.nejm.org  august 9, 2007
Immigrants and Health Care — At the Intersection of Two Broken Systems
make sure the resources are
there for the right people,” Mercado said. “Yet how can you
deny someone health access? If
we don’t treat and prevent illness  .  .  .  our whole community is going to suffer.”
Dr. Okie is a contributing editor of the
1. Mohanty SA, Woolhandler S, Himmelstein DU, Pati S, Carrasquillo O, Bor DH.
Health care expenditures of immigrants in
the United States: a nationally representative
analysis. Am J Public Health 2005;95:1431-8.
2. Goldman DP, Smith JP, Sood N. Immigrants and the cost of medical care. Health
Aff (Millwood) 2006;25:1700-11.
3. DuBard CA, Massing MW. Trends in emergency Medicaid expenditures for recent and
undocumented immigrants. JAMA 2007;297:
1085-92. [Erratum, JAMA 2007;297:1774.]
4. Brown ER, Pourat N, Wallace SP. Undocumented residents make up small share of
California’s uninsured population. Los Angeles: UCLA Center for Health Policy Research,
March 2007.
5. Pear R. Lacking papers, citizens are cut
from Medicaid. New York Times. March 12,
Copyright В© 2007 Massachusetts Medical Society.
Terra Firma — A Journey from Migrant Farm Labor
to Neurosurgery
Alfredo QuiГ±ones-Hinojosa, M.D.
You will spend the rest of
your life working in the
fields,” my cousin told me when
I arrived in the United States in
the mid-1980s. This fate indeed
appeared likely: a 19year-old illegal migrant farm worker, I
had no English language skills and no
dependable means of
support. I had grown
up in a small Mexican farming community, where I began
working at my father’s gas station at
the age of 5. Our
family was poor, and
we were subject to the diseases
of poverty: my earliest memory
is of my infant sister’s death
from diarrhea when I was 3 years
old. But my parents worked long
hours and had always made
enough money to feed us, until
an economic crisis hit our country in the 1970s. Then they could
no longer support the family,
and although I trained to be a
teacher, I could not put enough
food on the table either.
Desperate for a livable in-
come, I packed my few belongings and, with $65 in my pocket,
crossed the U.S. border illegally.
The first time I hopped the fence
into California, I was caught
eating anything I could get, with
hands bloodied from pulling
weeds — the very same hands
that today perform brain surgery.
My days as a farm worker
and sent back to Mexico, but I
tried again and succeeded. I am
not condoning illegal immigration; honestly, at the time, the
law was far from the front of my
mind. I was merely responding
to the dream of a better life, the
hope of escaping poverty so that
one day I could return home triumphant. Reality, however, posed
a stark contrast to the dream. I
spent long days in the fields picking fruits and vegetables, sleeping under leaky camper shells,
taught me a great deal about
economics, politics, and society.
I learned that being illegal and
poor in a foreign country could
be more painful than any poverty I had previously experienced.
I learned that our society sometimes treats us differently depending on the places we have
been and the education we have
obtained. When my cousin told
me I would never escape that life
of poverty, I became determined
to prove him wrong. I took night
n engl j med 357;6  www.nejm.org  august 9, 2007
Terra Firma — A Journey from Migrant Farm Labor to Neurosurgery
jobs as a janitor and subsequently as a welder that allowed me to
attend a community college where
I could learn English.
In 1989, while I was working
for a railroad company as a welder and high-pressure valve specialist, I had an accident that
caused me to reevaluate my life
once again. I fell into a tank car
that was used to carry liquefied
petroleum gas. My father was
working at the same company.
Hearing a coworker’s cry for help,
he tried to get into the tank; fortunately, someone stopped him. It
was my brother-in-law, Ramon,
who climbed in and saved my
life. He was taken out of the
tank unconscious but regained
consciousness quickly. By the time
I was rescued, my heart rate had
slowed almost to zero, but I was
resuscitated in time. When I
awoke, I saw a person dressed all
in white and was flooded with a
sense of security, confidence, and
protection, knowing that a doc-
tor was taking care of me. Although it was clear to me that our
poverty and inability to speak English usually translated into suboptimal health care for my community, the moment I saw this
physician at my bedside, I felt I
had reached terra firma, that I had
a guardian.
After community college, I was
accepted at the University of California, Berkeley, where a combination of excellent mentorship,
scholarships, and my own passion
for math and science led me to
research in the neurosciences. One
of my mentors there convinced
me, despite my skepticism, that
I could go anywhere I wanted for
medical school. Thanks to such
support and encouragement, I
eventually went to Harvard Medical School. As I pursued my own
education, I became increasingly
aware of the need and responsibility we have to educate our country’s poor.
It is no secret that minority
communities have the highest
dropout rates and the lowest
educational achievement levels
in the country. The pathway to
higher education and professional training programs is not
“primed” for minority students.
In 1994, when I started medical
school, members of minority
groups made up about 18% of
the U.S. population but accounted for only 3.7% of the faculty in
U.S. medical schools. I was very
fortunate to find outstanding
minority role models, but though
their quality was high, their numbers were low.
Given my background, perhaps it is not surprising that I
did not discover the field of neurosurgery until I was a medical
student. I vividly remember when,
in my third year of medical
school, I first witnessed neurosurgeons peeling back the dura
and exposing a real, live, throbbing human brain. I recall feeling absolute awe and humility —
and an immediate and deep
recognition of the intimacy between a patient and a doctor.
That year, one of my professors strongly encouraged me to
go into primary care, arguing
that it was the best way for me
to serve my Hispanic immigrant
community. Although I had initially intended to return to Mexico triumphant, I had since fallen in love with this country, and
I soon found myself immersed
in and committed to the betterment of U.S. society. With my
sights set on neurosurgery after
medical school, I followed my
heart and instincts and have tried
to contribute to my community
and the larger society in my own
way. I see a career in academic
medicine as an opportunity not
only to improve our understanding and treatment of human diseases but also to provide leadership within medicine and support
to future scientists, medical students, and physician scientists
from minority and nonminority
groups alike.
n engl j med 357;6  www.nejm.org  august 9, 2007
Terra Firma — A Journey from Migrant Farm Labor to Neurosurgery
My grandmother was the medicine woman in the small town in
rural Mexico where I grew up.
As I have gotten older, I have
come to recognize the crucial
role she played not only in instilling in me the value of healing but also in determining the
fate and future of others. She was
my first role model, and throughout my life I have depended on
the help of my mentors in pursuing my dreams. Like many other
illegal immigrants, I arrived in
the United States able only to
contemplate those dreams — I
was not at that point on solid
ground. From the fields of the
San Joaquin Valley in California
to the field of neurosurgery, it
has been quite a journey. Today,
as a neurosurgeon and researcher, I am taking part in the larger
journey of medicine, both caring for patients and conducting
clinical and translational research
on brain cancer that I hope will
lead to innovative ways of fighting devastating disease. And as
a citizen of the United States, I am
also participating in the great
journey of this country. For immigrants like me, this voyage still
means the pursuit of a better
life — and the opportunity to
give back to society.
An interview with Dr. QuiГ±ones-Hinojosa can be heard at www.nejm.org.
Dr. QuiГ±ones-Hinojosa is an assistant professor of neurosurgery and oncology and
director of the brain-tumor stem-cell laboratory at Johns Hopkins School of Medicine, Baltimore, and director of the braintumor program at the Johns Hopkins
Bayview campus.
Copyright В© 2007 Massachusetts Medical Society.
Pay for Performance, Version 2.0?
Thomas H. Lee, M.D.
Old wine in a new bottle.” “A
financial gamble.” “An early
glimpse of the next generation of
pay for performance.” All these
appraisals have been applied to
Geisinger Health System’s new approach to elective coronary-artery
bypass grafting (CABG), which
has been described with words
rarely invoked in health care, such
as “promise” and “guarantee.”
Geisinger, an integrated health
care delivery system in northeastern Pennsylvania, promises
that 40 key processes will be
completed for every patient who
undergoes elective CABG — even
though several of the “benchmarks” are to be reached before
or after hospitalization. And although Geisinger cannot guarantee good clinical outcomes, it
charges a standard flat rate that
covers care for related complications during the 90 days after
As a member of Geisinger’s
board of directors, I have watched
this program evolve over the past
year, and I see truth in all three
of the above assessments. Many
of the core components of the
program are familiar, but this
sort of application of those components represents a foray into
the unknown. Since a front-page
article in the New York Times on
May 17, 2007, drew national attention to the Geisinger program, other hospitals have been
watching closely and wondering
whether they, too, should go
down this road. Those who examine it closely will quickly discover that the program is less
about cardiac surgery than about
the search for an alternative to
traditional fee-for-service care.
The basic concept is far from
radical. The seven cardiac surgeons in the Geisinger delivery
system agreed on 40 processes
that should be completed during
the care of every patient undergoing elective CABG. Most of
the “Proven Care Benchmarks”
come directly from guidelines
established by the American College of Cardiology and the American Heart Association (ACC–AHA)
(see box). These steps (such as
the administration of preoperative
antibiotics at a specified time)
are prominent in the critical pathways in use for cardiac surgery
at many other hospitals.
The list does not force the surgeons to practice “cookbook medicine.” For example, they do not
necessarily have to use epiaortic
echocardiography to screen for
atheromata before manipulating
the aorta. But the protocol requires that they consider this test
and document the reason if they
decide not to use it.
Closer inspection reveals some
other items on the list that would
be new to most critical pathways
for CABG. The first benchmark
that must be documented is a
statement of the indication for
CABG according to the ACC–AHA
guidelines.1 These guidelines de-
n engl j med 357;6  www.nejm.org  august 9, 2007
Pay for Performance, Version 2.0?
Selected Key Processes in the Geisinger “Proven Care” Elective CABG Program.*
Preadmission documentation
American College of Cardiology–American Heart Association indication
for surgery
Explanation of treatment options to patient
Indication of whether patient is a current user of clopidogrel or warfarin
Screening for stroke risk
Screening for use of epiaortic echocardiography
Operative documentation
Patient receives correct dose of beta-blocker
Patient receives preoperative antibiotics (within 60 minutes of incision;
with vancomycin within 120 minutes)
Left internal thoracic artery is used for grafting of the left anterior descending
Postoperative documentation
Antibiotics are administered (postoperatively, for 24 to 48 hours)
Beta-blocker is administered (within 24 hours after surgery)
Tobacco screening and counseling are provided
Discharge documentation
Referral to cardiac rehabilitation is provided
Discharge medications (aspirin, beta-blockers, statin) are prescribed
Postdischarge documentation
Patient is taking medications correctly
Patient has or has not resumed smoking
Patient is enrolled in cardiac rehabilitation
*CABG denotes coronary-artery bypass grafting.
scribe 22 class I indications, for
which there is strong evidence in
support of the appropriateness of
CABG. At the other end of the
spectrum are 9 class III indications, for which the evidence actually argues against performing
surgery. In between are 11 class
IIa indications, for which the evidence generally favors the use of
CABG, and 3 class IIb indications,
for which there is less support.
The surgeons can proceed directly to surgery if the patient has
a class I or IIa indication. However,
if the strongest indication for surgery is class IIb, the case must be
reviewed by a colleague. For patients with class III indications
only, surgery is not an option.
Another preadmission benchmark is documentation that the
surgeon has reviewed the treatment options and their risks and
benefits with the patient and
that the patient prefers the surgical approach. Completion of this
step, along with the appropriateness assessment, provides assurance that the patient both desires
and needs the operation.
A third innovation is the requirement of postdischarge follow-up to ensure that patients
are taking their medications correctly, participating in a rehabilitation program, and (one hopes)
refraining from smoking. In between these preadmission and
postdischarge steps, the Geising-
er program looks very much like
other CABG critical pathways.
Of course, one key difference
is that Geisinger guarantees that
all 40 benchmarks will be
achieved for every elective CABG.
Delivering on that guarantee
turned out to be easier said than
done. After the Geisinger surgeons agreed on the benchmarks
in the spring of 2006, they looked
at a series of cases and found that
slightly more than half of patients
were going through every one of
the processes. The surgical teams
began creating systems to ensure
that all 40 processes would be
completed, without disrupting the
flow of care. Within a few months,
100% of patients were hitting
100% of the benchmarks. Today,
if any of the preoperative benchmarks are overlooked, surgery is
delayed until the unfinished task
has been completed.
Thus far, the Geisinger program sounds like a no-nonsense
critical pathway with some extra
bells and whistles — but now we
get to the risky part. For patients
who have surgery as part of this
program, Geisinger will not
charge for related care within
90 days. For example, there are
no additional charges for treatment of sternal wound infections or heart failure due to a
perioperative infarction, as long
as patients receive their care at a
Geisinger facility. On the other
hand, the usual charges would
apply to care for preexisting heart
failure or unrelated problems,
such as diverticulitis or a hip
To calculate the case rate for
CABG, Geisinger examined its
historical costs for related care
during the first 90 days after
surgery — and then set itself a
target of reducing those costs by
n engl j med 357;6  www.nejm.org  august 9, 2007
Pay for Performance, Version 2.0?
half. Geisinger has some encouraging early data suggesting that
its complication rates may have
decreased since the program was
introduced. But the numbers are
small, and no one knows for sure
whether this approach will really
reduce complication rates or lower postdischarge costs by anything approaching 50%.
This uncertainty explains why
most chief financial officers get
nervous when they hear about
Geisinger’s program. They point
out that Geisinger is a unique
organization. It has its own insurance company, and its physicians are salaried employees of
the Geisinger Health System. It
has a spectacular management
information system that records
and guides most aspects of both
inpatient and outpatient care. And
it has a tradition of innovation
and collaboration that makes it
easier to get physicians to participate in a program like this one.
On the basis of my interactions with Geisinger’s physicians,
I would say they are pretty much
like the doctors I know everywhere — willing to agree to
things that will improve patient
care but in no particular hurry
to compromise their profession-
al autonomy. I also know that
Geisinger’s financial leaders (and
its board) do not like uncertainty any more than those of other
But the fact is that elective
CABG is not a common procedure in the era of statins, betablockers, aspirin therapy, and
angioplasty. If the revenue from
CABG performed at the case rate
is too low to cover costs, Geisinger will be able to adjust its
prices upward without causing
too much damage. So the gamble,
though real, is not huge. And the
experience gained by learning
how clinicians need to collaborate to succeed within this framework seems worth the risk.
The real question for Geisinger and for the rest of the health
care system is whether this caserate approach might emerge as a
new form of pay for performance.
Many current models of pay for
performance (involving, for example, quality-of-care measures
for patients with diabetes) focus
on populations of patients whose
care is managed by primary care
physicians. For most specialists
and hospitals, existing incentive
systems put only a modest amount
of revenue at stake, and as would
be expected, resulting changes in
care have been modest as well.
But the drumbeat is growing
stronger for health care financing
models that go beyond rewarding volume alone. Case rates and
critical pathways are not foreign
concepts at many hospitals —
they just have not been married
so explicitly before. Geisinger is
actively working to extend this
approach to other surgical procedures, and diseases treated on
an outpatient basis, such as diabetes and hypertension, could be
next. A reasonable guess is that
models that work for organized
delivery systems such as Geisinger will spread over time to the
rest of U.S. health care. So this
experiment bears watching.
Dr. Lee is network president at Partners
Healthcare System, Boston, and an associate editor of the Journal. He is a member of
the Geisinger board of directors.
1. Eagle KA, Guyton RA, Davidoff R, et al.
ACC/AHA 2004 guideline update for coronary artery bypass graft surgery: a report of
the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Update the
1999 Guidelines for Coronary Artery Bypass
Graft Surgery). Washington, DC: American
College of Cardiology, 2004. (Accessed
July 19, 2007, at http://www.acc.org/
Copyright В© 2007 Massachusetts Medical Society.
n engl j med 357;6  www.nejm.org  august 9, 2007
Clinical Practice
Acute Ischemic Stroke
A 62-year-old man has sudden weakness of the left arm
and leg and slurred speech. Except for untreated hypertension, his medical history is unremarkable. He is
original article
a current smoker with a smoking history of 45 packProphylaxis versus Episodic Treatment to Prevent
years. On arrival at the emergency department 1 hour
Joint Disease in Boys with Severe Hemophilia
15 minutes after the onset of symptoms, he reports no
This randomized trial involving young boys with severe headache or vomiting. His blood pressure is 180/100
hemophilia showed that prophylaxis with regular infu- mm Hg; his pulse is 76 beats per minute and is regular.
sions of recombinant factor VIII was associated with How should this patient be evaluated and treated in
clinically and statistically significant reductions in joint the short term?
damage, as compared with episodic infusions at the
See P. 572; CME, P. 629
time of a clinically evident hemarthrosis. Because of
the high cost of recombinant factor VIII, its widespread Current Concepts
use for prophylaxis may be impractical.
Drug-Induced Immune Thrombocytopenia
See P. 535; Editorial, P. 603; CME, P. 630
Drug-induced thrombocytopenia should be suspected
in any patient with acute thrombocytopenia of unknown
Original Article
cause. Although the incidence is low, more than 100
Rituximab for the Treatment of Severe Pemphigus
drugs have been implicated in thrombocytopenia, inIn this study of 21 patients with severe pemphigus cluding quinine, sulfonamides, abciximab, carbamazewhose disease was refractory to or dependent on sys- pine, and vancomycin, as well as herbal remedies and
temic corticosteroids or who had contraindications to several nonprescription drugs. This review summarizes
corticosteroids, 18 patients (86%) had a complete re- the current understanding of pathogenesis and provides
mission after a single cycle of rituximab treatment. Two a guide for diagnosis and management of this potenpatients had serious infections, one of which resulted in tially dangerous disorder.
death. The efficacy of rituximab for pemphigus must be
See p. 580; CME, P. 631
weighed against the risk of severe adverse events.
See P. 545; Editorial, P. 605
A Pregnant Woman with Altered Mental Status
Original Article
A 20-year-old pregnant woman was admitted at 26 weeks’
gestation because of dizziness, confusion, and difficulty
walking. Six weeks earlier, she was confused and had
A variant on complement factor 3 is associated with odd head movements. Four days before admission,
age-related macular degeneration, with a population she had dizziness and weakness; she began to fall to
attributable risk of 22%. This finding underlines the her left and vomited. On admission, examination of the
importance of complement activation in the pathogen- cerebrospinal fluid showed a lymphocytic pleocytosis
esis of the disease.
with mildly elevated protein and normal glucose levels.
See P. 553
During the next 18 days, her condition worsened.
Complement C3 and Age-Related Macular
Original Article
See P. 589
NXY-059 for Acute Ischemic Stroke
Sounding Board
Two phase 3 clinical trials (SAINT I and SAINT II) evaluated the free-radical–trapping agent NXY-059 for the
treatment of acute ischemic stroke. The SAINT I trial,
reported last year, suggested that NXY-059 might be
effective. The authors now report the results of the
SAINT II trial, which clearly shows that NXY-059 is not
effective for ischemic stroke. The discrepancy in the
findings of the two trials is best explained by chance
false positive findings in the SAINT I trial.
See P. 562
Needing to Improve Care
versus Knowing How to Do It
The authors argue that interventions to improve health
care quality should be held to the same standards that
we apply to medical treatments. They believe qualityimprovement initiatives should not be widely disseminated unless studies have demonstrated that they are
safe and effective.
See P. 608
n engl j med 357;6  www.nejm.org  august 9, 2007
new england
journal of medicine
established in 1812
august 9, 2007
vol. 357  no. 6
Prophylaxis versus Episodic Treatment to Prevent Joint Disease
in Boys with Severe Hemophilia
Marilyn J. Manco-Johnson, M.D., Thomas C. Abshire, M.D., Amy D. Shapiro, M.D.,
Brenda Riske, M.S., M.B.A., M.P.A., Michele R. Hacker, Sc.D., Ray Kilcoyne, M.D., J. David Ingram, M.D.,
Michael L. Manco-Johnson, M.D., Sharon Funk, B.Sc., P.T., Linda Jacobson, B.S., Leonard A. Valentino, M.D.,
W. Keith Hoots, M.D., George R. Buchanan, M.D., Donna DiMichele, M.D., Michael Recht, M.D., Ph.D.,
Deborah Brown, M.D., Cindy Leissinger, M.D., Shirley Bleak, M.S.N., Alan Cohen, M.D., Prasad Mathew, M.D.,
Alison Matsunaga, M.D., Desiree Medeiros, M.D., Diane Nugent, M.D., Gregory A. Thomas, M.D.,
Alexis A. Thompson, M.D., Kevin McRedmond, M.D., J. Michael Soucie, Ph.D., Harlan Austin, Ph.D.,
and Bruce L. Evatt, M.D.
A bs t r ac t
Effective ways to prevent arthropathy in severe hemophilia are unknown.
We randomly assigned young boys with severe hemophilia A to regular infusions of
recombinant factor VIII (prophylaxis) or to an enhanced episodic infusion schedule
of at least three doses totaling a minimum of 80 IU of factor VIII per kilogram of
body weight at the time of a joint hemorrhage. The primary outcome was the incidence of bone or cartilage damage as detected in index joints (ankles, knees, and
elbows) by radiography or magnetic resonance imaging (MRI).
Authors’ affiliations are listed in the Appendix. Address reprint requests to Dr.
M.J. Manco-Johnson, Mountain States Regional Hemophilia and Thrombosis Center, MS F-416, PO Box 6507, Aurora, CO
80045, or at [email protected]
N Engl J Med 2007;357:535-44.
Copyright В© 2007 Massachusetts Medical Society.
Sixty-five boys younger than 30 months of age were randomly assigned to prophylaxis
(32 boys) or enhanced episodic therapy (33 boys). When the boys reached 6 years of
age, 93% of those in the prophylaxis group and 55% of those in the episodic-therapy group were considered to have normal index-joint structure on MRI (P = 0.006).
The relative risk of MRI-detected joint damage with episodic therapy as compared with
prophylaxis was 6.1 (95% confidence interval, 1.5 to 24.4). The mean annual numbers
of joint and total hemorrhages were higher at study exit in the episodic-therapy group
than in the prophylaxis group (P<0.001 for both comparisons). High titers of inhibitors of factor VIII developed in two boys who received prophylaxis; three boys
in the episodic-therapy group had a life-threatening hemorrhage. Hospitalizations
and infections associated with central-catheter placement did not differ significantВ­
ly between the two groups.
Prophylaxis with recombinant factor VIII can prevent joint damage and decrease the
frequency of joint and other hemorrhages in young boys with severe hemophilia A.
(ClinicalTrials.gov number, NCT00207597.)
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
efore the development of cryoprecipitate, a plasma fraction that contains concentrated factor VIII, boys with severe hemophilia A had a diminished life expectancy.1-3
These children are at risk for many types of hemorrhages, but the predominant source of chronic
coexisting disease is crippling, painful arthritis
due to hemarthrosis.4 Small trials were conducted
in the 1960s to determine whether routine administration of factor VIII concentrate was effective
as prophylaxis against hemophilic arthropathy.5‑8
Clinically effective prophylactic schedules were developed empirically, without the benefit of data
from controlled trials,9 and many clinicians began
to recommend prophylaxis with factor VIII.10
In the 1980s, when it was discovered that
plasma-derived factor VIII concentrates were contaminated by human immunodeficiency and hepatitis viruses, the use of prophylaxis was severely
curtailed.4 In 1992, approval of the first recombinant factor VIII molecule for replacement therapy
in the United States allowed for safe prophylaxis in
patients with hemophilia.11 Petrini and colleagues
reported the prevention of hemophilic arthropathy
when prophylaxis was initiated before patients
reached 2 years of age.12 Aledort and others reportВ­
ed that prophylaxis slowed the progression of established joint damage.13 Nevertheless, questions
remained as to when prophylaxis should begin,
what dose of recombinant factor VIII should be
administered, and how long prophylaxis should
be provided. An important question that could be
answered by a clinical trial was whether prophylaxis prevents joint hemorrhage and damage.14
The aim of our randomized trial was to determine whether prophylactic factor VIII infusions,
given every other day, are more effective in preventing joint damage than an intensive replacement regimen given at the time of a hemarthrosis. The study focused on the index joints — ankles,
knees, and elbows — because these joints are the
most susceptible to hemophilic arthropathy. This
trial was conducted in the context of a national
hemophilia comprehensive care system.15
Me thods
Study Design
We conducted a multicenter, randomized, openlabel trial, with written informed consent obtained
from the parents or guardians of all patients. Enrollment began in August 1996, and the last sub536
m e dic i n e
ject to be enrolled completed the study in April
2005. The power calculation was based on pilot
data indicating that normal joint structure would
be maintained in 70% of children receiving prophylaxis and 20% of those receiving enhanced episodic
therapy. Estimated proportions of loss of participants were 10% for the assessment of early joint
damage, 7% for the development of high-titer factor VIII antibodies, 7% for the assessment of lifethreatening hemorrhage, and 10% for follow-up.
Thus, 64 participants were needed to detect a significant difference between the two treatments
with a two-sided test (0.05 alpha level and 95%
power). Randomization was performed centrally
and stratified by site in permuted blocks of 2, 4,
or 6. The radiologists who reviewed joint images,
the physiotherapists who performed joint examinations, and the laboratory technologists who
performed assays were unaware of the patients’
treatment assignments and status with respect to
a history of bleeding.
Eligibility and Exclusion Criteria
Eligibility criteria were an age of less than 30
months, a factor VIII activity level of 2 U per deciliter or less, a history of two or fewer hemorrhages
into each index joint, normal baseline joint imaging, undetectable levels of factor VIII inhibitor,
a normal platelet count, and normal joint motion.
Children in the prophylaxis group received infusions of 25 IU of factor VIII (Kogenate or Kogenate FS, Bayer HealthCare) per kilogram of body
weight every other day to prevent bleeding. The
dose and the frequency of administration were
based on pharmacokinetic studies and clinical experience.9,16 Hemarthroses were defined as acute
episodes of joint pain with decreased joint motion.
When hemarthroses occurred during prophylaxis,
patients were treated with 40 IU per kilogram, and
the assigned prophylaxis schedule was resumed
the next day.
Children assigned to receive enhanced episodic
therapy were treated only at the time of clinically
recognized joint hemorrhage. The rationale for
this treatment was to decrease inflammation
and prevent joint damage by preventing rebleeding
after a joint hemorrhage. Children in this group
received 40 IU of factor VIII per kilogram at the
time of joint hemorrhage and 20 IU at 24 hours
and 72 hours after the first dose. Parents were
n engl j med 357;6  www.nejm.org  august 9, 2007
Preventing Joint Disease in Hemophilia
encouraged to continue infusions of 20 IU of factor VIII per kilogram every other day until joint
pain and impairment of mobility had completely
resolved, for a maximum of 4 weeks. All other
therapies, including surgery, and all bleeding
events other than hemarthroses, including nasal,
muscle, parenchymal, gastrointestinal, and intracranial hemorrhages, were managed according
to local standards of practice. In both groups,
the protocol allowed for two dose escalations of
5 IU of factor VIII per kilogram in the case of an
inadequate response. The protocol did not require the use of central-venous-access devices, and
all decisions regarding placement of the devices
were made according to local standards.
Outcome Measures
laxis group only), and serologic tests for hepatitis
B and C, human immunodeficiency virus, and parvovirus. Titers of factor VIII inhibitors were determined with the use of the Bethesda assay.19
Factor VIII trough levels were not used to alter
Clinical Assessment of Joints
Clinical examination of joints, with assessment of
swelling, strength, range of motion, pain, and gait,
was performed semiannually, as previously described, and videotaped for central review at study
entry, midpoint, and completion.20,21
Protocol Failure before Study Completion
The protocol allowed for early termination of participation if the assigned treatment was deemed
inadequate for the child as evidenced by the development of factor VIII inhibitors, life-threatening
hemorrhage, or bone or cartilage damage on joint
imaging. If an inhibitory titer exceeded 25 BU in
duplicate testing of the sample or if it exceeded
10 BU for more than 3 months, the child was withdrawn from the study. These thresholds were chosen to avoid the withdrawal of a child with a transient factor VIII inhibitor (Lusher JM: personal
Life-threatening hemorrhages were treated in
accordance with local standards. After the resolution of the first such event, the assigned treatment
was resumed. In the event of recurrence, the child
was removed from the study, but data were retained for inclusion in intention-to-treat analyses.
Participants with clinically suspected early joint
failure were eligible for an early joint evaluation.
The joint (or joints) in question were evaluated by
means of MRI, radiography, or both if the child
had had 8 hemorrhages into an index joint within
12 consecutive months or 20 hemorrhages into
an index joint since study enrollment or if the
highest score obtainable on any one item of the
joint physical examination had been recorded at
least 2 weeks after hemarthrosis. If the imaging
evaluation showed bone or cartilage damage, the
child was removed from the study.
The primary outcome was preservation of indexjoint structure, as determined by means of magnetic resonance imaging (MRI) and plain-film radiography at the completion of the study, when
participants were 6 years old. Secondary outcomes
were number of joint and other bleeding events,
number of infusions, and total units of factor VIII
administered. MRI and plain-film radiography
were performed as described previously.17,18 Joint
failure was defined as an MRI or radiograph
score that indicated a subchondral cyst, surface
erosion, or joint-space narrowing. MRIs and radiographs were read independently by two radiologists; discrepant readings were adjudicated by a
third radiologist.
Reports of infusions of factor VIII and emergency-room and clinic visits were collected monthly. At quarterly visits, data were collected on hospitalizations, port placements, port removals, and
infections. Each child was examined quarterly and
weighed for calculation of the dose of factor VIII.
Race and ethnic group were reported by the parent or guardian of each child.
Compliance was monitored by a review of infusion logs. However, no child was removed from
the study for any level of noncompliance. Death,
recurrent life-threatening hemorrhage, an inhibitory titer of 10 or more Bethesda units (BU), and
hospitalization were classified as serious adverse
Statistical Analysis
We used Fisher’s exact test to compare the two
Laboratory Assays
groups with respect to the primary outcome —
Blood was collected quarterly for the detection and the proportion of children in whom normal joint
measurement of factor VIII inhibitors, measure- structure was maintained, as determined by MRI
ment of factor VIII trough levels (in the prophy- or radiography. The relative risk of joint damage
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
119 Patients were screened
54 Were not eligible or did
not give consent
65 Underwent randomization
33 Were assigned to enhanced
episodic therapy
32 Were assigned to prophylaxis
3 Were removed from
protocol because of lifethreatening hemorrhage
6 Were removed from
protocol because of
joint damage
1 Was withdrawn
1 Was lost to follow-up
2 Were removed from
protocol because of hightiter inhibitor
1 Was removed from
protocol because of
joint damage
2 Were withdrawn
27 Completed protocol
22 Completed protocol
Figure 1. Randomization and Follow-up of Study Participants.
Although just 27 boys in the prophylaxis group and 22 boys in the episodic-therapy group remained on the protocol
unВ­til the age of 6 years, primary outcome
data wereManco-Johnson
available for boys who
were removed
from the protocol before
the age of 6 years.
REG F FIGURE: 1 of 3
and 95% confidence intervals were
joint imaging studies in children
Enon calculated for
the episodic-therapy group as compared with the who were withdrawn from the study because of
prophylaxis group. Differences inFigure
has been outredrawn early
and type
has been
reset. and joint imaging studies perPlease check carefully.
comes were evaluated with the t-test or the Mann–
formed in the remaining children at the age of
Whitney U test, as appropriate.
The Spearman cor- 6 years. For
the secondary analyses, data were inJOB: 35706
ISSUE: 08-09-07
relation coefficient was calculated for data that cluded until withdrawal from the study, loss to
were not normally distributed. Two-sided P values follow-up, early protocol failure, or completion of
of less than 0.05 were considered to indicate sta- the study at the age of 6 years.
tistical significance.
The proportion of data collected was calculated
Two interim analyses were planned and con- by dividing the number of data forms received by
ducted by an independent data and safety moni- the number of forms expected. Compliance was
toring board after one third and two thirds of determined by calculating the proportion of preparticipants had undergone evaluation of the out- scribed infusions that were actually administered.
come measures. Data used for interim analyses
included MRI and radiographic findings, the numR e sult s
ber of joint hemorrhages, the occurrence of lifethreatening hemorrhages, and the total number Sixty-five children were enrolled in the study beof hemorrhages and hospitalizations. All partici- tween August 1996 and March 2000; 32 children
pants randomly assigned to a treatment group were randomly assigned to prophylaxis and 33 to
were included in the intention-to-treat analysis of enhanced episodic treatment (Fig 1). The two
the primary outcome. Data used for this analysis groups showed no differences in baseline demo-
n engl j med 357;6  www.nejm.org  august 9, 2007
Preventing Joint Disease in Hemophilia
graphic characteristics (Table 1). The median factor VIII activity level for all the children was 0.6 U
per deciliter, with a range of 0.3 to 2.0; 31 of the
65 participants (48%) had one or more hemarthroses into index joints before enrollment.
The mean period of participation in the study
was 49 months (interquartile range, 48 to 58).
Primary outcome data from both MRI and radiographic studies were obtained for 50 of 65 participants (77%); partial data (with either MRI or
radiography) were obtained for 11 participants
(17%); and there were no data available for 4 participants (6%). Mean compliance was 96% (interquartile range, 96 to 100) in the prophylaxis group
and 98% (interquartile range, 98 to 100) in the
episodic therapy group. Among all participants,
an average of 94% of data forms were received.
Outcome results are shown in Table 2. According to the findings on MRI, the proportion of
participants in whom all six index joints were normal at 6 years of age was 25 of 27 (93%) in the
prophylaxis group and 16 of 29 (55%) in the enhanced episodic-therapy group (P = 0.002). As compared with the prophylaxis group, the episodictherapy group had a relative risk of damage to one
or more joints, as shown by MRI, of 6.1 (95% confidence [CI], 1.5 to 24.4). The corresponding relative risk for the prophylaxis group, as compared
with the episodic-therapy group, was 0.17, indicating an 83% reduction in the risk of joint damage as determined by MRI. With the use of radiography to assess joint damage, the relative risk
was 5.2 (95% CI, 0.65 to 41.5) with episodic therВ­
apy as compared with prophylaxis. Radiographic
and MRI readings were concordant in 97% of
index joints.
A total of 18 abnormal joints (13 ankles, 3 elbows, and 2 knees) were detected in 15 children
— 2 in the prophylaxis group and 13 in the episodic-therapy group. Six of the abnormalities
were detected by both MRI and radiography, seven by MRI alone, and one by radiography alone.
Only one type of imaging was available for the
four remaining abnormal joints.
For each joint, the MRI score was compared
with the total number of hemarthroses. As shown
in Figure 2, some joints had abnormal MRI scores
but no hemarthrosis, and some had normal MRI
scores despite many hemarthroses. Bone and cartilage damage detected on MRI was not correlated with hemarthroses (P = 0.63), and overall the
correlation of hemarthroses with MRI scores was
Table 1. Baseline Demographic and Clinical Characteristics of All Randomized
Episodic Therapy
(N = 33)
P Value
(N = 32)
Mean age (yr)
24 (75)
25 (76)
Race or ethnic group — no. (%)*
3 (9)
4 (13)
4 (12)
Asian or Pacific Islander
1 (3)
1 (3)
American Indian or
Alaskan native
1 (3)
2 (6)
Educational level of parent or
guardian — no. (%)
≤12 yr
20 (63)
13 (39)
>12 yr
12 (37)
20 (61)
First index-joint hemorrhage before
enrollment — no. (%)
18 (56)
13 (39)
14 (44)
20 (61)
No. of previous index-joint
No. of previous total hemorrhages
*Race and ethnic group were reported by the parent or guardian of each child.
weak (r = 0.14, P = 0.02). Joint physical-examination
scores showed a weak correlation with MRI scores
(r = 0.26, P<0.001).
Table 2 shows secondary outcomes. Table 3
shows serious adverse events. Average monthly factor VIII use and hemorrhages, as well as joint
physical examination scores, stratified by year of
age, are shown in Figure 3. No statistically significant differences between the two treatment groups
were found with respect to joint scores on physical
examination (Fig. 3A).
A central-venous-access device was placed in
54 children (83%). In 12 of these boys (22%), at
least one infection associated with the device developed. The median number of hospitalizations
per year was similar for both study groups. Most
hemophilia-related hospitalizations were for place-
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
Table 2. Outcome Data.*
(N = 32)
Episodic Therapy
(N = 33)
P Value
2 (7)
13 (45)
25 (93)
16 (55)
MRI findings
No. of participants with primary outcome data
Joint damage — no. (%)
No joint damage — no. (%)
Radiographic findings
No. of participants with primary outcome data
1 (4)
  5 (19)
27 (96)
22 (81)
Joint damage — no. (%)
No joint damage — no. (%)
No. of days in study
Reported no. of factor VIII infusions
Reported no. of factor VIII units infused
Joint hemorrhages (no./participant/yr)
Total hemorrhages (no./participant/yr)
*Plus–minus values are means ±SD. The data on MRI and radiographic findings include interim-analysis data for children
who were removed from the study because of early joint failure.
ment and removal of central-venous-access de- and was based on a history of joint hemorrhage
rather than age. In the prophylaxis group, radiologic evidence of preserved joint architecture was
found in 93% of participants at 6 years of age.
Dis cus sion
In this group, 18 of 32 (56%) of the children had
We found that prophylaxis with recombinant fac- one or two hemarthroses into one or more index
tor VIII was effective in preventing hemarthroses joints before prophylaxis, and 17 (53%) had one
and structural joint damage (as detected by MRI) to five hemorrhages into one or more index joints
in young boys with hemoВ­philia A.23 Reported sug- during prophylaxis. Prophylaxis was efficacious
gestions for the best time to begin prophylaxis in decreasing bleeding and joint damage after
range from before the first joint hemorrhage9 to up to five hemarthroses.
before 1 to 2 years of age to before the occurrence
More than half of the joint abnormalities that
of five hemarthroses.24 In our trial, prophylaxis were detected by MRI were not apparent in radiowas initiated between the ages of 6 and 30 months graphic studies, whereas only one joint abnormal-
n engl j med 357;6  www.nejm.org  august 9, 2007
Preventing Joint Disease in Hemophilia
MRI Score
ity that was detected by radiography was not detectВ­
ed by MRI, indicating that MRI is more sensitive
than radiography. We believe that MRI is the preВ­
ferable imaging technique for young boys with
Surprisingly, the number of clinically evident
hemarthroses correlated weakly with the outcome
as determined by MRI. In addition, joint abnormalities were not apparent on physical examination in the very young children in our study. It is
possible that the joint score we used was insufficiently sensitive for the detection of early arthropathy, even though our physical-examination
scoring system is more sensitive for the detection
of mild abnormalities of gait, joint swelling,
muscle strength, and atrophy than is that of the
World Federation of Hemophilia.9,12,13,20 Thus,
the absence of overt hemarthroses and abnormalities of joints on physical examination can
lead to the erroneous assumption that episodic
therapy in young children with hemophilia is effective. We propose that chronic microhemorrhage into the joints or subchondral bone in
young boys with hemophilia causes deterioration
of joints without clinical evidence of hemarthroses and that prophylaxis prevents this subclinical
The enhanced episodic therapy used in this
trial was experimental because it involved higher
doses and more infusions of factor VIII than are
provided in standard care. Enhanced episodic
therapy was used because the outcome of standard care is poor.13 Clearly, however, the results
of enhanced episodic therapy were inferior to
those of alternate-day prophylaxis.
Children who received enhanced episodic therapy had extra-articular bleeding in addition to
hemarthroses; 10% had recurrent, life-threatening
hemorrhage, including intracranial and gastrointestinal hemorrhage. Two children in the prophylaxis group were found to have high titers of
factor VIII inhibitors. This finding was not unexpected, since inhibitors develop in 30% of children with severe hemophilia, usually within the
first 50 exposures to factor VIII, and most of the
children in our study had fewer than 50 factor VIII
exposures at the time of enrollment.
Use of recombinant factor VIII has been estimated to account for more than 90% of the cost
of hemophilia care.25,26 By the age of 6 years, the
children in the prophylaxis group in our study
No. of Clinically Evident Index-Joint Hemorrhages
Figure 2. MRI Score for Index Joint According to
the ICM
of Hemorrhages
in That JointRETAKE
for Both 1st
Groups.2 of 3
MRI score is the average of the two concordant
Line a normal
4-C joint and
EMail a score of 0 indicates
H/T to H/T
of 7 to 10 indicates damage
joint bone16p6
or carEnon
tilage. The size of each circle corresponds to the numPLEASE of
ber of cases. TheAUTHOR,
largest number
index joints had
Figure has been redrawn and type has been reset.
no hemorrhage andPlease
an MRI
0. The shaded area
check carefully.
of the graph shows the 14 joints with bone or cartilage
on MRI. The larger bubbles represent
JOB: 35706
ISSUE: more
than one joint. Some joints with no hemorrhages had
high MRI scores, whereas some joints with more than
10 hemorrhages did not show bone or cartilage damage
on MRI.
Table 3. Serious Adverse Events.*
(N = 32)
(N = 33)
P Value
Detection of high-titer inhibitor
(no. of participants)
Life-threatening hemorrhage
(no. of participants)
Hemophilia-related hospitalization
CVAD (no. of participants)
≥1 CVAD-related infection
(no. of participants)
*Plus–minus values are means ±SD. CVAD denotes central-venous-access device.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
Episodic therapy
Average Joint Physical-Examination Score
Age (yr)
Average No. of Hemorrhages per Mo
Joint hemorrhages
Other hemorrhages
Episodic therapy:
Joint hemorrhages
Episodic therapy:
Other hemorrhages
Age (yr)
Episodic therapy
Average Monthly Factor VIII Use (IU/kg)
Age (yr)
were receiving 6000 IU of factor VIII per kilogram
Prophylaxis has not been widely used in the
per year, as compared with approximately 2500 IU care of patients with hemophilia. In 1995, when
per kilogram in the enhanced ICM
At the currentRETAKE
study was
1st conceived, only 33% of U.S.
a price of $1 per unit of recombinant
received prophylaxis.27
the cost of prophylaxis for a child
The Centers Revised
for Disease Control and Prevention
CASE weighing 50 kg
could reach $300,000 per year.EMail ARTIST: ts
that SIZE
51.5% of children with severe
n engl j med 357;6  www.nejm.org  august 9, 2007
Figure has been redrawn and type has been reset.
Please check carefully.
JOB: 35706
ISSUE: 08-09-07
Preventing Joint Disease in Hemophilia
Figure 3 (facing page). Joint Scores on Physical Examination, Frequency of Bleeding Events, and Factor VIII
Use According to Age and Study Group.
Panel A shows the mean joint score on physical examination. The scale sums the scores of all six index joints.
Mean joint scores for the two study groups were not
significantly different at any age. In Panel B, mean joint
and other hemorrhages increased progressively throughout the study in children receiving enhanced episodic
therapy, whereas the mean numbers of joint and other
hemorrhages remained at a low level in children receiving prophylaxis. As shown in Panel C, factor VIII use per
kilogram rose progressively throughout the study period
in the episode-therapy group; overall, however, there was
greater use in the prophylaxis group than in the episodictherapy group (P<0.001 for each year of age). T bars indicate standard errors. N denotes the number of participants at risk in each age group.
hemophilia who were younger than 6 years of
age received prophylaxis during 2004.28 We previously reported that the time required for infusions, unwillingness on the part of the child,
limitations in venous access, and difficulty in
balancing prophylaxis with other family needs
were major barriers to the implementation of
prophylaxis.29 Even in the present group of highly motivated, intensively supported families, the
infusion schedule was inadequate for 2 of the 32
participants in the prophylaxis group.
This study demonstrates the efficacy of prophylaxis with recombinant factor VIII in reducing the
incidence of joint hemorrhages, life-threatening
hemorrhages, and other hemorrhages and in lowering the risk of joint damage among young boys
with severe factor VIII deficiency. However, the
high cost of recombinant factor VIII is a barrier to widespread acceptance of prophylaxis.
Supported by grants from the Centers for Disease Control and
Prevention (U27/CCU812106) and the National Institutes of Health
(R00069). Bayer HealthCare donated the factor VIII used in the
study but had no role in study design, data accrual, data analysis,
or manuscript preparation. The Hemophilia and Thrombosis
Research Society recruited sites for participation.
Dr. Manco-Johnson reports receiving consulting fees from
Baxter BioScience, Bayer HealthCare, CSL Behring, and Wyeth,
speaking fees from Baxter BioScience, Bayer HealthCare, CSL
Behring, and Novo Nordisk, licensure study support from Baxter
BioScience, Bayer HealthCare, CSL Behring, Wyeth, and Novo
Nordisk, and grant support from Bayer HealthCare; Dr. Abshire,
consulting fees from Bayer HealthCare, CSL Behring, and Novo
Nordisk and licensure study support from Baxter BioScience,
Bayer HealthCare, CSL Behring, Octagen, and Wyeth; Dr. Shapiro,
consulting fees from Baxter BioScience, Bayer HealthCare, Inspiration Biopharmaceuticals, Pro Metic, Syntonix Pharmaceuticals, and Wyeth, speaking fees from Baxter BioScience, Novo
Nordisk, and Wyeth, licensure study support from Baxter BioScience, Bayer HealthCare, CSL Behring, Novo Nordisk, Octagen, and Wyeth, and grant support from Novo Nordisk; Ms.
Riske, consulting and lecture fees from Bayer HealthCare; Dr.
Valentino, consulting fees from Baxter BioScience, CSL Behring,
and Wyeth, speaking fees from Baxter HealthCare, Novo Nordisk,
and Wyeth, licensure study support from Baxter HealthCare,
Novo Nordisk, and Wyeth, and grant support from Baxter BioScience; Dr. Hoots, consulting fees from Baxter BioScience,
Bayer HealthCare, CSL Behring, Novo Nordisk, and Wyeth, speaking fees from Baxter Bioscience, Bayer HealthCare, Novo Nordisk,
and Wyeth, licensure study support from Avigen, Baxter BioВ­
Science, Bayer HealthCare, CSL Behring, Grifols, Novo Nordisk,
Octagen, Rho, and Wyeth, and grant support from Wyeth; Dr.
DiMichele, consulting fees from Bayer HealthCare, speaking fees
from Baxter BioScience, Bayer HealthCare, and Novo Nordisk,
licensure study support from Baxter BioScience, CSL Behring,
and Wyeth, and grant support from Baxter BioScience, Bayer
HealthCare, CSL Behring, Grifols Biologics, and Wyeth; Dr. Recht,
consulting fees from Baxter BioScience and licensure study support from Baxter BioScience and Wyeth; Dr. Leissinger, consulting fees from Baxter BioScience and CSL Behring, speaking fees
from Sanofi-Aventis and Grifols, and licensure study support
from Baxter BioScience, Bayer HealthCare, and Wyeth; Dr.
Mathew, consulting fees from Bayer HealthCare and CSL Behring,
speaking fees from Baxter BioScience, Bayer HealthCare, CSL
Behring, Novo Nordisk, and Wyeth, and licensure study support
from Baxter BioScience, Bayer HealthCare, Novo Nordisk, and
Wyeth; Dr. Nugent, consulting fees from Bayer HealthCare and
Novo Nordisk and licensure study support from Bayer HealthCare
and CSL Behring; Dr. Thompson, consulting and speaking fees
from Novartis and grant support from Baxter BioScience and
Novartis; and Dr. Brown, licensure study support from Grifols,
Baxter HealthCare, Novo Nordisk, Octagen, and Wyeth and grant
support from CSL Behring. No other potential conflict of interest relevant to this article was reported.
We thank the Data and Safety Monitoring Board members —
Dr. Louis Aledort (chair), Dr. David Tubergen, Dr. Gary Cutter, Dr.
Mark Yarborough, Mary Jo Cleveland, and Susan Havens — for
their valuable contributions throughout the study; the study nurses, especially Sheryl Giambartolomei, for their diligence in data
collection; Dr. Neil Goldenberg for his thoughtful comments and
support in the conduct and analysis of the study; Dr. Eduard Gorina, Dr. Peter Larsen, Dr. Richard Lutes, and Chris Cheney of
Bayer HealthCare; and the parents and participants who made
this study possible. This article is dedicated to Dr. Ray Kilcoyne,
pioneer in magnetic resonance joint imaging in hemophilia, who
died shortly after completion of the Joint Outcome Study.
The authors are affiliated with the following institutions: the University of Colorado and Health Sciences Center (M.J.M.-J., B.R., R.K.,
M.L.M.-J., S.F., L.J.,) and Children’s Hospital (M.J.M.-J., J.D.I.), Denver; Emory University, Atlanta (T.C.A.); Indiana Hemophilia and
Thrombosis Center, Indianapolis (A.D.S.); Beth Israel Deaconess Medical Center, Harvard Medical School, Boston (M.R.H.); Rush
Children’s Hospital, Chicago (L.A.V.); University of Texas, Houston (W.K.H., D.B.); University of Texas Southwestern Medical Center
and Children’s Medical Center at Dallas (G.R.B.); Weill Medical College of Cornell University, New York (D.D.); Phoenix Children’s
Hospital, Phoenix, AZ (M.R.); Tulane University, New Orleans (C.L.); Primary Children’s Hospital, Salt Lake City (S.B.); University of
Pennsylvania School of Medicine, Children’s Hospital of Philadelphia (A.C.); University of New Mexico, Albuquerque (P.M.); Oakland
Children’s Hospital, Oakland, CA (A.M.); University of Hawaii, Honolulu (D.M.); Children’s Hospital of Orange County, CA (D.N.);
Oregon Health and Science University, Portland (G.A.T.); Children’s Memorial Hospital and Northwestern University, Chicago (A.A.T.);
Palmetto Health Richland, Columbia, SC (K.M.); and Centers for Disease Control and Prevention, Atlanta (J.M.S., H.A., B.L.E.).
n engl j med 357;6  www.nejm.org  august 9, 2007
Preventing Joint Disease in Hemophilia
1. Mejia-Carvajal C, Czapek EE, Valenti-
no LA. Life expectancy in hemophilia outcome. J Thromb Haemost 2006;4:507-9.
2. Chorba TL, Holman RC, Strine TW,
Clarke MJ, Evatt BL. Changes in longevity
and causes of death among persons with
hemophilia A. Am J Hematol 1994;45:
3. Ikkala E, Heilske T, Myllylä G, Nevanlinna HR, Pitkänen P, Rasi V. Changes in
the life expectancy of patients with severe
haemophilia A in Finland in 1930-79. Br J
Haematol 1982;52:7-12.
4. Manco-Johnson MJ, Riske B, Kasper
CK. Advances in care of children with
hemoВ­philia. Semin Thromb Hemost 2003;
5. Ahlberg A. Haemophilia in Sweden.
VII. Incidence, treatment and prophylaxis
of arthropathy and other musculo-skeletal
manifestations of haemophlia A and B.
Acta Orthop Scand Suppl 1965;77:3-132.
6. Robinson PM, Tittley P, Smiley RK.
Prophylactic therapy in classical hemophilia: a preliminary report. Can Med Assoc J 1967;97:559-61.
7. Shanbrom E, Thelin GM. Experimental
prophylaxis of severe hemophilia with a
factor VIII concentrate. JAMA 1969;208:
8. Van Creveld S. Prophylaxis of joint
hemorrhages in hemophilia. Acta Haematol 1969;41:206-14.
9. Nilsson IM, Berntorp E, Löfqvist T, Pettersson H. Twenty-five years’ experience of
prophylactic treatment in severe haemophiВ­
lia A and B. J Intern Med 1992;232:25-32.
10. Medical and Scientific Advisory Council (MASAC) recommendations concerning
prophylaxis. Medical bulletin #193. New
York: National Hemophilia Foundation,
March 11, 1994.
11. Recombinate: licensed on December
10, 1992. FDA bulletin P92-39. Rockville,
MD: Food and Drug Administration, 1992.
(Accessed July 13, 2007, at http://www.fda.
12. Petrini P, Lindvall N, Egberg N, Blomback M. Prophylaxis with factor concentrates in preventing hemophilic arthropathy. Am J Pediatr Hematol Oncol 1991;
13. Aledort LM, Haschmeyer RH, Pettersson H. A longitudinal study of orthopaedic
outcomes for severe factor-VIII-deficient
haemophiliacs. J Intern Med 1994;236:
14. Stobart K, Iorio A, Wu JK. Clotting
factor concentrates given to prevent bleeding and bleeding-related complications in
people with hemophilia A or B. Cochrane
Database Syst Rev 2006;2:CD003429.
15. Baker JR, Crudder SO, Riske B, Bias V,
Forsberg A. A model for a regional system
of care to promote the health and wellbeing of people with rare chronic genetic
disorders. Am J Public Health 2005;95:
16. Hirschman RJ, Itscoitz SB, Shulman
NR. Prophylactic treatment of factor VIII
deficiency. Blood 1970;35:189-94.
17. Nuss R, Kilcoyne RF, Geraghty S, et
al. MRI findings in haemophilic joints
treated with radiosynoviorthesis with development of an MRI scale of joint damage. Haemophilia 2000;6:162-9.
18. Pettersson H, Ahlberg A, Nilsson IM.
A radiologic classification of hemophilic
arthropathy. Clin Orthop Relat Res 1980;
19. Kasper CK, Aledort L, Aronson D, et
al. Proceedings: a more uniform measurement of factor VIII inhibitors. Thromb
Diath Haemorr 1975;34:612.
20. Manco-Johnson MJ, Nuss R, Funk S,
Murphy J. Joint evaluation instruments for
children and adults with haemophilia.
Haemophilia 2000;6:649-57.
21. Hacker MR, Funk SM, Manco-Johnson MJ. The Colorado Haemophilia Pedi-
atric Joint Physical Examination Scale:
normal values and interrater reliability.
Haemophilia 2007;13:71-8.
22. Lusher JM. Natural history of inhibitor development in children with severe
hemophilia A treated with factor VIII products. In: Lee CA, Berntorp EE, Hoots WK,
eds. Textbook of hemophilia. Oxford, EnВ­
gland: Blackwell, 2005:34-8.
23. Soucie JM, Cianfrini C, Janco RL, et al.
Joint range-of-motion limitations among
young males with hemophilia: prevalence
and risk factors. Blood 2004;103:2467-73.
24. Kreuz W, Escuriola-Ettinghausen C,
Funk M, Schmidt H, Kornnhuber B. When
should prophylactic treatment in patients
with haemophilia A and B start? — the
German experience. Haemophilia 1998;4:
25. Smith PS, Teutsch SM, Shaffer PA,
Rolka H, Evatt B. Episodic versus prophylactic infusions for hemophilia A: a costeffectiveness analysis. J Pediatr 1996;129:
26. Globe DR, Curtis RG, Koerper MA.
Utilization of care in haemophilia: a resource-based method for cost analysis from
the Haemophilia Utilization Group Study
(HUGS). Haemophilia 2004;10:Suppl 1:
27. Blanchette VS, McCready M, Achonu
C, et al. A survey of factor prophylaxis in
boys with haemophilia followed in North
American haemophilia treatment centers.
Haemophilia 2003;9:Suppl 1:19-26.
28. Report on the Universal Data Collection Program. Atlanta: Centers for Disease
Control and Prevention, 2005;7(1):28. (Accessed July 18, 2007, at http://www.cdc.
29. Hacker MR, Geraghty S, Manco-Johnson M. Barriers to compliance with prophylaxis therapy in haemophilia. Haemophilia 2001;7:392-6.
Copyright В© 2007 Massachusetts Medical Society.
receive immediate notification when
a journal article is released early
To be notified when an article is released early
on the Web and to receive the table of contents
of the Journal by e-mail every Wednesday evening,
sign up through our Web site at
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
original article
A Single Cycle of Rituximab
for the Treatment of Severe Pemphigus
Pascal Joly, M.D., Ph.D., Hugo Mouquet, Ph.D., Jean-Claude Roujeau, M.D.,
Michel D’Incan, M.D., Ph.D., Danièle Gilbert, Ph.D., Serge Jacquot, M.D., Ph.D.,
Marie-Lise Gougeon, Ph.D., Christophe Bedane, M.D., Ph.D., Ralf Muller, Ph.D.,
Brigitte Dreno, M.D., Ph.D., Marie-Sylvie Doutre, M.D., Ph.D.,
Emmanuel Delaporte, M.D., Ph.D., Christine Pauwels, M.D.,
Nathalie Franck, M.D., FrГ©dГ©ric Caux, M.D., Ph.D., Catherine Picard, M.D., Ph.D.,
Emmanuelle Tancrede-Bohin, M.D., Philippe Bernard, M.D., Ph.D.,
François Tron, M.D., Ph.D., Michael Hertl, M.D., Ph.D.,
and Philippe Musette, M.D., Ph.D.
A bs t r ac t
The combination of multiple cycles of rituximab and intravenous immune globulins
has been reported to be effective in patients with severe pemphigus. The aim of this
study was to assess the efficacy of a single cycle of rituximab in severe types of pemphigus.
We studied 21 patients with pemphigus whose disease had not responded to an 8-week
course of 1.5 mg of prednisone per kilogram of body weight per day (corticosteroidrefractory disease), who had had at least two relapses despite doses of prednisone
higher than 20 mg per day (corticosteroid-dependent disease), or who had severe
contraindications to corticosteroids. The patients were treated with four weekly infusions of 375 mg of rituximab per square meter of body-surface area. The primary
end point was complete remission 3 months after the end of rituximab treatment;
complete remission was defined as epithelialization of all skin and mucosal lesions.
Eighteen of 21 patients (86%; 95% confidence interval, 64 to 97%) had a complete remission at 3 months. The disease relapsed in nine patients after a mean of 18.9В±7.9
months. After a median follow-up of 34 months, 18 patients (86%) were free of disease, including 8 who were not receiving corticosteroids; the mean prednisone dose
decreased from 94.0±10.2 to 12.0±7.5 mg per day (P = 0.04) in patients with corticosteroid-refractory disease and from 29.1±12.4 to 10.9±16.5 mg per day (P = 0.007) in
patients with corticosteroid-dependent disease. Pyelonephritis developed in one patient 12 months after rituximab treatment, and one patient died of septicemia 18
months after rituximab treatment. These patients had a profound decrease in the number of circulating B lymphocytes but normal serum levels of IgG.
From Rouen University Hospital, Rouen
(P.J., P.M.); INSERM UnitГ© 519, Rouen
UniВ­versity Hospital, Rouen (P.J., H.M.,
D.G., S.J., F.T., P.M.); Henri Mondor University Hospital, CrГ©teil (J.-C.R.); ClermontFerrand University Hospital, ClermontFerrand (M.D.); Pasteur Institute, Paris
(M.-L.G.); Limoges University Hospital,
Limoges (C.B.); Nantes University Hospital, Nantes (B.D.); Bordeaux University
Hospital, Bordeaux (M.-S.D.); Lille University Hospital, Lille (E.D.); Poissy–Saint
Germain-en-Laye Intercommunal Hospital, Saint Germain-en-Laye (C. Pauwels);
Paris V University, Paris (N.F.); Paris XIII
University Hospital, Bobigny (F.C.); Bichat
University Hospital, Paris (C. Picard);
Paris X University Hospital, Paris (E.T.-B.);
and Reims University Hospital, Reims
(P.B.) — all in France; and the Faculty
of Medicine, Philipp University, Marburg,
Germany (R.M., M.H.). Address reprint
requests to Dr. Joly at the Clinique
В­Dermatologique, HГґpital Charles Nicolle,
1 rue de Germont, 76031 Rouen CEDEX,
France, or at [email protected]
N Engl J Med 2007;357:545-52.
Copyright В© 2007 Massachusetts Medical Society.
A single cycle of rituximab is an effective treatment for pemphigus. Because of its potentially severe side effects, its use should be limited to the most severe types of the
disease. (ClinicalTrials.gov number, NCT00213512.)
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
emphigus is a life-threatening autoimmune blistering disease affecting the
skin and mucosa. It is mediated by pathogenic autoantibodies directed against desmoglein 1 and desmoglein 3, adhesion molecules of
the epidermis that are responsible for the cohesion between keratinocytes in skin and mucosa,
respectively.1-3 Patients with severe pemphigus require long-term treatment with corticosteroids and
other immunosuppressive drugs, which can lead
to serious adverse events.4,5
Rituximab, a monoclonal antibody directed
against the CD20 antigen of B lymphocytes, has
been demonstrated to be effective in various autoimmune diseases6-12 and in occasional cases of
life-threatening pemphigus.13-22 Recently, the combination of multiple cycles of rituximab and intravenous immune globulin, another potentially
active agent in pemphigus, was reported to be effective in a single-center study of 11 patients with
pemphigus.23 We report a multicenter series of 21
patients with severe pemphigus treated by a simple
regimen of one cycle of rituximab. Immunologic
evaluations explored the mechanism of the shortterm and long-term effects of rituximab on these
Me thods
m e dic i n e
The patients were treated with one cycle of four
weekly infusions of rituximab at a dose of 375 mg
per square meter of body-surface area on days 1,
8, 15, and 22. Corticosteroids were maintained at
the initial dose until the disease was controlled,
and the corticosteroid dose was then reduced by
10% twice a month. Patients with contraindications to corticosteroids were treated with rituximab
alone. Complete remission was defined as the epithelialization of all skin and mucosal lesions, partial remission as the epithelialization of more than
50% of lesions but not of all lesions, and relapse
as the occurrence of new cutaneous or mucosal
End Points
The patients were evaluated clinically and biologically every month during the first year of followup and every other month during the second year.
The primary end point was the rate of complete
remission 3 months after the last infusion of rituximab. The secondary end points were the rate of
complete remission during the study period, the
time from the start of rituximab treatment to complete remission, the number of relapses and the
length of time to each relapse, and adverse effects
of treatment.
Immunologic evaluation
Thirteen centers in France participated in this prospective, open trial. The study was approved by
the ethics committee of Seine-Maritime, and written informed consent was obtained from each patient.
Consecutive patients with severe mucosal erosions, superficial blisters, or both suggestive of
pemphigus vulgaris or pemphigus foliaceus; a histologic picture of intraepidermal acantholysis; and
deposition of IgG, complement component 3 (C3),
or both on the keratinocyte membrane detected
by direct immunofluorescence24 were included if
additional criteria were met. These were the absence of response to an 8-week course of treatment
with 1.5 mg of prednisone per kilogram of body
weight per day (corticosteroid-refractory cases); the
occurrence of at least two cutaneous or mucosal
relapses during the period when corticosteroid
doses were being decreased, despite prednisone
doses higher than 20 mg per day (corticosteroiddependent cases); and contraindication to corticosteroids because of severe associated medical
Blood samples were collected at each evaluation.
Titers of antibodies against desmoglein 1 and desmoglein 3 were measured by a desmoglein enzyme-linked immunosorbent assay (ELISA) test
(MESACUP Desmoglein Test, MBL Medical and Biological Laboratories) with 1:100 diluted serum.
For determination of IgG subclass, mouse antihuman IgG1, IgG2, IgG3, and IgG4 antibodies
(Sigma) and peroxidase-conjugated goat antimouse
IgG (Rockland) were used. IgG reactivity of pemphigus vulgaris serum against the NH2 terminal
domains of desmoglein 3 was characterized with
the use of ELISA with baculovirus-derived recombinant proteins of desmoglein 3.25
The phenotype of peripheral-blood mononuclear cells was determined by three-color flow cytometry with the use of murine monoclonal antibodies against CD3, CD4, CD5, CD19, CD20,
CD21, CD22, CD23, CD24, CD27, CD38, CD56, and
CD86 (Coulter). The IgG, IgA, and IgM repertoire
of peripheral-blood B lymphocytes was determined with the use of the immunoscope technique, a method that evaluates the diversity of the
n engl j med 357;6  www.nejm.org  august 9, 2007
Rituximab for the Treatment of Severe Pemphigus
B-lymphocyte repertoire and monitors its evolution. First, a polymerase chain reaction amplified
the heavy-chain variable (HV) region of the IgG,
IgA, and IgM molecules. In a second step, the
lengths of the complementary determining region 3 (CDR3) were analyzed on an automatic sequencer.26 The serum levels of antibodies against
pneumococcal capsule polysaccharide and tetanus
toxin were determined on days 0, 90, and 180.27
Statistical Analysis
Two-sided P values less than 0.05 were considered
to indicate statistical significance. Continuous variables are expressed as means В±SD. Mean doses
of prednisone were compared with the use of the
nonparametric Wilcoxon test. Roche, France (the
maker of rituximab) had no role in the design of
the study, in data accrual or analysis, or in manuscript preparation.
R e sult s
Characteristics of the Patients
Twenty-one patients (14 men and 7 women; 14 with
pemphigus vulgaris and 7 with pemphigus foliaceus) were enrolled between January 2003 and December 2004 and followed through March 2007.
Their mean (В±SD) age was 53.7В±15.6 years. The
baseline clinical characteristics of the patients are
shown in Table 1. The mean proportion of the
body-surface area involved was 27%. All the patients had severe oral involvement and weight loss
Table 1. Baseline Characteristics of Patients with Pemphigus Treated with Rituximab.*
Sex (M/F)
Age (yr)
CorticosteroidRefractory Disease
(N = 5)†CorticosteroidDependent Disease
(N = 11)
to Corticosteroids
(N = 5)
Type of pemphigus (no. of patients)
Pemphigus vulgaris
Pemphigus foliaceus
Duration of mucosal lesions (mo)
Duration of cutaneous lesions (mo)
Diabetes mellitus
Associated medical conditions (no. of patients)
Previous treatment failure (no. of patients)
Intravenous immune globulin
Mycophenolate mofetil
Mean body-surface area involved (%)
Prednisone dose at time of rituximab infusion
*Plus–minus values are means ±SD.
†Corticosteroid-refractory pemphigus did not respond to an 8-week course of treatment with 1.5 mg of prednisone per
kilogram of body weight per day.
‡Other medical conditions included asthma in one patient with corticosteroid-refractory pemphigus; rheumatoid arthritis
(one patient), myopathy (one), bone fracture (one), pyelonephritis (one), and depression (two) among patients with
corticosteroid-dependent disease; and blindness, bacterial meningitis, bone tuberculosis, gram-negative pneumonia,
osteonecrosis of the hip, and cardiac insufficiency, each in one patient, among those with contraindications to corticosteroids.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
Table 2. Complete Remission 3 Months after Rituximab Treatment in Patients with Severe Pemphigus.*
Pemphigus Vulgaris
(N = 14)
Pemphigus Foliaceus
(N = 7)
(N = 21)
no. with remission/total no.
Corticosteroid-refractory disease
Corticosteroid-dependent disease
Contraindication to corticosteroids
*Complete remission was defined as epithelialization of all skin and mucosal lesions.
of up to 10 kg. In addition, seven patients had severe involvement of one or more other mucosae,
including the genital, anal, conjunctival, or pharyngeal mucosae or a combination of these. All patients with corticosteroid-refractory disease and all
patients with corticosteroid-dependent disease had
been treated previously with various immunosuppressive drugs, intravenous immune globulins, or
both without success. The duration of all systemic
therapy before study entry ranged from 4 to 168
months (mean, 70.2). Severe associated medical
conditions or side effects of corticosteroids were
present at baseline in 14 patients (67%). Five patients were treated with rituximab alone because
they had contraindications to corticosteroid use.
Corticosteroids were contraindicated in two patients because of severe diabetes mellitus, arterial
hypertension, or both; in two patients because of
a history of severe infection, osteonecrosis of the
hip, or both; and in one patient because of old age
(84 years) and poor general condition (Table 1).
Response to rituximab treatment
Eighteen of 21 patients (86%; 95% confidence
interval, 64 to 94) had a complete remission at
3 months, including 12 of 14 patients with pemphigus vulgaris and 6 of 7 with pemphigus foliaceus (Table 2). Two patients with pemphigus vulgaris had a delayed complete remission on days 180
and 360 (Fig. 1). One patient with pemphigus foliaceus with lesions that initially involved the entire body surface was only slightly improved after
rituximab, and treatment of this patient was considered to have failed. All the other patients (95%)
had a complete remission at some time during the
study period. The median delays to partial and
complete remission are shown in Table 3.
Of the 20 patients who had a complete remission, 9 (6 with pemphigus vulgaris and 3 with
pemphigus foliaceus) had a relapse after a mean
period of 18.9В±7.9 months. Three patients with
corticosteroid-dependent disease had a total of six
relapses. Of the nine patients who had a relapse,
three were treated with topical corticosteroids
only, four with a moderately increased dose of oral
corticosteroids, and two with a second course of
rituximab. The second course of rituximab was
chosen because one of the two patients was 84
years old and was thought to be at high risk for
complications from corticosteroids, and the other
had a severe relapse. Both patients had a complete
remission again. After a median follow-up time of
34 months (range, 26 to 45), 18 patients (86%)
were free of disease, including 8 patients (38%)
who received no more corticosteroids.
The mean dose of prednisone for patients with
corticosteroid-refractory disease decreased from
94.0В±10.2 mg per day at baseline to 12.0В±7.5 mg
per day at the end of the study (P = 0.04) (Fig. 1).
The mean dose of prednisone for patients with
corticosteroid-dependent disease decreased from
29.1В±12.4 mg per day at baseline to 10.9В±16.5 mg
Figure 1 (facing page). Changes in Clinical Lesions,
Antidesmoglein Antibodies, and Doses of Prednisone
in Patients with Pemphigus Treated with Rituximab.
Panel A shows cutaneous lesions (blue) and mucosal
lesions (orange) in patients with pemphigus vulgaris;
Panel B shows cutaneous lesions (blue) and no mucosal lesions in patients with pemphigus foliaceus. Arrows
point to the number of days after rituximab treatment
at which a relapse occurred; patients were treated on
day 0. Panel C shows serum levels of anti–desmoglein
1 antibodies (blue) and anti–desmoglein 3 antibodies
(orange) in patients with pemphigus vulgaris. Panel D
shows serum levels of anti–desmoglein 1 antibodies
(blue) and no anti–desmoglein 3 antibodies in patients
with pemphigus foliaceus. Panels E and F show doses
of prednisone in patients with pemphigus vulgaris and
in those with pemphigus foliaceus, respectively.
n engl j med 357;6  www.nejm.org  august 9, 2007
120 180 240 300 360 420 480 540 600 660 720
FIGURE: 1 of 1
120 240 360 480 600 720 840 960 1080 1200 1320 1440
E Pemphigus vulgaris
120 240 360 480 600 720 840 960 1080 1200 1320 1440
C Pemphigus vulgaris
Mucosal Lesions
Anti–Desmoglein 3 Antibody
Titer (U/ml)
Skin Lesions
120 180 240 300 360 420 480 540 600 660 720
120 240 360 480 600 720 840 960 1080 1200 1320 1440
F Pemphigus foliaceus
Mucosal Lesions
120 240 360 480 600 720 840 960 1080 1200 1320 1440
D Pemphigus foliaceus
Skin Lesions
B Pemphigus foliaceus
Body-Surface Area Involved (%)
Anti–Desmoglein 1 Antibody
Titer (U/ml)
Body-Surface Area Involved (%)
Anti–Desmoglein 1 Antibody
Titer (U/ml)
Prednisone Dose (mg/day)
n engl j med 357;6  www.nejm.org  august 9, 2007
Prednisone Dose (mg/day)
A Pemphigus vulgaris
Rituximab for the Treatment of Severe Pemphigus
n e w e ng l a n d j o u r na l
Table 3. Median Time to Partial and Complete Remission in Patients
with Pemphigus Treated with Rituximab.
Pemphigus Vulgaris
(N = 14)
Type of Remission
Pemphigus Foliaceus
(N = 7)
median no. of days to remission
(interquartile range)
Partial remission
Cutaneous lesions
15 (15–30)
30 (30–56)
Mucosal lesions
60 (30–60)
Cutaneous lesions
30 (30–60)
60 (60–90)
Mucosal lesions
90 (60–90)
Complete remission
per day at the end of the study (P = 0.007). Nine
patients reported minor and transient side effects
during rituximab infusion: headache in three patients, asthenia in three, fever in one, chills in one,
and nausea in one. In addition, two severe side
effects were observed: one patient had pyeloВ­
nephritis 12 months after rituximab treatment,
and one patient died from septicemia 18 months
after rituximab treatment. The patient who died
also had rheumatoid arthritis and was concomitantly treated with the anti–tumor necrosis factor
agent etanercept.
Immunologic evaluations
Peripheral-blood B- and T-lymphocyte subpopulations and T-lymphocyte cytokine production were
analyzed longitudinally. The B-cell count decreased
dramatically from a median of 275 cells per cubic
millimeter (range, 29 to 752) to 0 cells per cubic
millimeter (range, 0 to 1) at day 21 and remained
undetectable until day 180 in all but two patients.
Reappearance of B cells began between day 180
and day 270. Ninety percent of them expressed
a CD19+CD27в€’ phenotype suggestive of naive B
lymphocytes. Twenty-nine percent had a CD19+
CD38highCD24high transitional phenotype suggestive of the migration of B cells from the bone
marrow to the periphery. No major changes in
T cells or natural killer cells or in T-cell cytokine production were detected after rituximab
Using the immunoscope method, we followed
the evolution of the repertoire of IgG, IgA, and
IgM in blood B lymphocytes in two patients. Before treatment, immunoscope profiles showed
some peak expansions in the immunoglobulin HV550
m e dic i n e
region genes (IgVH3a, IgVH3b, and IgVH4), the most
represented ones, reflecting in vivo antigen-driven
responses. Six months later, reappearing blood
B cells showed a typical gaussian distribution of
the CDR3 lengths for these IgVH families, similar
to that found in the naive B cells of cord blood,
suggesting the reconstitution of a diverse B-cell
repertoire (data not shown).
Antidesmoglein autoantibody response was then
analyzed. A dramatic decrease of IgG and IgG4
anti–desmoglein 1 and anti–desmoglein 3 antibodies was observed in 15 of the 18 patients who
had a complete remission 3 months after rituximab treatment, whereas persistent high titers
were detected in the 2 patients who had a delayed
complete remission and in the 1 patient in whom
treatment failed. Persistent high titers or rises in
anti–desmoglein 1 and anti–desmoglein 3 antibody levels were detected in patients with relapsing disease and, surprisingly, in five patients with
pemphigus vulgaris in whom a complete remission was maintained (Fig. 1). To disentangle this
discrepancy, we retested serum from these patients on recombinant proteins of five extracellular
(EC) domains of desmoglein 3 (EC1 to EC5). Antibodies reacting with epitopes of the EC1 to EC2
domains, which are considered the main pathogenic antibodies,28,29 decreased dramatically after
rituximab treatment and remained undetectable
until day 540, a result that was in accordance
with the absence of mucosal lesions in these
Finally, to assess the influence of rituximab on
antimicrobial response, we first determined the
serum level of IgG and IgM antibodies. No significant change in mean IgG levels was observed
after rituximab treatment (P = 0.79). On the contrary, mean IgM levels decreased from 1.5±1.1 g
per liter at baseline to 1.1В±0.9 g per liter on day
180 and 1.0±0.7 g per liter on day 360 (P = 0.003).
We then determined that the levels of antibodies
against pneumococcal capsule polysaccharide and
the levels of antibodies against tetanus toxin were
not significantly altered between baseline, day 21,
and day 90. The two patients who had severe infections, one at 12 months after rituximab treatment and one at 18 months, had normal serum
IgG levels (21 g per liter and 13.5 g per liter,
respectively), despite low numbers of circulating
B lymphocytes (5 per cubic millimeter and 9 per
cubic millimeter, respectively) at the times of their
n engl j med 357;6  www.nejm.org  august 9, 2007
Rituximab for the Treatment of Severe Pemphigus
Dis cus sion
Our study demonstrated the efficacy of rituximab
for severe pemphigus; 86% of patients were in complete remission 3 months after receiving four weekly infusions of rituximab. After a 34-month followup period, 18 patients were free of disease, and
8 of these patients were not receiving any systemic therapy. Overall, treatment with rituximab both
resulted in major clinical improvement and permitted a large decrease in the doses of corticosteroids. Our results were similar to those recently
reported by Ahmed et al.,23 who obtained a complete remission in 9 of 11 patients (82%). Our patients were treated with one cycle of four weekly
infusions of rituximab, and the patients in the
study by Ahmed et al. were treated with two induction cycles of rituximab, followed by consolidation
therapy and six infusions of intravenous immune
globulins, another potentially active agent in pemphigus.23,32,33 Therefore, we believe multiple cycles
of rituximab may not be necessary. Accordingly, we
suggest restricting the use of additional cycles of
rituximab to the few relapses that cannot be adequately controlled with conventional treatments.
Two severe side effects were observed in our
study: pyelonephritis developed in one patient 12
months after rituximab treatment, and one patient
died from septicemia 18 months after rituximab
treatment. It is important to emphasize that rituximab treatment is associated with a risk of death
from severe side effects, such as pneumocystis infection, toxic epidermal necrolysis, and progressive
multifocal leukoencephalopathy.34,35 The risk of
serious infection led Ahmed et al. to propose combining intravenous immune globulins with rituximab during the first 6 months of treatment with
rituximab.23 We do not think that 6 months of
treatment with immune globulins would have prevented the two serious infections observed in our
study, which occurred at 12 and 18 months in
patients who had no defect in IgG serum levels or
in preformed antibacterial antibodies. It is not
known whether the addition of intravenous imReferences
1. Stanley JR. Pemphigus and pemphigoid as paradigms of organ-specific, autoantibody-mediated diseases. J Clin Invest
2. Stanley JR. Cell adhesion molecules as
targets of autoantibodies in pemphigus
and pemphigoid, bullous diseases due to
mune globulins to rituximab reduces the risk of
serious infections.36,37
Immunologic investigations demonstrated that
one cycle of rituximab induced a prolonged depletion of peripheral-blood B lymphocytes, followed
by the reappearance of B cells with a naive phenotype similar to that found in cord blood from
neonates. Rituximab induced a modification of the
repertoire of B cells, with the disappearance of the
initially expanded populations and the reconstitution of a diverse B-cell repertoire that might
account for its long-lasting effect.
The clinical response to rituximab in patients
with pemphigus foliaceus was closely related to
the evolution of anti–desmoglein 1 antibodies,
which dramatically decreased in patients who had
a complete remission, whereas we were surprised
to observe persistently high levels of anti–desmoglein 3 antibodies in a few patients in whom pemphigus vulgaris was in complete remission. The
dissociation between anti–desmoglein 1 and anti–
desmoglein 3 antibody responses, as well as the
absence of major modification of serum levels of
antibodies against pneumococcal and tetanustoxin antigens, suggested some specificity of rituximab for the autoreactive B-cell response.38,39 One
hypothesis is that rituximab would differentially
affect the turnover of short-life and long-life plasma cells that produce antibodies directed against
desmoglein 1 and desmoglein 3.40
Overall, our study suggests that rituximab is a
very effective treatment for severe pemphigus.
Larger series with a longer follow-up are needed
to assess the long-term risks of this treatment.
Supported by the French Society of Dermatology, Paris, and
by Roche, France.
Presented in part at the December 2006 Congress of the
French Society of Dermatology and at the September 2006 Congress of the European Society of Dermatological Research
(ESDR), both in Paris.
No potential conflict of interest relevant to this article was
We thank Marie-France Hellot for her help with the statistical
analysis, Vincent Ferranti for monitoring the study, Ken Wood
for assistance in editing an earlier version of the manuscript,
and Nathalie Jourdain for excellent secretarial assistance.
defective epidermal cell adhesion. Adv
Immunol 1993;53:291-325.
3. Amagai M, Klaus-Kovtun V, Stanley
JR. Autoantibodies against a novel epithelial cadherin in pemphigus vulgaris,
a disease of cell adhesion. Cell 1991;67:
4. Bystryn JC, Steinman NM. The adju-
vant therapy of pemphigus: an update. Arch
Dermatol 1996;132:203-12.
5. McDonald CJ. Cytotoxic agents for
use in dermatology. J Am Acad Dermatol
6. Kazkaz H, Isenberg D. Anti B cell ther-
n engl j med 357;6  www.nejm.org  august 9, 2007
Rituximab for the Treatment of Severe Pemphigus
apy (rituximab) in the treatment of autoimmune diseases. Curr Opin Pharmacol 2004;
7. Stasi R, Stipa E, Forte V, Meo P, Amadori S. Variable patterns of response to
rituximab treatment in adults with chronic idiopathic thrombocytopenic purpura.
Blood 2002;99:3872-3.
8. Narat S, Gandla J, Hoffbrand AV,
Hughes RG, Mehta AB. Rituximab in the
treatment of refractory autoimmune cytopenias in adults. Haematologica 2005;90:
9. Zaja F, De Vita S, Russo D, et al.
Rituximab for the treatment of type II
mixed cryoglobulinemia. Arthritis Rheum
10. Specks U, Fervenza FC, McDonald TJ,
Hogan MC. Response of Wegener’s granulomatosis to anti-CD20 chimeric monoclonal antibody therapy. Arthritis Rheum
11. Levine TD. Rituximab in the treatment
of dermatomyositis: an open-label pilot
study. Arthritis Rheum 2005;52:601-7.
12. Arzoo K, Sadeghi S, Liebman HA.
Treatment of refractory antibody mediated autoimmune disorders with an antiCD20 monoclonal antibody (rituximab).
Ann Rheum Dis 2002;61:922-4.
13. Dupuy A, Viguier M, Bedane C, et al.
Treatment of refractory pemphigus vulgaris with rituximab (anti-CD20 monoclonal antibody). Arch Dermatol 2004;140:
14. Espana A, Fernandez-Galar M, Lloret P,
Sanchez-Ibarrola A, Panizo C. Long-term
complete remission of severe pemphigus
vulgaris with monoclonal anti-CD20 antibody therapy and immunophenotype correlations. J Am Acad Dermatol 2004;50:
15. Morrison LH. Therapy of refractory
pemphigus vulgaris with monoclonal antiCD20 antibody (rituximab). J Am Acad
Dermatol 2004;51:817-9.
16. Wenzel J, Bauer R, Bieber T, Tuting T.
Successful rituximab treatment of severe
pemphigus vulgaris resistant to multiple
immunosuppressants. Acta Derm Venereol 2005;85:185-6.
17. Schmidt E, Herzog S, Brocker EB, Zillikens D, Goebeler M. Long-standing remission of recalcitrant juvenile pemphigus
vulgaris after adjuvant therapy with rituximab. Br J Dermatol 2005;153:449-51.
18. Arin MJ, Engert A, Krieg T, Hunzel-
mann N. Anti-CD20 monoclonal antibody
(rituximab) in the treatment of pemphigus. Br J Dermatol 2005;153:620-5.
19. Kong HH, Prose NS, Ware RE, Hall
RP III. Successful treatment of refractory
childhood pemphigus vulgaris with antiCD20 monoclonal antibody (rituximab).
Pediatr Dermatol 2005;22:461-4.
20. Herrmann G, Hunzelmann N, Engert
A. Treatment of pemphigus vulgaris with
anti-CD20 monoclonal antibody (rituximab). Br J Dermatol 2003;148:602-3.
21. Goebeler M, Herzog S, Brocker EB,
Zillikens D. Rapid response of treatmentresistant pemphigus foliaceus to the antiCD20 antibody rituximab. Br J Dermatol
22. El Tal AK, Posner MR, Spigelman Z,
Ahmed AR. Rituximab: a monoclonal antibody to CD20 used in the treatment of
pemphigus vulgaris. J Am Acad Dermatol
23. Ahmed AR, Spigelman Z, Cavacini
LA, Posner MR. Treatment of pemphigus
vulgaris with rituximab and intravenous
immune globulin. N Engl J Med 2006;
24. Lever WF. Pemphigus and pemphigoid: a review of the advances made since
1964. J Am Acad Dermatol 1979;1:2-31.
25. MГјller R, Svoboda V, Wenzel E, et al.
IgG reactivity against non-conformational
NH2-terminal epitopes of the desmoglein
3 ectodomain relates to clinical activity
and phenotype of pemphigus vulgaris.
Exp Dermatol 2006;15:606-14.
26. Lim A, Lemercier B, Werth X, Lesjean
Pottier S, Huetz F, Kourilsky P. Most human peripheral blood B cells display a
unique heavy chain rearrangement. Int
Immunol (in press).
27. Ballet JJ, Sulcebe G, Couderc LJ, et al.
Impaired anti-pneumococcal antibody response in patients with AIDS-related persistent generalized lymphadenopathy. Clin
Exp Immunol 1987;68:479-87.
28. Anzai H, Fujii Y, Nishifuji K, et al.
Conformational epitope mapping of antibodies against desmoglein 3 in experimental murine pemphigus vulgaris. J Dermatol Sci 2004;35:133-42.
29. Sekiguchi M, Futei Y, Fujii Y, Iwasaki
T, Nishikawa T, Amagai M. Dominant
autoimmune epitopes recognized by pemphigus antibodies map to the N-terminal
adhesive region of desmogleins. J Immunol 2001;167:5439-48.
30. Hacker MK, Janson M, Fairley JA, Lin
MS. Isotypes and antigenic profiles of
pemphigus foliaceus and pemphigus vulgaris. Clin Immunol 2002;105:64-74.
31. Amagai M, Komai A, Hashimoto T,
et al. Usefulness of enzyme-linked immunosorbent assay using recombinant desmoВ­
gleins 1 and 3 for serodiagnosis of pemphigus. Br J Dermatol 1999;140:351-7.
32. Ahmed AR. Treatment of autoimmune
mucocutaneous blistering diseases with
intravenous immunoglobulin therapy. Expert Opin Investig Drugs 2004;13:1019-32.
33. Ahmed AR, Dahl MV. Consensus
statement on the use of intravenous immunoglobulin therapy in the treatment
of autoimmune mucocutaneous blistering diseases. Arch Dermatol 2003;139:10519.
34. Bermudez A, Marco F, Conde E, Mazo
E, Recio M, Zubizaretta A. Fatal varicellazoster infection following rituximab and
chemotherapy treatment in a patient with
follicular lymphoma. Haematologica 2000;
35. Quartier P, Tournilhac O, Archimbaud
C, et al. Enteroviral meningoencephalitis
after anti-CD20 (rituximab) treatment.
Clin Infect Dis 2003;36:e47-e49.
36. Lehrnbecher T. Intravenous immunoglobulins in the prevention of infection in
children with hematologic-oncologic diseases. Klin Padiatr 2001;213:Suppl 1:A103A105. (In German.)
37. Sullivan KM, Kopecky KJ, Jocom J, et
al. Immunomodulatory and antimicrobial
efficacy of intravenous immunoglobulin
in bone marrow transplantation. N Engl J
Med 1990;323:705-12.
38. Smith MR. Rituximab (monoclonal
anti-CD20 antibody): mechanism of action and resistance. Oncogene 2003;22:
39. Maloney DG, Smith B, Rose A. Rituximab: mechanism of action and resistance. Semin Oncol 2002;29:Suppl 2:2-9.
40. Hoyer BF, Manz RA, Radbruch A, Hiepe
F. Long-lived plasma cells and their contribution to autoimmunity. Ann N Y Acad
Sci 2005;1050:124-33.
Copyright В© 2007 Massachusetts Medical Society.
full text of all journal articles on the world wide web
Access to the complete text of the Journal on the Internet is free to all subscribers. To use this Web site, subscribers should go
to the Journal’s home page (www.nejm.org) and register by entering their names and subscriber numbers as they appear on
their mailing labels. After this one-time registration, subscribers can use their passwords to log on for electronic access to the
entire Journal from any computer that is connected to the Internet. Features include a library of all issues since January 1993
and abstracts since January 1975, a full-text search capacity, and a personal archive for saving articles and search results of
interest. All articles can be printed in a format that is virtually identical to that of the typeset pages. Beginning 6 months after
publication, the full text of all Original Articles and Special Articles is available free to nonsubscribers who have completed a
brief registration.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
original article
Complement C3 Variant and the Risk
of Age-Related Macular Degeneration
John R.W. Yates, F.R.C.P., Tiina Sepp, Ph.D., Baljinder K. Matharu, M.Sc.,
Jane C. Khan, F.R.C.Ophth., Deborah A. Thurlby, R.G.N., M.Sc.,
Humma Shahid, M.R.C.Ophth., David G. Clayton, M.A.,
Caroline Hayward, Ph.D., Joanne Morgan, B.Sc., Alan F. Wright, Ph.D., F.R.C.P.,
Ana Maria Armbrecht, Ph.D., F.R.C.S., Baljean Dhillon, F.R.C.S., F.R.C.Ophth.,
Ian J. Deary, Ph.D., F.R.C.P.E., Elizabeth Redmond, R.G.N., M.Sc.,
Alan C. Bird, M.D., F.R.C.S., and Anthony T. Moore, F.R.C.S., F.R.C.Ophth.,
for the Genetic Factors in AMD Study Group*
A bs t r ac t
Age-related macular degeneration is the most common cause of blindness in Western populations. Susceptibility is influenced by age and by genetic and environmental factors. Complement activation is implicated in the pathogenesis.
We tested for an association between age-related macular degeneration and 13 singlenucleotide polymorphisms (SNPs) spanning the complement genes C3 and C5 in
case subjects and control subjects from the southeastern region of England. All
subjects were examined by an ophthalmologist and had independent grading of
fundus photographs to confirm their disease status. To test for replication of the
most significant findings, we genotyped a set of Scottish cases and controls.
The common functional polymorphism rs2230199 (Arg80Gly) in the C3 gene, corresponding to the electrophoretic variants C3S (slow) and C3F (fast), was strongly
associated with age-related macular degeneration in both the English group (603
cases and 350 controls, P = 5.9×10–5) and the Scottish group (244 cases and 351
controls, P = 5.0×10–5). The odds ratio for age-related macular degeneration in C3
S/F heterozygotes as compared with S/S homozygotes was 1.7 (95% confidence interval [CI], 1.3 to 2.1); for F/F homozygotes, the odds ratio was 2.6 (95% CI, 1.6 to
4.1). The estimated population attributable risk for C3F was 22%.
Complement C3 is important in the pathogenesis of age-related macular degeneration. This finding further underscores the influence of the complement pathway in
the pathogenesis of this disease.
n engl j med 357;6  www.nejm.org  august 9, 2007
From the Cambridge Institute for Medical Research, University of Cambridge,
Cambridge (J.R.W.Y., T.S., B.K.M., J.C.K.,
D.A.T., H.S., D.G.C.); the Medical Research
Council Human Genetics Unit, Edinburgh
(C.H., J.M., A.F.W.); the Princess Alexandra Eye Pavilion, Edinburgh (A.M.A., B.D.);
the University of Edinburgh, Edinburgh
(I.J.D.); the Institute of Ophthalmology,
University College London (E.R., A.C.B.,
A.T.M.); and Moorfields Eye Hospital,
London (A.C.B., A.T.M.) — all in the United Kingdom. Address reprint requests to
Dr. Yates at the Department of Medical
Genetics, University of Cambridge, Cambridge Institute for Medical Research,
Wellcome Trust/MRC Bldg., Box 139,
Addenbrooke’s Hospital, Cambridge CB2
0XY, United Kingdom, or at [email protected]
Drs. Yates and Sepp contributed equally to
this article.
*Other members of the Genetic Factors
in Age-Related Macular Degeneration
(AMD) Study Group are listed in the
This article (10.1056/NEJMoa072618) was
published at www.nejm.org on July 18,
N Engl J Med 2007;357:553-61.
Copyright В© 2007 Massachusetts Medical Society.
n e w e ng l a n d j o u r na l
ge-related macular degeneration
is the leading cause of visual impairment
in the elderly and the most common cause
of blindness in Western countries.1 It affects the
macular region of the retina. The macula has a
high density of photoreceptors and provides detailed central vision. In the early stages of the
disease (referred to as age-related maculopathy),
deposits called drusen develop between the retinal pigment epithelium and underlying choroid.1
Later, the disease is manifested as either extensive atrophy of the retinal pigment epithelium
and overlying photoreceptor cells (geographic atrophy) or aberrant choroidal angiogenesis (choroidal neovascularization).1 Both of these conditions can lead to a loss of central vision. The
pathogenesis of age-related macular degeneration
is poorly understood. As with other late-onset
chronic diseases, susceptibility is influenced by
age, ethnic background, and a combination of environmental and genetic factors.1,2 Smoking status and family history are well-established determinants of risk.1,2
Recently, polymorphisms in the genes coding
for complement factor H (CFH) and complement
factor B (CFB) have been shown to be predictors
of risk for age-related macular degeneration.3-11
Another susceptibility locus has been mapped to
chromosome 10q26; the causative variation probably lies in a hypothetical gene called LOC387715
or in the promoter of the neighboring gene
HTRA1.11-14 The population attributable risk associated with variants in CFH, CFB, and LOC387715/
HTRA1 is at least 50%.11
CFH and CFB are key components of the alternative complement pathway. Their involvement
in age-related macular degeneration, together with
the finding that drusen contain proteins associated with inflammation and immune-mediated
processes,15 supports the hypothesis that inflammation and complement activation influence the
pathogenesis of age-related macular degeneration.
To test whether variants in other genes encoding
proteins in the complement pathway influence
susceptibility to age-related macular degeneration, we genotyped single-nucleotide polymorphisms (SNPs) spanning the complement genes
C3 and C5, encoding central proteins in the complement cascade, in subjects with age-related
macular degeneration and in control subjects.
m e dic i n e
Me thods
Cases and Controls
We studied three case–control groups, two in En­
gland and one in Scotland. English group 1 comprised 446 case subjects with end-stage age-related macular degeneration (geographic atrophy or
choroidal neovascularization) and 267 control
subjects, who were spouses of the index patients.
All subjects were recruited from ophthalmic clinics in eight hospitals in southeastern England
from 2002 to 2004.7 English group 2 comprised
157 case subjects with end-stage age-related macular degeneration and 83 controls (67 spouses
and 16 friends of index patients) recruited from
2003 to 2005, the majority from Moorfields Eye
Hospital in London and the remainder from southeastern England. All subjects described themselves as “white” rather than “other” on a recruitment questionnaire.
The Scottish group comprised 505 case subjects with age-related maculopathy or end-stage
age-related macular degeneration and 351 control
subjects. A total of 337 case subjects from the
Lothian region were recruited from ophthalmic
clinics in Edinburgh and 46 case subjects from
hospitals in Dundee and Inverness from 2004 to
2006. Control subjects, who were recruited from
the same sources in similar proportions, comprised 32 spouses and 174 subjects who had undergone cataract surgery. Another 122 case subjects and 145 controls came from the 1921
Lothian birth cohort.16
Written informed consent was obtained from
all subjects. The research protocol was in keeping
with the provisions of the Declaration of Helsinki,
and approval was obtained from a multicenter
research ethics committee and from research
ethics committees for each institution. Subjects
were examined by an ophthalmologist, and data
were collected regarding medical history, lifestyle, and smoking history. Color, stereoscopic
fundus photography of the macular region was
performed in all subjects. For English subjects,
the images were graded at the Reading Centre,
Moorfields Eye Hospital, with the use of the
International Classification of Age-Related Maculopathy and Macular Degeneration.17 For Scottish subjects, a study investigator graded images;
for validation, images from 100 case subjects and
n engl j med 357;6  www.nejm.org  august 9, 2007
Complement C3 Variant and Age-Related Macular Degener ation
controls were independently graded at the Moorfields Reading Centre (kappa statistic, 0.84). Eight
prospective English controls with age-related macular degeneration and 60 prospective Scottish
controls with age-related maculopathy were reclassified as case subjects. Data on disease status, sex, age, and smoking history of subjects are
provided in Table 1.
We extracted genomic DNA from peripheralblood leukocytes. We selected SNPs spanning the
C3 and C5 genes from the International HapMap
Project18 data (release 19) for the Centre d’Étude
du Polymorphisme Humain (CEPH) population
(Utah residents with ancestry from northern and
western Europe). Criteria for the selection of
SNPs were high heterozygosity with a minor allele frequency of at least 10%, tagging of the most
common haplotypes, and coverage of the main
blocks of linkage disequilibrium. The C3 SNP
rs2230199 — which is predicted to result in a
substitution of a glycine residue for arginine at
position 80 (Arg80Gly) — generates the “fast”
electrophoretic allotype of C3 (called C3F); the
alternative allotype is “slow” (C3S).19,20 We included this SNP in the analysis to provide extra
coverage and because of evidence of a functional
difference between the two alleles. On the basis
of our initial analysis, we included rs1047286
(Pro292Leu), which has a known association
with rs2230199.20,21 Initial genotyping was carried out in English group 1. Markers of interest
were genotyped in group 2 when samples became available. Data from the Scottish group
were used for replication.
We performed genotyping in English subjects
with the use of a single-nucleotide primer extension assay (ABI Prism SNaPshot Multiplex Kit,
Applied Biosystems) and a genetic analyzer (ABI
Table 1. Disease Status, Sex, Age, and Smoking History of Subjects.*
English Subjects (Groups 1 and 2)
(N = 350)
Case Subjects
(N = 603)
Scottish Subjects
(N = 351)
Case Subjects
(N = 505)
Disease status — no.
Age-related maculopathy
Geographic atrophy
Choroidal neovascularization
Geographic atrophy and choroidal
Sex — no. (%)
151 (43)
270 (45)
152 (43)
199 (57)
333 (55)
199 (57)
315 (62)
Never smoked
141 (41)
214 (36)
161 (46)
240 (48)
Current or former smoker
206 (59)
381 (64)
186 (54)
259 (52)
0 pack-yr
141 (41)
214 (36)
0.1–20.0 pack-yr
126 (36)
166 (28)
20.1–40.0 pack-yr
61 (18)
136 (23)
>40.0 pack-yr
19 (5)
79 (13)
Mean age — yr†190 (38)
Smoking history — no. (%)
No. of subjects
Quantity of cigarettes smoked†*Plus–minus values are means ±SD.
†For the English subjects, P<0.001 for the comparisons between case and control subjects. Data on the quantity of cigarettes smoked were not collected for the Scottish subjects.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
Prism 3100, Applied Biosystems) and in Scottish
subjects — for rs2230199 and rs1047286 — with
the use of competitive allele-specific polymerasechain-reaction assays (Taqman SNP Genotyping
Assay, Applied Biosystems and KASPar SNP Genotyping System, KBiosciences, respectively). Manufacturers’ protocols were followed.
Statistical Analysis
We used the chi-square test for comparisons of
categorical variables and allele and genotype frequencies and to check for Hardy–Weinberg
equilibrium. All P values were calculated with
two-sided tests, and no correction was made for
multiple testing. The Mann–Whitney U test was
used to compare the ages of case subjects and
controls. Logistic-regression analysis was used
to investigate interactions between genotype and
other variables and to estimate odds ratios and
95% confidence intervals. The covariables of age
and smoking history were included in the logistic model if univariate analysis had shown a significant difference. Odds ratios for categorical
variables were estimated in relation to a reference
category. Data were analyzed with the use of the
SPSS statistical software package, version 11.0.
The population attributable risk was calculated from the formula 100D ÷ (1 + D), in which
D was equal to P1(RR1 – 1) + P2(RR2 – 1), where P1
and P2 are the frequencies of the at-risk genotypes, and RR1 and RR2 their associated relative
risks, as compared with the low-risk genotype.
For the purposes of estimation, odds ratios were
equated to relative risks, since the disease prevalence is low.
R e sult s
In the initial screening, 12 SNPs spanning C3
and C5 (those listed in Table 2, excluding
rs1047286) were genotyped in 446 case subjects
with late-stage age-related macular degeneration
and 267 control subjects (English group 1). No
evidence of an association was found with variants in C5 (Table 2). In C3, the expressed SNP
rs2230199 showed strong evidence of an association (P<0.001) and was genotyped in an additional 157 case subjects and 83 controls (English
group 2). The enlarged sample also provided
strong evidence of an association (P = 5.9×10–5)
(Table 2).
To test for replication of this finding, rs2230199
m e dic i n e
was genotyped by a different laboratory in 244
case subjects with late-stage age-related macular
degeneration, 261 case subjects with age-related
maculopathy, and 351 controls (Scottish group).
Again, there was a highly significant association
between the minor allele and age-related macular degeneration (P = 5.0×10–5) (Table 3). International HapMap Project18 data for the CEPH population showed that rs2230199 had an r2 value of
0.75 with rs2230203 but a low r2 value with other
C3 SNPs in our marker panel and with other C3
SNPs in the HapMap data set. SNP rs2230203
did not show a significant association with agerelated macular degeneration in group 1 alone,
but there was weak evidence of an association in
groups 1 and 2 combined (Table 2).
Because of the known association between
the allotypes of rs2230199 and the expressed C3
SNP rs1047286, the English and Scottish subjects
were genotyped for this marker (Tables 2 and 3).
The minor allele frequency was significantly higher in case subjects than in controls in both groups,
but the association was not as strong as for
rs2230199. Stepwise logistic-regression analysis
confirmed that rs2230199 is a significantly better predictor of risk for age-related macular degeneration. With this SNP in the model, adding
rs1047286 made no contribution (P = 0.90). With
rs1047286 in the model, adding rs2230199 produced a significant improvement in fit (P = 0.02).
Odds ratios for age-related macular degeneration as a function of rs2230199 genotype are given
in Table 4. Results for the English and Scottish
groups were similar. In the combined data set,
with the common CC genotype as the reference,
the odds ratio was 1.7 for CG heterozygotes and
2.6 for GG homozygotes. The estimated population attributable risk for this variant was 22%.
Subgroup analysis that was confined to case
subjects with only choroidal neovascularization
showed a highly significant association in both
case–control groups. For case subjects with only
geographic atrophy, the association was significant in the English group (P = 4.6×10–4) but not
in the Scottish group, which had fewer subjects
with geographic atrophy. The Scottish group included case subjects with age-related maculopathy, and in this subgroup the association fell just
short of significance (Table 3).
Data on other susceptibility loci for age-related
macular degeneration were available for English
group 1. Results for CFH Y402H have been pub-
n engl j med 357;6  www.nejm.org  august 9, 2007
rs2230203 (6650782 bp, exon 13)
rs2230204 (6649848 bp, exon 14)
94 (36)
109 (45)
68 (27)
rs344550 (6622953 bp, intron 33)
rs2277984 (6619511 bp, intron 36)
105 (41)
rs344542 (6627517 bp, intron 27)
rs2241393 (6625304 bp, intron 29)
97 (37)
126 (47)
235 (67)
rs2287846 (6636557 bp, intron 22)
Pro292Leu (HAV4-1 allotype)
rs1047286 (6653262 bp, exon 9)
227 (65)
128 (50)
223 (64)
n engl j med 357;6  www.nejm.org  august 9, 2007
82 (33)
132 (53)
100 (40)
126 (50)
106 (44)
122 (47)
119 (46)
121 (47)
113 (42)
99 (28)
106 (30)
109 (32)
108 (42)
10 (4)
40 (16)
13 (5)
60 (24)
28 (12)
45 (17)
35 (14)
42 (16)
27 (10)
16 (5)
17 (5)
14 (4)
20 (8)
269 (63)
141 (33)
233 (59)
118 (28)
186 (44)
183 (42)
158 (36)
164 (38)
233 (52)
354 (59)
323 (54)
303 (51)
138 (33)
204 (47)
140 (35)
201 (47)
184 (43)
196 (45)
227 (52)
208 (48)
177 (40)
207 (34)
230 (38)
242 (41)
159 (36)
17 (4)
86 (20)
24 (6)
108 (25)
53 (13)
57 (13)
54 (12)
64 (15)
34 (8)
41 (7)
49 (8)
45 (8)
30 (7)
Case Subjects
number (percent)
248 (57)
P Value†Allele
*Because of genotyping failures, data were not available for all subjects. Genotype frequencies are given as percentages of subjects genotyped. The location of single-nucleotide polyВ­
morphisms (SNPs) in C3 and C5 is specified by the position on a contig map of chromosomes 19 and 9, respectively, from the National Center for Biotechnology Information’s SNP
Database, build 127. The numbering of the C3 amino acid sequence excludes the 22 residues of the signal peptide present in precursor C3 (pro-C3). Percentages may not total 100 because of rounding. “A” denotes common allele, “a” minor allele, and MAF minor allele frequency.
†P values were calculated with the chi-square test and are for comparisons of allele and genotype frequencies between case subjects and controls.
160 (63)
79 (31)
rs17611 (31090405 bp, exon 19)
rs7026551 (31054338 bp, intron 31)
137 (55)
rs7033790 (31109597 bp, intron 8)
number (percent)
Arg80Gly (slow/fast allotype)
rs2230199 (6658387 bp, exon 3)
Amino Acid Data
rs2250656 (6658534 bp, intron 2)
Gene and SNP ID
Table 2. Genotyping Results for English Subjects.*
Complement C3 Variant and Age-Related Macular Degener ation
18 (8)
13 (5)
94 (41)
116 (51)
13 (4)
Age-related macular degeneration
Pro292Leu (HAV4-1 allotype)
Arg80Gly (slow/fast allotype)
210 (64) 103 (32)
141 (56)
96 (38)
35 (7)
215 (65) 103 (31)
All subjects
14 (4)
215 (65) 103 (31)
Age-related maculopathy
14 (4)
117 (49) 100 (42)
215 (65) 103 (31)
Age-related macular degeneration
14 (4)
258 (53) 196 (40)
22 (9)
number (percent)
number (percent)
P Value†Case Subjects
Amino Acid Data
C3 Gene SNP ID
Table 3. Genotyping Results for Scottish Replication Group.*
*Because of genotyping failures, data were not available for all subjects. Genotype frequencies are given as percentages of subjects genotyped. The numbering of the C3 amino acid
­sequence excludes the 22 residues of the signal peptide present in precursor C3 (pro-C3). “A” denotes common allele, “a” minor allele, and MAF minor allele frequency.
†P values were calculated with the chi-square test and are for comparisons of allele and genotype frequencies between case subjects and controls.
n e w e ng l a n d j o u r na l
m e dic i n e
lished previously and are in agreement with other
reports.7 Odds ratios and population attributable
risks for LOC387715 (rs10490924) and CFB (rs641153)
are given in Table 5. The results are similar to
those of other studies, except that we found a
lower odds ratio for rs10490924 homozygotes.
When these variables were included in the stepwise logistic model, C3 rs2230199 remained significant, with an odds ratio of 1.4 for CG heterozygotes and 3.3 for GG homozygotes (with CC
genotype as the reference), confirming that these
susceptibility loci are independent risk factors.
Dis cus sion
Our study showed a strong association between
the complement C3 S/F (Arg80Gly) polymorphism
and age-related macular degeneration, with similar findings for geographic atrophy and choroidal neovascularization. The C3F allele frequency
is approximately 20% in white populations but
lower in other ethnic groups. For age-related
maculopathy, the association fell just short of
significance, raising the possibility that this polymorphism has less influence on the earlier stages
of the disease.
The complement system comprises more than
30 plasma and cell-surface proteins. It mediates
the host defense against pathogens and the elimination of immune complexes and apoptotic cells;
it also facilitates adaptive immune responses.22
C3 is the most abundant complement component,
synthesized predominantly in the liver but to a
lesser extent in other cells and tissues. Significant
C3 messenger RNA is detectable in the neural
retina, choroid, retinal pigment epithelium, and
cultured retinal-pigment-epithelium cells.15
Cleavage of C3 into C3a and C3b is the central step in complement activation and can be
initiated by the classic antibody-mediated pathway, the lectin pathway, or the alternative complement pathway.22 C3b attaches to pathogens or
other target surfaces and binds factor B, which
is then cleaved. The resulting C3bBb complex has
C3 convertase activity, which amplifies the response by further cleavage of C3 and leads to the
formation of C3b2Bb complexes with C5 convertase activity. This brings about cleavage of component C5 and recruitment by C5b of components C6 through C9 to form a large molecular
pore on target membranes (the membrane attack
complex), resulting in cell lysis.22
n engl j med 357;6  www.nejm.org  august 9, 2007
Complement C3 Variant and Age-Related Macular Degener ation
Drusen contain C3 and its activation products, as well as C5, membrane attack complex,
and CFH,6,15 supporting the hypothesis that local inflammation and activation of the complement cascade contribute to the pathogenesis of
age-related macular degeneration. Further support for this hypothesis comes from conclusive
evidence that variants in CFH influence susceptibility to age-related macular degeneration.3-9
CFH is a key regulator of the alternative complement pathway and prevents uncontrolled complement activation. Variants in factor B also apВ­
pear to influence susceptibility to age-related
macular degeneration.10,11 In mice, activation of
complement and formation of the membrane
attack complex are essential for the development
of laser-induced choroidal neovascularization.
Indeed, the finding that choroidal neovascularization cannot be induced by laser coagulation in
C3в€’/в€’ mice demonstrates the key role of C3 in this
As a result of cleavage of C3 to form C3b, the
molecule undergoes conformational changes that
expose several binding sites, including the thioВ­
ester moiety, which is essential for C3b binding
to target surfaces.24 Exposure of this activated
acyl-imidazole intermediate requires a substantial
relocation of the thioester-containing domain to
a position adjacent to the first macroglobulin
domain.24 Arg80 together with Arg72 and Lys82
forms a positively charged patch on the surface
of this domain, which, in C3b, is brought into
close proximity with the negatively charged carboxyl groups of several amino acids on the surface of the thioester-containing domain (Fig. 1).
Substitution of an uncharged glycine for the positively charged Arg80 is predicted to weaken the
interaction between these oppositely charged surfaces and could potentially influence thioester activity or other binding interactions of the thioВ­
ester-containing domain, including a probable
C3b/C3d binding site with CFH.25 It follows that
there could well be functional differences between the C3 S/F variants.
Direct experimental evidence of functional
differences in vitro between the C3 S/F allotypes
is not conclusive. Arvilommi26 reported that erythrocytes coated with C3F showed greater rosetting with peripheral-blood mononuclear cells than
those coated with C3S. Welch et al.27 studied uptake on sheep erythrocytes, hemolytic activity,
conversion to inactive C3b, and capacity to solu-
Table 4. Complement C3 rs2230199 Genotype (C3 S/F Allotype) and Odds Ratios
for Age-Related Macular Degeneration.*
Odds Ratio (95% CI)
CG (C3 S/F Allotype)
GG (C3 F/F Allotype)
English subjects
1.6 (1.2–2.2)
2.4 (1.3–4.4)
Scottish subjects
1.8 (1.2–2.6)
2.9 (1.4–5.9)
Combined groups
1.7 (1.3–2.1)
2.6 (1.6–4.1)
*Odds ratios are for the comparison with the CC genotype (C3 S/S allotype).
Table 5. Odds Ratios for Age-Related Macular Degeneration and Population
Attributable Risk for Variants at the Susceptibility Loci CFH, CFB, and LOC387715
in English Group 1.*
Locus (Variant)
Odds Ratio (95% CI)
Attributable Risk
CFH (Y402H)7
3.1 (2.0–4.6)
6.3 (3.8–10.4)
CFB (rs641153)
0.5 (0.3–0.8)
0.2 (0.02–1.8)
LOC387715 (rs10490924)
2.4 (1.6–3.6)
2.4 (1.5–3.8)
*Odds ratios are for the comparison with the AA genotype. “A” denotes common
allele, and “a” minor allele.
bilize preformed immune complexes. The only
significant difference was that C3F had lower
activity than C3S in a hemolytic assay using sensitized sheep erythrocytes as a result of a small
difference in cell-surface binding. BartГіk and
Walport28 found no differences between binding
of C3S and C3F and the major complement receptor types 1, 2, and 3.
On the other hand, there is compelling indirect evidence of a functional difference between
C3S and C3F. A recent study has shown that the
C3 S/F genotype is an important determinant of
the long-term outcome of renal transplantation.29
In recipients who were C3S homozygotes, graft
survival was substantially prolonged and renal
function significantly better with C3 F/F and C3
F/S donor kidneys than with C3 S/S kidneys.
Several associations of disease with C3F have
been reported, including IgA nephropathy,30 systemic vasculitis,31 partial lipodystrophy, and
membranoproliferative glomerulonephritis type
II (MPGNII).32,33 The association with MPGNII
is particularly relevant. This is a rare disease characterized by complement-containing dense deposits in the glomerular basement membrane of the
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
ated with serum C3 nephritic factor,34 an autoantibody directed against the C3bBb complex,
but rare cases associated with mutations in CFH
have been reported.35 A similar form of glomerulonephritis develops in CFH-deficient pigs36 and
CFH knockout mice.37 The interface of the capillary tuft, the glomerular basement membrane,
and the glomerular epithelial cells in the kidney
is similar in structure to the interface involving
choriocapillaris, Bruch’s membrane, and retinal
pigment epithelium in the eye, and macular drusen similar to those in age-related macular degeneration develop in patients with MPGNII, but
at a much younger age.38 These lesions are structurally and compositionally identical to those in
patients with age-related macular degeneration
and show immunoreactivity to complement C5
and C5b-9 complexes.39 Drusen have also been
reported in patients with partial lipodystrophy.40
The association of MPGNII and partial lipodystrophy with C3F fits well with our current findings.
In summary, our study shows a strong association between the C3F variant and age-related
macular degeneration, and there is evidence of
functional differences between the C3 S/F allotypes. It follows that C3F is likely to have a causal
role in the disorder. The estimated population
attributable risk in the white population is 22%.
These findings add to our growing understanding
of the genetics of age-related macular degeneration
and provide conclusive evidence that the complement pathway has a key role in the pathogenesis
of this common and debilitating condition.
Figure 1. Structure of Complement C3b, Showing the Location of Arg80.
Ribbon representation of the structure of complement C3b, as proposed
by Janssen et al.,24 shows the interface region between the macroglobulin 1
domain (MG1) (light blue) and the thioester-containing domain (orange).
The residues participating in the formation of the thioester bond are shown
in yellow. Arg80 (red) is located in MG1, adjacent to two other positively
charged amino acids, Arg72 and Lys82 (purple). These residues are approximately 4 Г… from the negatively charged amino acids Asp1007, Glu1008,
Glu1010, and Glu1013 (dark blue) in the thioester-containing domain. The
first three of these residues contribute to a probable C3b/C3d binding site
with complement factor H.25 Arg80 may also have interactions with negatively charged residues (not shown) in the complement C1r/C1s–Uegf–
Bmp1–containing domain, adjacent to the thioester-containing domain.
kidney.34 The condition is caused by uncontrolled
activation of the alternative complement pathway.
In the majority of patients, the condition is associ-
Supported by grants from the Medical Research Council,
United Kingdom (to Drs. Yates, Clayton, Bird, and Moore), the
Chief Scientist Office, Scotland (to Dr. Dhillon), the Macular
Vision Research Foundation (to Dr. Wright), and the Wellcome
Trust and the Juvenile Diabetes Research Foundation (to Dr.
Clayton). The Lothian Birth Cohort collection was supported by
the Biotechnology and Biological Sciences Research Council.
Dr. Deary is the recipient of a Royal Society–Wolfson Research
Merit Award.
Dr. Dhillon reports receiving consulting fees from Novartis
and Pfizer. No other potential conflict of interest relevant to this
article was reported.
We thank members of the Scottish Macula Society Study
Group (M. Gavin, F. Imrie, N. Lois, R. Murray, A. Purdie, A.
Pyott, S. Roxburgh, C. Styles, M. Virdi, and W. Wykes) for their
help in the recruitment of patients for our study; clinic staff and
medical photographers at the participating clinics for their help;
Tunde Peto and colleagues at the Reading Centre, Moorfields
Eye Hospital, London, for grading the fundus photographs; the
staff at Tepnel Life Sciences for performing the DNA extractions; Roger Williams for his helpful discussion about C3b protein structure; the International HapMap Consortium for the
use of data; and all the patients and their families who participated in the study.
n engl j med 357;6  www.nejm.org  august 9, 2007
Complement C3 Variant and Age-Related Macular Degener ation
The following investigators are members of the Genetic Factors in Age-Related Macular Degeneration Study Group: S.S. Bhattacharya,
P. Bishop, P. Black, Z. Butt, V. Chong, N.E. Day, C. Edelsten, A. Fitt, D.W. Flanagan, A. Glenn, S. Harding, C. Jakeman, C. Jones, R.J.
Lamb, A. Lotery, V. Moffatt, C.M. Moorman, T. Peto, R.J. Pushpanathan, and T. Rimmer.
1. de Jong PTMV. Age-related macular de-
generation. N Engl J Med 2006;355:147485.
2. Haddad S, Chen CA, Santangelo SL,
Seddon JM. The genetics of age-related
macular degeneration: a review of progress
to date. Surv Ophthalmol 2006;51:316-63.
3. Klein RJ, Zeiss C, Chew EY, et al.
Complement factor H polymorphism in
age-related macular degeneration. Science
4. Haines JL, Hauser MA, Schmidt S, et
al. Complement factor H variant increases
the risk of age-related macular degeneration. Science 2005;308:419-21.
5. Edwards AO, Ritter R III, Abel KJ,
Manning A, Panhuysen C, Farrer LA.
Complement factor H polymorphism and
age-related macular degeneration. Science
6. Hageman GS, Anderson DH, Johnson
LV, et al. A common haplotype in the complement regulatory gene factor H (HF1/
CFH) predisposes individuals to age-related macular degeneration. Proc Natl Acad
Sci U S A 2005;102:7227-32.
7. Sepp T, Khan JC, Thurlby DA, et al.
Complement factor H variant Y402H is
a major risk determinant for geographic
atrophy and choroidal neovascularization
in smokers and nonsmokers. Invest Ophthalmol Vis Sci 2006;47:536-40.
8. Li M, Atmaca-Sonmez P, Othman M,
et al. CFH haplotypes without the Y402H
coding variant show strong association
with susceptibility to age-related macular
degeneration. Nat Genet 2006;38:1049-54.
9. Hughes AE, Orr N, Esfandiary H, DiazTorres M, Goodship T, Chakravarthy U.
A common CFH haplotype, with deletion
of CFHR1 and CFHR3, is associated with
lower risk of age-related macular degeneration. Nat Genet 2006;38:1173-7. [Erratum, Nat Genet 2007;39:567.]
10. Gold B, Merriam JE, Zernant J, et al.
Variation in factor B (BF) and complement
component 2 (C2) genes is associated with
age-related macular degeneration. Nat
Genet 2006;38:458-62.
11. Maller J, George S, Purcell S, et al.
Common variation in three genes, including a noncoding variant in CFH, strongly
influences risk of age-related macular degeneration. Nat Genet 2006;38:1055-9.
12. Jakobsdottir J, Conley YP, Weeks DE,
Mah TS, Ferrell RE, Gorin MB. Susceptibility genes for age-related maculopathy
on chromosome 10q26. Am J Hum Genet
13. Dewan A, Liu M, Hartman S, et al.
HTRA1 promoter polymorphism in wet
age-related macular degeneration. Science
14. Yang Z, Camp NJ, Sun H, et al. A vari-
ant of the HTRA1 gene increases susceptibility to age-related macular degeneration. Science 2006;314:992-3.
15. Mullins RF, Russell SR, Anderson DH,
Hageman GS. Drusen associated with aging and age-related macular degeneration
contain proteins common to extracellular
deposits associated with atherosclerosis,
elastosis, amyloidosis, and dense deposit
disease. FASEB J 2000;14:835-46.
16. Deary IJ, Whiteman MC, Starr JM,
Whalley LJ, Fox HC. The impact of childhood intelligence on later life: following
up the Scottish mental surveys of 1932 and
1947. J Pers Soc Psychol 2004;86:130-47.
17. Bird AC, Bressler NM, Bressler SB, et
al. An international classification and
grading system for age-related maculopathy and age-related macular degeneration.
Surv Ophthalmol 1995;39:367-74.
18. The International HapMap Consortium. The International HapMap Project.
Nature 2003;426:789-96.
19. Azen EA, Smithies O. Genetic polymorphism of C’3(beta1C-globulin) in human serum. Science 1968;162:905-7.
20. Botto M, Fong KY, So AK, Koch C,
Walport MJ. Molecular basis of polymorphisms of human complement component C3. J Exp Med 1990;172:1011-7.
21. Koch C, Behrendt N. A novel polymorphism of human complement component
C3 detected by means of a monoclonal antibody. Immunogenetics 1986;23:322-5.
22. Walport MJ. Complement (first of
two parts). N Engl J Med 2001;344:105866.
23. Bora PS, Hu Z, Tezel TH, et al. Immunotherapy for choroidal neovascularization in a laser-induced mouse model
simulating exudative (wet) macular degeneration. Proc Natl Acad Sci U S A
24. Janssen BJ, Christodoulidou A, McCarthy A, Lambris JD, Gros P. Structure of
C3b reveals conformational changes that
underlie complement activity. Nature 2006;
25. Jokiranta TS, Jaakola VP, Lehtinen MJ,
Pärepalo M, Meri S, Goldman A. Structure of complement factor H carboxyl-terminus reveals molecular basis of atypical
haemolytic uremic syndrome. EMBO J
26. Arvilommi H. Capacity of complement
c3 phenotypes to bind on to mononuclear
cells in man. Nature 1974;251:740-1.
27. Welch TR, Beischel L, Kleesattel A.
Functional consequences of the genetic
polymorphism of the third component of
complement. J Pediatr 1990;116:S92-S97.
28. BartГіk I, Walport MJ. Comparison of
the binding of C3S and C3F to complement receptors types 1, 2, and 3. J Immunol 1995;154:5367-75.
29. Brown KM, Kondeatis E, Vaughan
RW, et al. Influence of donor C3 allotype
on late renal-transplantation outcome.
N Engl J Med 2006;354:2014-23.
30. Rambausek M, van den Wall Bake AW,
Schumacher-Ach R, et al. Genetic polymorphism of C3 and Bf in IgA nephropathy. Nephrol Dial Transplant 1987;2:
31. Finn JE, Zhang L, Agrawal S, Jayne
DR, Oliveira DB, Mathieson PW. Molecular analysis of C3 allotypes in patients with
systemic vasculitis. Nephrol Dial Transplant 1994;9:1564-7.
32. Finn JE, Mathieson PW. Molecular
analysis of C3 allotypes in patients with
nephritic factor. Clin Exp Immunol 1993;
33. McLean RH, Winkelstein JA. Genetically determined variation in the complement system: relationship to disease.
J Pediatr 1984;105:179-88.
34. Appel GB, Cook HT, Hageman G, et
al. Membranoproliferative glomerulonephritis type II (dense deposit disease):
an update. J Am Soc Nephrol 2005;16:
35. Abrera-Abeleda MA, Nishimura C,
Smith JL, et al. Variations in the complement regulatory genes factor H (CFH) and
factor H related 5 (CFHR5) are associated
with membranoproliferative glomerulonephritis type II (dense deposit disease).
J Med Genet 2006;43:582-9.
36. HГёgГҐsen K, Jansen JH, Mollnes TE,
Hovdenes J, Harboe M. Hereditary porcine membranoproliferative glomerulonephritis type II is caused by factor H deficiency. J Clin Invest 1995;95:1054-61.
37. Pickering MC, Cook HT, Warren J, et al.
Uncontrolled C3 activation causes membranoproliferative glomerulonephritis in
mice deficient in complement factor H.
Nat Genet 2002;31:424-8.
38. Colville D, Guymer R, Sinclair RA,
Savige J. Visual impairment caused by
retinal abnormalities in mesangiocapillary (membranoproliferative) glomerulonephritis type II (“dense deposit disease”).
Am J Kidney Dis 2003;42:E2-E5.
39. Mullins RF, Aptsiauri N, Hageman
GS. Structure and composition of drusen
associated with glomerulonephritis: implications for the role of complement activation in drusen biogenesis. Eye 2001;15:
40. Patel D, Page B. Ocular complications
in acquired partial lipodystrophy. Postgrad
Med J 2006;82:774.
Copyright В© 2007 Massachusetts Medical Society.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
original article
NXY-059 for the Treatment of Acute
Ischemic Stroke
Ashfaq Shuaib, M.D., Kennedy R. Lees, M.D., Patrick Lyden, M.D.,
James Grotta, M.D., Antonio Davalos, M.D., Stephen M. Davis, M.D.,
Hans-Christoph Diener, M.D., Tim Ashwood, Ph.D., Warren W. Wasiewski, M.D.,
and Ugochi Emeribe, Ph.D., for the SAINT II Trial Investigators*
A bs t r ac t
From the Division of Neurology, University of Alberta, Edmonton, Canada (A.S.);
the Acute Stroke Unit and Cerebrovascular Clinic, University Department of Medicine and Therapeutics, Gardiner Institute, Western Infirmary, Glasgow, United
Kingdom (K.R.L.); the University of California San Diego Stroke Center, San Diego
(P.L.); the Department of Neurology, University of Texas–Houston Medical School,
Houston (J.G.); the Department of Neurosciences, Hospital Universitari Germans
Trias i Pujol, Universitat AutГІnoma de
Barcelona, Badalona, Spain (A.D.); the
Department of Neurology, Royal Melbourne
Hospital, University of Melbourne, ParkВ­
ville, Australia (S.M.D.); the Department
of Neurology, University Duisburg–Essen,
Essen, Germany (H.-C.D.); AstraZeneca
R&D Södertälje, Medical Neuroscience,
Södertälje, Sweden (T.A.); and Astra­
Zeneca, Wilmington, DE (W.W.W., U.E.).
Address reprint requests to Dr. Lees at
the Acute Stroke Unit and Cerebrovascular Clinic, University Department of Medicine and Therapeutics, Gardiner Institute, Western Infirmary, 44 Church St.,
Glasgow G11 6NT, United Kingdom, or at
[email protected]
*The investigators participating in the
Stroke–Acute Ischemic NXY Treatment
II (SAINT II) trial are listed in the Appendix.
N Engl J Med 2007;357:562-71.
The free-radical–trapping agent NXY-059 showed promise as a neuroprotectant in
the Stroke–Acute Ischemic NXY Treatment I (SAINT I) trial, reducing disability
when given to patients who had acute ischemic stroke. We sought confirmation of
efficacy in a second, larger trial.
We enrolled 3306 patients with acute ischemic stroke in a randomized, doubleblind trial to receive a 72-hour infusion of intravenous NXY-059 or placebo within
6 hours after the onset of stroke symptoms. Our primary end point was the distribution of disability scores on the modified Rankin scale at 90 days. We examined
scores on neurologic and activities-of-daily-living scales as secondary end points.
We also tested the hypothesis that NXY-059 would reduce alteplase-related intracranial hemorrhages.
The efficacy analysis was based on 3195 patients. Prognostic factors were well balanced between the treatment groups. Mortality was equal in the two groups, and
adverse-event rates were similar. The distribution of scores on the modified Rankin
scale did not differ between the group treated with NXY-059 (1588 patients) and the
placebo group (1607 patients; P = 0.33 by the Cochran–Mantel–Haenszel test; odds
ratio for limiting disability, 0.94; 95% confidence interval [CI], 0.83 to 1.06). Analysis of categorized scores on the modified Rankin scale confirmed the lack of benefit: the odds ratio for trichotomization into modified Rankin scale scores of 0 to 1
versus 2 to 3 versus 4 to 6 was 0.92 (95% CI, 0.80 to 1.06). There was no evidence
of efficacy for any of the secondary end points. Among patients treated with alteplase,
there was no difference between the NXY-059 group and the placebo group in the
frequency of symptomatic or asymptomatic hemorrhage.
Copyright В© 2007 Massachusetts Medical Society.
NXY-059 is ineffective for the treatment of acute ischemic stroke within 6 hours
after the onset of symptoms. (ClinicalTrials.gov number, NCT00061022.)
n engl j med 357;6  www.nejm.org  august 9, 2007
NXY-059 for Acute Ischemic Stroke
urrently, thrombolysis with alteВ­
plase (tissue plasminogen activator [rt-PA])
is the only widely approved treatment for
acute stroke, and it is underused. There is an urgent need for new therapies that are safer and can
be offered to a higher percentage of patients.
Cerebral tissue can be protected in animal
models by a variety of agents that attenuate neuronal injury after ischemia,1 but none of these
putative neuroprotectants have been confirmed as
an effective therapy in clinical trials. NXY-059, a
free-radical–trapping agent, has been extensively
tested in animal models of focal ischemic stroke
and has been shown to improve functional recovery and reduce the size of the cerebral infarction.2
A large study reported recently showed that NXY059 was significantly better than placebo in improving the outcome in patients with ischemic
stroke treated within 6 hours after the onset of
symptoms.3 Further support for the suggestion
that NXY-059 is a true biologic signal came from
post hoc analysis of the data, which also revealed
that treatment with NXY-059 significantly reduced
the incidence of intracranial hemorrhage among
patients in whom alteplase was also used.3 We
sought to confirm the results of the Stroke–Acute
Ischemic NXY Treatment I (SAINT I) trial with a
larger trial, SAINT II.
Me thods
Study design
This randomized, double-blind, placebo-controlled
study enrolled 3306 patients from May 2003
through June 2006. The study involved 362 centers from 31 countries and was approved by local
or national institutional review boards as appropriate. Patients were assigned to treatment only
after they had given informed consent or, for patients who were unable to do so, after consent
had been obtained from an acceptable surrogate.
The study was part of the SAINT program,
which consists of two large, independent, doubleblind studies. Its design and conduct were developed by a steering committee of stroke experts
from Europe, North America, and Australia. The
steering committee had complete access to all
data and was responsible for writing the manuscript. An independent data and safety monitoring
committee was responsible for safety reviews and
futility analysis. The sponsor, AstraZeneca, was
responsible for operational aspects of the trial,
including collecting and storing the data, and
performing analyses according to the approved
plan. The academic authors vouch for the completeness and veracity of the data and analyses.
We originally planned to enroll 1700 patients
in SAINT II, but on the basis of the results of
SAINT I,3 we planned to increase the power of the
trial by enlarging the study population to 3200 paВ­
tients without unblinding of any SAINT II data.4
Patients were eligible for enrollment if they were
18 years of age or older and had a clinical diagnosis of an acute ischemic stroke with an onset within the previous 6 hours. They had to score at least
six points on the National Institutes of Health
Stroke Scale (NIHSS),5 with at least two points
for limb weakness. All patients received appropriate routine stroke care according to local treatment practice, including alteplase for eligible patients presenting within 3 hours after the onset of
the stroke. For patients receiving alteplase, treatment with the study drug had to be started within 30 minutes after completion of the alteplase
Study intervention
Patients were randomly assigned by a computergenerated coding system to receive an intravenous
infusion of either NXY-059 or placebo. The treating
centers were required to maintain an average time
of no more than 4 hours to the start of infusion
of the study drug. Randomization was stratified
according to country, NIHSS score at baseline,
side of the infarction, and intention to treat with
alteplase. AstraZeneca supplied the study drug as
a concentrate to be diluted to 15 mg per milliliter
in 500 ml of 0.9% saline solution. The initial infusion rate was 2270 mg per hour, and the rate
was reduced after an hour to 480 to 960 mg (32 to
64 ml) per hour for a further 71 hours, with the
aim of maintaining a target serum concentration
of 260 Ојmol of unbound study drug per liter. The
infusion rate was guided by the estimated rate of
creatinine clearance, based on the serum creatinine concentration.6 The infusion rate was adjusted to 32 ml per hour for clearance rates of 30 to
50 ml per minute, 44 ml per hour for clearance
rates of 51 to 80 ml per minute, and 64 ml per
hour for clearance rates greater than 80 ml per
minute. For patients with clearance rates of less
than 30 ml per minute, treatment was withdrawn.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
Clinical Assessment
Patients were assessed at various times throughout the study, including at the time of enrollment,
24 and 72 hours after the start of infusion of the
study drug, and on days 7, 30, and 90. Initial assessments included a physical examination, neuroВ­
imaging, and an NIHSS assessment to determine
the severity of the stroke. The examiners were
trained and certified in the use of the NIHSS
examination (scores range from 0 to 42, with a
higher score indicating greater stroke severity).5
Assessments after completion of infusion of
the study drug were primarily functional or neuroВ­
logic, including the score on the modified Rankin
scale7 (assessed on days 7, 30, and 90), the NIHSS
(days 7 and 90), and the Barthel index8 (days 7,
30, and 90). The modified Rankin scale is a
global disability scale with a range from 0, indicating no residual symptoms, to 5, for patients
who are bedridden and require constant care.
In this study, patients who died were assigned a
score of 5 on the modified Rankin scale. The investigators were trained, tested, and certified in
the use of the modified Rankin scale according
to a method involving the use of a DVD developed
specifically for this trial.9 The NIHSS quantifies
the level of neurologic deficit, with higher scores
at day 90 predictive of dependence. The Barthel
index measures activities of daily living; scores
range from 0 to 100, with 0 indicating complete
dependence and 100 indicating independence.
m e dic i n e
Figure 1 (facing page). Enrollment, Group Assignment,
and Follow-up.
After randomization, 24 patients received the wrong
treatment kit. Fourteen patients assigned to placebo
received at least some NXY-059; these patients were
considered in the efficacy analysis to have been treated
with placebo and were considered in the safety analysis to have been treated with NXY-059. Ten patients
В­assigned to NXY-059 were treated only with placebo;
these patients were considered in the efficacy analysis
to have been treated with NXY-059 and were considered in the safety analysis to have been treated with
placebo. The efficacy analysis included all patients
who underwent randomization and received any investigational treatment and for whom stratification data
and any post-treatment assessment data were available. The last observation was carried forward for survivors who were unable to undergo assessment with
the modified Rankin scale at 90 days. Patients who
died were assigned a score of 5 on the scale (worst
outcome) and were considered to have a valid assessment; 3126 patients (97.8%) had a valid 90-day assessment, and 69 (2.2%) had an earlier observation carried
forward. The per protocol population (2858 patients,
88.2% of all treated patients) consisted of patients for
whom there were no major protocol violations and
who received at least 75% of the target dose for their
rate of creatinine clearance. Patients in whom the study
drug was stopped because of progressive stroke or
death were retained in the per protocol population.
rhage was defined as the presence of any blood
on imaging after alteplase treatment in the absence of such worsening of the NIHSS score.3
Patients meeting the criterion for progressive
stroke (an increase of at least four points on the
Safety Assessments
NIHSS score) or new stroke in the first week also
Vital signs were recorded at enrollment and at underwent repeated imaging studies. Follow-up
specified times throughout the infusion and dur- scans were read centrally by readers who were
ing the follow-up period. Routine laboratory data unaware of treatment assignments.
were obtained at the time of enrollment and at 24
hours, 72 hours, and 7 days. Electrocardiograms Statistical Analysis
were obtained at enrollment and at 24 hours, 72 The analyses were prespecified. For the efficacy
hours, and (if the findings at 72 hours were analysis, we included all patients with baseline
В­abnormal) 7 days. Laboratory and electrocardio- data who had commenced any treatment and had
graphic results were analyzed centrally. To deter- undergone any post-treatment assessment. The
mine whether NXY-059 had any effect on hemor- patients were analyzed according to the treatment
rhagic transformation (intracerebral hemorrhage) assignment for efficacy outcomes and according
after treatment with alteplase, brain imaging was to the treatment actually received for safety outrepeated after 72 hours in patients receiving con- comes. The primary outcome measure was the
comitant alteplase. On the basis of the results of score on the modified Rankin scale at 90 days or
SAINT I, we defined symptomatic intracerebral the last rating, analyzed across the whole distrihemorrhage as a worsening in the NIHSS score bution of scores with the use of the Cochran–
by four or more points within 36 hours plus the Mantel–Haenszel test, with adjustment for stratipresence of any blood on imaging after alteplase fication variables (NIHSS score, side of the infarct,
treatment; asymptomatic intracerebral hemor- and use or nonuse of alteplase) and with the use
n engl j med 357;6  www.nejm.org  august 9, 2007
NXY-059 for Acute Ischemic Stroke
3306 Patients underwent
1660 Were assigned to
1646 Were assigned to
65 Did not start treatment and
were excluded from analysis
1635 Started treatment
1606 Started treatment
14 Who were assigned to placebo
10 Who were assigned to NXY-059
crossed over to the alternative
treatment group
1631 Received only
placebo and were
included in the
safety population
1610 Received some
NXY-059 and were
included in the
safety population
45 Withdrew or were lost to follow-up
before first assessment
1 Was not included in efficacy analysis
because of incomplete stratification
1313 Underwent modified Rankin
scale assessment at day 90
27 Withdrew or were lost to
follow-up by day 90, with last
observation carried forward
267 Died, assigned score of 5 on
the modified Rankin scale
1279 Underwent modified Rankin
scale assessment at day 90
42 Withdrew or were lost to
follow-up by day 90, with last
observation carried forward
267 Died, assigned score of 5 on
the modified Rankin scale
1607 Were included in the efficacy
1588 Were included in the efficacy
337 Received <75% infusion or had
protocol exclusion
1441 Were included in the
per protocol analysis
1417 Were included in the
per protocol analysis
of modified ridit scores10 (i.e., the midrank score linked to the van Elteren test that allowed us to
Г· [the number of observations + 1]), to account analyze the modified Rankin scale as an ordinal
AUTHOR: Shuaib (Lees)
for ordered categories. The generalized
Cochran– rather thanRETAKE
a binary1st
outcome, without assuming
REG F FIGURE: 1 of 3
Mantel–Haenszel test is a nonparametric method proportional odds.113rdDichotomized and trichotoCASE
n engl j med 357;6  www.nejm.org  august 9, 2007
Figure has been redrawn and type has been reset.
Please check carefully.
JOB: 35606
ISSUE: 08-09-07
n e w e ng l a n d j o u r na l
mized scores on the modified Rankin scale were
also analyzed.
The sample size of 3200 patients was chosen
to provide at least 80% power to detect a common
odds ratio of 1.2 (across all cutoff points of the
modified Rankin scale), the result that was seen
in the SAINT I study.3 Additional measurements
of the modified Rankin scale were completed at
7 and 30 days. Neurologic function was assessed
on the basis of the total NIHSS score, the first
secondary end point, at 90 days or at the time of
the last rating, and it was also analyzed by means
of the Cochran–Mantel–Haenszel test with a mod­
ified ridit score and adjustment for the baseline
NIHSS score, side of the infarct, and use or nonuse of alteplase.10,11
We ordered the analysis of efficacy outcomes
hierarchically, avoiding the need for further adjustment for multiplicity, because formal statistical testing was performed only if the preceding
Table 1. Baseline Characteristics of the Patients.*
Mean age — yr
Male sex — no. (%)
(N = 1631)
(N = 1610)
870 (53.3)
898 (55.8)
Mean time from onset of stroke to treatment — hr:min
Mean NIHSS score
Treatment or intention to treat with
alteplase — no. (%)
716 (43.9)
710 (44.1)
Mean time from onset of stroke to
administration of alteplase —
Mean age — yr
Mean NIHSS score at trial entry
1246 (76.4)
1243 (77.2)
356 (21.8)
357 (22.2)
Transient ischemic attack
161 (9.9)
135 (8.4)
Ischemic heart disease
536 (32.9)
531 (33.0)
Atrial fibrillation
499 (30.6)
438 (27.2)
Diabetes mellitus
406 (24.9)
386 (24.0)
Use of antiplatelet drugs
566 (34.7)
590 (36.6)
Cardioembolic stroke†775 (47.6)
738 (46.1)
History — no. (%)
*Scores on the NIHSS (National Institutes of Health Stroke Scale) range from
0, indicating normal functioning, to 42, indicating most severe impairment.
†Percentages are based on 1627 patients in the placebo group and 1600 in the
NXY-059 group.
m e dic i n e
end point was significant. The primary end point,
the score on the modified Rankin scale, was the
first outcome, and a positive trial would have
been declared if this end point had been significant, irrespective of end points lower in the hierarchy.
Safety end points included death, serious and
nonserious adverse events, laboratory measurements, vital signs, and neuroimaging data. The
incidence of intracranial hemorrhage in patients
who were treated concomitantly with alteplase
was prospectively analyzed with the chi-square
test. Descriptive statistics (number and frequency)
were used to summarize all types of intracranial
hemorrhage for each treatment group.
R e sult s
Baseline characteristics
Of the 3306 patients randomly assigned to treatment, 3241 were treated with NXY-059 or placebo
(Fig. 1). In 65 patients, the infusion was not started, most commonly because of delayed recognition that the patient did not meet the eligibility
criteria (20 patients assigned to NXY-059 and 19
assigned to placebo). The investigators became
aware of the study-drug assignment in five cases.
End-of-study assessment of drug infusion was
completed in 3074 patients (1542 [95.8%] of those
receiving NXY-059 and 1532 [93.9%] of those receiving placebo), and 2593 patients (80.0%) completed the 90-day follow-up assessment. Among
the 3241 patients treated, modified-Rankin-scale
outcome data were unavailable for 46 (23 assigned
to NXY-059 and 23 assigned to placebo).
The baseline characteristics of the patients are
shown in Table 1. The mean time from the onset
of symptoms to the start of infusion of the study
drug was 3 hours 48 minutes. A total of 1426
patients (44.0%) received treatment with alteplase.
According to extrapolation from measured concentrations, in 96.6% of NXY-059–treated patients, a target plasma unbound concentration
of 150 Ојmol per liter was reached, which is well
above the levels that have been shown to be neuroprotective in animal models of stroke.
Clinical Outcomes
The distribution of scores on the modified Rankin
scale at 90 days was similar in the two groups
(P = 0.33 by the Cochran–Mantel–Haenszel test;
number of patients, 3195; odds ratio for a favor-
n engl j med 357;6  www.nejm.org  august 9, 2007
NXY-059 for Acute Ischemic Stroke
able outcome, 0.94; 95% confidence interval [CI],
0.83 to 1.06) (Fig. 2). We evaluated all possible
dichotomizations of the scale as well as the trichotomization into categories 0 to 1, 2 to 3, and
4 to 5, as has been tested elsewhere.3 There were
no significant differences between the two groups
in any of the cutoff points: the P value for the
trichotomization was 0.23 (odds ratio, 0.92; 95%
CI, 0.80 to 1.06). Analysis of the group of patients who received at least 75% of the infusion
and who were in full compliance with the protocol also did not show a significant difference between the two groups (P = 0.25 by the Cochran–
Mantel–Haenszel test; odds ratio, 0.93; 95% CI,
0.81 to 1.06). There was no improvement in disability among survivors in the NXY-059 group at
7, 30, or 90 days, as measured by the Cochran–
Mantel–Haenszel test (day 90, P = 0.53; odds ratio,
0.96; 95% CI, 0.83 to 1.10).
NXY-059 had no effect on any of the prespecified secondary end points. The total NIHSS
score at the last rating in the NXY-059 group was
not significantly different from that in the placebo group (P = 0.73 by the Cochran–Mantel–Haen­
szel test; Mann–Whitney U statistic, 0.50; 95% CI,
0.48 to 0.52).
In addition, there was no significant difference between the groups in the percentage of
patients who had a complete recovery, as measured by an NIHSS score of 0 versus 1 to 42
(16.3% in the NXY-059 group and 16.6% in the
placebo group), or a nearly complete recovery, as
measured by an NIHSS score of 0 to 1 versus 2 to
42 (27.6% in both groups). Finally, there was no
significant difference between the groups in the
percentage of patients with a Barthel index score
of 95 or more (40.9% in the NXY-059 group and
42.3% in the placebo group).
Placebo 10.3
5 or Death
Proportion of Patients in the
Efficacy Population (%)
Figure 2. Primary Outcome at 90 Days According to
Shuaib (Lees)
the ICM
Score on
the Modified
Rankin Scale. RETAKE
FIGURE: 2 of 3
There were no significant interactions between
the treatment effect of NXY-059 and the time from
the onset of symptoms to treatment (4 hours or
less vs. more than 4 hours), presence of diabetes
or hypertension, severity of stroke, or use of alteВ­
plase (Fig. 3).
Safety Analysis
There were 534 deaths, 267 in each treatment
group (16.5% of patients). The mean (В±SE) time
from randomization to death was 24.3В±1.4 days
in the NXY-059 group and 21.9В±1.4 days in the
placebo group (P = 0.98 by the log-rank test). Few
deaths were reported during the drug infusion:
18 in the NXY-059 group (1.1%) and 18 in the placebo group (1.1%). The most common causes of
death were neurologic damage from the initial
stroke (in 83 patients in the NXY-059 group
[31.1%] and 107 in the placebo group [40.1%])
and bronchopneumonia (41 in the NXY-059 group
[15.4%] and 34 in the placebo group [12.7%]).
Adverse events were reported in 84.2% of patients
in the NXY-059 group and 84.5% of patients in
Odds Ratio (95% CI)
≤4 hr
>4 hr
0.91 (0.78–1.06)
0.98 (0.79–1.22)
Use of alteplase
No use of alteplase
0.90 (0.74–1.08)
0.97 (0.82–1.15)
NIHSS 10–14
NIHSS 15–19
0.85 (0.69–1.05)
0.92 (0.74–1.13)
1.14 (0.83–1.52)
1.00 (0.71–1.41)
Placebo Better
NXY-059 Better
Figure 3. Treatment Interactions with Important Covariates.
There was no significant interaction with any stratification variable or preRETAKE
AUTHOR: Shuaib (Lees)
specified covariate.
primary outcome measure was the score on the
3 of 3or the last rating, as analyzed2nd
modified Rankin
at 90 days
with the
test with a modified ridit score,
Revised adjusted for
baseline stratification
variables. The Cochran–Mantel–Haenszel
test conSIZE
H/T scale,
H/Twith deaths
22p3 assigned a
siders the full Enon
range of the modified Rankin
modified Rankin score of 5. Odds ratios were estimated by logistic regresAUTHOR, PLEASE NOTE:
sion, adjusted forFigure
stratification variables, and are shown with their
has been redrawn and type has been reset.
estimated 95% confidence intervals.
Please check
carefully. between covariates and
treatment effect were analyzed with the Cochran–Mantel–Haenszel test:
P = 0.82 for JOB:
time 35706
to treatment, P = 0.61 for use of alteplase,
and P = 0.70
ISSUE: 08-09-07
for severity of stroke according to the National Institutes of Health Stroke
Scale (NIHSS) category. Since none of these were significant, the individual
odds ratios and their 95% confidence intervals are provided only for illustration. The size of the squares corresponds to the size of each subgroup.
n engl j med 357;6  www.nejm.org  august 9, 2007
Figure has been redrawn and type has been reset.
Please check carefully.
JOB: 35706
ISSUE: 08-09-07
n e w e ng l a n d j o u r na l
the placebo group. Serious adverse events were
reported in 39.6% of patients in the NXY-059
group and 40.2% of patients in the placebo group
(Table 2). The percentage of patients who discontinued medication was 4.5% in the NXY-059
group and 6.3% in the placebo group. The only
adverse event that occurred significantly more
frequently in the NXY-059 group than in the placebo group was hypokalemia. Hypokalemia occurred in 12.1% of NXY-059–treated and 9.2% of
placebo-treated patients at any stage and in 9.1%
and 6.3% of patients, respectively, during the infusion but was not associated with any cardiac or
Table 2. Safety Outcomes Recorded during 90 Days of Follow-up.*
(N = 1631)
(N = 1610)
no. (%)
Serious adverse events
Stroke in evolution
157 (9.6)
150 (9.3)
Ischemic stroke
62 (3.8)
55 (3.4)
37 (2.3)
39 (2.4)
Hemorrhagic transformation stroke
28 (1.7)
30 (1.9)
Aspiration pneumonia
20 (1.2)
26 (1.6)
Brain edema
28 (1.7)
24 (1.5)
Cerebral hemorrhage
25 (1.5)
23 (1.4)
9 (0.6)
22 (1.4)
25 (1.5)
18 (1.1)
7 (0.4)
16 (1.0)
268 (16.4)
289 (17.9)
227 (13.9)
256 (15.9)
150 (9.2)
195 (12.1)
151 (9.3)
170 (10.6)
Urinary tract infection
161 (9.9)
164 (10.2)
Stroke in evolution
164 (10.1)
152 (9.4)
Atrial fibrillation
103 (6.3)
111 (6.9)
113 (6.9)
111 (6.9)
90 (5.5)
103 (6.4)
79 (4.8)
101 (6.3)
Myocardial infarction
Atrial fibrillation
Adverse events
*The events included are the 10 most common serious adverse events and the
10 most common adverse events in the NXY-059 group (adverse events include those that were also coded as serious). In the analysis of safety, patients were considered as treated and not as randomly assigned to treatment.
Because of the large number of potential adverse events that could be examined (>100), formal statistical testing was not conducted; as a guide, a difference of 1.5% or more between groups will occur with a probability of approximately 0.05 before adjustment for multiplicity.
m e dic i n e
other complications. Hypokalemia resolved within 7 days.
In our prespecified analysis of patients treated with alteplase, there were no significant differences between the NXY-059 group and the
placebo group in the percentage of patients with
symptomatic cerebral hemorrhage (4.6% and 5.3%,
respectively; P = 0.57), asymptomatic cerebral hem­
orrhage (17.9% and 16.1%, post hoc analysis), or
overall cerebral hemorrhage (22.5% vs. 21.4%,
P = 0.60).
Dis cus sion
This second trial of NXY-059 included a generous
sample of patients with acute ischemic stroke.
Although the trial was adequately powered to detect a clinically useful difference in the primary
end point, the modified Rankin score, we failed
to confirm the efficacy of NXY-059.
SAINT I showed promising results, with reducВ­
tion of global disability.3 The effect on disability
was moderate but was thought to be consistent
with a neuroprotective action. Although there
were no significant effects on the prespecified
secondary end points, post hoc analyses showed
supportive trends for other end points.4 Moreover, patients treated with alteplase and NXY-059
had fewer cerebral hemorrhages.3 We had refined
the SAINT II protocol after the successful SAINT I
trial; changes included an increase in the sample
size from 1700 to 3200 patients, a revised approach to analysis of the NIHSS score, and a proВ­
spective analysis of intracerebral hemorrhage.
However, none of the benefits demonstrated in
SAINT I were confirmed in the SAINT II study.
With the exception of mild, asymptomatic hypokalemia, the numbers and types of adverse events
and serious adverse events, including neurologic
events, were similar in the two groups.
Faced with conflicting results from the two
pivotal trials of NXY-059, we must examine possible explanations. These trials followed nearly
identical protocols, with minimal differences in
statistical analysis. With the exception of a higher
frequency of alteplase use in SAINT II (44% vs.
29%), the baseline characteristics and demographВ­
ic features of the study populations were similar.
In both trials, the average time to treatment was
less than 4 hours and target plasma drug concentrations were achieved and maintained in more
than 95% of patients.
n engl j med 357;6  www.nejm.org  august 9, 2007
NXY-059 for Acute Ischemic Stroke
We first considered whether the conflicting
results of the two trials might be related to the
higher rate of alteplase use in SAINT II. However, we found no evidence of an interaction between alteplase use and the effect of NXY-059 in
either trial. Although we cannot completely rule
out a ceiling effect in SAINT II resulting from an
already maximal improvement due to alteplase
use, monotherapy with NXY-059 was also ineffective. From these data, we cannot draw any conclusion about the arguments for or against includВ­
ing patients treated with alteplase in future trials
of putative neuroprotectants.
We did not use perfusion imaging as a selection criterion for this trial. It is conceivable that
any treatment effect might have been diluted by
the inclusion of patients with established infarction or without a perfusion deficit at the time of
randomization. However, because of the large
sample size and the expectedly high proportion
of patients in whom penumbra would still have
been present within 4 hours after the onset of
stroke, we do not think that the use of imaging
would have transformed the outcome of this
trial. In addition, we did not collect data on the
subtype of stroke, but we find no evidence from
our data on the severity of stroke that the treatment effect was likely to have been influenced by
the subtype of stroke.
We do not consider that our statistical approach was responsible for the false positive result in SAINT I. In both trials, the same method
was used to analyze the results for the primary
end point — namely, comparison of the distribution of disability scores rather than simple dichotomy. Our analysis takes into account any
potential harm by requiring that any deleterious
effect at one end of the scale be more than balanced by a benefit at other levels if significance
is to be achieved. We consider that the disparity
between the two studies occurred by chance, possibly because of the inclusion of patients with a
latent poorer prognosis in the placebo group in
SAINT I. However, there was no notable imbalance between the treatment groups with respect
to any single measured variable or combination
of variables. The likelihood that our prior trial
was positive simply because of the play of chance
underscores the need for replication by a second
pivotal trial.
The observed reduction in the rate of hemorrhagic transformation associated with NXY-059
after thrombolysis in SAINT I was biologically
plausible, given the role of free-radical–mediated
disruption of the blood–brain barrier. Vasculoprotection with NXY-059 during thrombolysis
was consistent with the putative mechanism of
the drug.12,13 The absence of this finding in
SAINT II provides further confirmation of the
discrepancy between the trials and suggests that
this was also a chance finding or the result of
differences in the study populations.
Stroke Therapy Academic Industry Roundtable
(STAIR) criteria were designed to help in the
development of drug therapy for acute stroke.14‑16
Even though it met all these criteria, the SAINT II
study showed no significant effect in patients
with acute ischemic stroke. It is possible that the
animal models of acute focal infarction are not
relevant to the patient population; they certainly
are insufficient to guarantee a positive clinicaltrial result. The molecular heterogeneity and
pharmacodynamics in elderly patients may differ from those in the healthy and relatively
young animals used for preclinical studies. The
clear failure to demonstrate a benefit of NXY-059
in the SAINT II study suggests that we need to
reevaluate the strategies that have been used in
the development of drugs for neuroprotection.
More emphasis on phase 2 studies in patients,
involving a surrogate outcome measure that is
more sensitive to treatment effects even if it is
not acceptable for drug-registration purposes,
may be necessary to supplement the preclinical
In summary, SAINT II provides clear evidence
that the promise offered by the preclinical data
and the positive findings of SAINT I has been a
false dawn. The most reasonable interpretation
is that NXY-059 is safe but ineffective for the
treatment of acute ischemic stroke.
Drs. Lees, Grotta, and Davis report receiving fees and expenses from AstraZeneca for steering-committee work and
lectures. Dr. Davis reports receiving consulting or speaking
fees from Novo Nordisk, Pfizer, Sanofi-Aventis, Bristol-Myers
Squibb, Boehringer Ingelheim, Paion, and Servier; Dr. Davalos,
consulting or speaking fees AstraZeneca, Boehringer Ingelheim,
Pfizer, Merck Sharpe & Dohme, Sanofi-Synthelabo, Bristol-Myers
Squibb, Bayer, Paion, Forest Pharmaceuticals, Daiichi Asubio,
Eli Lilly, Fujisawa, Novo Nordisk, and Ferrer International; Dr.
Diener, consulting or speaking fees from AstraZeneca, GlaxoSmithKline, Boehringer Ingelheim, BASF, Abbott, Novartis,
Parke-Davis, Merck Sharpe & Dohme, Servier, Sanofi-Synthelabo, Bayer, Fresenius, and Janssen-Cilag; Dr. Lyden, consulting
or speaking fees from AstraZeneca, Bayer, Mitsubishi, Pfizer,
Lilly, and Merck and research contracts with AstraZeneca and
Bayer; Dr. Grotta, research support from AstraZeneca, NovoNordisk, and Boehringer Ingleheim; and Dr. Shuaib, consulting or
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
speaking fees from AstraZeneca, GlaxoSmithKline, Boehringer
Ingelheim, Pfizer, Roche, Merck, and Sanofi-Synthelabo. Drs.
Ashwood, WasiewВ­ski, and Emeribe report being employees of
AstraZeneca and holding stock in AstraZeneca. The SAINT trials
were sponsored by AstraZeneca. NXY-059 is subject to a partnership agreement between AstraZeneca and Renovis; Renovis had
no influence on the conduct, analysis, or interpretation of the
study but was given an opportunity to review and comment on
m e dic i n e
the manuscript. No other potential conflict of interest relevant
to this article was reported.
We thank the patients and their relatives; the clinical and research teams of the stroke units; the AstraZeneca coordinating
and monitoring staff who participated in this trial; Dr. Tomas
Odergren for his substantial contributions to the development
of NXY-059 throughout AstraZeneca’s program; and the scientific, data-management, and statistical teams.
The following investigators participated in the SAINT II trial: Steering committee — K.R. Lees, Glasgow, United Kingdom (chair); A. Shuaib
(principal investigator), Edmonton, AB, Canada; T. Ashwood, Södertälje, Sweden (sponsor representative); A. Davalos, Barcelona;
S. Davis, Melbourne, Australia; H.C. Diener, Essen, Germany; J. Grotta, Houston; P. Lyden, San Diego, CA; W. Wasiewski, Wilmington,
DE (sponsor representative). Data and safety monitoring board — S. Pocock, London (chair); H. Adams, Iowa City, IA; P. Bath, Nottingham,
United Kingdom; D. Oakes, Rochester, NY; N.G. Wahlgren, Stockholm. Study team leader — C. Baker, Wilmington, DE. Study team physicians — W.W. Wasiewski, Wilmington, DE; L. Rodichok, Wilmington, DE; H.G. Hardemark, Södertälje, Sweden. Study team statisticians
— V. Alderfer, Wilmington, DE; U. Emeribe, Wilmington, DE. Contract research organizations — PPD Development, Covance Central
Laboratory Services, Perceptive Informatics, Parexel International, eResearch Technology, Fisher Clinical Services, TEVA Pharmaceutical
Industries. Medical science director — L. O’Dowd. Clinical centers: Argentina — R. Rey, S. Ameriso, E. Reich, F. Buonanotte, R.R. Escalante,
M.M. Esnaola y Rojas, E. Gatto; Austria — M. Brainin, W. Grisold, F. Aichner, G. Ladurner, F. Gruber; Australia — S. Davis, C. Levi,
C. Bladin, G. Donnan, J. Sturm, D. Schultz, C. Anderson, A. Barber; Belgium — G. Vanhooren, P. De Deyn, V. Thijs; Bulgaria — P. Stamenova, D. Baldaranov, D. Minchev, A. Tunev, V. Platikanov, T. Notcheva; Brazil — A. Massaro, E. Evaristo, A. Gomes Neto, L. Barea,
M.A. Friedrich, S.C. Martina, J. Fernandes; Canada — A. Shuaib, D. Selchen, P. Teal, S. Phillips, C. Voll, D. Howse, J. Teitelbaum, D.
Rivest, H. Desai, R. Arts, D. Stewart, A. Durocher, M. Lamothe, L. Berger, D. Gladstone, P. Bailey, M. Hill; Chile — V. Díaz Tapia, A.
Hoppe, R. Maturana Dasori; China — C. Lu, X. Wang, S. Ding, Y. Wang, L. Cui, Y. Huang, J. Sheng Zeng, E. Xu, Z. Zhang, L. Miao;
Czech Republic — G. Waberzinek, K. Urbánek, J. Bauer, M. Bar, D. Václavík, E. Ehler, H. Lachmann, O. Škoda; France — A. Bonafe, J.
Boulliat, P. Amarenco, F. Ziegler, T. Moulin, D. Sablot, M. Bataillard, J.F. Albucher, F. Rouanet, F. Nicoli, I. Crassard, F. Pico, H. Hosseini; Germany — J. Glahn, D. Schneider, M. Rosenkranz, A. Hetzel, P. Vogel, G. Hamann, P. Ringleb; Greece — K. Vemmos, K. Karageorgio, G. Georgiadis; Hong Kong, China — L. Wong, T. Hong Tsoi; Hungary — S. Horváth, A. Csányi, C. Béla, N. Szegedi; Israel — D.
Yarnitsky, J. Streifler, R. Milo, B. Weller, B. Gross, D. Tanne, T. Lampl, J. Rabey, N. Bornstein, S. Honigman, I. Wirguin, R. Leker;
Korea — J.S. Kim, B.C. Lee; Mexico — C. Brito Cantu, S. Ruiz, M. Hector, J. Rodriguez, L. Rivera, G. Aguayo, F. Leon, B. Vega, G. Osorno; Philippines — A. Aquino, C. Chua, M.C. San Jose, A. Díaz, J. Navarro, A. Baroque; Poland — A. Członkowska, A. Kuczyńska, B.
Gaweł, A. Buczek, M. Jackowski, M. Glonek, P. Nowacki, G. Chabik, Z. Stelmasiak, H. Kwieciski; Portugal — A. Vasco Salgado, L.
Cunha, M. Tuna, G. Goncalves, T. Pinho e Melo, M. Rodigues; Russia — V. Skvortsova, Z. Suslina, A. Lebedeva, Y. Kolomoyetz, V.
Shmirev, A. Skoromets, A. Fedin; Singapore — R. Gan, B. Chan, H.M. Chang; Slovakia — M. Dvorák, M. Brozman, E. Kurča, J. Vyletelka,
J. Herén­yiuvá; South Africa — M. Bassoon, J. Thorne, J.S. Roos, M. Isaacs; Spain — F. Rubio, J.S. Leal, J. Roquer, A. Chamorro, J.
Vivancos, J. Castillo, M. Lozano, C. Molina; Switzerland — H. Mattle, B. Tettenborn, J. Bogousslavsky, H. Hungerbühler, R. Baumgartner, P. Lyrer; Taiwan — H.H. Hu, S.T. Chen; United States — M. Harris, M. Sauter, N. Culligan, W. Truax, D. Koons, R. Hull, W. Holt,
K. Madden, F. Abbott, S. Braheny, C. Chang, B. Cucchiara, C. Graffagnino, J. Gross, J. Grotta, P. Katz, R. Kelley, N. Papamitsakis, J.
Freiberg, J. Schim, J. Harris, V. Salanga, A. Bernstein, M. Brody, C. Gomez, A. Ahuja, F. Chang, D. Chiu, M. Concha, J. Couch, B.
Dandapani, B. Demaerschalk, F. Diamond, B. Dihenia, J. Dissin, R. Jackel, M. Jacoby, K. John, D. Krieger, L. Labiche, L. Larsen, D.
Leifer, E. Wilson, M. Young, R. Reichwein, M. Reiss, R. Ringel, N. Rodgers-Neame, J. Sander, D. Saur, R. Schechter, C. Villar, D. Chen,
I. Altafullah, P. Ash, J. Wulff, C. Imbus, A. Felix, C. Kososky, T. Habiger, C.H. Liu, S. Mallenbaum, K. Ng, R. Brooks, K. Levin, H.
Sachdev, J. Bertoni, R. Stephens, S. Fox, D. Brock, E. Crisostomo, J. Wang, M. Nash, P. Akins, N. Badjatia, D. Wright, D. Lefkowitz, J.
Liss, P. Mazzeo, J. McDowell, M. Tremwel, E. Albakri, J. Belden, T. Bell, P. Blachman, C. Boutwell, F. Campanella, A. Walker, V. Erlich,
J. Gebel, L. Glass, B. Grayum, J. Hanna, D. Heiselman, W. Hickling, J. Hollander, N. Iannuzzi, T. Henmen, A. Majid, M. Moonis, C.
Morgan, G. Newman, R. O’Connor, E. Oribe, T. Wei, M. Schneck, I. Silverman, R. Smith, S. Starkman, M. Tabbaa, D. Tamulonis, M.
Torbey, L. Wechsler, A. Turel, J. Waxler, V. Mangeshkumar, R. Libman, S. Martino, D. Morris, H. Edmonds, S. Cohen, F. Munschauer,
J. Bobenhouse, R. Taylor, W. Honeycutt, C. Brondos, S. Wolf, J. Jacobsen, R. Paschall, D. Hill, L. Brass, D. Dietrich, H. Kirshner, C.
Fanale, W. Justiz, G. Albers, W. Clark, T. Ashizawa, J. Stafford, M. Rubin, G. Ferencz, W. Likosky, J. Frey, C. Perkins, E. Hekler, G.
Howell, M. Alberts; United Kingdom: K.R. Lees, G.A. Ford, K. Muir.
1. O’Collins VE, Macleod MR, Donnan
GA, Horky LL, van der Worp BH, Howells
DW. 1026 Experimental treatments in
acute stroke. Ann Neurol 2006;59:467-77.
2. Green AR, Shuaib A. Therapeutic strategies for the treatment of stroke. Drug
Discov Today 2006;11:681-93.
3. Lees KR, Zivin JA, Ashwood T, et al.
NXY-057 for acute ischemic stroke. N Engl
J Med 2006;354:588-600.
4. Lees KR, Davalos A, Davis SM, et al.
Additional outcomes and subgroup analyses of NXY-059 for acute ischemic stroke in
the SAINT I Trial. Stroke 2006;37:2970-8.
5. Brott T, Adams HP Jr, Olinger CP, et
al. Measurements of acute cerebral infarction: a clinical examination scale. Stroke
6. Cockcroft DW, Gault MH. Prediction
of creatinine clearance from serum creatinine. Nephron 1976;16:31-41.
7. van Swieten JC, Koudstaal PJ, Visser
MC, Schouten HJ, van Gijn J. Interobserver
agreement for the assessment of handicap
in stroke patients. Stroke 1988;19:604-7.
8. Mahoney FI, Barthel DW. Functional
evaluation: the Barthel Index. Md State
Med J 1965;14:61-5.
9. Quinn TJ, Lees KR, Hardemark HG,
Dawson J, Walters MR. Initial experience
of a digital training resource for modified
Rankin scale assessment in clinical trials.
Stroke (in press).
10. Koch GG, Edwards S. Clinical efficacy
trials with ordinal data. In: Peace KK, ed.
Biopharmaceutical statistics for drug development. New York: Marcel Dekker,
11. Stokes ME, Davis CS, Koch GG. Categorical data analysis using the SAS system. 2nd ed. Cary, NC: SAS Institute,
n engl j med 357;6  www.nejm.org  august 9, 2007
NXY-059 for Acute Ischemic Stroke
12. Kuroda S, Tsuchidate R, Smith M-L,
Maples KR, Siesjö BK. Neuroprotective effects of a novel nitrone, NXY-059, after
transient focal cerebral ischemia. J Cereb
Blood Flow Metab 1999;19:778-87.
13. Lapchak PA, Araujo DM, Song D, Wei J,
Purdy R, Zivin JA. Effects of the spin trap
agent disodium-[tert-butylimino)methyl]
benzene-1,3-disulfonate N-Oxide (generic
NXY-059) on intracerebral hemorrhage
in a rabbit large clot embolic stroke model: combination studies with tissue plas-
minogen activator. Stroke 2002;33:166570.
14. Stroke Therapy Academic Industry
Roundtable (STAIR). Recommendations
for standards regarding preclinical neuroprotective and restorative drug development. Stroke 1999;30:2752-8.
15. Stroke Therapy Academic Industry
Roundtable II (STAIR-II). Recommendations for clinical trial evaluation of acute
stroke therapies. Stroke 2001;32:1598-606.
16. Fisher M, Stroke Therapy Academic
Industry Roundtable. RecommendaВ­tions
for advancing development of acute stroke
therapies: Stroke Therapy Academic Industry Roundtable 3. Stroke 2003;34:153946.
17. MR Stroke Collaborative Group, Phan
TG, Donnan GA, Davis SM, Byrnes G.
Proof of principle phase II MRI studies in
stroke: sample size estimates from dichotomous and continuous data. Stroke 2006;
Copyright В© 2007 Massachusetts Medical Society.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
clinical practice
Acute Ischemic Stroke
H. Bart van der Worp, M.D., Ph.D., and Jan van Gijn, F.R.C.P.
This Journal feature begins with a case vignette highlighting a common clinical problem.
Evidence supporting various strategies is then presented, followed by a review of formal guidelines,
when they exist. The article ends with the authors’ clinical recommendations.
A 62-year-old man has sudden weakness of the left arm and leg and slurred speech.
Except for untreated hypertension, his medical history is unremarkable. He is a current smoker with a smoking history of 45 pack-years. On arrival at the emergency
department 1 hour 15 minutes after the onset of symptoms, he reports no headache or
vomiting. His blood pressure is 180/100 mm Hg, and his pulse is 76 beats per minute
and is regular. Neurologic examination shows dysarthria, a left homonymous hemianopia, severe left-sided weakness, and a failure to register light touch on the left side
of the body when both sides are touched simultaneously (left tactile extinction). How
should this patient be evaluated and treated in the short term?
The Cl inic a l Probl e m
From the Department of Neurology, Rudolf
Magnus Institute of Neuroscience, University Medical Center Utrecht, Utrecht,
the Netherlands. Address reprint requests to Dr. van der Worp at the Department of Neurology, Rudolf Magnus
Institute of Neuroscience, University MedВ­
ical Center Utrecht, Heidelberglaan 100,
3584 CX Utrecht, the Netherlands, or at
[email protected]
N Engl J Med 2007;357:572-9.
Copyright В© 2007 Massachusetts Medical Society.
Stroke ranks second after ischemic heart disease as a cause of lost disability-adjusted
life-years in high-income countries and as a cause of death worldwide.1 The incidence of stroke varies among countries and increases exponentially with age. In
Western societies, about 80% of strokes are caused by focal cerebral ischemia due
to arterial occlusion, and the remaining 20% are caused by hemorrhages.2
Ischemic brain injury is thought to result from a cascade of events from energy
depletion to cell death. Intermediate factors include an excess of extracellular excitatory amino acids, free-radical formation, and inflammation.3 Initially after
arterial occlusion, a central core of very low perfusion is surrounded by an area of
dysfunction caused by metabolic and ionic disturbances but in which structural
integrity is preserved (the ischemic penumbra). In the first minutes to hours, therefore, clinical deficits do not necessarily reflect irreversible damage. Depending on
the rate of residual blood flow and the duration of ischemia, the penumbra will
eventually be incorporated into the infarct if reperfusion is not achieved (Fig. 1).3
Thirty-day case fatality rates for ischemic stroke in Western societies generally
range between 10 and 17%.2 The likelihood of a poor outcome after stroke increases with increasing age, with the coexistence of diseases such as ischemic
heart disease and diabetes mellitus, and with increasing size of the infarct. The
likelihood also varies according to the infarct site. Mortality in the first month
after stroke has been reported to range from 2.5% in patients with lacunar infarcts4 to 78% in patients with space-occupying hemispheric infarction.5
S t r ategie s a nd E v idence
Acute stroke is typically characterized by the sudden onset of a focal neurologic
deficit, though some patients have a stepwise or gradual progression of symptoms.
Common deficits include dysphasia, dysarthria, hemianopia, weakness, ataxia,
sensory loss, and neglect. Symptoms and signs are unilateral, and consciousness is
n engl j med 357;6  www.nejm.org  august 9, 2007
clinical pr actice
generally normal or impaired only slightly, except rare cases, such as if infective endocarditis is
in the case of some infarcts in the posterior cir- suspected. In the days thereafter, transthoracic
echocardiography or, preferably, transesophageal
echocardiography may be indicated to rule out
Initial Assessment
In the majority of cases of stroke, making the
diagnosis is straightforward. However, especially Imaging
in patients with unusual features (e.g., gradual on- Cerebral infarction cannot be distinguished with
set, seizure at the onset of symptoms, or impaired certainty from intracerebral hemorrhage on the
consciousness), the differential diagnosis should basis of symptoms and signs alone. In all patients
include migraine, postictal paresis, hypoglyce- with suspected ischemic stroke, computed tomogmia, conversion disorder, subdural hematoma, raphy (CT) or magnetic resonance imaging (MRI)
and brain tumors.
of the brain is therefore required. Noncontrast
Atherosclerosis (leading to thromboembolism CT may suffice (Fig. 2); as compared with MRI,
or local occlusion) and cardioembolism are the it is more widely available, faster, less susceptible
leading causes of brain ischemia. However, un- to motion artifacts, and less expensive. Both CT
usual causes should be considered, especially if and MRI have a high sensitivity for acute intrapatients are younger (e.g., below 50 years of age) cranial hemorrhage, but MRI has a much higher
and have no apparent cardiovascular risk factors. sensitivity than CT for acute ischemic changes,
Some clinical clues that suggest alternative diag- especially in the posterior fossa and in the first
noses are ptosis and miosis contralateral to the
deficit (carotid-artery dissection), fever and a cardiac murmur (infective endocarditis), and headache and an elevated erythrocyte sedimentation
rate in patients older than 50 years of age (giantcell arteritis).
Deficits should be assessed by careful neurologic examination. Several scales have been developed to quantify the severity of the neurologic
Figure 1. Progression over Time (Left to Right) of the Infarct Core (Red),
with Irreversible Damage at the Expense of the Ischemic Penumbra (Green).
deficit, mainly for use in research studies; the
National Institutes of Health Stroke Scale6 is most
often used. An irregular pulse suggests atrial
fibrillation. A very high blood pressure may signal hypertensive encephalopathy and precludes
thrombolysis if sustained at or above 185/110
mm Hg. Carotid bruits lack sufficient sensitivity
AUTHOR Van De Worp
and specificity for a diagnosis of severe carotid
Laboratory testing during the acute phase
ARTIST: mleahy
should include measurement of the glucose level
(since hypoglycemia may also cause focal neuroAUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
logic deficits), a complete blood count, and meaPlease check carefully.
surement of the prothrombin time and partialJOB: 35706
ISSUE: 08-09-07
thromboplastin time, particularly if thrombolysis
Figure 2. CT Scans Obtained 1 Hour 40 Minutes after the Onset of Symptoms
is considered. An electrocardiogram may reveal
Suggestive of Cortical Stroke in the Territory of the Right Middle Cerebral
atrial fibrillation or an acute or previous myoArtery.
cardial infarction as potential causes of thromboВ­
An unenhanced CT scan (Panel A) shows a slight loss of differentiation of
embolism. Because stroke may be complicated by
gray and white matter in the basal ganglia (arrows). A CT angiographic image
myocardial ischemia and arrhythmias, cardiac
shows occlusion of the first segment of the right middle cerebral artery
monitoring is recommended for at least the first
(Panel B, arrow) and atherosclerotic lesions in the carotid bifurcation
AUTHOR carotid
Worp is not shown.
ICMThe external
В­(Panel C, arrow).
24 hours.8 Echocardiography in the first hours
after the onset of stroke is necessary only in
n engl j med 357;6  www.nejm.org 
ARTIST: mleahy
9, 2007
Figure has been redrawn and type has been reset.
Please check carefully.
n e w e ng l a n d j o u r na l
hours after an ischeВ­mic stroke.9 Cytotoxic edema
is detectable within minutes after the onset of
ischemia, with a reduced apparent diffusion coefficient on diffusion-weighted imaging (Fig. 3).10
However, it remains unclear whether early visualization of ischemia has important implications
for management.
For patients in whom acute invasive treatment
strategies (such as intraarterial thrombolysis or
mechanical clot retrieval) are considered, urgent
CT or magnetic resonance angiography is useful
to identify the site of arterial occlusion (Fig. 2).
Either method can provide complete visualization
from the aortic arch to the circle of Willis and
beyond.10 Carotid duplex ultrasonography and
transcranial Doppler ultrasonography have also
been used to detect the site of occlusion.10
intravenous Thrombolysis
The National Institute of Neurological Disorders
and Stroke Recombinant Tissue Plasminogen Activator (NINDS rt-PA) Stroke Study, a multicenter,
randomized trial, has demonstrated the efficacy
of treatment with intravenous rt-PA (alteplase)
started within 3 hours after the onset of symptoms.11 Among patients treated with rt-PA (0.9 mg
per kilogram of body weight, with 10% of the
dose administered as a bolus and the rest infused
over 1 hour and a maximum total dose of 90 mg),
31 to 50% had a favorable neurologic or functional outcome at 3 months (depending on the scale
used), as compared with 20 to 38% of patients
given placebo; mortality rates were similar in the
two groups. Symptomatic intracranial hemorrhage
Figure 3. MRI Scans Obtained 2 Days after the Onset of Ischemic Stroke
in the Territory of the Right Middle Cerebral Artery.
AUTHOR Van De Worp
A hyperintense ICM
lesion in
the temporal and frontal lobes and in the basal gan2nd
glia is shown on fluid-attenuated inversion recovery (Panel A) and
weighted imaging (Panel
B), corresponding to a reduced
apparent diffusion
coefficient (Panel C). Similar changes may
be observed
on diffusion-weighted
ARTIST: mleahy
imaging in the FILL
first hours
after the onset
of symptoms.
Figure has been redrawn and type has been reset.
Please check carefully.
m e dic i n e
occurred in 6.4% of patients treated with intravenous rt-PA and in 0.6% of controls. Four other
trials of intravenous rt-PA therapy given within
6 hours after the onset of symptoms (with few
patients treated within 3 hours) failed to find a
benefit of thrombolysis separately, but if analyzed in combination, they provided support for
a benefit of treatment administered within the first
3 hours after stroke.12,13 Even within the 3-hour
time frame, the benefit of rt-PA is greater the
sooner treatment is started.13
The risk of symptomatic intracranial hemorrhage after thrombolysis is higher in patients with
more severe strokes and with increased age.14
However, a post hoc subgroup analysis of the
NINDS rt-PA Stroke Study found no significant
differences in the benefit from rt-PA therapy
across these and other subgroups,15 but the numbers of patients in each subgroup were small.
Similar concerns have been raised about the efficacy and safety of rt-PA in patients with early
ischemic changes on CT. Other post hoc analyses of data from the NINDS rt-PA Stroke Study
showed that in the first 3 hours after the onset of
symptoms, the appearance of ischemic changes
on CT was not an independent predictor of an
increased risk of symptomatic intracranial hemorВ­
rhage or other adverse outcomes after treatment
with rt-PA.16 Several observational studies have
suggested that intravenous thrombolysis with
rt-PA can be used in the community setting with
efficacy and safety similar to that found in the
randomized trials.17,18
Other Treatments
In two large randomized trials, the use of aspirin
(160 or 300 mg per day), initiated within 48 hours
after the onset of stroke and continued for 2 weeks
or until discharge, led to reduced rates of death
or dependency at discharge or at 6 months,19,20
probably by means of reducing the risk of recurrent ischemic stroke. In both trials, the routine
use of aspirin was recommended as secondary
prevention after the first few weeks. Although
the benefit was small (77 patients would need to
be treated to prevent a poor outcome in 1 patient),
aspirin is inexpensive, has a good safety profile,
and appears to be effective across the range of
patients with ischemic stroke.21 Because the effect
of aspirin in combination with rt-PA is uncertain,
it seems wise to withhold aspirin for 24 hours in
n engl j med 357;6  www.nejm.org  august 9, 2007
clinical pr actice
patients treated with the use of intravenous thrombolysis. The use of dipyridamole or clopidogrel in
the acute phase of ischemic stroke has not been
tested in randomized trials.
A meta-analysis of six randomized trials involving
21,966 patients found no evidence that the use
of anticoagulants (unfractionated heparin, lowmolecular-weight heparins, heparinoids, thrombin
inhibitors, or oral anticoagulants) in the acute
phase of stroke improves functional outcomes.22
According to this analysis, nine fewer cases of
recurrent ischemic stroke would be expected per
1000 patients treated, but so would nine more
cases of symptomatic intracranial hemorrhage.22
A meta-analysis of seven trials similarly failed to
show improvement in functional outcome with
the use of anticoagulant therapy in patients with
acute cardioembolic stroke.23
ment in 93 patients 60 years of age or younger
with space-occupying infarction in the territory
of the middle cerebral artery, surgical treatment
in the first 48 hours after the onset of stroke reduced both the case fatality rate (22%, vs. 71%
in the medical-management group) and the rate
of moderately severe or severe disability or death
(57% vs. 79%).29 Surgery appeared to be less beneficial for patients with aphasia (vs. those without aphasia), patients older than 50 years of age
(vs. those 50 years of age or younger), and patients in whom surgery was performed on the
second day after the onset of stroke (vs. the first
day after onset); however, the numbers of patients
in these subgroups were small.
Data from randomized and other trials indicate that patients who receive care in a stroke unit
are more likely to survive, regain independence,
and return home than are those who do not receive such organized care.30
Prevention and Management
of Complications
Strategies to Reduce Risk of Recurrent
Stroke or Other Cardiovascular Events
Nutrition is often compromised in patients admitted to the hospital with stroke. However, in randomized trials, neither the routine use of oral
nutritional supplements24 nor early tube feeding25
to prevent or treat undernutrition in hospitalized
patients with stroke resulted in improved longterm functional outcome.
Patients with acute stroke are at increased risk
for deep venous thrombosis and pulmonary embolism, and the risk increases with increasing age
and stroke severity.26 Although the use of anticoagulants does not improve overall functional
outcomes, the use of subcutaneously administered
low-dose unfractionated heparin or low-molecularweight heparin has been recommended in patients
at high risk for deep venous thrombosis, such as
patients who are immobile (e.g., due to paralysis
of a leg).8,27
In patients with large supratentorial infarcts,
space-occupying brain edema may lead to transtentorial or uncal herniation, usually between the
second and fifth days after the onset of stroke.5
Case series of such patients in intensive care units
have reported early case fatality rates of up to
78%.5 No medical therapy has proved effective.28
In a pooled analysis of three randomized trials
comparing surgical treatment (hemicraniectomy
and duraplasty, the insertion of a dural patch to
enlarge the intradural space) with medical treat-
In patients presenting with stroke, attention to
secondary prevention of stroke and other cardiovascular complications is routinely warranted. Although space limitations preclude a detailed discussion of recommended strategies, they include
the use of low-dose aspirin and dipyridamole in
patients with ischemic stroke of arterial origin31;
oral anticoagulation in patients with cardiac embolism; treatment of hypertension; statin therapy
for the lowering of lipid levels; glucose control in
patients with diabetes; smoking cessation; and
carotid endarterectomy in patients with substantial ipsilateral carotid stenosis. These issues have
been discussed in detail elsewhere.32,33
Anticoagulant Therapy
A r e a s of Uncer ta in t y
Even in high-income countries such as the United
States, only a small minority of patients with acute
ischemic stroke receive intravenous rt-PA.34 Its use
is currently restricted to a 3-hour time window
after the onset of symptoms, on the basis of results of the NINDS rt-PA Stroke Study,11 but a
pooled analysis of six randomized trials has suggested a potential benefit within up to 6 hours
after the onset of stroke.13 Trials assessing treatment in this extended time frame among broad
populations of patients with ischemic stroke are
under way.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
Preliminary data have suggested that the identification of patients who would benefit from
thrombolysis beyond a 3-hour interval might be
improved by quantification of the ischemic penumbra with the use of diffusion–perfusion MRI
or perfusion CT techniques (Fig. 4).35-37 This suggestion requires further study.
Although the intent of intravenous thrombolysis is to recanalize occluded arteries, none of the
pivotal clinical trials tested whether recanalization actually occurred. Other studies have shown
that complete recanalization of an occluded middle cerebral artery 2 hours after the start of
thrombolysis was achieved in only up to one third
of patients.38,39 In one controlled trial, continuous
2-MHz transcranial Doppler ultrasonography applied for 2 hours augmented the rate of rt-PA–
induced arterial recanalization.38 Limited data
suggest that the addition of intravenous galactose–based microbubbles to this treatment strategy may further increase rates of recanalization.39
Because it is still uncertain whether additional
measures to improve perfusion also improve
functional outcome, these techniques cannot be
recommended for use outside clinical trials.
As compared with intravenous thrombolysis,
intraarterial thrombolysis may increase the likelihood of recanalization, but the two strategies
have not been directly compared in a sufficiently
large randomized trial. In a small randomized
trial, the administration of both intraarterial recombinant prourokinase and intravenous heparin, as compared with intravenous heparin alone,
within 6 hours after the onset of stroke resulted
in a higher rate of recanalization of the middle
cerebral artery (66% vs. 18%) and a higher rate
Figure has been redrawn and type has been reset.
Please check carefully.
m e dic i n e
of a favorable functional outcome (no disability
to slight disability) at 3 months (40% vs. 25%,
P = 0.04).40 However, the procedures required to
deliver the thrombolytic agent to the site of vascular occlusion involve more time than does intraВ­
venous therapy. Thrombolytic “bridging therapy,”
in which intravenous thrombolysis is followed by
intraarterial thrombolysis,41 could permit more
rapid treatment and improved rates of recanalization but is resource intensive, limiting widespread application. Mechanical thrombectomy in
patients with acute intracranial occlusion of the
intracranial carotid artery has resulted in a high
rate of recanalization in case series,42 but controlled trials are lacking.
Other Treatments
High blood pressure,43 a high serum glucose level,44 and a high body temperature45 in the first
hours to days after ischemic stroke have all been
associated with poor long-term outcomes. The effects of the early lowering of blood pressure and
maintenance of normothermia and normoglycemia are currently being tested in large randomized trials.43,46,47
Data from randomized trials are needed to
guide the management of blood pressure in the
context of acute stroke. Given concerns about adВ­
verse effects of the short-term lowering of blood
pressure on cerebral perfusion, current guidelines
based on consensus opinion recommend withholding antihypertensive therapy during the acute
phase of stroke unless the diastolic blood pressure exceeds 120 mm Hg or the systolic blood
pressure exceeds 220 mm Hg in patients who are
not candidates for rt-PA.8 Blood-pressure monitoring is recommended before, during, and after
rt-PA therapy, and intravenous antihypertensive
therapy is recommended to maintain the systolic
blood pressure below 180 mm Hg and the diastolic blood pressure below 105 mm Hg.
Figure 4. Perfusion CT Scans Obtained 1 Hour 45 Minutes after the Onset
of Ischemia in the Territory of the Right Middle Cerebral Artery.
De mean
Worp transit time
ICM prolongation
A large area shows
of the
(in seconds)
4 a_c a reduction in cerebral blood
(Panel A), and aREG
area shows
(in milliliters perCASE
100 g)TITLE
(Panel B). These two maps suggest
4-C shown in green and
and a small infarct
core (Panel C, with the
the suggested infarct
in red).
Hundreds of neuroprotective strategies have been
shown to improve outcome in animal models of
focal cerebral ischemia,48 but thus far only rt-PA
and aspirin have been shown to be clearly efficacious in patients. Although early data suggested a
possible benefit of the free-radical–trapping agent
NXY-059 in acute ischemic stroke,49 a large multicenter trial reported on by Shuaib et al. in this
issue of the Journal showed no improvement in
n engl j med 357;6  www.nejm.org  august 9, 2007
clinical pr actice
functional outcomes of patients who were treated
with this agent within 6 hours after the onset of
Hypothermia has been shown to reduce infarct volume and improve neurologic outcomes
in animal models of focal cerebral ischemia51; it
has also improved functional outcomes in randomized clinical trials involving patients with
global cerebral ischemia after cardiac arrest,52,53
but the improvement was not consistent among
those with traumatic brain injury.54 Large clinical trials testing the effect of hypothermia in paВ­
tients with acute ischemic stroke are warranted.
Guidel ine s from
Profe s siona l S o cie t ie s
Table 1. Main Contraindications to Intravenous Thrombolysis in Patients
with Acute Ischemic Stroke.*
Onset of symptoms >3 hr before start of treatment
Intracranial hemorrhage on CT or MRI
Head trauma or stroke in previous 3 mo
Myocardial infarction in previous 3 mo
Gastrointestinal or urinary tract hemorrhage in previous 21 days
Major surgery in previous 14 days
History of intracranial hemorrhage
Systolic blood pressure ≥185 mm Hg or diastolic blood pressure
≥110 mm Hg
Evidence of active bleeding or acute trauma on examination
Use of oral anticoagulants and an INR ≥1.7
Use of heparin in previous 48 hr and a currently prolonged aPTT
Platelet count <100,000 per cubic millimeter
Blood glucose level <50 mg/dl (2.7 mmol/liter)
Seizure with postictal residual neurologic impairments
from Adams et al., which provides a more complete overview of inPractice guidelines have been issued by the Stroke *Adapted
dications and contraindications. INR denotes international normalized ratio,
Council of the American Heart Association and
and aPTT activated partial-thromboplastin time.
the American Stroke Association8 and by the European Stroke Initiative.55 The recommendations
in this article are generally consistent with those tient presented within 3 hours after the onset of
symptoms, we would recommend therapy with
intravenous rt-PA. We would start aspirin after 24
hours (300 mg daily for the first 2 weeks) and
C onclusions
would then administer lower-dose aspirin and dia nd R ec om mendat ions
pyridamole for secondary prevention. Aggressive
The patient described in the vignette had a sudden management of other cardiovascular risk factors
left-sided hemiparesis, strongly suggestive of a — including encouraging the patient to stop smok­
right hemisphere stroke. CT or MRI of the brain ing, treating his hypertension, and initiating statin
should be performed promptly; MRI is more sen- therapy — is also warranted.
sitive for early ischemic changes, but either methDr. van der Worp reports receiving lecture fees from GlaxoВ­
od can fully rule out hemorrhage. In the absence SmithKline, Pfizer, and Servier; and Dr. van Gijn, consulting and
of bleeding or other contraindications to throm- lecture fees from Sanofi-Aventis. No other potential conflict of
bolysis (e.g., spontaneous, complete clearing of interest relevant to this article was reported.
We thank Audrey Tiehuis, M.D., for providing the scans shown
the deficits or an increase in blood pressure to in Figures 2 and 4 and L. Jaap Kappelle, M.D., for his valuable
185/110 mm Hg or more) (Table 1), since the pa- comments on an early version of the manuscript.
1. Lopez AD, Mathers CD, Ezzati M, JamiВ­
son DT, Murray CJ. Global and regional
burden of disease and risk factors, 2001:
systematic analysis of population health
data. Lancet 2006;367:1747-57.
2. Feigin VL, Lawes CM, Bennett DA,
Anderson CS. Stroke epidemiology: a review of population-based studies of incidence, prevalence, and case-fatality in the
late 20th century. Lancet Neurol 2003;2:
3. Dirnagl U, Iadecola C, Moskowitz MA.
Pathobiology of ischaemic stroke: an inВ­
tegrated view. Trends Neurosci 1999;22:
4. Norrving B. Long-term prognosis after
lacunar infarction. Lancet Neurol 2003;
5. Hacke W, Schwab S, Horn M, Spranger
M, De Georgia M, von Kummer R. �Malignant’ middle cerebral artery infarction:
clinical course and prognostic signs. Arch
Neurol 1996;53:309-15.
6. Brott T, Adams HP Jr, Olinger CP, et
al. Measurements of acute cerebral infarction: a clinical examination scale. Stroke
7. Hankey GJ, Warlow CP. Symptomatic
carotid ischaemic events: safest and most
cost effective way of selecting patients for
angiography, before carotid endarterectomy. BMJ 1990;300:1485-91.
8. Adams HP Jr, del Zoppo G, Alberts MJ,
et al. Guidelines for the early management
of adults with ischemic stroke: a guideline from the American Heart Association/
American Stroke Association Stroke Council, Clinical Cardiology Council, Cardiovascular Radiology and Intervention Council, and the Atherosclerotic Peripheral
Vascular Disease and Quality of Care Outcomes in Research Interdisciplinary Working Groups: the American Academy of
Neurology affirms the value of this guideline as an educational tool for neurologists. Stroke 2007;386:1655-711. [Erratum,
Stroke 2007;38(6):e38.]
9. Chalela JA, Kidwell CS, Nentwich LM,
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
et al. Magnetic resonance imaging and
computed tomography in emergency assessment of patients with suspected acute
stroke: a prospective comparison. Lancet
10. Muir KW, Buchan A, von Kummer R,
Rother J, Baron JC. Imaging of acute
stroke. Lancet Neurol 2006;5:755-68.
11. The National Institute of Neurological Disorders and Stroke rt-PA Stroke
Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J
Med 1995;333:1581-7.
12. Wardlaw JM, Zoppo G, Yamaguchi T,
Berge E. Thrombolysis for acute ischaeВ­
mic stroke. Cochrane Database Syst Rev
13. Hacke W, Donnan G, Fieschi C, et al.
Association of outcome with early stroke
treatment: pooled analysis of ATLANTIS,
ECASS, and NINDS rt-PA stroke trials.
Lancet 2004;363:768-74.
14. Khatri P, Wechsler LR, Broderick JP.
Intracranial hemorrhage associated with
revascularization therapies. Stroke 2007;
15. NINDS t-PA Stroke Study Group. Generalized efficacy of t-PA for acute stroke:
subgroup analysis of the NINDS t-PA
Stroke Trial. Stroke 1997;28:2119-25.
16. Patel SC, Levine SR, Tilley BC, et al.
Lack of clinical significance of early ischeВ­
mic changes on computed tomography in
acute stroke. JAMA 2001;286:2830-8.
17. Graham GD. Tissue plasminogen activator for acute ischemic stroke in clinical
practice: a meta-analysis of safety data.
Stroke 2003;34:2847-50.
18. Wahlgren N, Ahmed N, DГЎvalos A, et
al. Thrombolysis with alteplase for acute
ischaemic stroke in the Safe Implementation of Thrombolysis in Stroke-Monitoring
Study (SITS-MOST): an observational study.
Lancet 2007;369:275-82. [Erratum, Lancet
19. International Stroke Trial Collaborative Group. The International Stroke Trial
(IST): a randomised trial of aspirin, subcutaneous heparin, both, or neither among
19,435 patients with acute ischaemic
stroke. Lancet 1997;349:1569-81.
20. CAST (Chinese Acute Stroke Trial)
Collaborative Group. CAST: randomised
placebo-controlled trial of early aspirin
use in 20,000 patients with acute ischaeВ­
mic stroke. Lancet 1997;349:1641-9.
21. Chen ZM, Sandercock P, Pan HC, et al.
Indications for early aspirin use in acute
ischemic stroke: a combined analysis of
40 000 randomized patients from the Chinese Acute Stroke Trial and the International Stroke Trial. Stroke 2000;31:1240-9.
22. Gubitz G, Sandercock P, Counsell
C. Anticoagulants for acute ischaemic
stroke. Cochrane Database Syst Rev 2004;
23. Paciaroni M, Agnelli G, Micheli S,
Caso V. Efficacy and safety of anticoagulant treatment in acute cardioembolic
m e dic i n e
stroke: a meta-analysis of randomized
controlled trials. Stroke 2007;38:423-30.
24. Dennis MS, Lewis SC, Warlow C. Routine oral nutritional supplementation for
stroke patients in hospital (FOOD): a multiВ­
centre randomised controlled trial. Lancet
25. Idem. Effect of timing and method of
enteral tube feeding for dysphagic stroke
patients (FOOD): a multicentre randomised
controlled trial. Lancet 2005;365:764-72.
26. Kelly J, Rudd A, Lewis R, Hunt BJ. Venous thromboembolism after acute stroke.
Stroke 2001;32:262-7.
27. Francis CW. Prophylaxis for thromboembolism in hospitalized medical patients.
N Engl J Med 2007;356:1438-44.
28. Hofmeijer J, van der Worp HB, Kappelle LJ. Treatment of space-occupying
hemispheric infarction. Crit Care Med
29. Vahedi K, Hofmeijer J, Juettler E, et al.
Early decompressive surgery in malignant
infarction of the middle cerebral artery:
a pooled analysis of three randomised controlled trials. Lancet Neurol 2007;6:215-22.
30. Stroke Unit Trialists’ Collaboration.
Organised inpatient (stroke unit) care for
stroke. Cochrane Database Syst Rev 2002;
31. Halkes PH, van Gijn J, Kappelle LJ,
Koudstaal PJ, Algra A. Aspirin plus dipyridВ­
amole versus aspirin alone after cerebral
ischaemia of arterial origin (ESPRIT):
randomised controlled trial. Lancet 2006;
367:1665-73. [Erratum, Lancet 2007;369:
32. Johnston SC. Transient ischemic attack. N Engl J Med 2002;347:1687-92.
33. Sacco RL, Adams R, Albers G, et al.
Guidelines for prevention of stroke in patients with ischemic stroke or transient
ischemic attack: a statement for healthcare professionals from the American
Heart Association/American Stroke Association Council on Stroke: co-sponsored
by the Council on Cardiovascular Radiology and Intervention: the American Academy of Neurology affirms the value of
this guideline. Stroke 2006;37:577-617.
34. Douglas VC, Tong DC, Gillum LA, et
al. Do the Brain Attack Coalition’s criteria
for stroke centers improve care for ischeВ­
mic stroke? Neurology 2005;64:422-7.
35. Kidwell CS, Alger JR, Saver JL. Beyond
mismatch: evolving paradigms in imaging
the ischemic penumbra with multimodal
magnetic resonance imaging. Stroke 2003;
36. Wintermark M, Flanders AE, Velthuis
B, et al. Perfusion-CT assessment of infarct core and penumbra: receiver operating characteristic curve analysis in 130
patients suspected of acute hemispheric
stroke. Stroke 2006;37:979-85.
37. Albers GW, Thijs VN, Wechsler L, et
al. Magnetic resonance imaging profiles
predict clinical response to early reperfusion: the diffusion and perfusion imaging
evaluation for understanding stroke evolution (DEFUSE) study. Ann Neurol 2006;
38. Alexandrov AV, Molina CA, Grotta
JC, et al. Ultrasound-enhanced systemic
thrombolysis for acute ischemic stroke.
N Engl J Med 2004;351:2170-8.
39. Molina CA, Ribo M, Rubiera M, et al.
Microbubble administration accelerates
clot lysis during continuous 2-MHz ultrasound monitoring in stroke patients treatВ­
ed with intravenous tissue plasminogen
activator. Stroke 2006;37:425-9.
40. Furlan A, Higashida R, Wechsler L,
et al. Intra-arterial prourokinase for acute
ischemic stroke — the PROACT II Study:
a randomized controlled trial. JAMA 1999;
41. The IMS Study Investigators. Combined intravenous and intra-arterial recanalization for acute ischemic stroke:
the Interventional Management of Stroke
Study. Stroke 2004;35:904-11.
42. Flint AC, Duckwiler GR, Budzik RF,
Liebeskind DS, Smith WS. Mechanical
thrombectomy of intracranial internal
carotid occlusion: pooled results of the
MERCI and Multi MERCI Part I trials.
Stroke 2007;38:1274-80.
43. Willmot M, Leonardi-Bee J, Bath PM.
High blood pressure in acute stroke and
subsequent outcome: a systematic review.
Hypertension 2004;43:18-24.
44. Capes SE, Hunt D, Malmberg K,
Pathak P, Gerstein HC. Stress hyperglycemia and prognosis of stroke in nondiabetic and diabetic patients: a systematic
overview. Stroke 2001;32:2426-32.
45. Reith J, JГёrgensen S, Pedersen PM, et al.
Body temperature in acute stroke: relation
to stroke severity, infarct size, mortality,
and outcome. Lancet 1996;347:422-5.
46. van Breda EJ, van der Worp HB, van
Gemert HM, et al. PAIS: Paracetamol
(Acetaminophen) In Stroke; protocol for
a randomized, double blind clinical trial.
BMC Cardiovasc Disord 2005;5:24.
47. Gray CS, Hildreth AJ, Alberti GK,
O’Connell JE. Poststroke hyperglycemia:
natural history and immediate management. Stroke 2004;35:122-6. [Erratum,
Stroke 2004;35:1229.]
48. O’Collins VE, Macleod MR, Donnan
GA, Horky LL, van der Worp BH, Howells
DW. 1,026 Experimental treatments in
acute stroke. Ann Neurol 2006;59:46777.
49. Lees KR, Zivin JA, Ashwood T, et al.
NXY-059 for acute ischemic stroke. N Engl
J Med 2006;354:588-600.
50. Shuaib A, Lees KR, Lyden P, et al.
NXY-059 for the treatment of acute ischemic stroke. N Engl J Med 2007;357:56271.
51. van der Worp HB, Sena ES, Donnan
GA, Howells DW, Macleod MR. Hypothermia in animal models of acute ischaemic
stroke: a systematic review and meta-analyВ­
sis. Brain (in press).
n engl j med 357;6  www.nejm.org  august 9, 2007
clinical pr actice
52. Bernard SA, Gray TW, Buist MD, et al.
Treatment of comatose survivors of outof-hospital cardiac arrest with induced
hypothermia. N Engl J Med 2002;346:55763.
53. Hypothermia after Cardiac Arrest
Study Group. Mild therapeutic hypother-
mia to improve the neurologic outcome
after cardiac arrest. N Engl J Med 2002;
346:549-56. [Erratum, N Engl J Med 2002;
54. McIntyre LA, Fergusson DA, HГ©bert
PC, Moher D, Hutchison JS. Prolonged
therapeutic hypothermia after traumatic
brain injury in adults: a systematic review.
JAMA 2003;289:2992-9.
55. Hack W, Kaste M, Bogousslavsky J, et
al. European Stroke Initiative recommendations for stroke management — update
2003. Cerebrovasc Dis 2003;16:311-37.
Copyright В© 2007 Massachusetts Medical Society.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
review article
current concepts
Drug-Induced Immune Thrombocytopenia
Richard H. Aster, M.D., and Daniel W. Bougie, Ph.D.
From the Department of Medicine, Medical College of Wisconsin (R.H.A.), and the
Blood Research Institute, BloodCenter of
Wisconsin (R.H.A., D.W.B.) — both in
Milwaukee. Address reprint requests to
Dr. Aster at the Blood Research Institute,
BloodCenter of Wisconsin, 8727 Watertown Plank Rd., Milwaukee, WI 53226-3548,
or at [email protected]
N Engl J Med 2007;357:580-7.
Copyright В© 2007 Massachusetts Medical Society.
rug-induced thrombocytopenia can be caused by dozens, perhaps hundreds, of medications. Because thrombocytopenia can have many
other causes, the diagnosis of drug-induced thrombocytopenia can easily
be overlooked. On occasion, outpatients with drug-induced thrombocytopenia are
treated for autoimmune thrombocytopenia and can have two or three recurrences
before the drug causing the disorder is identified.1 In acutely ill, hospitalized patients, drug-induced thrombocytopenia can be overlooked because thrombocytopenia is attributed to sepsis, the effect of coronary-artery bypass surgery, or some other
underlying condition. Although drug-induced thrombocytopenia is uncommon, it can
have devastating, even fatal consequences that can usually be prevented simply by
discontinuing the causative drug. It is therefore important that clinicians have a
general understanding of this condition and the drugs that can cause it.
In this review, we focus on conditions in which exposure to a drug leads to the
destruction of circulating platelets, often accompanied by bleeding symptoms. We
do not consider thrombocytopenia resulting from dose-dependent hematosuppression, which often occurs after treatment with chemotherapeutic and immunosuppressive agents such as cisplatin and cyclophosphamide.2 Although heparin-induced
thrombocytopenia is the most common drug-related cause of a drop in the platelet
count, we do not discuss this condition because of its complexity and because
thrombosis, rather than thrombocytopenia, is the major threat to an affected patient. Because heparin is often given together with certain drugs that are likely to
cause drug-induced thrombocytopenia (platelet inhibitors and vancomycin), it is
important to distinguish between heparin-induced thrombocytopenia and druginduced thrombocytopenia. Heparin-induced thrombocytopenia was recently reviewed in the Journal.3
Drug-induced platelet destruction is usually caused by drug-induced antibodies,
but this can be difficult to prove. In this review, we include many conditions for
which an immune cause has not yet been fully documented. Although platelets are
the preferred targets of drug-induced antibodies, drugs can also cause immune
hemolytic anemia4 and neutropenia5 through similar mechanisms.
Pr e sen tat ion
Acute thrombocytopenia after exposure to quinine was recognized as a clinical
entity about 140 years ago.6 Subsequently, many other medications were implicated
in the development of this condition. Quinine, which is rarely used now as an antimalarial drug but is often prescribed for nocturnal muscle cramps, may still be the
most common trigger. People of any age and either sex can be affected.
The course of drug-induced thrombocytopenia in a representative patient is
shown in Figure 1. Typically, a patient will have taken the sensitizing drug for
about 1 week or intermittently over a longer period before presenting with petechial
n engl j med 357;6  www.nejm.org  august 9, 2007
current concepts
The incidence of drug-induced thrombocytopenia is not well defined, in part because reporting
is voluntary and is not critically reviewed. On
the basis of several epidemiologic studies in the
United States and Europe, the estimated minimum
incidence is about 10 cases per million population
per year but the number could be higher in selectВ­
ed groups, such as hospitalized patients and
elderly people.10 A case–control study of patients
in Massachusetts, Rhode Island, and Philadelphia
showed that during each week of exposure, trimethoprim–sulfamethoxazole and quinine–quin­
idine caused thrombocytopenia in 38 and 26 of
every 1 million users, respectively.11 Since these
drugs carry relatively high risks of drug-induced
thrombocytopenia, the rate at which most drugs
cause the condition is probably lower. However,
a few drugs (including abciximab and gold salts)
cause immune thrombocytopenia in about 1% of
Platelets (Г—10в€’5/mm3)
hemorrhages and ecchymoses that are indicative
of thrombocytopenia. Occasionally, symptoms deВ­
velop within 1 or 2 days after what is apparently
the first exposure to a drug, particularly in patients given platelet inhibitors such as abciximab
who may have preexisting, perhaps naturally occurring, antibodies.7 Systemic symptoms such as
lightheadedness, chills, fever, nausea, and vomiting often precede bleeding symptoms. Severely
affected patients have florid purpura and bleeding from the nose, gums, and gastrointestinal
or urinary tract (“wet purpura”). In such cases,
thrombocytopenia is invariably severe (<20,000
platelets per cubic millimeter).
If the causative medication is stopped, symptoms usually resolve within 1 or 2 days, and the
platelet count returns to normal in less than a
week. For reasons that are poorly understood,
patients with drug-induced thrombocytopenia ocВ­
casionally present with disseminated intravascular coagulation8 or renal failure and other findings indicative of the hemolytic–uremic syndrome
or thrombotic thrombocytopenic purpura.9
Petechiae and
oral and urinary
Figure 1. Course of Immune Thrombocytopenia in a Patient Treated
with Sulfamethoxazole.
Trimethoprim–sulfamethoxazole was prescribed for treatment of a urinary
tract infection in a 29-year-old
woman. On day 7 of treatment,
and ecchymoses,
and gross urinary bleeding
3rd 4000 per
and the antibiotic treatment was stopped. Her platelet count was
cubic millimeter.
A platelet transfusion given on day 9 Revised
had little effect on
the platelet count.
occurred during theSIZE
next few days.
ARTIST: recovery
A strong sulfamethoxazole-dependent,
platelet-reactive antibody
was detectCombo
ed in a blood sample obtained
3 months
and after 5 years. Figure
No trimethoprim-dependent
were detected.
has been redrawn and type has
been reset.
Please check carefully.
JOB: 356xx
ISSUE: 08-02-07
have been implicated.2,12,13 Many of the drugs that
are common triggers for this disorder also cause
sensitivity reactions involving the skin and other
George and colleagues critically analyzed reports of drug-induced thrombocytopenia published through 200513,15 and identified 85 medications for which a cause-and-effect relationship
was considered to be “definite” (58 agents) or
“probable” (27 agents) on the basis of clinical
criteria (Table 2).13 (A compendium of implicated
drugs and case reports updated through August
2004 is available at http://moon.ouhsc.edu/jgeorge/
DITP.html.) In the analysis by George et al., no
weight was given to laboratory demonstration of
drug-dependent antibodies. As improved techniques are developed, serologic testing may become increasingly useful for identifying the specific cause of thrombocytopenia in individual
cases — albeit after the fact.
C aus at i v e Agen t s
Rare but convincing examples of drug-induced
Many of the drugs shown in multiple studies to thrombocytopenia induced by herbal remedies16,17
be capable of causing drug-induced thrombocy- and foods18 have been described, and there are
topenia are listed in Table 1, but at least 100 others numerous reports of acute, severe thrombocyton engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
Table 1. Drugs Commonly Implicated as Triggers of Drug-Induced Thrombocytopenia.*
Drug Category
Drugs Implicated in Five or More Reports
Unfractionated heparin, low-molecular-weight heparin
Other Drugs
Cinchona alkaloids
Quinine, quinidine
Platelet inhibitors
Abciximab, eptifibatide, tirofiban
Antirheumatic agents
Gold salts
Antimicrobial agents
Linezolid, rifampin, sulfonamides, vancomycin
Sedatives and anticonvulsant agents
Carbamazepine, phenytoin, valproic acid
Histamine-receptor antagonists
Analgesic agents
Acetaminophen, diclofenac, naproxen
Diuretic agents
Chemotherapeutic and immunoВ­
suppressant agents
Fludarabine, oxaliplatin
Cyclosporine, rituximab
*For a more extensive list, see Aster,2 Warkentin,12 and George et al.13 and the University of Oklahoma Web site (http://
Table 2. Criteria and Level of Evidence for Establishing a Causative Relationship in Drug-Induced Thrombocytopenic
Criterion and Level of Evidence
Therapy with the candidate drug preceded thrombocytopenia, and recovery from
thrombocytopenia was complete and sustained after discontinuation of therapy.
The candidate drug was the only drug used before the onset of thrombocytopenia,
or other drugs were continued or reintroduced after discontinuation of therapy
with the candidate drug, with a sustained normal platelet count.
Other causes of thrombocytopenia were ruled out.
Re-exposure to the candidate drug resulted in recurrent thrombocytopenia.
Level of evidence
Definite — criteria 1, 2, 3, and 4 are met.
Probable — criteria 1, 2, and 3 are met.
Possible — criterion 1 is met.
Unlikely — criterion 1 is not met.
*The information is adapted from George et al.13
penia after injection of iodinated contrast medium for radiographic studies.12,13 Whether immune
mechanisms are involved is unknown.
Severe thrombocytopenia and other signs and
symptoms of thrombotic thrombocytopenic purpura develop in approximately 1 of every 2500
patients treated with the platelet inhibitor ticlopidine and a much smaller fraction of those given
the closely related drug clopidogrel, usually after
1 to 2 weeks of treatment.19 The responsible
mechanisms have not yet been defined. Acute,
severe, usually self-limited thrombocytopenia has
been described in patients treated with recently
developed monoclonal antibodies such as infliximab (anti–tumor necrosis factor-α antibody),
efalizumab (anti-CD11a antibody), and rituximab
(anti-CD20 antibody).2 No causative mechanism
has yet been identified.
Certain drugs, such as the antiepileptic agent
valproate,20 the cardiac agent amrinone,21 and the
antibiotic linezolid,22 induce low-grade thrombocytopenia in up to 30% of patients receiving long-
n engl j med 357;6  www.nejm.org  august 9, 2007
current concepts
term treatment; the mechanisms of action may
be nonimmune but are poorly understood. The
decrease in platelets is rarely severe enough to
require treatment.
Although chemotherapeutic and immunosuppressive agents typically cause thrombocytopenia by suppressing hematopoiesis, they can also
cause immune thrombocytopenia.23 Drug-induced
thrombocytopenia should be suspected, therefore,
in patients treated with such drugs if there is an
acute drop in the platelet level after exposure.
Immune thrombocytopenia caused by vancomycin is probably more common than is generally
recognized, and it is easily overlooked in seriousВ­
ly ill patients because the low platelet count can
be attributed to other causes.24 In rare cases, the
quinine in tonic water or an aperitif causes immune thrombocytopenia (“cocktail purpura”).25
Patho gene sis
Drug-induced thrombocytopenia, like other idiosyncratic drug-sensitivity reactions, affects only
a small fraction of patients taking medications
that can trigger the disorder. No predisposing
genetic or environmental factors have been identified, and no suitable animal models are available. Mechanisms by which drugs are thought to
cause immune thrombocytopenia are summarized
in Table 3.
Thrombocytopenia Induced by Quinine
and Other Drugs
The hallmark of thrombocytopenia induced by
quinine and many other drugs is a remarkable
antibody that binds tightly to normal platelets
only in the presence of the sensitizing drug. The
epitopes targeted by these antibodies usually reside on glycoprotein IIb/IIIa or Ib/V/IX complexes,
the major platelet receptors for fibrinogen and
von Willebrand factor, respectively.26,27
Small molecules, like drugs, are thought to be
immunogenic only when linked covalently to a
large carrier molecule, usually a protein. Antibodies induced by drug–protein adducts are largely
specific for the drug (or the small molecule, called
the hapten), although some antibodies recognize
the drug and its carrier molecule.28 Accordingly,
early investigators assumed that drug-dependent
antibodies found in patients with drug-induced
thrombocytopenia were hapten-specific (or drugspecific). This may be true of antibodies that
cause hemolytic anemia in patients treated with
massive doses of penicillin29 and perhaps some
of those that cause thrombocytopenia in patients
given penicillin-like drugs.30 However, serologic
studies in patients with drug-induced thrombocytopenia caused by other drugs failed to support this concept.31 An alternative hypothesis was
that the drug reacts directly with the antibody to
produce immune complexes that somehow target
Table 3. Mechanisms Underlying Drug-Induced Immune Thrombocytopenia.*
Examples of Drugs
Hapten-dependent Hapten links covalently to membrane protein Very rare
and induces drug-specific immune response
Penicillin, possibly some cephalosporin antibiotics
Quinine-type drug Drug induces antibody that binds to membrane 26 cases per 1 million users of
В­ uinine per week, probably
protein in presence of soluble drug
В­fewer cases with other drugs
Quinine, sulfonamide antibiotics,
nonsteroidal antiinflammatory
Fiban-type drug
Drug reacts with glycoprotein IIb/IIIa to induce 0.2–0.5%
a conformational change (neoepitope) recognized by antibody (not yet confirmed)
Tirofiban, eptifibatide
Antibody recognizes murine component of
В­chimeric Fab fragment specific for platelet
membrane glycoprotein IIIa
Drug induces antibody that reacts with autolo- 1.0% with gold, very rare with progous platelets in absence of drug
cainamide and other drugs
Immune complex
Drug binds to platelet factor 4, producing im- 3–6% among patients treated with Heparins
mune complex for which antibody is specific;
unfractionated heparin for 7 days,
immune complex activates platelets through
rare with low-molecular-weight
Fc receptors
0.5–1.0% after first exposure,
10–14% after second exposure
Gold salts, procainamide
*The information is adapted from Aster.2
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
platelets and cause their destruction.31 However,
these hypothetical drug–antibody complexes were
never demonstrated experimentally, and it was
later found that drug-dependent antibodies, like
other antibodies, react with platelets through
their Fab domains rather than through their Fc
domains, as would be expected of immune complexes.32,33 Other possibilities considered were
that the drug reacts with the target protein to
produce a compound epitope (part drug, part proВ­
tein) for which the antibody is specific34,35 and
that the drug induces a conformational change
in the protein, creating a new target epitope elsewhere in the molecule.2,36
A recently proposed model aimed at reconciling several of these hypotheses37 suggests that
drug-dependent antibodies are derived from a
pool of naturally occurring antibodies with weak
affinity for self antigens,38 residing in this case
on certain platelet membrane glycoproteins. According to this model, the interaction between
these low-affinity antibodies and their target
antigens is too weak to affect blood cells under
normal circumstances. However, certain drugs
affect both antibody and antigen in such a way
that the strength of the interaction is greatly increased (Fig. 2). When a B cell expressing such
an antibody is induced to proliferate and undergo
affinity maturation in a patient taking the medication, the resulting antibody can destroy the
targeted blood cell if the drug is present.37 This
model suggests that whether the drug binds first
to the antibody or to the targeted membrane protein depends simply on its relative affinity for
one component or the other.
Thrombocytopenia Induced by Platelet
Acute thrombocytopenia, usually mild but occasionally life-threatening, is a common complication of treatment with the platelet inhibitors tirofiban and eptifibatide, which are widely used to
prevent restenosis after coronary angioplasty.7
These ligand-mimetic drugs (“fibans”) inhibit
thrombosis by binding to a specific site on the
platelet О±IIb/ОІ3 integrin (glycoprotein IIb/IIIa) and
competitively inhibiting platelet–fibrinogen interaction.7 Antibodies causing thrombocytopenia in
patients given these agents probably recognize
structural changes (neoepitopes) induced in the
О±IIb/ОІ3 integrin when a fiban binds to it, but this
theory has not been formally proved. Curiously,
m e dic i n e
these antibodies can occur naturally, creating the
possibility that thrombocytopenia will arise within a few hours after the patient’s first exposure to
the drug.7
Thrombocytopenia Induced by Abciximab
Abciximab is a widely used chimeric (human–
mouse) Fab fragment that is specific for ОІ3 integrin (glycoprotein IIIa). Like the fibans, it blocks
platelet–fibrinogen interaction. Abciximab itself
does not cause thrombocytopenia because it lacks
the Fc domain required for recognition of antibody-coated platelets by phagocytes. However, in
about 1% of patients given abciximab for the
first time, and in more than 10% of those treated
a second time, acute thrombocytopenia develops
within a few hours of starting an infusion.39 In
some patients the onset of thrombocytopenia is
delayed until 5 to 8 days after the initial 24- to
48-hour period of exposure to the drug.40 AbcixiВ­
mab-induced thrombocytopenia is often mild, but
fatalities have been recorded.41,42
Acute thrombocytopenia after first exposure
to abciximab appears to be caused by preexisting
antibodies specific for murine structural elements
in the abciximab molecule.40,42 Antibodies that
cause delayed thrombocytopenia are newly induced
but recognize the same target; they cause thrombocytopenia because abciximab-coated platelets
are still present in the circulation 10 to 14 days
after treatment.40
Thrombocytopenia Due to Drug-Induced
In rare cases, drugs induce true autoantibodies
that are capable of destroying platelets in the absence of the sensitizing agent.43 About 1% of patients treated with gold salts for rheumatoid arthritis have this complication.44 Autoantibodies
induced by gold may be unique in having specificity for platelet membrane glycoprotein V.45 Other
drugs that are probably capable of inducing autoimmune thrombocytopenia include procainamide,
sulfonamide antibiotics, and interferons alfa and
beta.2,12,13,43 Acute, sometimes severe, but usualВ­
ly transient thrombocytopenia can occur several
weeks after the vaccination of children or adults
for various infectious diseases, but it is rare.46,47
This condition resembles acute idiopathic thrombocytopena, which sometimes develops in children after a viral infection, but its cause has not
been established.
n engl j med 357;6  www.nejm.org  august 9, 2007
current concepts
antibody CDR
antibody CDR
Platelet antigen
Platelet antigen
Low-affinity fit
High-affinity fit
Figure 2. Model for the Binding of a Drug-Dependent Antibody to an Epitope on a Platelet Glycoprotein.
Antibodies capable of causing drug-dependent thrombocytopenia react weakly with an epitope on a target glycoprotein. The binding affinity
Draft 5
(K A) for this interaction is too low to allow a sufficient number of antibody molecules to bind in the absence of the drug (“low-affinity
fit”). The drug contains structural elements that are complementary to a negatively charged site on the glycoprotein
Fig #
site (H) on the complementarity-determining region (CDR) of the antibody. The drug interacts with these sites toTitle
the fit between
antibody binding
the two proteins, increasing the K A to a value that permits binding to occur at levels of antibody, antigen, and ME
drug achieved in the cirВ­
culation after ingestion of the drug (“high-affinity fit”). Adapted from Bougie et al.37
Figure has been redrawn and type has been reset
Please check carefully
Platelet-specific autoimmunity induced by drugs
is clinically similar to acute idiopathic autoimmune thrombocytopenia.43 The possibility that
drug-induced autoimmune thrombocytopenia is
not uncommon remains speculative.
Di agnosis
Drug-induced thrombocytopenia should be suspected in any patient who presents with acute
thrombocytopenia of unknown cause. In considering this diagnosis, the clinician should keep in
mind that 5 to 7 days of exposure is usually needed to produce sensitization in a patient given a
drug for the first time. As previously noted, platelet inhibitors are exceptions to this general rule.
In adults, the presence of severe thrombocytopenia (<20,000 platelets per cubic millimeter) increases the likelihood that a patient has druginduced thrombocytopenia, and it should be
strongly suspected in any patient with a history
of acute, transient thrombocytopenia. Because patients sometimes do not report exposure to drugs
later found to be the responsible agents,1 a detailed, careful history of drug exposure is essential. Patients should be asked specifically about
quinine, quinidine, sulfonamides, herbal remedies,
folk medicines, common nonprescription drugs
such as acetaminophen, and recent vaccinations. 08/02/07
Thrombocytopenia caused by undisclosed drug
use has been described.48
In patients with sensitivity to quinine, quinidine, sulfonamides, and many other drugs, it is
often possible to identify antibodies that react
with normal platelets in the presence of the drug
but not in its absence.23,42,49 However, testing is
technically demanding and not widely available
(except for heparin) and is therefore not useful
in the immediate care of a patient. Testing can
be helpful in documenting the cause of thrombocytopenia after the fact and, more generally, in
determining which drugs can cause drug-induced
thrombocytopenia. Unfortunately, in patients with
a history that is typical of drug-induced thrombocytopenia, antibody tests may be negative.2,36
One important reason for this is that a drug metabolite produced in vivo can be the sensitizing
agent.34,50 The range of drug metabolites capable
of inducing drug-induced thrombocytopenia is
not well defined, but this type of sensitivity may
be more common than has been thought.34,50
If there is a strong suspicion that thrombocytopenia was drug-induced and documentation of
drug sensitivity is critical for diagnosis or management, a diagnostic challenge can be considered. Just 1 or 2 mg of a drug can cause a substan-
n engl j med 357;6  www.nejm.org  august 9, 2007
Issue date
n e w e ng l a n d j o u r na l
tial drop in platelet levels,51 and a conventional
dose can cause severe thrombocytopenia and
bleeding.52 Therefore, it is important to start with
a few milligrams of the drug and to monitor
platelet counts closely for 24 hours. Antibodies
sometimes become undetectable after a few
months, in which case the drug may initially
have no effect on the platelet count.
T r e atmen t a nd Pro gnosis
Many patients with drug-induced thrombocytopenia have only petechial hemorrhages and occasional ecchymoses and require no specific treatment other than discontinuation of the sensitizing
medication. When there is uncertainty about the
causative drug, all medications should be discontinued, and pharmacologic equivalents with differВ­
ent chemical structures substituted as necessary.
Patients who have severe thrombocytopenia and
“wet purpura” should be aggressively treated with
1. Reddy JC, Shuman MA, Aster RH.
Quinine/quinidine-induced thrombocytopenia: a great imitator. Arch Intern Med
2. Aster R. Drug-induced thrombocytopenia. In: Michelson AD, ed. Platelets.
New York: Academic Press, 2007:887-902.
3. Arepally GM, Ortel TL. HeparinВ­induced thrombocytopenia. N Engl J Med
4. Arndt PA, Garratty G. The changing
spectrum of drug-induced immune hemolytic anemia. Semin Hematol 2005;42:
5. Stroncek DF. Drug-induced immune
neutropenia. Transfus Med Rev 1993;7:26874.
6. Vipan W. Quinine as a cause of purpura. Lancet 1865;2:37.
7. Aster RH. Immune thrombocytopenia
caused by glycoprotein IIb/IIIa inhibitors.
Chest 2005;127:Suppl 2:53S-59S.
8. Knower MT, Bowton DL, Owen J, Dunagan DP. Quinine-induced disseminated
intravascular coagulation: case report and
review of the literature. Intensive Care Med
9. Kojouri K, Vesely SK, George JN. Quinine-associated thrombotic thrombocytopenic purpura–hemolytic uremic syndrome:
frequency, clinical features, and long-term
outcomes. Ann Intern Med 2001;135:104751.
10. van den Bemt PM, Meyboom RH, Egberts AC. Drug-induced immune thrombocytopenia. Drug Saf 2004;27:1243-52.
11. Kaufman DW, Kelly JP, Johannes CB,
m e dic i n e
platelet transfusions because of the risk of fatal
intracranial or intrapulmonary hemorrhage.2,53,54
Corticosteroids are often given, but there is no
evidence that they are helpful if the thrombocytopenia is drug-induced. Intravenous immune globulin55 and plasma exchange56 have been used in
acutely ill patients, but the benefit of these treatments is uncertain.2
Once established, drug sensitivity probably perВ­
sists indefinitely. Therefore, patients should be
advised to avoid permanently the medication
thought to be the cause of thrombocytopenia.
Fortunately, drug-induced antibodies tend to be
specific for the sensitizing drug,57 and patients
usually tolerate pharmacologic equivalents, even
those with quite similar structures.
Supported by grants from the National Heart, Lung, and Blood
Institute (HL-13629 and HL-44612) and the Northland affiliate
of the American Heart Association (0235419Z).
No potential conflict of interest relevant to this article was
et al. Acute thrombocytopenic purpura in
relation to the use of drugs. Blood 1993;
12. Warkentin TE. Thrombocytopenia due
to platelet destruction and hypersplenism.
In: Hoffman R, Benz EJ Jr, Shattil SJ, et al.,
eds. Hematology: basic principles and pracВ­
tice. 4th ed. Philadelphia: Elsevier, 2005:
13. George JN, Raskob GE, Shah SR, et al.
Drug-induced thrombocytopenia: a systematic review of published case reports.
Ann Intern Med 1998;129:886-90.
14. Roujeau JC, Kelly JP, Naldi L, et al.
Medication use and the risk of Stevens–
Johnson syndrome or toxic epidermal
necrolysis. N Engl J Med 1995;333:16007.
15. Li X, Hunt L, Vesely SK. Drug-induced
thrombocytopenia: an updated systematic
review. Ann Intern Med 2005;142:474-5.
16. Arnold J, Ouwehand WH, Smith GA,
Cohen H. A young woman with petechiae.
Lancet 1998;352:618.
17. Azuno Y, Yaga K, Sasayama T, Kimoto
K. Thrombocytopenia induced by Jui, a traditional Chinese herbal medicine. Lancet
18. Lavy R. Thrombocytopenic purpura
due to lupinus termis bean. J Allergy Clin
Immunol 1964;35:386-9.
19. Bennett CL, Connors JM, Carwile JM,
et al. Thrombotic thrombocytopenic purpura associated with clopidogrel. N Engl J
Med 2000;342:1773-7.
20. Trannel TJ, Ahmed I, Goebert D. Occurrence of thrombocytopenia in psychi-
atric patients taking valproate. Am J Psychiatry 2001;158:128-30.
21. Ross MP, Allen-Webb EM, Pappas JB,
McGough EC. Amrinone-associated thromВ­
bocytopenia: pharmacokinetic analysis.
Clin Pharmacol Ther 1993;53:661-7.
22. Attassi K, Hershberger E, Alam R,
Zervos MJ. Thrombocytopenia associated
with linezolid therapy. Clin Infect Dis
23. Curtis BR, Kaliszewski J, Marques MB,
et al. Immune-mediated thrombocytopenia resulting from sensitivity to oxaliplatin. Am J Hematol 2006;81:193-8.
24. Von Drygalski A, Curtis BR, Bougie
DW, et al. Vancomycin-induced immune
thrombocytopenia. N Engl J Med 2007;
25. Belkin GA. Cocktail purpura: an unusual case of quinine sensitivity. Ann Intern Med 1967;66:583-6.
26. Visentin GP, Newman PJ, Aster RH.
Characteristics of quinine- and quinidineinduced antibodies specific for platelet
glycoproteins IIb and IIIa. Blood 1991;77:
27. Asvadi P, Ahmadi Z, Chong BH. Druginduced thrombocytopenia: localization of
the binding site of GPIX-specific quininedependent antibodies. Blood 2003;102:
28. Parker CW. Hapten immunology and
allergic reactions in humans. Arthritis
Rheum 1981;24:1024-36.
29. Garratty G. Immune cytopenia associated with antibiotics. Transfus Med Rev
n engl j med 357;6  www.nejm.org  august 9, 2007
current concepts
30. Salamon DJ, Nusbacher J, Stroupe T,
Wilson JH, Hanrahan JB. Red cell and
platelet-bound IgG penicillin antibodies in a patient with thrombocytopenia.
Transfusion 1984;24:395-8.
31. Shulman NR. A mechanism of cell
destruction in individuals sensitized to
foreign antigens and its implications in
auto-immunity: combined Clinical Staff
Conference at the National Institutes of
Health. Ann Intern Med 1964;60:506-21.
32. Christie DJ, Mullen PC, Aster RH.
Fab-mediated binding of drug-dependent
antibodies to platelets in quinidine- and
quinine-induced thrombocytopenia. J Clin
Invest 1985;75:310-4.
33. Smith ME, Reid DM, Jones CE, Jordan
JV, Kautz CA, Shulman NR. Binding of
quinine- and quinidine-dependent drug
antibodies to platelets is mediated by the
Fab domain of the immunoglobulin G and
is not Fc dependent. J Clin Invest 1987;79:
34. Mueller-Eckhardt C, Salama A. Druginduced immune cytopenias: a unifying
pathogenetic concept with special emphasis on the role of drug metabolites. Transfus Med Rev 1990;4:69-77.
35. Shulman NR, Reid DM. Mechanisms
of drug-induced immunologically mediated cytopenias. Transfus Med Rev 1993;7:
36. Aster RH. Drug-induced immune cytopenias. Toxicology 2005;209:149-53.
37. Bougie DW, Wilker PR, Aster RH.
Patients with quinine-induced immune
thrombocytopenia have both “drug-depen­
dent” and “drug-specific” antibodies. Blood
38. Pillai S. Two lymphoid roads diverge
— but does antigen bade B cells to take
the road less traveled? Immunity 2005;23:
39. Tcheng JE, Kereiakes DJ, Lincoff AM,
et al. Abciximab readministration: results
of the ReoPro Readministration Registry.
Circulation 2001;104:870-5.
40. Curtis BR, Divgi A, Garritty M, Aster
RH. Delayed thrombocytopenia after treatment with abciximab: a distinct clinical
entity associated with the immune response to the drug. J Thromb Haemost
41. McCorry RB, Johnston P. Fatal delayed thrombocytopenia following abciximab therapy. J Invasive Cardiol 2006;18:
42. Curtis BR, Swyers J, Divgi A, McFarland JG, Aster RH. Thrombocytopenia
after second exposure to abciximab is
caused by antibodies that recognize abciximab-coated platelets. Blood 2002;99:
43. Aster RH. Can drugs cause autoimmune thrombocytopenic purpura? Semin
Hematol 2000;37:229-38.
44. von dem Borne AE, Pegels JG, van der
Stadt RJ, van der Plas-van Dalen CM, Helmerhorst FM. Thrombocytopenia associated with gold therapy: a drug-induced
autoimmune disease? Br J Haematol 1986;
45. Garner SF, Campbell K, Metcalfe P,
et al. Glycoprotein V: the predominant
target antigen in gold-induced autoimmune thrombocytopenia. Blood 2002;100:
46. Neau D, Bonnet F, Michaud M, et al.
Immune thrombocytopenic purpura after
recombinant hepatitis B vaccine: retrospective study of seven cases. Scand J Infect Dis 1998;30:115-8.
47. Nieminen U, Peltola H, Syrjala MT,
Makipernaa A, Kekomaki R. Acute thrombocytopenic purpura following measles,
mumps and rubella vaccination: a report
on 23 patients. Acta Paediatr 1993;82:26770.
48. Reid DM, Shulman NR. Drug purpura
due to surreptitious quinidine intake.
Ann Intern Med 1988;108:206-8.
49. Bougie DW, Wilker PR, Wuitschick ED,
et al. Acute thrombocytopenia after treatment with tirofiban or eptifibatide is assoВ­
ciated with antibodies specific for ligandoccupied GPIIb/IIIa. Blood 2002;100:
50. Bougie D, Aster R. Immune thrombocytopenia resulting from sensitivity to metabolites of naproxen and acetaminophen.
Blood 2001;97:3846-50.
51. Shulman NR. Immunoreactions involvВ­
ing platelets. IV. Studies on the pathogenesis of thrombocytopenia in drug purpura
using test doses of quinidine in sensitized
individuals; their implications in idiopathic thrombocytopenic purpura. J Exp
Med 1958;107:711-29.
52. Schmitt SK, Tomford JW. Quinineinduced pancytopenia and coagulopathy.
Ann Intern Med 1994;120:90-1.
53. Freiman JP. Fatal quinine-induced
thrombocytopenia. Ann Intern Med 1990;
54. Fireman Z, Yust I, Abramov AL. Lethal
occult pulmonary hemorrhage in druginduced thrombocytopenia. Chest 1981;
55. Ray JB, Brereton WF, Nullet FR. Intravenous immune globulin for the treatment
of presumed quinidine-induced thrombocytopenia. DICP 1990;24:693-5.
56. Pourrat O. Treatment of drug-related
diseases by plasma exchanges. Ann Med
Interne (Paris) 1994;145:357-60.
57. Christie DJ, Weber RW, Mullen PC,
Cook JM, Aster RH. Structural features of
the quinidine and quinine molecules necessary for binding of drug-induced antibodies to human platelets. J Lab Clin Med
Copyright В© 2007 Massachusetts Medical Society.
images in clinical medicine
The Journal welcomes consideration of new submissions for Images in Clinical
Medicine. Instructions for authors and procedures for submissions can be found
on the Journal’s Web site at www.nejm.org. At the discretion of the editor, images
that are accepted for publication may appear in the print version of the Journal,
the electronic version, or both.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
images in clinical medicine
Intravesical Foreign Body
Nicola Mondaini, M.D.
Riccardo Bartoletti, M.D.
University of Florence
Florence, Italy
[email protected]
n 82-year-old woman presented with abdominal colic. she had
a 20-year history of type 2 diabetes and a 2-year history of recurrent urinary
despite many courses of
antimicrobial therapy. She lived in a
showed a large stone in her bladder,
with EMail
a foreign body in theLine
of theRevised
stone (Panel A). On cystotomy (Panel B), an
Enon bladder
mleahy measuring
6 by 2 by
0.5 cm was removed (Panel C), and a pen
cap was found at the center (Panel D). The patient was discharged 2 days later and
had a complete
further urinary tract infections. The patient
Figure has been
and type
did not know howPlease
cap carefully.
made its way into her bladder. A foreign body should be
considered in a patient with persistent urinary tract problems, especially if the patient
JOB: 35706
ISSUE: 8-09-07
is frail.
Copyright В© 2007 Massachusetts Medical Society.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
case records of the massachusetts general hospital
Founded by Richard C. Cabot
Nancy Lee Harris, m.d., Editor Eric S. Rosenberg, m.d., Associate Editor
Jo-Anne O. Shepard, m.d., Associate Editor
Alice M. Cort, m.d., Associate Editor
Sally H. Ebeling, Assistant Editor
Christine C. Peters, Assistant Editor
Case 24-2007: A 20-Year-Old Pregnant
Woman with Altered Mental Status
Andrew J. Cole, M.D., John W. Henson, M.D., Michael H.A. Roehrl, M.D., Ph.D.,
and Matthew P. Frosch, M.D., Ph.D.
Pr e sen tat ion of C a se
A 20-year-old pregnant woman was admitted to this hospital at 26 weeks of gestation because of dizziness, confusion, and difficulty walking.
Ten weeks before admission, the patient had a positive result on a home pregnancy test and presented to a neighborhood health center for prenatal screening.
Tests for sickle cell trait, syphilis, and human immunodeficiency virus (HIV) and
hepatitis B and C virus antibodies were negative. Serologic tests for varicella–zoster virus and rubella IgG were positive. Two weeks later, an endocervical specimen
was positive for Chlamydia trachomatis infection and negative for gonorrhea. The
patient missed follow-up appointments, and treatment with azithromycin was
initiated 4 weeks later.
Six weeks before admission, she moved into a shelter for pregnant women. Staff
members described her as happy, with a childlike affect, a poor memory, confusion, and odd movements of her head. During the next 2 weeks, nausea and vomiting occurred daily and were controlled with metoclopramide. Four days before
admission, dizziness and weakness on the left side developed; she began to fall to
her left and vomited several times. The next day, she went to the emergency department of another hospital. On evaluation, the patient was oriented to location
but not to date, day, or month, and she provided inconsistent information about
her medical history. The uterus was gravid, and the remainder of the physical
examination was normal. An electrocardiogram revealed sinus tachycardia and
counterclockwise rotation with T waves in the right precordial leads. Urinalysis
showed a protein level of 30 mg per deciliter and a glucose level of 100 mg per
deciliter (5.6 mmol per liter). Tricyclic metabolites were present on toxicology
screening of a urine specimen. Computed tomographic (CT) scanning of the head
revealed a slight, diffuse prominence of the ventricular system. There was no intracranial mass or other focal brain lesion.
On the second hospital day, the patient was alert, calm, cooperative, and oriented to person, location, and current events but was not aware of details of her
life. There were no tremors or extrapyramidal signs. Ultrasonographic examination revealed normal fetal anatomy and growth, corresponding to a gestation of
25 weeks 6 days. On the third day, the weakness, nausea, and vomiting had resolved,
n engl j med 357;6  www.nejm.org  august 9, 2007
From the Neurology Service (A.J.C., J.W.H.,
M.P.F.) and the Departments of RadiolВ­
ogy (J.W.H.) and Pathology (M.H.A.R.,
M.P.F.), Massachusetts General Hospital; and the Departments of Neurology
(A.J.C., J.W.H.) and Pathology (M.H.A.R.,
M.P.F.), Harvard Medical School.
N Engl J Med 2007;357:589-600.
Copyright В© 2007 Massachusetts Medical Society.
n e w e ng l a n d j o u r na l
and the patient was thought to have returned to
her baseline mental status. She was discharged
to the shelter with a recommendation to schedule a follow-up neurologic evaluation. At the
shelter, she was dizzy, had difficulty walking,
and fell into a chair. That evening, she was
brought to the emergency department of this
In the emergency department, the patient reported feeling “woozy” and nauseated. She noted
a mild headache of gradual onset, extending bandВ­
like across the brow. The history as given by the
patient was inconsistent; the history was then
provided by staff members of the shelter, and
many details were lacking. The patient was a
native of Cape Verde who had immigrated to this
country 3 years previously. She had had measles
at 4 months of age and varicella infection in childВ­
hood. At 7 years of age, she injured her head in
a fall but was said to have recovered fully. ImmuВ­
nizations included polio vaccine and diphtheria,
pertussis, and tetanus vaccine series; measles vaccine (at 11 months of age); measles, mumps, and
rubella vaccine combination; and hepatitis B vaccine (between 2 and 3 years before admission, on
enrollment in high school).
The patient had attended school through the
10th grade and was unemployed. During the
3 years before admission, she had lived with relatives, friends, and a boyfriend, as well as in shelters. She was single, and she no longer maintained a social relationship with the father of the
fetus. Her parents and seven siblings were alive
but not in contact with her at the time of admission. No family medical history was available. She
had no known allergies and did not use alcohol,
illicit drugs, or tobacco.
On examination in the emergency department,
the patient was alert but somewhat uncooperative, with involuntary head movements. Her mental status was not formally assessed, but her level
of cognitive function was said by a friend to be at
baseline. The blood pressure was 107/81 mm Hg,
the pulse 84 beats per minute, and the temperature 36.3В°C; the respirations were 18 per minute,
and the oxygen saturation was 100% while the paВ­
tient was breathing ambient air. Acneiform lesions
were present on her face. The abdomen was soft,
gravid, and not tender; the fetus appeared to be
healthy. The 1st cranial nerve was not tested, and
the 2nd through 12th nerves were intact. Strength
was intact, and the gait was unsteady. The remainВ­
der of the examination was normal.
m e dic i n e
Results of laboratory tests are shown in Table 1.
After premedication with lorazepam at a dose of
1 mg to control involuntary movements, magnetВ­
ic resonance imaging (MRI) of the brain was
performed without administration of contrast
material. On T2-weighted, fluid-attenuated inverВ­
sion recovery (FLAIR) images, hyperintense signal
was seen in the left hippocampus and parahippoВ­
campal gyrus as well as in the posterior limb of
the left internal capsule. There was no evidence of
restricted diffusion.
Examination by a neurology consultant showed
that the patient was oriented to person and place,
with a childlike affect. Her speech was fluent,
and naming was intact. She could read a short
sentence and do simple addition. She was lefthanded, could write her name but not a sentence,
and followed simple and complex commands.
Her attention was variable, and testing of her
memory showed recollection of zero of three
items at 5 minutes on repeated examination.
There was mild asymmetry of the face with flattening of the right nasolabial fold. Smell and
taste were not tested. The function of the other
cranial nerves was intact. There were choreiform
movements of the head and neck, poor performance of rapid alternating movements, and apraxВ­
ia. Hypertonia and hyperreflexia with clonus were
noted in the right leg. The gait was wide-based,
with postural instability and leaning toward the
left. She was unable to stand on one foot. She was
admitted to the neurology service.
On the second hospital day, a lumbar puncture was performed. Results of cerebrospinal
fluid analysis are shown in Table 2; other test
results are listed in Table 1. An enzyme-linked imВ­
munosorbent assay for serum antibodies against
HIV was negative. An electroencephalogram
showed diffuse theta slowing and frontal intermittent rhythmic delta activity, which was more
prominent in the right hemisphere than in the
left. There was no epileptiform activity (Fig. 1).
Repeated MRI of the brain after the administration of gadolinium showed no changes and no
evidence of abnormal enhancement. Acyclovir was
administered intravenously.
The next day, a serum Lyme antibody test, a test
of a throat swab for Mycoplasma pneumoniae nucleic acid, and cultures of blood and urine were
negative; results of other tests are listed in Table 1.
On the fifth day, the patient’s condition appeared
to be improved. She was oriented and remembered details of her past; dysmetria and truncal
n engl j med 357;6  www.nejm.org  august 9, 2007
case records of the massachuset ts gener al hospital
ataxia were reduced. Results on a repeated elecВ­
troencephalogram were unchanged. The next
day, a repeated lumbar puncture was performed
(Table 2).
Between the 7th and 18th hospital days, the
patient’s motor function gradually worsened,
right-sided neglect developed, she became unable
to feed herself, her responsiveness and ability to
follow commands decreased, and she became
incontinent. She began lying in a fetal position,
moaning and crying out unintelligible sounds.
A skin test for tuberculosis, a test of a nasopharyngeal specimen for respiratory viral antigens,
and a viral culture of a stool specimen were negative. Nucleic acid testing for HIV RNA and tests
for antinuclear antibodies were negative. Levels
of free and total thyroxine were normal, and the
thyroglobulin level was elevated (54.7 ng per milliВ­
liter; normal range, 4 to 40). On the 12th day, the
acyclovir was discontinued, and ceftriaxone, at
a dose of 2 g, was administered intravenously.
A repeated electroencephalographic study showed
increased attenuation of background activity and
less abundant frontal intermittent rhythmic delta
activity. MRI on the 13th day showed new hyperintense signal in the pons and middle cerebellar
peduncles with associated restricted diffusion of
water on T2-weighted FLAIR images. Restricted
diffusion was also noted in the posterior limb of
the left internal capsule. There was atrophy in the
left medial temporal lobe, with resolution of the
abnormal hyperintense signal on FLAIR images.
On the 14th day, a third lumbar puncture was
On the 18th hospital day, a test result was
Differ en t i a l Di agnosis
Dr. Andrew J. Cole: I was involved in this patient’s
care from the time of her admission and am therefore aware of the diagnosis. I will discuss the
case as it unfolded in order to illustrate the diagnostic process and therapeutic decision making
that took place. The patient lived semi-independently until she became pregnant 27 weeks before
admission. Her level of function at that time was
unknown, and because of the lack of information, it was not possible to determine either her
level of function before her illness or the tempo
of her disease.
Neurologic differential diagnosis relies primarily on the physical examination for localization
of lesions and on the history, especially the nature
of onset and pace of progression, to identify the
disease process. This patient’s neurologic examination showed abnormal cognitive function indicating dysfunction of the cortical and subcortical
gray matter, abnormal motor function indicating dysfunction of the pyramidal motor system,
and choreiform movements indicating dysfunction of the extrapyramidal motor systems. This
examination also showed a clumsy gait and difficulty performing rapid alternating movements,
indicating dysfunction of the cerebellum or its
connections. With the limited information about
the pace of her disease, we needed to consider
inherited, congenital, and acquired diseases that
could be acute, subacute, or chronic, with static,
episodic, or proВ­gressive tempos. We had to base
our differential diagnosis on the neurologic examination, initial laboratory testing, and electroencephalographic and MRI studies.
Cerebrospinal Fluid Examination
The results of the cerebrospinal fluid analysis in
this patient showed a lymphocytic pleocytosis
with few red cells, a mildly elevated protein level,
and a normal glucose level. These findings are
characteristic of aseptic meningitis. We were thus
concerned about viruses, rickettsia, spirochetes,
partially treated bacterial infection, a parameninВ­
geal focus of infection, certain autoimmune illnesses such as systemic lupus erythematosus or
Behçet’s disease, vasculitides, carcinoma, a reaction to the toxic effects of certain medications
such as nonsteroidal antiinflammatory drugs,
and chemical meningitis related to the rupture
of a cyst. Although they were nonspecific, the
cerebrospinal fluid findings provided support for
the possibility of acute or subacute infection or
inflammatory illness. The presence of an inflammatory response made chronic degenerative diseasВ­
es such as Huntington’s disease, Wilson’s disease,
and systems abiotrophies such as multisystem
atrophy unlikely.
Electroencephalographic Studies
The initial electroencephalogram was markedly
abnormal, but the findings were nonspecific
(Fig. 1). The slow and attenuated posterior dominant rhythm suggests cortical gray-matter disease, whereas the intermittent frontal rhythmic
delta activity suggests subcortical gray-matter
disease. The monomorphic slow waves also suggest that initially the subcortical white matter
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
Table 1. Results of Laboratory Tests.*
Reference Range
for Adults†Variable
On Admission
Hematocrit (%)
36.0–46.0 (in women)
Hemoglobin (g/dl)
12.0–16.0 (in women)
White-cell count (per mm3)
On Hospital
Day 2
Differential count (%)
Platelet count (per mm3)
Mean corpuscular volume (Ојm3)
Erythrocyte sedimentation rate (mm/hr)
Glucose (mg/dl)
Sodium (mmol/liter)
Potassium (mmol/liter)
Chloride (mmol/liter)
Carbon dioxide (mmol/liter)
Phosphorus (mg/dl)
Magnesium (mmol/liter)
Calcium (mg/dl)
Urea nitrogen (mg/dl)
Creatinine (mg/dl)
Bilirubin (mg/dl)
Protein (g/dl)
Creatine kinase (U/liter)
40–150 (in women)
Alkaline phosphatase (U/liter)
Aspartate aminotransferase (U/liter)
Alanine aminotransferase (U/liter)
Lipase (U/dl)
Amylase (U/liter)
Rapid plasma reagin
Human chorionic gonadotropin, quantitative
<6 (in nonpregnant women);
6–15 (borderline)
Toxicology screen
Partial-thromboplastin time lupus anticoagulant
Anticardiolipin IgG antibodies (GPL units)
n engl j med 357;6  www.nejm.org  august 9, 2007
case records of the massachuset ts gener al hospital
Table 1. (Continued.)
Reference Range
for Adults†Variable
Anticardiolipin IgM antibodies (MPL units)
On Admission
On Hospital
Day 2
Ceruloplasmin (mg/dl)
Iron (Вµg/dl)
Iron-binding capacity (Вµg/dl)
Vitamin B12 (pg/ml)
Ferritin (ng/ml)
Transferrin (mg/dl)
Antistreptolysin O (IU/ml)
Thyrotropin (ВµU/ml)
Serum protein electrophoresis (mg/dl)
Herpes simplex virus type 1 antibody IgG
Positive (>6.00)
Herpes simplex virus type 2 antibody IgG
*To convert the values for glucose to millimoles per liter, multiply by 0.05551. To convert the values for urea nitrogen
to millimoles per liter, multiply by 0.357. To convert the values for creatinine to micromoles per liter, multiply by 88.4.
To convert the values for total and direct bilirubin to micromoles per liter, multiply by 17.1. To convert the values for
phosphorus to millimoles per liter, multiply by 0.3229. To convert the values for magnesium to milliequivalents per
В­liter, multiply by 2. To convert the values for calcium to millimoles per liter, multiply by 0.250. To convert the values
for iron and iron-binding capacity to micromoles per liter, multiply by 0.1791. To convert the values for vitamin B12 to
picomoles per liter, multiply by 0.7378.
†Reference values are affected by many variables, including the patient population and the laboratory methods used.
The ranges used at Massachusetts General Hospital are for adults who are not pregnant and do not have medical
В­conditions that could affect the results. The ranges therefore may not be appropriate for all patients.
was relatively spared. Dr. Henson, may we review
the radiologic studies?
Dr. John W. Henson: Axial T2-weighted FLAIR
images from the MRI studies of the brain on the
day of admission, performed without the administration of gadolinium, revealed a region of hyperintense signal in the left medial temporal lobe
(Fig. 2A) and subtle increased signal in the posterior limb of the left internal capsule corresponding to the location of the corticospinal
tract. These foci did not show restricted diffusion or abnormal enhancement on a gadoliniumenhanced study performed the next day. The appearance of the pons was unremarkable, and no
other clinically significant findings were noted.
Magnetic resonance venography of the head was
By day 13, there had been marked changes.
There was a region of abnormal signal in the
pons (Fig. 2B), with areas of restricted diffusion
on the diffusion-weighted image and apparent-
diffusion-coefficient maps. The hyperintensity of
the left medial temporal lobe had resolved, and
there was volume loss in the region of the hippoВ­
campal formation. There was restricted diffusion
in the left corticospinal tract (Fig. 2C); no abnormal enhancement was detected. These findings
were interpreted as resulting from a subacute enВ­
cephalitis caused by an infection or an autoimmune disorder.
Dr. Cole: In summary, this patient has a disturbance of cognitive function, pyramidal tract and
cerebellar dysfunction, and a choreiform-movement disorder, and both laboratory tests and electroencephalographic and imaging studies suggest
an infectious or autoimmune encephalitis.
Disorders of Movement
Chorea is a hyperkinetic movement disorder that
may result from a number of neurologic diseases;
it may appear or worsen during pregnancy, a condition known as chorea gravidarum. Most patients
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
Table 2. Results of Cerebrospinal Fluid Tests.
Normal Range
Opening pressure (mm H20)
Hospital Day 2
Hospital Day 6
Hospital Day 14
Pink, slightly turbid
Slightly pink, clear
Colorless, slightly
Red-cell count (per mm3)
Tube 1
Tube 4
Tube 1
Tube 4
Tube 1
Tube 4
Tube 1
Tube 4
Tube 1
Tube 4
Tube 1
Tube 4
White-cell count (per
Differential count (%)
Reactive lymphocytes
Other hematic cells (%)
Large mononuclear cells
with abundant cytoplasm and nucleoli
Large mononuclear cells
with basophilic cytoplasm and prominent nucleoli
Tube 1
Tube 4
Tube 1
Tube 4
Unidentified cells (%)
Protein (mg/dl)
Glucose (mg/dl)
Venereal Disease Research Laboratory test
IgG (mg/dl)
Albumin (mg/dl)
Oligoclonal bands on agarose
None seen in 80 ×
Several seen in 63Г—
Gram’s stain
No organisms
Acid-fast bacilli stain
No organisms
No organisms
Varicella–zoster virus (PCR)
None detected
None detected
Enterovirus RNA (PCR)
None detected
Cytomegalovirus DNA
None detected
None detected
Epstein–Barr virus DNA
None detected
None detected
No organisms
n engl j med 357;6  www.nejm.org  august 9, 2007
No organisms
case records of the massachuset ts gener al hospital
Table 2. (Continued.)
Normal Range
Hospital Day 2
Human herpes virus type 6 DNA (PCR)
Herpes simplex virus (PCR)
Hospital Day 6
Hospital Day 14
None detected
None detected
None detected
None detected
None detected
Mycoplasma pneumoniae (PCR)
Encephalitis antibodies
Eastern equine encephalitis
None detected
West Nile virus IgM
None detected
No growth
No growth
No growth
No growth
No growth
No growth
No growth
No growth
*PCR denotes polymerase chain reaction.
with this condition present during the second
trimester with an isolated movement disorder that
resolves after delivery. The most common causes
are acute rheumatic fever (Sydenham’s chorea) and
the antiphospholipid-antibody syndrome. We also
considered other illnesses associated with chorea,
including systemic lupus erythematosus, HuntingВ­
ton’s chorea, and Wilson’s disease, although the
latter two illnesses were ruled out by the cerebrospinal fluid and other findings.
Sydenham’s chorea is a late complication of
infection with group A streptococcus; the onset
occurs months after acute infection. Most cases
occur in childhood, but up to 30% of patients
may have recurrent chorea months or years after
the initial episode.1 This patient had only a minimally elevated antistreptolysin-antibody titer, with
no other evidence of recent streptococcal infection
or cardiac disease. The antiphospholipid-antibody
syndrome2 may be primary or secondary to systemic lupus erythematosus, and it may beВ­come
manifest during pregnancy. This patient had no
symptoms or signs of systemic lupus, but lupus
confined to the central nervous system is well recognized and may worsen during the course of
pregnancy. Antiphospholipid-antibody testing and
all laboratory studies for lupus were negative.
Acute Viral Encephalitis
Herpes simplex encephalitis was initially considered as one of the acute infectious encephalitides
because of the abnormality detected in the left
hippocampus on MRI. Unlike the arboviral encephalitides and West Nile virus encephalitis,
which occur in the summer and fall, when mosquitoes are abundant, herpes simplex encephalitis occurs sporadically throughout the year. The
negative results on cerebrospinal fluid testing for
herpes simplex virus nucleic acid and the lack of
response to acyclovir made this diagnosis unlikely. Other common causes of viral encephalitis,
including enteroviral and echoviral infections, as
well as infection with coxsackievirus, typically proВ­
Figure 1. Electroencephalogram Obtained on the Second
Hospital Day.
The electroencephalogram is markedly abnormal,
Cole of the posterior RETAKE
rhythm1 of
bursts of frontal intermittent
activity (arrows), sometimes maximal
CASE delta
on the
relatively symmetric
bilatEMailright and other timesLine
are no mst
epileptiform features and no
ic discharges.
Figure has been redrawn and type has been reset.
Please check carefully.
n engl j med 357;6  www.nejm.org 
357069, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
Figure 2. Brain Imaging Studies.
An axial T2 -weighted FLAIR image (Panel A) obtained
on the day of admission showed a region of abnormal,
hyperintense signal in the left hippocampus (arrow) and
in the posterior limb of the left internal capsule (not
shown). On the 13th hospital day, there was a new region of В­hyperintense signal on T2-weighted FLAIR images
in the pons and middle cerebellar peduncles (Panel B,
arrow) and restricted diffusion in the middle cerebellar
peduncle and posterior limb of the internal capsule as
shown on the diffusion-weighted images (Panel C, arrow).
Subacute Encephalitides
Paraneoplastic Encephalitis
Paraneoplastic limbic encephalitis may precede
the appearance of a tumor by months or even years.
Psychiatric symptoms, memory failure, confusion
and drowsiness, disordered respiration, ataxia, and
cranial-nerve palsies have been reported. There
may be a modest increase in protein in the cerebrospinal fluid, but few cells are detected. This
patient’s clinical and cerebrospinal fluid findings
were not consistent with paraneoplastic encephalitis.
Postinfectious Encephalitis
duce prominent signs of meningeal irritation, with
photophobia, meningismus, nausea, and headAUTHOR
Cole signs of focalRETAKE
ache,ICMbut only
cerebral1stdysREG F which
2a-c ofbe
4 fleeting and are not2nd
Several infections may be associated with postinfectious encephalitis syndromes, including measles, mumps, and rubella; influenza; Epstein–Barr
virus; and varicella–zoster virus. Acute disseminated encephalomyelitis may occur after a variety
of viral infections and after the administration of
rabies and smallpox vaccines.3 It usually begins
with nonspecific symptoms such as fever, headache, stiff neck, vomiting, and anorexia. Neurologic examination may show optic neuritis, ataxia,
and focal weakness; seizures and decreased consciousness may develop. This patient did not have
a recent history of immunizations or a viral infection or evidence of optic neuritis, and the imaging findings were not typical of acute disseminatВ­
ed encephalomyelitis.
The cerebrospinal fluid findings were an
important clue to the diagnosis in this case. Although the protein level in the initial cerebrospinal fluid specimen was modestly elevated, the IgG
component was markedly elevated. Dr. Roehrl,
would you discuss the analysis and implications
of this finding?
Dr. Michael H.A. Roehrl: Cerebrospinal fluid levels of total protein, albumin, and IgG obtained
on the second hospital day are shown in Table 2.
Figure has been redrawn and type
has been
n engl
j med
357;6  www.nejm.org  august 9, 2007
Please check carefully.
case records of the massachuset ts gener al hospital
Agarose-gel electrophoresis of cerebrospinal fluid
revealed oligoclonal bands in the gamma region
(Fig. 3A). The cerebrospinal fluid to serum mass
concentration quotients for albumin and IgG
con­centrations were QAlb = 4.7×10 −3 and QIgG = 
26.1Г—10в€’3, corresponding to an IgG index of 5.6
(normal value, <0.85).4 On the basis of an analysis developed by Reiber,5‑7 this patient’s results
(Fig. 3B) indicated markedly increased intrathecal
IgG synthesis (intrathecal proВ­duction fraction,
87.7%) without evidence of clinВ­ically significant
blood–brain barrier dysfunction. The mass concentration ratio of IgG to total protein in the cerebrospinal fluid was 58.9%.
Dr. Cole: The diseases that can elicit an intrathecal response of this magnitude are syphilis,
chronic rubella panencephalitis, and subacute
sclerosing panencephalitis.8-11 This patient did
not have syphilis, as shown by negative results on
serologic testing and a symptom complex that
was inconsistent with the disease. Postrubella enВ­
cephalitis12,13 may affect patients with history of
remote or congenital rubella infection and pres-
A О±1 О±2
Normal cerebrospinal
Patient’s cerebrospinal
ea igh
Q Q %
Figure 3. Results of Cerebrospinal Fluid Electrophoresis.
Panel A shows the results of agarose-gel electrophoresis
of a specimen of the patient’s cerebrospinal fluid collect­
ed on the second hospital day (concentrated to 1/63 of
the original volume) and cerebrospinal fluid from a normal control (concentrated to 1/80 of the original volume).
P, A, О±1, О±2, ОІ, and Оі denote the electrophoretic preВ­
albumin, albumin, alpha-1, alpha-2, beta, and gamma
regions, respectively. The arrow shows the position of
several strong bands in the gamma region, indicating
the presence of multiple oligoclonal immunoglobulins.
There is also a relative decrease of the level of albumin
in the patient’s cerebrospinal fluid. Panel B shows a
double-logarithmic graph designed according to the
method proposed by Reiber (also called a Reibergram),
in which cerebrospinal fluid to serum mass concentration quotients for albumin (QAlb) and IgG (Q IgG) are
plotted along the abscissa and ordinate, respectively.
Upper limits (Q High) and lower limits (QLow) of normal
values are shown as solid lines, with the dotted line indicating mean normal values (QMean). The isopercentiles
(dashed lines) correspond to various relative amounts
of intrathecal IgG production (20 to 80%). The vertical
dashed line denotes the age-adjusted upper limit of
QAlb, separating normal function (left) and abnormal
function (right) of the blood–brain barrier. The patient
(red dot) had markedly increased intrathecal IgG synthesis
(intrathecal production fraction, 87.7%) without evidence
of significant dysfunction of the blood–brain barrier
(zone 4). Zone 1 denotes normal function, zone 2 denotes pure blood–brain barrier dysfunction, and zone 3
denotes a combination of increased intrathecal IgG
synthesis and blood–brain barrier dysfunction.
ents with progressive dementia, ataxia, chorea,
ICM degeneration, and seizures. Examination
FIGURE: 3 of 4
of the cerebroВ­spinal fluid shows pleocytosis and
a moderately
elevated protein
50% of theARTIST:
fluid protein
of immunoglobulins. TheCombo
diagnosis is confirmed
by a high
titer in the cereFigure has been redrawn and type has been reset.
brospinal fluid. Please check carefully.
JOB: 35706Sclerosing Panencephalitis
Measles causes three distinct diseases of the central nervous system: postinfectious encephalomyelitis, subacute measles encephalitis, and subВ­acute
sclerosing panencephalitis.14 Subacute sclerosing panencephalitis is typically seen 7 to 10 years
after infection with measles, and patients present
most commonly with declining performance in
school, behavioral changes, headache, adВ­ventitious
movements, and sometimes seizures.15 Characteristic findings include myoclonic jerks that are
often periodic and are associated with periodic
lateralized or bilateral epileptiform discharges on
an electroencephalogram. Although most cases
occur in childhood or adolescence, cases beginning as late as the fifth decade of life have been
described.16,17 The incidence of subacute scleros-
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
ing panencephalitis has decreased with widespread vaccination against measles; howВ­ever, it
persists in places where measles vaccination is
uncommon.18 The incidence of this condition is
increased as much as 10 times in patients in
whom measles develops before the age of
2 years; this patient had measles at 4 months of
age. The illness may present during pregnancy,
possibly as a result of altered immune status.8,19
In summary, this patient presented with a
subacute progressive neurologic disease characterized by widespread dysfunction of the central
nervous system, inflammatory features in the
cerebrospinal fluid, and an extremely high level
of cerebrospinal fluid IgG. My colleagues and I
favored the diagnosis of subacute sclerosing panencephalitis, a delayed consequence of her infection with measles at 4 months of age. This condition may have been exacerbated by her pregnancy.
Specimens of serum and cerebrospinal fluid from
the 14th hospital day were sent for testing of
levels of antibodies against measles.
Dr. Roehrl: The measles-specific cerebrospinal
fluid to serum IgG antibody index was elevated at
31.8 (normal value, <1.4)7 (Table 3). Measles-specific IgM antibodies were not detectВ­ed in the serum. This profile indicates a chronic immune response to measles infection in the cerebrospinal
fluid compartment. A comparison with mumps
titers shows the specificity of the immunologic
process. The results confirm the diagnosis of subacute sclerosing panencephalitis in the patient.
The disease was a late consequence of persistent
infection with measles virus.
Dis cus sion of M a nage men t
Dr. Cole: Cell-mediated immunity stimulated by
T-helper cell type 1 (Th1) inducing cytokines is
crucial for the clearance of measles virus in the
weeks after infection, whereas cytokines that induce type 2 helper T (Th2) cells are implicated in
antibody production. Several reports have suggested that treatment with intrathecal interferon
alfa-2 (a cytokine that promotes Th1 activity),
with or without treatment with the antiviral immunomodulatory agent inosine pranobex, may
slow or even arrest the progression of subacute
sclerosing panencephalitis.20-25 However, no definitive data are available to show the efficacy of
this treatment approach.
When we made the diagnosis in this patient,
m e dic i n e
Table 3. Results of Blood and Cerebrospinal Fluid Tests
for Measles and Mumps Antibodies on Hospital Day 14.
she was at 28 weeks of gestation, and her condition was deteriorating rapidly. We believed that an
attempt at treatment was in the best interest of
the patient and her fetus. We treated her with
interferon and inosine pranobex for 8 weeks,
without any clear clinical benefit. After the delivery of a healthy baby by elective cesarean section at 34 weeks’ gestation, a decision was made
to discontinue treatment, with the consent of the
patient’s mother. The patient died 6 weeks later.
Subacute sclerosing panencephalitis.
Pathol o gic a l Dis cus sion
Dr. Matthew P. Frosch: The autopsy revealed inflammatory infiltrates containing macrophages, plasma cells, and lymphocytes around vessels, with
neuronal destruction and reactive gliosis that were
most prominent in the brain stem (Fig. 4A). No
viral inclusions of the type usually seen in acute
measles encephalitis (and occasionally present in
subacute sclerosing panencephalitis) were seen in
the nuclei or cytoplasm of any cell types (neurons,
glia, or vascular endothelial cells). Unlike other
measles-associated diseases, subacute sclerosing
panencephalitis is caused by the persistence of
defective viruses that do not form complete viral
particles but may infect adjacent cells by direct
contact.26,27 In the cerebral cortex, neuronal populations were preserved, but there was marked
gliosis with reactive astrocytes (Fig. 4B). The brain
stem was softened by the pathologic process, but
the hemispheric white matter was firmer than
normal; this was most evident when the brain
was cut in the fresh state. This firmness corresponds to the sclerosis in the name, subacute
n engl j med 357;6  www.nejm.org  august 9, 2007
case records of the massachuset ts gener al hospital
nonspecific, but together with the serologic studies, they are consistent with a diagnosis of subA
acute sclerosing panencephalitis.
Fixed and frozen specimens of brain tissue
were sent to Dr. William Bellini at the Centers for
Disease Control and Prevention for immunohistochemical and molecular diagnostic studies.14,28‑30
Immunohistochemical analysis for measles virus
nucleoprotein antigen was negative. Molecular
diagnostic studies performed to detect portions
of the measles genome were negative in repeated
attempts at amplification of regions of the
measles nucleoprotein gene. Weak signals were
detected in a real-time reverse-transcriptase–
polymerase-chain-reaction assay, but there were
insufficient amounts of amplified DNA product
to perform a sequence analysis. As a result, it was
not possible to definitively establish the presence
of measles virus in this case.
Dr. Nancy Lee Harris (Pathology): Dr. Cort, who
prepares the case histories for these exercises,
was able to obtain some additional history, which
was not available to the patient’s caregivers.
Dr. Alice M. Cort (Internal Medicine): Two years
before admission, the patient was seen in the
emergency department of another hospital because of an episode of loss of consciousness that
was associated with rolling of the eyes and arching of the back. On examination, there were intermittent involuntary movements of the head and
neck and jerking movements of the arms. Deeptendon reflexes were brisk and symmetric. CT
scanning of the head and an electroencephalogram were normal. Further testing showed an
early pregnancy, which was electively terminated.
During the next 20 months, the patient was lost
to medical follow-up, and it is not known whether
Figure 4. Findings in the Brain at Autopsy.
the abnormal movements stopped.
A marked inflammatory infiltrate around the vessels and
Dr. Harris: Dr. Cole, is it possible that subacute
neuronal destruction are most prominent in this section
RETAKE plasma
Cole contains macrophages,
ICM brain
sclerosing panencephalitis began during this paof the
stem, which
F lymphocytes;
FIGURE 4a-c ofthere
is prominent perivascular
tient’s previous pregnancy, stabilized, and then
lymphocytes (Panel A, hematoxylinRevised
and eosin).
worsened with the second pregnancy?
4-C are preserved,
In the
cerebral cortex, neuronal
Dr. Cole: Subacute sclerosing panencephalitis
H/T (Panel
but a marked
is present
B, immunohisto16p6
always a chronically progressive disease,
chemical analysis for glial fibrillary acidic protein). Marked
although plateaus in the clinical course have been
microglial activation
is present
the subcortical white
Figure has been redrawn and type has been reset.
matter with rod-shaped
described. The duration of the disease can be
Please check
carefully. cells (Panel C,
arrows; hematoxylin and eosin) (inset, immunohistoyears, however, and in retrospect, her symptoms
JOB: staining
35706 for CD68).
ISSUE: 8-9-07
2 years earlier may have been an earlier manifestation of the same disease. The effect of the pasclerosing panencephalitis. The subcortical white tient’s treatment on the appearance of the brain
matter also showed reactive gliosis and extensive at autopsy and on the ability to detect viral antimicroglial activation (Fig. 4C). These findings are gens or DNA is not known.
n engl j med 357;6  www.nejm.org  august 9, 2007
case records of the massachuset ts gener al hospital
A nat omic a l Di agnosis
Subacute sclerosing panencephalitis secondary
to measles virus infection.
Dr. Cole reports receiving consulting fees from GlaxoSmithKline,
Abbott Laboratories, and Supernus Pharmaceuticals and lecture fees
from GlaxoSmithKline, Abbott Laboratories, and Ortho-McNeil;
1. al-Eissa A. Sydenham’s chorea: a new
look at an old disease. Br J Clin Pract 1993;
2. Levine JS, Branch DW, Rauch J. The
antiphospholipid syndrome. N Engl J Med
3. Menge T, Hemmer B, Nessler S, et al.
Acute disseminated encephalomyelitis:
an update. Arch Neurol 2005;62:1673-80.
4. Thompson EJ, Riches PG, Kohn J. AntiВ­
body synthesis within the central nervous
system: comparisons of CSF IgG indices
and electrophoresis. J Clin Pathol 1983;36:
5. Reiber H. The discrimination between
different blood-CSF barrier dysfunctions
and inflammatory reactions of the CNS
by a recent evaluation graph for the protein profile of cerebrospinal fluid. J Neurol 1980;224:89-99.
6. Idem. Flow rate of cerebrospinal fluid
(CSF) — a concept common to normal
blood-CSF barrier function and to dysfunction in neurological diseases. J Neurol Sci
7. Reiber H, Peter JB. Cerebrospinal fluid analysis: disease-related data patterns
and evaluation programs. J Neurol Sci
8. Case Records of the Massachusetts
General Hospital (Case 15-1998). N Engl J
Med 1998;338:1448-56.
9. Fishman RA. Cerebrospinal fluid in
diseases of the nervous system. 2nd ed.
Philadelphia: W.B. Saunders, 1992.
10. Norrby E, Vandvik B. Relationship between measles virus-specific antibody activities and oligoclonal IgG in the central
nervous system of patients with subacute
sclerosing panencephalitis and multiple
sclerosis. Med Microbiol Immunol 1975;
11. Wolinsky JS. Progressive rubella panencephalitis. In: Vinken PJ, Bruyn GW, Klawans HL, McKendall RR, eds. Viral disease. Amsterdam: Elsevier, 1989:405-16.
Dr. Henson, consulting fees from GlaxoSmithKline; and Dr. Frosch,
consulting fees from Biogen Idec and Bristol-Myers Squibb. No
other potential conflict of interest relevant to this article was reported.
We thank Dr. Mandakolathur R. Murali, director of the Clinical
Immunology Laboratory, Department of Pathology, Massachusetts
General Hospital, for the analysis of the immunologic findings
in the cerebrospinal fluid.
12. Townsend JJ, Baringer JR, Wolinsky JS,
et al. Progressive rubella panencephalitis:
late onset after congenital rubella. N Engl
J Med 1975;292:990-3.
13. Weil ML, Itabashi H, Cremer NE, OshiВ­
ro L, Lennette EH, Carnay L. Chronic progressive panencephalitis due to rubella
virus simulating subacute sclerosing panencephalitis. N Engl J Med 1975;292:994-8.
14. Honarmand S, Glaser CA, Chow E, et
al. Subacute sclerosing panencephalitis in
the differential diagnosis of encephalitis.
Neurology 2004;63:1489-93.
15. Swoveland PT, Johnson KP. Subacute
sclerosing panencephalitis and other paramyxovirus infections. In: Vinken PJ, Bruyn
GW, Klawans HL, McKendall RR, eds. Viral
disease. Amsterdam: Elsevier, 1989:417.
16. Gagnon A, Bouchard RW. Fulminating
adult-onset subacute sclerosing panencephalitis in a 49-year-old man. Arch Neurol 2003;60:1160-1.
17. Prashanth LK, Taly AB, Ravi V, Sinha S,
Arunodaya GR. Adult onset subacute sclerosing panencephalitis: clinical profile of
39 patients from a tertiary care centre. J NeuВ­
rol Neurosurg Psychiatry 2006;77:630-3.
18. Bellini WJ, Rota JS, Lowe LE, et al.
Subacute sclerosing panencephalitis: more
cases of this fatal disease are prevented by
measles immunization than was previously
recognized. J Infect Dis 2005;192:1686-93.
19. Wirguin I, Steiner I, Kidron D, et al.
Fulminant subacute sclerosing panencephВ­
alitis in association with pregnancy. Arch
Neurol 1988;45:1324-5.
20. Gokcil Z, Odabasi Z, Demirkaya S,
Eroglu E, Vural O. Alpha-interferon and
isoprinosine in adult-onset subacute sclerosing panencephalitis. J Neurol Sci 1999;
21. Anlar B, Yalaz K, Oktem F, Köse G.
Long-term follow-up of patients with subacute sclerosing panencephalitis treated
with intraventricular alpha-interferon.
Neurology 1997;48:526-8.
22. Gascon G, Yamani S, Crowell J, et al.
Combined oral isoprinosine-intraventricular alpha-interferon therapy for subacute
sclerosing panencephalitis. Brain Dev 1993;
23. Yalaz K, Anlar B, Oktem F, et al. Intraventricular interferon and oral inosiplex in
the treatment of subacute sclerosing panencephalitis. Neurology 1992;42:488-91.
24. Steiner I, Wirguin I, Morag A, AbramВ­
sky O. Intraventricular interferon treatment for subacute sclerosing panencephalitis. J Child Neurol 1989;4:20-4.
25. Panitch HS, Gomez-Plascencia J, Norris FH, Cantell K, Smith RA. Subacute
sclerosing panencephalitis: remission after
treatment with intraventricular interferon.
Neurology 1986;36:562-6.
26. Ueda S, Okuno Y, Hamamoto Y, Oya H.
Subacute sclerosing panencephalitis (SSPE):
isolation of a defective variant of measles
virus from brain obtained at autopsy.
Biken J 1975;18:113-22.
27. Hirano A, Ayata M, Wang AH, Wong
TC. Functional analysis of matrix proteins
expressed from cloned genes of measles
virus variants that cause subacute sclerosing panencephalitis reveals a common defect in nucleocapsid binding. J Virol 1993;
28. Zaki SR, Bellini WJ. Measles. In: Connor DH, Chandler FW, Schwartz DA, Manz
HJ, Lack EE, eds. Pathology of infectious
diseases. Stamford, CT: Appleton & Lange,
29. Bellini WJ, Rota JS, Lowe LE, et al.
Subacute sclerosing panencephalitis: more
cases of this fatal disease are prevented by
measles immunization than previously recognized. J Infect Dis 2005;192:1686-93.
30. Hummel KB, Lowe L, Bellini WJ, Rota
PA. Development of quantitative genespecific real-time RT-PCR assays for the deВ­
tection of measles virus in clinical specimens. J Virol Methods 2006;132:166-73.
Copyright В© 2007 Massachusetts Medical Society.
Lantern Slides Updated: Complete PowerPoint Slide Sets from the Clinicopathological Conferences
Any reader of the Journal who uses the Case Records of the Massachusetts General Hospital as a teaching exercise or reference material is now eligible to
receive a complete set of PowerPoint slides, including digital images, with identifying legends, shown at the live Clinicopathological Conference (CPC)
that is the basis of the Case Record. This slide set contains all of the images from the CPC, not only those published in the Journal. Radiographic, neuroВ­
logic, and cardiac studies, gross specimens, and photomicrographs, as well as unpublished text slides, tables, and diagrams, are included. Every year 40
sets are produced, averaging 50-60 slides per set. Each set is supplied on a compact disc and is mailed to coincide with the publication of the Case Record.
The cost of an annual subscription is $600, or individual sets may be purchased for $50 each. Application forms for the current subscription year, which
began in January, may be obtained from the Lantern Slides Service, Department of Pathology, Massachusetts General Hospital, Boston, MA 02114
(telephone 617-726-2974) or e-mail [email protected]
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
edi t or i a l s
Safer Drugs for the American People
Gregory D. Curfman, M.D., Stephen Morrissey, Ph.D., and Jeffrey M. Drazen, M.D.
By wide margins, both the House and the Senate
have now passed bills that aim to ensure the safety
of the drug supply in the United States.1,2 Given
the serious safety problems that have arisen with
drugs taken by millions of Americans, this legislation is long overdue. It is now up to both houses of Congress to resolve the differences between
the two versions and agree on a strong final bill
to send to the President.
This year, Congress had to consider reauthorization of the Prescription Drug User Fee Act
(PDUFA), which it must do every 5 years. Under
PDUFA, the major funding for the review of new
drugs by the Food and Drug Administration (FDA)
comes from user fees paid by pharmaceutical
companies. This mechanism has been controversial because, although it was designed to accelerate the drug-approval process and can make new
drugs available to patients without delay, it has
directed no money to the postmarketing assessment of drug safety. In addition, some believe that
user fees pose a conflict of interest for the FDA.
Nonetheless, to ensure that the FDA has adequate
financial resources, both the House and the Senate versions of the bill not only maintain these
user fees but increase them (with some of the
money now directed to safety assessment).
In reauthorizing PDUFA, Congress has wisely
taken the opportunity to strengthen the FDA’s authority over drug safety, and both versions of the
bill give the FDA new tools to accomplish this objective. At a minimum, we believe that the final
bill should contain the following essential components.
The FDA must have the authority to mandate
adequately powered postmarketing clinical trials
of the safety of approved drugs and to require specific timetables for their completion and reporting of results.
The FDA must also have the authority to conduct an annual review of drug safety for the first
3 years after a drug’s approval and again at 7 years.
To assist patients in reporting adverse effects
of the drugs they are taking, drug advertisements
and labels must include a toll-free telephone number and a Web address.
There must be substantial penalties for drug
advertising that overstates efficacy or understates
adverse effects.
The FDA must be able to mandate changes to
drug labels as new information about safety and
efficacy becomes available.
In order to provide safety surveillance, every
new drug should have a pharmacovigilance plan
at the time of its approval. Sophisticated pharmacoepidemiologic surveillance systems must be created to allow the FDA to monitor reports of adverse drug effects in large databases.
Clinical trials of drugs must be registered in a
public database.
In the wake of the landmark report on drug
safety issued by the Institute of Medicine almost
a year ago,3 Congress is now poised to pass the
most important drug-safety legislation in a century. Although we remain concerned about the
user-fee approach to funding of the FDA, we believe that the bills initiated by Congressmen
Henry Waxman and Edward Markey and by Senators Edward Kennedy and Michael Enzi are important steps in the right direction. Congress
should now speedily agree on a substantive drugsafety bill and give the FDA the tools it needs to
do its work. The crisis of confidence in the nation’s drug supply must be resolved.
This article (10.1056/NEJMe078154) was published at www.
nejm.org on July 18, 2007.
1. H.R. 2900, 110th Cong. (2007).
n engl j med 357;6  www.nejm.org  august 9, 2007
2. S. 1082, 110th Cong., 1st Sess. (2007).
3. Committee on the Assessment of the US Drug Safety System,
Baciu A, Stratton K, Burke SP, eds. The future of drug safety:
promoting and protecting the health of the public. Washington,
DC: National Academies Press, 2006. (Available at http://www.
Copyright В© 2007 Massachusetts Medical Society.
Prophylactic Treatment for Prevention of Joint Disease
in Hemophilia — Cost versus Benefit
Goris Roosendaal, M.D., Ph.D., and Floris Lafeber, Ph.D.
Patients with severe hemophilia number only
around 400,000 worldwide, but their lifespan has
increased because of improved treatment, and with
that the prevalence of the disease is increasing.
Despite the relatively small number of patients
with severe hemophilia, the disease has an important socioeconomic effect because of its distinctive feature: a level of clotting factor (factor
VIII or factor IX) so low that without replacement therapy, frequent hemorrhages occur, most
often in the ankle, knee, and elbow joints. These
joint hemorrhages culminate in a severe arthropathy, with signs of inflammation (as in rheumatoid arthritis) as well as degenerative damage (as
in osteoarthritis).1,2 As a result of these hemorrhages, patients with hemophilic arthropathy can
become disabled at a relatively young age.
Joint bleeding in hemophilia can be prevented by the administration of a concentrated clotting factor derived from plasma (cryoprecipitate)
or a recombinant clotting protein. Since these
products are expensive, the beneficial results they
have to offer should be balanced by a consideration of their cost.
In this issue of the Journal, Manco-Johnson and
colleagues report the results of a randomized
study that assessed the clinical benefit of a very
costly product, recombinant factor VIII, for the
prophylactic treatment of young boys (up to the
age of 6 years) with severe hemophilia.3 Prophylaxis, a scheduled intravenous injection of clotting factor in anticipation of bleeding, was compared with on-demand, or episodic, treatment,
the administration of clotting factor whenever
signs of joint bleeding appeared. The benefit of
prophylaxis as compared with episodic treatment
was clear. After a mean follow-up of 49 months,
93% of the patients in the prophylaxis group
showed no joint damage on magnetic resonance
imaging as compared with 55% in the episodic-
therapy group. The number of infused units of
factor VIII per participant was approximately
350,000 in the prophylaxis group (with 0.63 joint
hemorrhage per year) and approximately 113,000
in the episodic-therapy group (with 4.89 joint
hemorrhages per year). Several previous retrospective or uncontrolled studies have suggested the
superiority of prophylaxis,4,5 but the study by
Manco-Johnson et al. is a controlled randomized
trial comparing the effects of prophylaxis with
episodic treatment in hemophilia.
The study puts a harsh light on cost–benefit
issues in the management of hemophilia: the
estimated annual cost for prophylactic treatment
of one patient with recombinant factor VIII was
$300,000. Manco-Johnson and her colleagues
comment that “the high cost of recombinant
factor VIII is a barrier to widespread acceptance
of prophylaxis.” With the current challenges entailed in reducing health care costs, this is a reality we must deal with, a challenge to further
optimize the balance between cost and benefit.
In the balance of cost with the efficacy of
treatment in hemophilia, it is important to consider that inadequately treated patients make great
demands on the health care system. Such patients
often need expensive additional treatment, such
as joint replacement. Yet the kind of cost-effectiveness studies that have been performed for many
other chronic diseases have never been performed
for hemophilia. The dilemma in managing hemophilia is not whether to use prophylaxis or episodic treatment but how to manage prophylaxis
such that the optimal, most cost-effective treatment is provided.
The crux of the problem is this: How many
joint hemorrhages can be tolerated before disabling arthropathy occurs, and how can clottingfactor replacement therapy be managed such that
optimal prevention of joint bleeding is achieved?
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
The study by Manco-Johnson and colleagues provides little help in answering these questions.
The authors conclude that “Prophylaxis was efficacious in decreasing bleeding and joint damage after up to five hemarthroses.” At the population level, their conclusion is correct, but what
does it mean for the individual patient? Could it
be that patients who had no joint hemorrhages
before the start of prophylaxis were at a lower
risk for a subsequent joint hemorrhage than patients who started prophylaxis after having one
or more joint hemorrhages? In the prophylaxis
group, 18 patients had had at least one joint
hemorrhage before treatment and 14 patients had
no previous hemorrhage. This difference provides
a unique opportunity to evaluate these subgroups
for the development of joint damage despite prophylaxis.
The outcome of such an analysis might be
anticipated, considering the patients’ long experience with hemophilia. Nilsson and colleagues6
have noted progressive deterioration in joints
that are damaged before the start of prophylaxis,
whether or not there is further bleeding into
those joints. Petrini and colleagues7 and Fischer
et al.8 found that the early institution of prophylaxis — before the age of 3 years or before the
third joint hemorrhage — resulted in radiologic
and orthopedic scores that were significantly better than those associated with later commencement of prophylaxis (after the age of 5 years).
Progression of arthropathy continued in children
in both groups who had experienced more than
five joint hemorrhages before the start of prophylaxis, whether it was instituted early or late
in childhood. Kreuz and colleagues9 also found
a worsening of joint scores despite ongoing prophylaxis in patients who had a median of six joint
hemorrhages before starting prophylaxis.
These results fit well with the lack of correlation between the number of hemorrhages and
the severity (progression rate) of arthropathy in
the study by Manco-Johnson et al. It may be
that the first episode of bleeding or early episodes initiate a self-perpetuating arthropathy
and that subsequent episodes of bleeding only
accelerate the degenerative process. Even a small
number of joint hemorrhages can cause irreversible joint alterations, leading to an arthropathy
that can be slowed by prophylaxis but not prevented. In this regard, it is notable that culturing full-thickness articular cartilage explants in
m e dic i n e
the presence of blood, even for a short period of
time, has devastating effects on chondrocyte activity and cartilage integrity.10,11
For all these reasons, it seems that prophylaxis will be most effective if it can prevent joints
from reaching the point of no return. But determining when to start prophylaxis is difficult because severe hemophilia varies considerably in
terms of the number and frequency of joint hemorrhages.12 Starting prophylaxis at a very early
age could be cost-effective because young patients,
with their low body weight, require only small
amounts of clotting factor. Moreover, the preservation of joint structure with early initiation of
prophylaxis is thought to result in fewer episodes of bleeding later in life, less disability in
adults, and therefore lower health care costs. It
seems reasonable, in light of all these considerations, to begin prophylaxis after the occurrence
of one joint hemorrhage rather than at a fixed
age.13 Subsequent prophylaxis can be tailored to
the individual patient, on the basis of that patient’s clinical course, to achieve optimum costeffectiveness.14
The study by Manco-Johnson and colleagues
provides renewed and detailed evidence that prophylaxis in severe hemophilia with recombinant
factor VIII has clinical benefit despite its huge
costs. However, we must await prospective studies
that demonstrate the most balanced approach,
weighing possible adverse events and cost against
No potential conflict of interest relevant to this article was reported.
From the Department of Hematology and Van Creveld Clinic and
the Department of Rheumatology and Clinical Immunology, University Medical Center Utrecht, Utrecht, the Netherlands.
1. Roosendaal G, van Rinsum AC, Vianen ME, van den Berg
HM, Lafeber FP, Bijlsma JW. Haemophilic arthropathy resembles
degenerative rather than inflammatory joint disease. Histopathology 1999;34:144-53.
2. Wen FQ, Jabber AA, Chen YX, Kazarian T, Patel DA, Valentino LA. c-myc Proto-oncogene expression in hemophilic synovitis: in vitro studies of the effects of iron and ceramide. Blood
3. Manco-Johnson MJ, Abshire TC, Shapiro AD, et al. Prophylaxis versus episodic treatment to prevent joint disease in boys
with severe hemophilia. N Engl J Med 2007;357:535-44.
4. Liesner RJ, Khair K, Hann IM. The impact of prophyactic
treatment on children with severe haemophilia. Br J Haematol
5. Aledort LM, Haschmeyer RH, Pettersson H. A longitudinal
study of orthopaedic outcomes for severe factor-VIII-deficient
haemophiliacs. J Intern Med 1994;236:391-9.
6. Nilsson IM, Berntorp E, Löfqvist T, Pettersson H. Twentyfive years’ experience of prophylactic treatment in severe haemophilia A and B. J Intern Med 1992;232:25-32.
n engl j med 357;6  www.nejm.org  august 9, 2007
7. Petrini P, Lindvall N, Egberg N, Blombäck M. Prophylaxis
with factor concentrates in preventing hemophilic arthropathy.
Am J Pediatr Hematol Oncol 1991;13:280-7.
8. Fischer K, van der Bom JG, Mauser-Bunschoten EP, et al. The
effects of postponing prophylactic treatment on long-term outcome in patients with severe hemophilia. Blood 2002;99:2337-41.
9. Kreuz W, Escuriola-Ettingshausen C, Funk M, Schmidt H,
Kornhuber B. When should prophylactic treatment in patients
with haemophilia A and B start? The German experience. Haemophilia 1998;4:413-7.
10. Jansen NW, Roosendaal G, Bijlsma JW, Degroot J, Lafeber
FP. Exposure of human cartilage tissue to low concentrations of
blood for a short period of time leads to prolonged cartilage
damage: an in vitro study. Arthritis Rheum 2007;56:199-207.
11. Hooiveld M, Roosendaal G, Vianen M, van den Berg M,
Bijlsma J, Lafeber F. Blood-induced joint damage: longterm effects in vitro and in vivo. J Rheumatol 2003;30:339-44.
12. van Dijk K, van der Bom JG, Lenting PJ, et al. Factor VIII
half-life and clinical phenotype of severe hemophilia A. Haematologica 2005;90:494-8.
13. Roosendaal G, Lafeber FP. Pathogenesis of haemophilic arthropathy. Haemophilia 2006;12:Suppl 3:117-21.
14. Astermark J, Petrini P, Tengborn L, Schulman S, Ljung R,
Berntorp E. Primary prophylaxis in severe haemophilia should
be started at an early age but can be individualized. Br J Haematol 1999;105:1109-13.
Copyright В© 2007 Massachusetts Medical Society.
Rituximab and Pemphigus — A Therapeutic Advance
Luis A. Diaz, M.D.
Pemphigus vulgaris and pemphigus foliaceus are
rare autoimmune blistering diseases of obscure
causes. The scientific bases for the therapy of
these disorders have evolved in the past 60 years.
Landmark studies demonstrated that intraepidermal blisters were subcorneal in pemphigus foliaceus1 and suprabasilar in pemphigus vulgaris2,3
and suggested that an ongoing process of epidermal-cell detachment (acantholysis) led to blister formation. Armed with this limited information, dermatologists in the 1950s began to use
systemic steroids to treat these diseases. Before
the use of corticosteroids, the diagnosis of pemphigus carried a poor prognosis; it was believed
that if a patient survived for more than a year,
the diagnosis of pemphigus was probably incorrect.
The autoimmune nature of these diseases was
substantiated in the 1960s, when Beutner and
Jordon4 discovered that pemphigus vulgaris and
pemphigus foliaceus antiepidermal autoantibodies were present in patients with these disorders.
In the 1980s my colleagues and I showed by passive transfer experiments that pemphigus vulgaris and pemphigus foliaceus autoantibodies
faithfully reproduced the disease in the epidermis
of neonatal mice.5,6 Stanley and coworkers7,8
showed that the epidermal antigens recognized
by pemphigus vulgaris and pemphigus foliaceus
autoantibodies were desmoglein 3 and desmoglein 1, respectively, both members of the cadherin
superfamily of adhesion molecules. The molecular mechanisms of the epidermal acantholysis
triggered by pemphigus vulgaris and pemphigus
foliaceus autoantibodies remain unknown, but
they are the subject of intense investigation in several laboratories.9
The autoimmune nature of pemphigus was
used as a rationale for the introduction of immunosuppressive drugs and plasmapheresis, both
aimed at eliminating pathogenic autoantibodies
from the patient. The combined use of systemic
steroid and immunosuppressive agents such as
azathioprine, cyclophosphamide, methotrexate,
mycophenolate mofetil, and cyclosporine has become the standard of therapy for pemphigus vulgaris and pemphigus foliaceus with good results.
However, the morbidity associated with the longterm use of these drugs remains problematic for
the patients, especially those with diabetes, hypertension, or other medical conditions. Since the introduction of systemic steroids in the 1950s and
immunosuppressive agents in the 1960s, the prognosis for patients with the disease has been gradually improving. Recent reports estimate an overall 5% mortality rate due to complications from
therapy for patients with pemphigus vulgaris.10
During the past 10 years we have witnessed
the arrival of new and exciting therapies for pemphigus. Infusions of intravenous immune globulins and B-cell–targeted humanized monoclonal
antibodies are two examples. Rituximab is an
anti-CD20 chimeric monoclonal antibody approved
by the Food and Drug Administration for the
treatment of B-cell non-Hodgkin’s lymphoma and
rheumatoid arthritis that is refractory to anti–
tumor necrosis factor О±. These therapies are especially useful in cases of pemphigus that are
not responsive to conventional therapy with corticosteroids and immunosuppressive drugs. In this
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
issue of the Journal, Joly and colleagues11 report
a prospective open trial of rituximab in patients
with pemphigus vulgaris or pemphigus foliaceus
who had not had a response to prednisone at a
dose of 1.5 mg per kilogram of body weight per
day, who had active disease despite systemic doses of prednisone greater than 20 mg per day, or
in whom systemic corticosteroids were contraindicated. Treatment with adjuvant intravenous
immune globulins and immunosuppressive therapy was described as having failed in some of
these patients. Eighteen of 21 patients who were
treated with four weekly infusions of rituximab
had a complete clinical remission by 3 months.
In 8 of the 18 patients, this remission was maintained without corticosteroid or immunosuppressive therapy after a median follow-up of almost
3 years. One patient died of septicemia, and
pyelonephritis developed in another.
In addition to the clinical response, Joly and
colleagues also followed the progression of certain immunologic markers. Three weeks after the
onset of therapy, peripheral-blood B cells decreased by a dramatic 99% and remained undetectable for 6 months in two patients. B cells
reappeared after 6 months. No changes were
detected in T cells or in the titers of antibodies
against pneumococcal capsule polysaccharide, tetanus toxoid, or total IgG values. There was a significant decrease of anti–desmoglein 1 and anti–
desmoglein 3 autoantibodies 3 months after the
onset of therapy. These autoantibodies reappeared
in the serum of patients during relapses. Of note,
in five patients with pemphigus vulgaris who
were in complete clinical remission, the titers of
autoantibodies remained elevated. These studies
point toward the relevance of certain “pathogenic” epitopes on the ectodomain of desmoglein 1
and desmoglein 3, as previously reported.12
One of the conclusions of Joly and colleagues
is that pemphigus vulgaris and pemphigus foliaceus are life-threatening skin diseases that are
mediated by antigen-specific activated B cells and
their products — that is, anti–desmoglein 3 and
anti–desmoglein 1 autoantibodies. These autoantibodies are IgG4 restricted and pathogenic in
most patients; their effect may depend on the epidermal-cell epitopes they bind. Since the clinical
remission induced by rituximab in patients with
pemphigus vulgaris and pemphigus foliaceus was
associated with a decrease in the titers of anti–
desmoglein 3 and anti–desmoglein 1 autoanti606
m e dic i n e
bodies without changes in the titers of antipneumococcal and antitetanus antibodies, it may be
assumed that these antibodies eliminate desmoglein 1–specific and desmoglein 3–specific B-cell
precursors. It has been proposed that long-lived
plasma cells produce antigen-specific autoantibodies that show stable titers in the serum of
patients, whereas short-lived plasma cells and
plasmablasts produce antigen-specific autoantibodies in titers that fluctuate with disease activity.13 Long-lived and short-lived plasma cells do
not express the CD20 antigen; consequently, by
eliminating a large population of B cells with
rituximab therapy, it may be possible to eliminate a substantial number of desmoglein 1 and
desmoglein 3 B-cell precursors of short-lived
plasma cells. It is clear that further studies are
needed to disclose the molecular and cellular
mechanisms of action of rituximab.
In summary, the article by Joly et al. supports
previous reports that describe the beneficial effects of rituximab in the treatment of pemphigus without the need for infusions of intravenous
immune globulins; however, its use must be restricted to a limited number of patients with
pemphigus vulgaris or pemphigus foliaceus that
is not responsive to conventional therapy (e.g.,
systemic corticosteroids and immunosuppressive
agents) or to patients in whom these drugs may
be harmful. Caution should be exercised in the
use of rituximab because of the risk of serious
short-term complications such as viral and bacterial infections and the potential for as yet unknown long-term complications. Finally, this study
demonstrates the value of a multicenter approach
to accomplish relevant clinical research in orphan
diseases such as pemphigus. A U.S. national registry for pemphigus and standardized tools to
assess the disease activity of pemphigus are necessary to accelerate research progress.
Dr. Diaz reports being named as coinventor on a patent held
by the University of North Carolina for the inhibition of the
phosphorylation of HSP27 (heat-shock protein 27) for the treatment of blistering disorders. No other potential conflict of interest relevant to this article was reported.
From the Department of Dermatology, University of North
Carolina at Chapel Hill, Chapel Hill.
1. Vieira JP. Contribuição ao estudo do pemphigo no estado de
SГЈo Paulo. SГЈo Paulo: Empresa GrГЎfica da Revista dos Tribunais,
2. Civatte A. Diagnostic histopathologique de la dermatite polymorphe douloureuse ou maladie de Duhring-Brocq. Ann Dermatol Syphiligr 1943;3:1-30.
3. Lever WF. Pemphigus. Medicine (Baltimore) 1953;32:1-123.
n engl j med 357;6  www.nejm.org  august 9, 2007
4. Beutner EH, Jordon RE. Demonstration of skin antibodies in
sera of patients with pemphigus vulgaris by indirect immunofluorescent staining. Proc Soc Exp Biol Med 1964;117:505-10.
5. Anhalt GJ, Labib RS, Voorhees JJ, Beals TF, Diaz LA. Induction of pemphigus in neonatal mice by passive transfer of IgG
from patients with the disease. N Engl J Med 1982;306:1189-96.
6. Roscoe JT, Diaz L, Sampaio SA, et al. Brazilian pemphigus
foliaceus autoantibodies are pathogenic to BALB/c mice by passive transfer. J Invest Dermatol 1985;85:538-41.
7. Eyre RW, Stanley JR. Human autoantibodies against a desmosomal protein complex with a calcium-sensitive epitope are
characteristic of pemphigus foliaceus patients. J Exp Med 1987;
8. Amagai M, Klaus-Kovtun V, Stanley J. Autoantibodies
against a novel epithelial cadherin in pemphigus vulgaris, a disease of cell adhesion. Cell 1991;67:869-77.
9. Li N, Rubenstein DS, Liu Z, Diaz LA. Pemphigus foliaceus,
pemphigus vulgaris, paraneoplastic pemphigus, bullous pemphigoid herpes gestationis and cicatricial pemphigoid. In: Runge MS, Patterson WC, eds. Principles of molecular medicine. 2nd
ed. Totowa, NJ: Humana Press, 2006:959-69.
10. Herbst A, Bystryn J-C. Patterns of remission in pemphigus
vulgaris. J Am Acad Dermatol 2000;42:422-7.
11. Joly P, Mouquet H, Roujeau J-C, et al. A single cycle of rituximab for the treatment of severe pemphigus. N Engl J Med 2007;
12. Li N, Aoki V, Hans-Filho G, Rivitti EA, Diaz LA. The role of
intramolecular epitope spreading in the pathogenesis of endemic pemphigus foliaceus (fogo selvagem). J Exp Med 2003;197:
13. Radbruch A, Muehlinghaus G, Luger EO, et al. Competence
and competition: the challenge of becoming a long-lived plasma
cell. Nat Rev Immunol 2006;6:741-50.
Copyright В© 2007 Massachusetts Medical Society.
view current job postings at the nejm careercenter
Visit our online CareerCenter for physicians
at www.nejmjobs.org to see the expanded features and
services available. Physicians can conduct a quick search
of the public database by specialty and view hundreds
of current openings that are updated daily online
at the CareerCenter.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
s ounding boa r d
The Tension between Needing to Improve Care
and Knowing How to Do It
Andrew D. Auerbach, M.D., M.P.H., C. Seth Landefeld, M.D., and Kaveh G. Shojania, M.D.
The past 7 years have seen unprecedented interest
in patient safety and the quality of health care.1-3
As physicians whose careers are focused on improving quality and safety, we have welcomed this
change. However, we have also witnessed recent
initiatives that emphasize dissemination of innovative but unproven strategies, an approach that
runs counter to the principle of following the evidence4 in selecting interventions that meet quality and safety goals, as well as the idea that interventions should be tailored to local needs and
resources.5 These principles have been used as
safeguards in helping us pursue practices that have
clear benefits for patients and that can be implemented with local resources. This approach also
reflects the recognition of how little we know
about ways to improve care in a large number of
Our consideration of the rationale for rapid
dissemination of novel quality and safety strategies has led us to identify a number of weaknesses inherent in approaches that consistently favor
action over evidence. In this article, we outline the
arguments in favor of rapid dissemination and
the counterpoints to each of the arguments (Table 1). We conclude by proposing a framework
for evaluating interventions to improve the safety and effectiveness of health care.
Ar gument 1: We C annot Wait
The most common argument in favor of prioritizing action over evidence is that the need to address quality and safety problems is urgent. Often, this need is summed up by the question,
“How many times does outcome X need to occur
before we implement intervention Y?”
This question seems particularly compelling
because hundreds of thousands of patients (possibly millions) experience harm as a result of underuse, overuse, or misuse of medical therapies.8
However, similar claims about the scale of mor608
bidity and mortality could be made for heart disease, cancer, AIDS, depression, and many other
disorders. Medical error may be the eighth leading cause of death in the United States,2 but by
proceeding largely on the basis of urgency rather than evidence, we exempt the eighth cause of
death from standards applied to the top seven.
In addition, the question of how many instances of X outcome need to occur before we implement Y intervention assumes that we can define
Y and X accurately, as well as connect Y to a decreased risk of X. Donabedian pointed out that
Y can be either a structural element of health care
(e.g., staffing ratios) or a process (e.g., administration of a drug) and emphasized the importance
of establishing a connection between Y and the
outcome of interest, X.9 Unfortunately, connections between structural or process-based interventions and outcomes are usually presumptive,7,10
and defining problems and solutions with respect
to patient safety is generally difficult.10,11
For example, mandates to reduce residents’
work hours reflect the view that tired residents
cause errors that harm patients. However, evidence
linking patient harm directly to care provided by
a fatigued resident is indirect,12,13 and although
reductions in work hours do not appear to have
harmed patients, evidence that reforms have met
their goal of improving safety is tentative at
best.14-18 Furthermore, to be cost-effective, a reduction in work hours would have to result in
greater improvement in safety than that reported
for any other intervention.19 Regardless of whether an 80-hour workweek ultimately improves patient safety, an intervention with a number of
potential effects was introduced without a full
understanding of its risks and benefits and without a plan to evaluate its effectiveness after implementation.
Promising initiatives and bold efforts at improvement can consume tremendous resources yet
confer only a small benefit15,20,21 or a benefit that
n engl j med 357;6  www.nejm.org  august 9, 2007
sounding Board
Table 1. Arguments for and against Rapid Dissemination of Quality-Improvement Interventions.
Why Proceeding Quickly Is Critical
Why Evaluation Is Critical
We cannot wait — the need to
improve the quality of care
is urgent.
Thousands of patients are injured or
killed each year by medical errors.
The need to improve the treatment of
many diseases is equally urgent, yet
we demand rigorous evidence that a
therapy works before recommending
it widely.
Any effort to improve quality is
better than the current state
of affairs.
On balance, the harms of quality
improvement are likely to be far
less than those of the status quo.
Knowledge of the harms and opportunity
costs of quality improvement is important for an understanding of the
net benefit to patients and health
care systems, which is often small.
Emulating successful organizations Emulation and collaboration provide
can speed effective improvement.
an efficient means of disseminating potentially effective solutions.
Emulation and collaboration can incorrectly promote or even overlook interventions that have not worked.
The effectiveness of some qualityimprovement strategies is
Even though many quality-improvement
practices have a simple rationale,
they may be less effective than expected and can be difficult to implement fully.
Insistence on evidence may lead us
to underuse interventions that
are obviously effective.
Innovation can be catalyzed by dis- Preliminary data provide an imporsemination of strategies that
tant opportunity to speed innovahave promise but are unproven.
tion and improve care rapidly.
Flawed, biased, or incomplete data may
lead to adoption of interventions that
are ineffective or harmful.
The framework of evidence-based
medicine does not apply to
quality improvement.
The nature of quality improvement
exempts it from the usual strategies of assessment.
Given the complexity of quality and safety
problems, the complexity of their
causes, and how little we understand
them, we should use rigorous study
designs to evaluate them.
Developing evidence in quality
improvement is too costly.
The resources and expertise required
to evaluate quality and safety interventions rigorously make trials
impractical, particularly when the
field is moving so quickly.
As compared with the large opportunity
costs incurred by wide implementation of ineffective quality and safety
strategies, investments in better evaluation would be small.
is at best unclear.22 How many such examples of efforts to intensify the treatment of diabetes,
must we have before we decide to choose our ef- adverse consequences of many other efforts at
forts more wisely?
improvement of care have been less predictable,
including errors introduced by computerized entry of physicians’ orders,25,26 bar coding,27 and
Ar gument 2 : Any Effor t
to Impr ove Is Be t ter than
infection-control isolation protocols.28 Side efthe Current S tate of Affair s
fects may seem inherently less likely with quality-improvement interventions than with drugs
Multiple problems in our flawed health care sys- and devices. However, most quality-improvement
tem lead to the view that any attempt at improve- interventions involve changes in the organizament is better than the status quo. Although tion of complex systems, and the law of uninunderstandable, this view ignores the possibil- tended consequences — long recognized as a side
ity that quality-improvement efforts can cause effect of complex change — tends to apply to
harm.23 Unfortunately, few studies have assessed such interventions.29-31
this possibility. For example, only 12 of 66 reports on trials of strategies to improve care for
Ar gument 3: Emul ating
patients with diabetes included rates of hypoglySucce ssful Or g aniz ations
C an Speed Impr ovement
cemia.24 However, in 7 of those 12 studies, hypoglycemia was more frequent in the intervention group than in the control group. Although A recommendation to emulate successful organihypoglycemia is an easily anticipated consequence zations reflects the reasoning that adopting fea-
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
tures of these organizations — the institutional
culture, leadership styles, or specific improvement
practices — will result in similar successes. Unfortunately, this reasoning ignores the possibility that many unsuccessful organizations also
share these features, so that the truly critical determinants of success are not being targeted.
For instance, continuous quality improvement
and quality-improvement collaboratives are often
recommended on the basis of their adoption by
successful organizations. However, systematic
evaluations of these approaches have shown
that they result in only modest improvements at
best.20,21,23,32 These disappointing findings probably reflect the overemphasis on success that is
inherent in benchmarking and the collaborative
approach, which tend to neglect an examination
of unsuccessful organizations that share features
of successful ones.33
Successful organizations may also have a vested interest in promoting their services or preferred quality-improvement strategies, further
distorting the usefulness of emulating such organizations. Even when direct financial conflicts
of interest do not exist, any organization that has
undertaken a major campaign to improve the
quality of care has little incentive to invest resources in a rigorous evaluation of the effects of
its efforts. If anecdotal reports or superficial
analyses are positive, the organization will understandably focus on advertising these measures
of success rather than pursuing more rigorous
Ar gument 4 : The Effec tivene ss
of Some Qualit y- Impr ovement
S tr ategie s Is O bvious
Some solutions appear to be so obviously beneficial that requiring evidence seems like asking for
randomized trials of parachutes.34 However, anyone who has undertaken a quality-improvement
project understands that identifying an apparent
solution to a problem is only a first step. Even
with pilot testing and evaluative steps, implementing solutions in practice can present numerous
Hand washing is an example of a well-defined,
effective solution to a problem (nosocomial infections), but strategies that consistently result in
increased hand washing remain unestablished.35
Unfortunately, many initiatives fall into the hand610
m e dic i n e
washing category — that is, the case for improvement is obvious, but effective strategies for translating solutions into practice remain elusive.8
Changes in complex systems can have unanticipated consequences (as we note with respect
to Argument 2), such as new problems25,26,30,31,36
or simply the failure to achieve the desired goal.
Until we advance the basic sciences in quality improvement (e.g., organizational theory and ergonomics),4,7,23,32 we cannot assume that even the
most apparently straightforward solutions can
be seamlessly implemented. Without an understanding of not only what to do but also how to
help people actually do it, many apparently obvious quality-improvement interventions have more
in common with calls for world peace than with
parachutes — the goal is not in question, but
the path for achieving it is.
Ar gument 5: Pr omising B u t
Unpr oven S tr ategie s C an C ataly ze
Many quality-improvement interventions have
such strong face validity that their dissemination seems to be justified on the basis of early or
preliminary evidence. This strategy will certainly
speed dissemination, but it also carries substantial risks.
Early trials of medical emergency teams suggested a large potential benefit37-40 — to the point
that some observers regarded further study as
unethical.41 However, a large, randomized trial
subsequently showed that medical emergency
teams had no effect on patient outcomes.42 The
validity of the earlier positive studies has also
been questioned,43 but only after many hospitals
introduced medical emergency teams (and have
had no reason to switch from advertising the
adoption of an innovation to questioning its usefulness in the first place — Argument 3).
There are many examples of drugs or devices
that showed substantial promise on the basis of
early findings, which were then modified or refuted by later-phase research. These often represent therapies for disorders that affect millions
of people (as we note with respect to Argument
1). Yet we rarely sanction the widespread distribution of new drugs on the basis of preliminary
data alone. It is therefore not clear why we favor
approaches to quality improvement that foster
change over appropriate evaluation.
n engl j med 357;6  www.nejm.org  august 9, 2007
sounding Board
It is worth emphasizing that when studies show
no benefit of an intervention with strong face
validity, as has occurred with rapid-response
teams42 and more recently with teamwork training,44 one should not necessarily conclude that
the intervention has no value. The finding may
simply mean that the intervention had no effect
in the form and setting that were studied. The
crucial point is that without the randomized trial,
we would have no way of knowing that implementation of the intervention in its current form
confers no advantage over usual care (or confers
a much smaller advantage than that suggested by
preliminary studies) and that refinement is necessary.
tion and control groups. In clinical medicine, important confounders are often well known and
easily planned for, so that observational studies
can adjust for these factors, thereby producing
results that often agree with the results of randomized trials.48,49 However, outcomes of quality-improvement interventions depend on many
factors, related to patients, providers, and organizations, that remain poorly understood. Thus,
the complexity of health care and the dearth of
evidence with respect to how components of the
system interact to influence outcomes provide a
strong rationale for conducting randomized trials to evaluate quality and safety interventions
whenever feasible.
Argument 6: The Framework
of Evidence-based Medicine Does
Not Apply to Qualit y Improvement
Ar gument 7 : De veloping
E vidence in Qualit y
Impr ovement Is To o Cos tly
A recent commentary argued that we would not
require randomized trials to determine whether
we have solved problems or learned skills in our
daily lives.45 By extension, according to this argument, evidence-based medicine may not apply to
the processes that underlie many quality-improvement initiatives. Although it is true that we often
do not need trials to test our acquisition of knowledge or skills, we do need them when choosing
between alternative methods of acquisition —
particularly when training is costly or the skill is
of high value.
Rigorous evaluation does not always require
randomized trials. Alternative designs (e.g., before-and-after studies that include concurrent control groups and time-series designs involving
multiple preintervention and postintervention
measurements) can sometimes provide robust
results,32,46 as can research that combines quantitative and qualitative approaches.47 But anecdotal reports and simple before-and-after studies,
although sometimes adequate to justify local quality-improvement efforts, are probably never sufficient to support widespread initiatives because
of the risks of expending tremendous resources
without obtaining a true benefit and possibly
introducing new problems.
Randomized, controlled trials, although not
always necessary,46 remain highly relevant to quality improvement. The value of such trials lies in
the random assignment of subjects with unknown
characteristics that affect outcomes to interven-
Many people have argued that with limited resources available for quality-improvement efforts,
the costs of evaluation are untenable. However,
one could also argue that we should not spend
scarce resources on quality improvement unless
we know it is effective. More important, there are
tremendous opportunity costs. An institution that
invests millions of dollars or expends hundreds
of personnel hours in implementing an ineffective system almost certainly could have made
other investments that would have benefited its
patients. Moreover, if the investment at one hospital is multiplied by thousands of hospitals
across the country, then surely spending several
million dollars for evaluation is cost-effective,
given the billions of dollars at stake with widespread implementation. In this sense, it is the absence of evidence — with respect to efficacy, possible harms, and strategies for implementation
— that is too costly, not the efforts to generate
such evidence.
The urge to favor action over evidence in efforts
to improve the quality and safety of health care
is understandable. However, we have seen in recent years that progress in quality improvement
occurs just as it does in the rest of biomedicine:
interventions that appear to be promising on the
basis of preliminary studies often prove to have
no benefit, and those that are beneficial typi-
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
cally result in modest improvements, not monu- to harms, squander scarce resources, or delude
mental breakthroughs. And quality-improvement us about the effectiveness of our efforts.
interventions, like clinical therapies, can have unSupported by a Patient Safety Research and Training Grant
toward effects and both direct and indirect costs. from the Agency for Healthcare Research and Quality (K08
These commonalities compel us to argue that in- HS11416-02, to Dr. Auerbach), an Academic Leadership Award
from the National Institute on Aging (K07 AG00912, to Dr.
terventions to improve the quality and safety of Landefeld), and funding from the Government of Canada Rehealth care should meet the same standards that search Program (to Dr. Shojania).
No potential conflict of interest relevant to this article was reare applied to the adoption of all medical techported.
We thank Erin Hartman, M.S., Ralph Gonzales, M.D., M.S.P.H.,
In the rest of biomedicine, innovation begins and Chaim Bell, M.D., M.P.H., for their helpful comments on
with basic-science experimentation and proceeds an early version of the manuscript.
through evaluative trials in successive phases. From the University of California, San Francisco, Department
The basic sciences in quality improvement differ of Medicine, San Francisco (A.D.A., C.S.L.); San Francisco Vetfrom those in the rest of biomedicine, but the erans Affairs Medical Center, San Francisco (C.S.L.); and Ottawa Health Research Institute, Ottawa (K.G.S.).
framework for evaluating candidate interventions
is largely the same. Clinicians often make deci- 1. Institute of Medicine. Crossing the quality chasm: a new
sions about treatment in individual patients on health system for the 21st century. Washington, DC: National
Academy Press, 2001.
the basis of limited evidence or even just intu- 2. Kohn LT, Corrigan JM, Donaldson MS, eds. To err is human:
ition. Similarly, individual hospitals may pursue building a safer health system. Washington, DC: National Acadpromising quality-improvement strategies on the emy Press, 2000.
3. Berwick DM, Calkins DR, McCannon CJ, Hackbarth AD. The
basis of scant evidence, including anecdotal re- 100,000 Lives Campaign: setting a goal and a deadline for imports or face validity. However, clinical practices proving health care quality. JAMA 2006;295:324-7.
based on such limited evidence would never be- 4. Brennan TA, Gawande A, Thomas E, Studdert D. Accidental
deaths, saved lives, and improved quality. N Engl J Med 2005;
come broad standards of care, much less require- 353:1405-9.
ments for accreditation or reimbursement. Simi- 5. Nelson EC, Batalden PB, Ryer JC, eds. The clinical improvelarly, recommending or mandating the widespread ment action guide. Oakbrook Terrace, IL: Joint Commission on
Accreditation of Healthcare, 1998.
adoption of interventions to improve quality or 6. Forster AJ, Shojania KG, van Walraven C. Improving patient
safety requires rigorous testing to determine safety: moving beyond the “hype” of medical errors. CMAJ
whether, how, and where the intervention is ef- 2005;173:893-4.
7. Shojania KG, Duncan BW, McDonald KM, Wachter RM. Safe
fective — just as in the rest of medicine. Clarifi- but sound: patient safety meets evidence-based medicine. JAMA
cation of this picture is critical because a num- 2002;288:508-13.
ber of widely promulgated interventions are likely 8. Chassin MR, Galvin RW. The urgent need to improve health
care quality: Institute of Medicine National Roundtable on
to be wholly ineffective, even if they do not harm Health Care Quality. JAMA 1998;280:1000-5.
patients. Even worse, in the current environment, 9. Donabedian A. Evaluating the quality of medical care. Milwe will not know what these interventions are. bank Mem Fund Q 1966;44:Suppl:166-206.
10. Pronovost PJ, Miller MR, Wachter RM. Tracking progress in
The movement to improve quality and safety patient safety: an elusive target. JAMA 2006;296:696-9.
has achieved substantial momentum in recent 11. Hofer TP, Kerr EA, Hayward RA. What is an error? Eff Clin
years and has begun to address the many errors Pract 2000;3:261-9.
12. Gaba DM, Howard SK. Patient safety: fatigue among cliniof omission and commission that harm patients cians and the safety of patients. N Engl J Med 2002;347:1249-55.
each day. Moreover, the visible moral leadership 13. Weinger MB, Ancoli-Israel S. Sleep deprivation and clinical
associated with these efforts3 has played a crucial performance. JAMA 2002;287:955-7.
14. Fletcher KE, Davis SQ, Underwood W, Mangrulkar RS, Mcrole in maintaining public trust. Although the Mahon LF Jr, Saint S. Systematic review: effects of resident work
scope of the problems may seem to favor action hours on patient safety. Ann Intern Med 2004;141:851-7.
over knowledge, quality improvement is on com- 15. Laine C, Goldman L, Soukup JR, Hayes JG. The impact of a
regulation restricting medical house staff working hours on the
mon ground with the rest of biomedicine. The quality of patient care. JAMA 1993;269:374-8.
temptation is to circumvent traditional models 16. Shetty KD, Bhattacharya J. Changes in hospital mortality asof evidence when it comes to quality improve- sociated with residency work-hour regulations. Ann Intern Med
ment, but this temptation has always existed in 17. Horwitz LI, Kosiborod M, Lin Z, Krumholz HM. Changes in
medicine for those seeking cures to conditions outcomes for internal medicine inpatients after work-hour reguwith high morbidity. Just as in the rest of medi- lations. Ann Intern Med 2007;147:97-103.
18. Goldman L, Fiebach N. Hippocrates affirmed? Limiting
cine, we must pursue the solutions to quality and residents’ work hours does no harm to patients. Ann Intern Med
safety problems in a way that does not blind us 2007;147:143-4.
n engl j med 357;6  www.nejm.org  august 9, 2007
sounding Board
19. Nuckols TK, Escarce JJ. Residency work-hours reform: a cost
analysis including preventable adverse events. J Gen Intern Med
20. Landon BE, Hicks LS, O’Malley AJ, et al. Improving the management of chronic disease at community health centers. N Engl
J Med 2007;356:921-34.
21. Landon BE, Wilson IB, McInnes K, et al. Effects of a quality
improvement collaborative on the outcome of care of patients
with HIV infection: the EQHIV study. Ann Intern Med 2004;
22. Wachter RM, Pronovost PJ. The 100,000 Lives Campaign:
a scientific and policy review. Jt Comm J Qual Patient Saf 2006;
23. Mittman BS. Creating the evidence base for quality improvement collaboratives. Ann Intern Med 2004;140:897-901.
24. Shojania KG, Ranji SR, McDonald KM, et al. Effects of quality improvement strategies for type 2 diabetes on glycemic control: a meta-regression analysis. JAMA 2006;296:427-40.
25. Koppel R, Metlay JP, Cohen A, et al. Role of computerized
physician order entry systems in facilitating medication errors.
JAMA 2005;293:1197-203.
26. Han YY, Carcillo JA, Venkataraman ST, et al. Unexpected
increased mortality after implementation of a commercially sold
computerized physician order entry system. Pediatrics 2005;116:
1506-12. [Erratum, Pediatrics 2006;117:594.]
27. Patterson ES, Cook RI, Render ML. Improving patient safety
by identifying side effects from introducing bar coding in medication administration. J Am Med Inform Assoc 2002;9:540-53.
28. Stelfox HT, Bates DW, Redelmeier DA. Safety of patients isolated for infection control. JAMA 2003;290:1899-905.
29. Merton RK. The unanticipated consequences of purposive
social action. Am Sociol Rev 1936;1:894-904.
30. Perrow C. Normal accidents: living with high-risk technologies. New York: Basic Books, 1984.
31. Reason J. Managing the risks of organizational accidents.
Burlington, VT: Ashgate Publishing, 1997.
32. Shojania KG, Grimshaw JM. Evidence-based quality improvement: the state of the science. Health Aff (Millwood) 2005;
33. Denrell J. Selection bias and the perils of benchmarking.
Harvard Business Review. April 2005:114-9, 134.
34. Smith GC, Pell JP. Parachute use to prevent death and major
trauma related to gravitational challenge: systematic review of
randomised controlled trials. BMJ 2003;327:1459-61.
35. Gawande A. On washing hands. N Engl J Med 2004;350:
36. Schnipper JL, Kirwin JL, Cotugno MC, et al. Role of pharma-
cist counseling in preventing adverse drug events after hospitalization. Arch Intern Med 2006;166:565-71.
37. Bristow PJ, Hillman KM, Chey T, et al. Rates of in-hospital
arrests, deaths and intensive care admissions: the effect of a
medical emergency team. Med J Aust 2000;173:236-40.
38. Buist MD, Moore GE, Bernard SA, Waxman BP, Anderson
JN, Nguyen TV. Effects of a medical emergency team on reduction of incidence of and mortality from unexpected cardiac arrests in hospital: preliminary study. BMJ 2002;324:387-90.
39. Bellomo R, Goldsmith D, Uchino S, et al. Prospective controlled trial of effect of medical emergency team on postoperative morbidity and mortality rates. Crit Care Med 2004;32:91621.
40. DeVita MA, Braithwaite RS, Mahidhara R, Stuart S, Foraida
M, Simmons RL. Use of medical emergency team responses to
reduce hospital cardiopulmonary arrests. Qual Saf Health Care
41. Kerridge RK, Saul WP. The medical emergency team, evidence-based medicine and ethics. Med J Aust 2003;179:313-5.
42. Hillman K, Chen J, Cretikos M, et al. Introduction of the
medical emergency team (MET) system: a cluster-randomised
controlled trial. Lancet 2005;365:2091-7. [Erratum, Lancet 2005;
43. Winters BD, Pham J, Pronovost PJ. Rapid response teams —
walk, don’t run. JAMA 2006;296:1645-7.
44. Nielsen PE, Goldman MB, Mann S, et al. Effects of teamwork training on adverse outcomes and process of care in labor
and delivery: a randomized controlled trial. Obstet Gynecol
45. Berwick DM. Broadening the view of evidence-based medicine. Qual Saf Health Care 2005;14:315-6.
46. Glasziou P, Chalmers I, Rawlins M, McCulloch P. When are
randomised trials unnecessary? Picking signal from noise. BMJ
47. Bradley EH, Herrin J, Wang Y, et al. Strategies for reducing
the door-to-balloon time in acute myocardial infarction. N Engl
J Med 2006;355:2308-20.
48. Benson K, Hartz AJ. A comparison of observational studies
and randomized, controlled trials. N Engl J Med 2000;342:187886.
49. Concato J, Shah N, Horwitz RI. Randomized, controlled trials, observational studies, and the hierarchy of research designs.
N Engl J Med 2000;342:1887-92.
Copyright В© 2007 Massachusetts Medical Society.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
c or r e sp ondence
Adjunctive Antidepressant Treatment for Bipolar Depression
To the Editor: Sachs et al. (April 26 issue)1 report that adding an adjunctive antidepressant drug
offers no benefit over continued mood-stabilizer
monotherapy in the treatment of bipolar depression. This finding contradicts some previous studies and the experience of many clinicians. In his
thoughtful accompanying editorial, Belmaker2
identifies possible reasons for this discrepancy,
including diagnostic heterogeneity and changes
in the manifestation of bipolar disorder during
the past 20 years. These observations are perhaps
too circumspect.
The past two decades have seen a marked increase in the diagnosis of bipolar disorder by
North American psychiatrists.3 This trend has
many origins: a reaction to past underdiagnosis,
a belief that this disorder is more treatable and
less stigmatizing than some alternative diagnoses,
and an insufficiently critical application of the diagnostic criteria of the Diagnostic and Statistical
Manual of Mental Disorders, fourth edition (DSM-IV),
as well as a genuine conviction on the part of some
authorities that this is indeed a widely prevalent
condition.4 More recently, direct-to-consumer advertising by pharmaceutical companies has exacthis week’s letters
Adjunctive Treatment for Bipolar Depression
Glucose Regulation in Young Adults with Very
Low Birth Weight
Use of Physicians’ Services for Medicare
Video on Orotracheal Intubation
A Man with Swelling of Both Eyes and an Ear
Pregnancy in Congenital В­Erythropoietic
erbated the trend. To the extent that the diagnostic boundaries of bipolar disorder are allowed
to expand, the benefits of specific pharmacologic
treatments are likely to decline.
Lawrence H. Price, M.D.
Audrey R. Tyrka, M.D., Ph.D.
Butler Hospital
Providence, RI 02906
[email protected]
Dr. Price reports receiving lecture fees from AstraZeneca. No
other potential conflict of interest relevant to this letter was reported.
1. Sachs GS, Nierenberg AA, Calabrese JR, et al. Effectiveness
of adjunctive antidepressant treatment for bipolar depression.
N Engl J Med 2007;356:1711-22.
2. Belmaker RH. Treatment of bipolar depression. N Engl J Med
3. Patten SB. Does almost everybody suffer from a bipolar disorder? Can J Psychiatry 2006;51:6-8.
4. Akiskal HS. Validating �hard’ and �soft’ phenotypes within
the bipolar spectrum: continuity or discontinuity? J Affect Disord 2003;73:1-5.
To the Editor: Sachs et al. do not consider the
heterogeneity of bipolar depression, concluding
that adding antidepressants to mood stabilizers
for patients with this condition is not effective.
This finding, which contradicts the results of previous studies,1 could lead to possible deleterious
consequences for patients. As suggested in the
editorial by Belmaker, the low recruitment rate
(366 of 2689 patients with at least one major depressive episode) may reflect a selection bias.
There is considerable clinical heterogeneity in
bipolar depression.2 It is possible to distinguish
two types of bipolar depression, one characterized by a global inhibition and the other by activation features.3 Inhibited depressions appear to
share common mechanisms with unipolar depression,4 and failing to treat such conditions
with antidepressant drugs may increase the risk
of suicide. Activated depressions are closer to
mixed states and may be worsened by antidepressants. This differential response to treatment may
n engl j med 357;6  www.nejm.org  august 9, 2007
account for the lack of efficacy of antidepressants reported in the study by Sachs et al., which
did not assess heterogeneity.
Chantal Henry, M.D., Ph.D.
University Bordeaux 2
33076 Bordeaux, France
[email protected]
Jacques Demotes-Mainard, M.D., Ph.D.
75013 Paris, France
Marion Leboyer, M.D., Ph.D.
INSERM UnitГ© 841
94000 CrГ©teil, France
1. Gijsman HJ, Geddes JR, Rendell JM, Nolen WA, Goodwin
GM. Antidepressants for bipolar depression: a systematic review
of randomized, controlled trials. Am J Psychiatry 2004;161:
2. Benazzi F. Bipolar disorder — focus on bipolar II disorder
and mixed depression. Lancet 2007;369:935-45.
3. Henry C, M’Baïlara K, Poinsot R, et al. Evidence for two
types of bipolar depression using a dimensional approach. Psychother Psychosom (in press).
4. Oquendo MA, Hastings RS, Huang Y-Y, et al. Brain serotonin
transporter binding in depressed patients with bipolar disorder
using positron emission tomography. Arch Gen Psychiatry 2007;
To the Editor: The trial by Sachs and colleagues
is the latest in a string of studies showing that
the addition of antidepressants to mood stabilizers in depressed bipolar patients is of minimal
benefit.1,2 However, antidepressants given in the
absence of a mood stabilizer — either alone (in
type II illness)3,4 or added to a second-generation
antipsychotic drug5 — are effective in treating
bipolar depression. The major difference in the
design of these studies is the absence or presence
of mood stabilizers. When mood stabilizers (lithium or valproate, carbamazepine, or lamotrigine)
are used, antidepressants are of minimal benefit.
This finding has traditionally been interpreted as
showing that mood stabilizers have an intrinsic
antidepressant effect, which cannot be augmented by the antidepressant agent. An alternative
interpretation that would reconcile the discrepant data is that mood stabilizers may actually
interfere with or block the effect of antidepressant drugs. This distinction is important, since it
would alter the approach to treating depression
in patients with bipolar disorder. Studies that specifically examine this question need to be performed.
Rif S. El-Mallakh, M.D.
University of Louisville School of Medicine
Louisville, KY 40202
[email protected]
1. Nemeroff CB, Evans DL, Gyulai L, et al. Double-blind, pla-
cebo-controlled comparison of imipramine and paroxetine in
the treatment of bipolar depression. Am J Psychiatry 2001;158:
2. Post RM, Leverich GS, Nolen WA, et al. A re-evaluation of
the role of antidepressants in the treatment of bipolar depression: data from the Stanley Foundation Bipolar Network. Bipolar
Disord 2003;5:396-406.
3. Amsterdam J. Efficacy and safety of venlafaxine in treatment
of bipolar II major depressive episode. J Clin Psychopharmacol
4. Amsterdam JD, Garcia-EspaГ±a F, Fawcett J, et al. Efficacy
and safety of fluoxetine in treating bipolar II major depressive
episode. J Clin Psychopharmacol 1998;18:435-40.
5. Tohen M, Vieta E, Calabrese J, et al. Efficacy of olanzapine
and olanzapine-fluoxetine combination in the treatment of bipolar I depression. Arch Gen Psychiatry 2003;60:1079-88. [Erratum, Arch Gen Psychiatry 2004;61:176.]
The authors reply: Price and Tyrka comment
on several trends related to the diagnosis of bipolar disorder, which raise concerns we share. However, as we detailed in our article, the Systematic
Treatment Enhancement Program for Bipolar Disorder (STEP-BD) rigorously applied DSM-IV criteria both in the selection of subjects at initial entry
into the program and for inclusion in the randomized study of acute depression. We do not believe
our results necessarily contradict those of previous studies. Instead, we suggest that our conclusions differ from those of previous studies largely
because in our study, durable recovery was the
primary outcome variable. As we stated, the results
of our cross-sectional, last-observation-carriedforward analyses, as typically applied in most previous studies of short-term treatment, revealed a
proportion of subjects treated with antidepressants
who had rates of successful outcome similar to
those reported in previous studies. However, mood
stabilizers alone provided an equivalent benefit
in the depressed subjects with bipolar disorder
whom we studied.
Dr. Henry and colleagues raise the issue of
heterogeneity and suggest a differential response
on the basis of two principal subtypes of depression in bipolar disorder: global inhibition and
activation. Although their proposed typology is
potentially interesting, the existence of such heterogeneity is unlikely to account for the absence
of difference between study groups in our randomized study. Notably, no increase in suicidality
was observed in subjects treated without standard antidepressants. STEP-BD does plan to examine symptom profiles obtained at baseline as
predictors of treatment outcome.
Dr. El-Mallakh makes a scientific point that
cannot be adequately resolved currently. Treating
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
acute bipolar depression for short periods with
antidepressants alone could result in higher response rates than the use of antidepressants in
combination with mood stabilizers. However, a
blinded maintenance study1 reported lower response rates when patients with bipolar disorder
who were receiving placebo under double-blind
conditions became depressed and received adjunctive antidepressants, as compared with the response rates observed when the antidepressants
were added to mood stabilizers in a blinded fashion. To our knowledge, no published study has
had design features that directly address Dr. ElMallakh’s supposition, which requires that patients be randomly assigned to receive antidepressants in combination with a mood stabilizer or
as monotherapy. Since evidence-based treatment
guidelines consistently recommend the concurrent
use of a mood stabilizer, it is likely that ethical
concerns, as well as concerns about practical complexities and study costs, account for the lack of
such initiatives.
Dr. Belmaker’s concern about limited generalizability and low recruitment rates in our study
requires some clarification. Unlike most clinical
m e dic i n e
trials, STEP-BD enrolled a large number of patients
with bipolar disorder who were seeking treatment.
However, the relatively low percentage of depressed patients who underwent randomization
reflected the large number of patients who were
ineligible owing to previous treatment with both
bupropion and paroxetine, an unwillingness to
accept treatment with approved mood stabilizers,
or an unwillingness to taper the dose of a current
antidepressant medication. However, we acknowledge that results from our study may not apply to
all antidepressants, since our study examined only
bupropion and paroxetine.
Gary S. Sachs, M.D.
Harvard Medical School
Boston, MA 02115
Charles Bowden, M.D.
University of Texas Health Science Center
San Antonio, TX 78229-3900
Michael E. Thase, M.D.
University of Pennsylvania
Philadelphia, PA 19104
1. Gyulai L, Bowden CL, McElroy SL, et al. Maintenance effi-
cacy of divalproex in the prevention of bipolar depression. Neuropsychopharmacology 2003;28:1374-82.
Glucose Regulation in Young Adults with Very Low Birth Weight
To the Editor: Hovi et al. (May 17 issue)1 report
that young adults who had a very low birth weight
have higher indexes of glucose intolerance and
higher blood pressure than do those who were
born at term. Although the authors report that
maternal preeclampsia occurred significantly more
frequently in the very-low-birth-weight group (35
of 166) than in the term group (13 of 172,
P<0.001), they did not adjust their comparisons
for this variable.
In our tertiary neonatal intensive care unit in
an urban area in northern Italy, from 2000 to
2006, glucose intolerance (fasting glucose level,
>215 mg per deciliter)2 was recorded in 74 of 469
neonates with a very low birth weight. Hyperglycemia occurred in 22 of 92 (23.9%) of those with
a history of maternal preeclampsia,3 as compared
with 52 of 377 (13.8%) without maternal preeclampsia (95% confidence interval [CI], 1.12 to
3.44; P = 0.02).
Preeclampsia and insulin resistance may be
associated in pregnant women,4 putting stress on
the fetus and causing the onset of fetal hormonal
responses, possibly involving blood pressure and
glucose regulation. Since circumstances during
the fetal period may have lifelong programming
effects on various body functions and susceptibility to disease, our data suggest that preeclampsia — rather than very low birth weight itself
— might account for early disturbances in glucose regulation.
Paolo Manzoni, M.D.
Maria Grazia BaГ№, M.D.
Daniele Farina, M.D.
Sant’Anna Obstetrics and Neonatology Hospital
10126 Turin, Italy
[email protected]
1. Hovi P, Andersson S, Eriksson JG, et al. Glucose regulation
in young adults with very low birth weight. N Engl J Med 2007;
2. Cowett RM, Farrag HM. Selected principles of perinatalneonatal glucose metabolism. Semin Neonatol 2004;9:37-47.
3. Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy. Am
J Obstet Gynecol 2000;183:S1-S22.
4. Montoro MN, Kjos SL, Chandler M, Peters RK, Xiang AH,
Buchanan TA. Insulin resistance and preeclampsia in gestational diabetes mellitus. Diabetes Care 2005;28:1995-2000.
n engl j med 357;6  www.nejm.org  august 9, 2007
The authors reply: Manzoni et al. raise the important issue of maternal preeclampsia, which is
a frequent reason for prematurity. Previous reports
suggest that offspring of mothers with preeclampsia have higher-than-normal blood pressure in
childhood1 and impaired glucose regulation in
adulthood.2 In our study, we decided not to preВ­
sent results adjusted for preeclampsia because
such an adjustment would not have altered the
results. We here present numerical data to support this decision.
Young adults born to mothers with preeclampsia had glucose and insulin concentrations similar to those of the rest of the subjects. However,
their systolic blood pressure was increased by
3.8 mm Hg (95% CI, 0.2 to 7.4). After adjustment
for preeclampsia, very low birth weight was associated with an increase in the fasting insulin level
of 16.7% (95% CI, 4.2 to 30.6), an increase in the
2-hour glucose level of 6.4% (95% CI, 0.3 to 12.7),
and an increase in systolic blood pressure of 4.4
mm Hg (95% CI, 1.7 to 7.1). These estimates are
similar to those we reported without adjustment
for preeclampsia. Also, when we excluded subjects
with a history of maternal preeclampsia, the effects of very low birth weight remained similar
and statistically significant. Therefore, we conclude that maternal preeclampsia did not explain
the association of very low birth weight with
impaired glucose regulation and higher blood
Petteri Hovi, M.D.
National Public Health Institute
FI-00300 Helsinki, Finland
Sture Andersson, M.D., Ph.D.
Hospital for Children and Adolescents
FI-00029 HUS Helsinki, Finland
Eero Kajantie, M.D., Ph.D.
National Public Health Institute
FI-00300 Helsinki, Finland
1. Tenhola S, Rahiala E, Martikainen A, Halonen P, Voutilainen
R. Blood pressure, serum lipids, fasting insulin, and adrenal
hormones in 12-year-old children born with maternal preeclampsia. J Clin Endocrinol Metab 2003;88:1217-22.
2. Thomas C, Hyppönen E, Power C. Prenatal exposures and
glucose metabolism in adulthood: are effects mediated through
birth weight and adiposity? Diabetes Care 2007;30:918-24.
Use of Physicians’ Services for Medicare Beneficiaries
To the Editor: The Specialty Society Relative
Value Update Committee of the American Medical Association (AMA) has two primary concerns
regarding the article by Maxwell et al. (May 3 issue).1 First, we are unable to reconcile the data in
Table 1 of the article with our recommendations
in this time period. For example, our recommendations for “tests” represent a 2% increase, not a
68% increase, as asserted in the article. Second,
the authors did not consider major changes from
the most recent 5-year review of the resourcebased relative-value scale. A comparison of 1992
and 2007 work relative-value units (RVUs), with
the 2005 utilization data used as a common multiplier, reveals that evaluation and management
work valuation has increased 45%, while imaging
work valuation has decreased 6% since 1992.
We question the appropriateness of excluding
the most significant change and subsequent redistribution since the inception of the resourcebased relative-value scale — the revaluation of
evaluation and management services in 2007. We
agree that the scale’s improvements should be
ongoing, and we have submitted to the Centers
for Medicare and Medicaid Services (CMS) a number of significant recommendations related to the
practice-expense payments for consideration in
the 2008 Medicare resource-based relative-value
William L. Rich, III, M.D.
American Medical Association
Chicago, IL 60610
1. Maxwell S, Zuckerman S, Berenson RA. Use of physicians’
services under Medicare’s resource-based payments. N Engl J
Med 2007;356:1853-61.
To the Editor: Maxwell et al. analyzed the effect
of the resource-based relative-value scale, reporting that the share of Medicare total RVUs for evaluation and management services did not change
between 1992 and 2002. In an accompanying editorial, Newhouse describes concerns about potential underpayment for these services.1
My colleagues and I believe the Berenson–
Eggers Type of Service (BETOS) classification used
by Maxwell et al. does not provide sufficiently
accurate clinical-service-group delineation. For example, the BETOS evaluation and management
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
Table 1. Effect of the Resource-Based Relative-Value Scale on Evaluation and Management Services and Major Procedures, 1992–2002.
Service Category
Average Allowed RVUs
per Service
Average Annual Allowed
RVU per Beneficiary
Average Allowed Charge
per Service
Average Annual Allowed
Charges per Beneficiary
Evaluation and
Major procedures
category includes pathology services, and the
BETOS major procedures category excludes highvolume surgical procedures such as cataract extractions.
We analyzed Medicare data files during the
same period, defining evaluation and management
services as office, hospital, emergency-department, and nursing-facility services and home
visits, plus consultations and critical care. Major
procedures were defined as surgery with 90-day
global periods.
With the use of more refined service assignments, we identified dramatic growth in evaluation and management RVUs and payments, both
per service and per beneficiary (Table 1); this
growth was not shown by the proportional analysis of Maxwell et al. In addition, we identified
large reductions in payments for major procedures that were not noted by Maxwell and colleagues. An analysis of data from recent years is
necessary to define the current effect of the resource-based relative-value scale.
Thomas R. Russell, M.D.
American College of Surgeons
Chicago, IL 60611
1. Newhouse JP. Medicare spending on physicians — no easy
fix in sight. N Engl J Med 2007;356:1883-4.
To the Editor: The article by Maxwell and colleagues is a welcome addition to the literature on
Medicare’s resource-based relative-value scale payment system. Using similar methods, we likewise
found that the increase in the volume and intensity mix of physician services was the largest
source of growth over the first 5 years of this
system.1 In addition, we found that, at most, half
of the growth in volume and intensity was due to
demographic factors such as the aging of the
population, selective enrollment of healthier beneficiaries in managed-care plans, and changes in
the prevalence of health conditions. Also, increases in service use were surprisingly uniform
across conditions, indicating that technological
advances and changing standards of care are affecting most categories of patients. This finding
poses difficult challenges to policymakers seeking to control Medicare spending.
Melinda B. Buntin, Ph.D.
Jose J. Escarce, M.D., Ph.D.
Dana Goldman, Ph.D.
Santa Monica, CA 90407
[email protected]
1. Buntin MJB, EscarcГ© JJ, Goldman D, Kan H, Laugesen MJ,
Shekelle P. Increased Medicare expenditures for physicians’ services: what are the causes? Inquiry 2004;41:83-94.
The Authors Reply: We agree with Rich that the
most recent 5-year review of the resource-based
relative-value scale is relevant to our findings.
Data from this review were not available to us
during our study. We discuss this review in our
article, noting the increases in work RVUs of selected evaluation and management services and
factors that diluted the effect of those increases
in terms of actual payments to physicians.1
Rich reports that on the basis of a comparison of 1992 and 2007 RVUs, with 2005 utilization
data used as a common multiplier, tests reviewed
by the Specialty Society Relative Value Update
Committee of the AMA resulted in an estimated
2% increase in work volume. Since this is lower
than our estimate, we reviewed our data and
found that the discrepancy arose from four test
codes that were incorrectly categorized in the RVU
files provided to us. These tests had low 1992
work volumes and high 2002 work volumes,
which resulted in an overestimation of growth
due to RVU revisions and an underestimation of
growth due to service quantity and mix. When
we applied the calculation to the corrected data
file, we found an increase of 11% in work volume for all tests due to RVU revisions from the
5-year review and annual revision processes. Be-
n engl j med 357;6  www.nejm.org  august 9, 2007
cause tests are less than 3% of work volume, this
change had a very limited effect on our other
Russell questions the BETOS classification2
and finds different results when the service groupings of the American College of Surgeons are
used. BETOS is used by the CMS, the Medicare
Payment Advisory Commission, and many health
services researchers as the standard classification system for physicians’ services. On the basis
of an examination of only surgical procedures
with 90-day global periods, Russell reports declines in total RVUs and charges per beneficiary.
Our study also showed declines in the total volume for some groups within the major procedures
category. Russell found higher increases in work
volume when the cognitive services of some specialists were removed from the evaluation and
management category. Excluding those services,
we also find higher increases in work volume
among the remaining services in this BETOS
The study by Buntin and colleagues provided
a valuable early analysis of the sources of growth
in expenditures for physicians’ services during the
5 years after implementation of the physician fee
schedule.3 They used the Medicare Current Beneficiary Survey, which permitted them to conduct
a detailed investigation of beneficiary health status and sociodemographic factors as sources of
expenditure growth.
Stephanie Maxwell, Ph.D.
Stephen Zuckerman, Ph.D.
Robert A. Berenson, M.D.
Urban Institute
Washington, DC 20037
1. Centers for Medicare & Medicaid Services. Medicare pro-
gram: five-year review of work relative value units under the
physician fee schedule and proposed changes to the practice
expense methodology: notice. Fed Regist 2006;71(125):37255.
2. Idem. Berenson-Eggers Type of Service (BETOS). (Accessed
July 19, 2007, at http://www.cms.hhs.gov/HCPCSReleaseCodeSets/
3. Buntin MJB, EscarcГ© JJ, Goldman D, Kan H, Laugesen MJ,
Shekelle P. Increased Medicare expenditures for physicians’ services: what are the causes? Inquiry 2004;41:83-94.
Video on Orotracheal Intubation
To the Editor: In the Video in Clinical Medicine
about orotracheal intubation, presented by Kabrhel
et al. (April 26 issue),1 the authors state that the
“combination of flexion of the neck and extension of the head [the sniffing position] improves
the alignment of the axes of the oral cavity, pharynx, and larynx, facilitating optimal visualization
of the vocal cords.” This is an anatomical myth
that is not supported by clinical evidence.2 In a
randomized crossover study of 456 consecutive
patients, Adnet et al.3 found no differences in
glottic visualization by direct laryngoscopy or in
the score on the intubation-difficulty scale between the use of simple head extension and routine sniffing position. Furthermore, in 11% of patients, glottic exposure was lessened with the use
of the “three axes alignment.” However, a multivariable analysis showed that reduced neck mobility and obesity were independently related to
improvement in the laryngoscopic view with the
use of the sniffing position.3 These results are
consistent with studies of other populations4 and
with anatomical studies involving magnetic resonance imaging.5 In conclusion, routine use of the
sniffing position should not be a standard procedure in orotracheal intubation.
David A. RincГіn, M.D.
Universidad Nacional de Colombia
BogotГЎ 11001000, Colombia
[email protected]
1. Kabrhel C, Thomsen TW, Setnik GS, Walls RM. Orotracheal
intubation. N Engl J Med 2007;356:e15 (Web only). (Available at
http://content.nejm.org/cgi/content/full/356/17/e15.) [Erratum,
N Engl J Med 2007;356:2228.]
2. Adnet F, Borron SW, Lapostolle F, Lapandry C. The three
axis alignment theory and the “sniffing position”: perpetuation
of an anatomic myth? Anesthesiology 1999;91:1964-5.
3. Adnet F, Baillard C, Borron SW, et al. Randomized study
comparing the “sniffing position” with simple head extension
for laryngoscopic view in elective surgery patients. Anesthesiology 2001;95:836-41.
4. Collins JS, Lemmens HJM, Brodsky JB, Brock-Utne JG, Levitan RM. Laryngoscopy and morbid obesity: a comparison of the
“sniff” and “ramped” positions. Obes Surg 2004;14:1171-5.
5. Adnet F, Borron SW, Dumas JL, Lapostolle F, Cupa M, Lapandry C. Study of the “sniffing position” by magnetic resonance imaging. Anesthesiology 2001;94:83-6.
To the Editor: In their video on orotracheal intubation, Kabrhel et al. suggest that neuromuscular-blocking agents and sedatives will “prevent
the patient from vomiting.” Readers should not
be comforted that the use of these agents will
prevent emesis. Etomidate, the medication mentioned in the video, is known to be emetogenic
and has been shown to be associated with emesis
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
during intubation.1 In addition, we were disappointed that there was no mention of the difficult-airway algorithm of the American Society
of Anesthesiologists, an evidence-based practice
guideline on this important topic that has guided airway management for 15 years.2
Sachin Kheterpal, M.D.
George A. Mashour, M.D., Ph.D.
sure on the cricoid cartilage during emergency
intubation (the Sellick maneuver), even though
this maneuver is not always used. The usefulness
of cricoid pressure in preventing aspiration has
never been proved, and there is ample literature
on both sides of the issue. What is not in doubt
is that abnormally applied cricoid pressure can
lead to partial or complete obstruction of the airway, making ventilation and intubation difficult
University of Michigan
Ann Arbor, MI 48109
or impossible. Also not in doubt is the fact that
[email protected]
very few practitioners know where the cricoid
1. Bozeman WP, Young S. Etomidate as a sole agent for endo- cartilage is and how to hold 30 N of pressure.
tracheal intubation in the prehospital air medical setting. Air
One other indisputable fact is that aspiration,
Med J 2002;21:32-7.
2. Practice guidelines for management of the difficult airway: though potentially fatal, is very rare if intubation
an updated report by the American Society of Anesthesiologists is performed correctly. Thus, I would recommend
Task Force on Management of the Difficult Airway. Anesthesiolthat the emphasis on the Sellick maneuver be
ogy 2003;98:1269-77. [Erratum, Anesthesiology 2004;101:565.]
completely removed from this video, since it is
more likely to cause harm than good if the perTo the Editor: Kabrhel et al. recommend use of son performing the intubation is inexperienced.
a carbon dioxide detector to confirm the place- Nir Hoftman, M.D.
ment of the endotracheal tube. Furthermore, they University of California, Los Angeles
suggest that the device be used to detect carbon Los Angeles, CA 90049
dioxide consistently within the first six breaths
after intubation and during subsequent breaths. The Authors Reply: The concise format of proThey caution that false negative results with re- cedural videos, such as ours, limits what can be
gard to tube placement may occur in patients in addressed. Though the sniffing position is the
cardiac arrest, in whom carbon dioxide may not most commonly recommended position to facilibe present in the lungs.
tate orotracheal intubation and is considered the
It would have been relevant to point out the standard of care,1 its usefulness has been debatpossibility of color change in the device due to ed. The crossover study cited by Dr. RincГіn comagents other than exhaled carbon dioxide. Gastric pared the sniffing position with head extension
contents, mucus, and drugs such as epinephrine but not with neutral position.1 Considering the
can cause false positive results.1,2 It is imperative study design, it is somewhat misleading to focus
that clinicians using these devices be aware of only on the 11% of patients in whom neck extenthis limitation. One way to avoid this pitfall is to sion alone provided better glottic exposure withobserve the change in color in the device with out mentioning the 18% of patients in whom the
each breath. A false positive result causes a per- sniffing position provided a better view. Studies
manent color change in the device; hence, the have demonstrated an improved laryngoscopic
color does not vary with ventilation.
view with head elevation as compared with neutral position.2 However, we agree that for most
Venkatesh Srinivasa, M.D.
patients, the benefit of the sniffing position is
VA Boston Healthcare System
West Roxbury, MA 02132
probably conferred by the head extension it [email protected]
vides. As such, we do not find a strong reason to
Bhavani Shankar Kodali, M.D.
abandon a long-standing practice, and we agree
Brigham and Women’s Hospital
with Adnet et al.,1 who state that their study is
Boston, MA 02115
“less an indictment of the sniffing position than
1. Brackney SM, Nellcor Puritan Bennett. Caution when using reinforcement of the importance of head extencolorimetry to confirm endotracheal intubation. Anesth Analg
sion in all patients.”
We agree with Drs. Kheterpal and Mashour
2. Srinivasa V, Kodali BS. Caution when using colorimetry to
confirm endotracheal intubation. Anesth Analg 2007;104:738.
that the difficult-airway algorithm of the American Society of Anesthesiologists is a very useful
To the Editor: The video on orotracheal intuba- guide. However, it is perhaps most applicable to
tion places special emphasis on the use of pres- the planning of operating-room cases and is less
n engl j med 357;6  www.nejm.org  august 9, 2007
helpful when intubation is emergent or unavoidable. They are correct in stating that etomidate
can cause emesis, but this is not an issue when
neuromuscular-blocking agents are also used.
Unfortunately, detailed discussions of the American Society of Anesthesiologists algorithm and
pharmaceutical adjuncts were beyond the scope
of our video. We also agree with Drs. Srinivasa
and Kodali that during assessment of end-tidal
carbon dioxide levels, the detector should be free
of foreign material (e.g., vomit, mucus, or drugs)
and that the color should vary during the respiratory cycle.
Regarding Dr. Hoftman’s comments: we acknowledge in the text accompanying the video
that cricoid pressure can distort the airway and
that the benefit of decreased aspiration is theoretical. The maneuver is also frequently performed
improperly. Although the evidence is mixed as to
whether cricoid pressure improves or degrades
intubating conditions, surveys show that it is al-
most universally performed,3 so we thought it best
to describe the technique. No doubt, future studies will clarify the precise role, if any, for this
Christopher Kabrhel, M.D.
Massachusetts General Hospital
Boston, MA 02114
Todd W. Thomsen, M.D.
Mount Auburn Hospital
Cambridge, MA 02138
Ron Walls, M.D.
Brigham and Women’s Hospital
Boston, MA 02115
1. Adnet F, Baillard C, Borron SW, et al. Randomized study
comparing the “sniffing position” with simple head extension
for laryngoscopic view in elective surgery patients. Anesthesiology 2001;95:836-41.
2. Levitan RM, Mechem CC, Ochroch EA, Shofer FS, Hollander
JE. Head-elevated laryngoscopy position: improving laryngeal
exposure during laryngoscopy by increasing head elevation. Ann
Emerg Med 2003;41:322-30.
3. Morris J, Cook TM. Rapid sequence induction: a national
survey of practice. Anaesthesia 2001;56:1090-7.
Case 14-2007: A Man with Pain
and Swelling of Both Eyes and the Right Ear
To the Editor: In the Case Record of a man
with fever and pain and swelling of both eyes and
the right ear, presented by Butterton and colleagues (May 10 issue),1 the right ear is described
as “tender, with erythema of the external canal
involving the preauricular and postauricular soft
tissue and pinna.” This is repeated in the discussion. The point that is not emphasized but should
be is that, although the external ear was inflamed,
the earlobe was spared. This is evident even in
the clinical photograph. Sparing of the earlobe is
in itself diagnostic of polychondritis.2,3 This clinical feature alone helps in differentiating the case
from other conditions such as cellulitis or malignant otitis externa, in which the earlobe is also
inflamed.2 The earlobe is spared in polychondritis because it is devoid of cartilage, and only cartilaginous structures are involved. Therefore,
while one is evaluating a patient for red ear, it is
pertinent to look specifically for inflammation of
the earlobe, the absence of involvement of which
is diagnostic of polychondritis. However, under
certain circumstances, leprosy, leishmaniasis,
frostbite, and trauma can be considered in the
differential diagnosis of auricular chondritis.2,3
Sandeep Chauhan, M.D.
Government Medical College & Hospital
Chandigarh 160030, India
[email protected]
Vikas Agarwal, M.D., D.M.
Sanjay Gandhi Postgraduate Institute of Medical Sciences
Lucknow 226014, India
Sanjay D’Cruz, M.D., D.M.
Government Medical College & Hospital
Chandigarh 160030, India
1. Case Records of the Massachusetts General Hospital (Case
14-2007). N Engl J Med 2007;356:1980-8.
2. Piette J-C, Vinceneux P. Relapsing polychondritis. In: Harris
ED, Budd RC, Firestein GS, et al., eds. Kelley’s textbook of rheumatology. 7th ed. Philadelphia: Elsevier Saunders, 2005:1541-6.
3. Gilliland BC. Relapsing polychondritis. In: Kasper DL,
Braunwald E, Fauci AS, Hauser SL, Longo DL, Jameson JL, eds.
Harrison’s principles of internal medicine. 16th ed. New York:
McGraw-Hill, 2005:2015-6.
The Discussants Reply: Chauhan et al. highlight the specificity of sparing of the earlobe in
the diagnosis of relapsing polychondritis. In this
case, the observation that swelling involved only
the cartilaginous helix, and not the lobe, was not
made until we first saw the patient on the fourth
hospital day. Therefore, the text reflects how the
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
case was presented to us, with ambiguity as to
which part of the ear was swollen. The sparing of
the lobe is described at the end of the case history. We wanted to permit readers to make this
observation on their own from Figure 1 of the
article and to draw their own conclusions.
The swelling of the helix with sparing of the
earlobe, in the context of this patient’s history,
led us immediately to the diagnosis of relapsing
polychondritis and decreased our suspicion of
infectious causes. We thus started corticosteroid
treatment before the results of the biopsy were
m e dic i n e
available and before this case met all the diagnostic criteria listed in Table 3 of the Case Record.
However, we emphasize that there is no laboratory, radiographic, or physical finding that is
pathognomonic of relapsing polychondritis. Auricular chondritis alone, even with sparing of the
lobe, is not absolutely diagnostic of this disease.
Deborah S. Collier, M.D.
Joan R. Butterton, M.D.
Massachusetts General Hospital
Boston, MA 02114
Pregnancy in a Patient with Congenital Erythropoietic Porphyria
To the Editor: Congenital erythropoietic porphyria, a rare autosomal recessive disorder of heme
biosynthesis, results from markedly deficient uroporphyrinogen III synthase activity.1,2 Data regarding pregnancy in patients with this condition are lacking, which may be explained by the
rarity of the disease, as well as social restrictions
due to associated photosensitivity or disfiguring
effects in more severely affected patients. We describe a successful pregnancy in a patient with
congenital erythropoietic porphyria.
The patient presented at 3 months of age with
pink urine and severe photosensitivity; congenital erythropoietic porphyria was diagnosed on the
basis of analysis of urinary, fecal, and blood porphyrin. Bone marrow transplantation, an established treatment option,3 was considered at the
age of 18 months but was not pursued, owing to
concerns about risks and the lack of a suitable
donor. At the age of 9 years, in an attempt to
suppress erythropoiesis and thus reduce the overproduction of porphyrin, hypertransfusion was
initiated in combination with subcutaneous administration of deferoxamine but was discontinued because of its minimal efficacy. Analysis of
the uroporphyrinogen III synthase gene identified Cys73→Arg and IVS8-23A→G mutations, a
combination previously described in a patient
moderately affected with congenital erythropoietic porphyria.4
At 19 years of age, our patient became pregnant. There were no antenatal complications. The
plasma porphyrin level increased during the second trimester and fell during the third trimester
(Fig. 1), changes that were consistent with physi622
ologic changes in erythropoiesis and fluid balance
during pregnancy.5 At 39 weeks of gestation,
elective caesarean section for breech presentation was performed under dimmed lights with
minimal exposure of the patient’s skin, without
complication. A female infant with normal Apgar scores and birth weight of 3.52 kg was delivered. The plasma porphyrin level in the infant
was 231 nmol per liter at the time of the birth
(adult reference range, 0 to 11.2 nmol per liter)
and decreased to 104 nmol per liter at 48 hours.
No follow-up blood samples were available.
Teeth erupted when the infant was 6 months
of age; they showed distal brown pigmentation,
with a sharply defined margin — consistent with
porphyrin staining — where calcification had occurred before birth in association with exposure
to maternal porphyrin (i.e., erythrodontia) (Fig.
2). This condition has been recognized in patients with congenital erythropoietic porphyria.
The child has had normal development and no
photosensitivity during 3 years of follow-up.
The severity of congenital erythropoietic porphyria correlates with the degree of residual uroporphyrinogen III synthase activity, with more
activity resulting in less severe disease.1,2 The
patient’s early clinical presentation was assumed
to indicate severe disease. In retrospect, the clinical course suggested a moderate clinical phenotype, confirmed by subsequent genotype studies.
Cys73в†’Arg is the most frequently reported mutation in patients with congenital erythropoietic
porphyria, accounting for one third of affected
alleles,1,2 and leads to complete loss of function
of uroporphyrinogen III synthase. In contrast, the
n engl j med 357;6  www.nejm.org  august 9, 2007
Plasma Porphyrin Level (nmol/liter)
congenital erythropoietic porphyria who are not
severely affected may do well with conservative
management and that successful pregnancy is possible in such patients.
Nirmala Hallai, M.R.C.P.
Royal Hallamshire Hospital
Sheffield S10 2JF, United Kingdom
Alexander Anstey, F.R.C.P.
Royal Gwent Hospital
Newport NP20 2UB, United Kingdom
Gestation (wk)
Figure 1. Plasma Porphyrin Levels during Pregnancy.
range Badminton
is 0 to 11.2 nmol per RETAKE
liter. On the
x axis,
indicates1 18
of 2weeks before conception; the2nd
between values vary.
Figure has been redrawn and type has been reset.
Please check carefully.
JOB: 35706
ISSUE: 08-09-07
Figure 2. Brown Staining of the Infant’s Teeth.
IVS8-23A→G mutation occurs at a branch point,
and some function is retained.3
The present case indicates that patients with
Shevaun Mendelsohn, F.R.C.P.
John Williams, F.R.C.O.G.
Gareth Evans-Jones, F.R.C.P.H.
Sadia Malick, M.B., B.S.
Countess of Chester Hospital
Chester CH2 1UL, United Kingdom
Michael N. Badminton, Ph.D., F.R.C.Path.
Cardiff University
Cardiff CF14 4XN, United Kingdom
[email protected]
1. Desnick RJ, Astrin KH. Congenital erythropoietic porphyria:
advances in pathogenesis and treatment. Br J Haematol
2. De Verneuil H, Ged C, Moreau-Gaudry F. Congenital erythropoietic porphyria. In: Kadish KM, Smith KM, Guilard R, eds.
The porphyrin handbook. Vol. 14. Medical aspects of porphyrins. San Diego, CA: Academic Press, 2003:43-63.
3. Tezcan I, Xu W, Gurgey A, et al. Congenital erythropoietic
porphyria successfully treated by allogeneic bone marrow transplantation. Blood 1998;92:4053-8.
4. Fontanellas A, Bensidhoum M, Enriquez de Salamanca R,
Moruno Tirado A, de Verneuil H, Ged C. A systematic analysis of
the mutations of the uroporphyrinogen III synthase gene in congenital erythropoietic porphyria. Eur J Hum Genet 1996;4:27482.
5. Milman, Byg KE, Agger AO. Haemoglobin and erythrocyte
indices during normal pregnancy and postpartum in 206 women
with and without iron supplementation. Acta Obstet Gynecol
Scand 2000;79:89-98.
Correspondence Copyright В© 2007 Massachusetts Medical Society.
instructions for letters to the editor
Letters to the Editor are considered for publication, subject to editing and abridgment, provided they do not contain material
that has been submitted or published elsewhere. Please note the following: •Letters in reference to a Journal article must not
exceed 175 words (excluding references) and must be received within 3 weeks after publication of the article. Letters not
related to a Journal article must not exceed 400 words. All letters must be submitted over the Internet at http://authors.nejm.org.
•A letter can have no more than five references and one figure or table. •A letter can be signed by no more than three authors.
•Financial associations or other possible conflicts of interest must be disclosed. (Such disclosures will be published with the
letters. For authors of Journal articles who are responding to letters, this information appears in the published articles.)
•Include your full mailing address, telephone number, fax number, and e-mail address with your letter.
Our Web site: http://authors.nejm.org
We cannot acknowledge receipt of your letter, but we will notify you when we have made a decision about publication. Letters
that do not adhere to these instructions will not be considered. Rejected letters and figures will not be returned. We are unable
to provide prepublication proofs. Submission of a letter constitutes permission for the Massachusetts Medical Society, its
licensees, and its assignees to use it in the Journal’s various print and electronic publications and in collections, revisions, and
any other form or medium.
n engl j med 357;6  www.nejm.org  august 9, 2007
n e w e ng l a n d j o u r na l
m e dic i n e
b o ok r e v ie w s
Preventing Medication Errors
(Quality Chasm Series.) By the Committee on Identifying
and Preventing Medication Errors and the Board on Health
Care Services. Edited by Philip Aspden, Julie A. Wolcott,
J. Lyle Bootman, and Linda R. Cronenwett. 463 pp.
Washington, DC, National Academies Press, 2007.
$54.95. ISBN 978-0-309-10147-9.
Medication Errors
Second edition. Edited by Michael R. Cohen. 680 pp.,
illustrated. Washington, DC, American Pharmacists
Association, 2007. $89.95. ISBN 978-1-58212-092-8.
ealth care is inherently risky, and
filling a handwritten prescription could be
the most dangerous of all medical procedures.
Medication errors, which are often caused by illegible handwriting, are a subgroup of medical
errors and may cause as many as 7000 of the
98,000 deaths in the United States each year that
the Institute of Medicine (IOM) attributes to medical errors. Now the IOM, in its new report in the
Quality Chasm Series, Preventing Medication Errors,
estimates that in the United States, the average
hospital patient is subject to at least one medication error per day and that each year more than
1.5 million patients have injuries resulting from
preventable adverse drug events. The new study
covers the development of an evidence-based review of drug safety as well as guidelines for policymakers and government agencies.
The IOM authorship committee (disclosure: two
of its members are employees of Intermountain
Healthcare) has clearly fulfilled its charge concerning prescription and over-the-counter drugs
and complementary and alternative medications.
However, the book has little practical advice for
clinicians. Indeed, the major purpose of the committee was to recommend agendas for consumers,
health care organizations, and the industries and
regulatory agencies involved with drugs, medical
devices, and health care information technology.
I was surprised that the evidence for the most
highly touted prevention strategies, such as computerized provider order entry, computer-based
decision support systems, smart pumps, and bar
coding, was reviewed only in the last appendix.
The second edition of the book Medication Errors, edited by Michael R. Cohen, takes up where
the IOM report leaves off and complements it
with detailed information that is useful for clinicians and clinical pharmacists. A pharmacist and
the president of the nonprofit Institute for Safe
Medication Practices (ISMP), Cohen has been a
voice crying out in the wilderness for three deВ­
cades and was a 2005 recipient of a John D. and
Catherine T. MacArthur Foundation “genius
grant.” Cohen is coeditor of a widely read periodical published by the ISMP that focuses on
“high-alert medications” and recent problems with
particular drugs. He devised a method that uses
“tall man” lettering to prevent errors in the use of
drugs with look-alike, sound-alike names, and he
also created a voluntary system for reporting errors, now called the United States Pharmacopeia–
ISMP Medication Errors Reporting Program.
Cohen’s ISMP proficiency tests have shown the
inability of many pharmacy computer systems to
detect unsafe and potentially fatal medication
orders and have been an important impetus for
the design of safer systems.
Cohen defines a medication error as “any preventable event that may cause or lead to inappropriate medication use or patient harm while the
medication is in the control of the health care
professional, patient, or consumer.” This definition includes errors at any stage of medication use
and not just during the process of drug administration. Cohen believes that the “five rights” of
safe medication use (the right patient, right drug,
right time, right dose, and right route of administration) place too much focus on individual performance and overlook the systemic problems
that underlie the human errors. He emphasizes
that “finding out who was involved is less important than learning what went wrong, how,
and why.”
The IOM report represents, so far as possible,
a consensus of the committee and looks toward
fundamental reforms in the health care system,
including the pharmaceutical industries and regulatory agencies. In addition to systemic reforms,
its authors mention some of the things that individual physicians, nurses, pharmacists, and caregivers can do to prevent errors. Cohen’s book, on
the other hand, contains chapters by authorities
from a variety of disciplines, several of whom
n engl j med 357;6  www.nejm.org  august 9, 2007
— including Cohen — also contributed to the
IOM report. Cohen’s book better describes specific medication errors and how to prevent them.
The centerpiece is an excellent chapter on “highalert medications,” which Cohen defines as drugs
with a high risk of causing patient injury or
death if they are misused; the chapter includes a
comprehensive table of safety measures for various drug classes. Many physicians who trained
in the old days will be surprised to learn that
there are good reasons for using trade names as
well as generic names when prescribing certain
medications, such as lipid-based amphotericin B
Both books call for the elimination of handwritten prescriptions and for the use of electronic
decision-support systems by prescribers and pharmacies by the year 2010. Yet the authors of both
books also recognize that electronic prescribing
is not a panacea. The IOM committee also calls
for a gradual shift of the research agenda from
defining incidence rates to prevention strategies
and suggests a starting budget of $50 million to
$100 million to fund these studies through the
federal Agency for Healthcare Research and Quality. However, the committee does not further explain its rationale.
Medication errors are a complex epidemiologic
problem as well as a quality problem, and those
who study such errors are saddled with a profusion of poorly defined concepts, such as “near
miss,” “intercepted error,” and “potential adverse
drug events.” The focus on errors has been useful because it draws attention to preventability.
However, the definitions that have been used for
various types of errors and adverse events vary
widely, and the IOM committee calls for an international consensus conference to define terms.
This effort is analogous to the methods used by
intensivists to define the sepsis syndrome. Most
medication errors, such as a missed dose, do not
cause harm to patients, and conversely, most adverse drug events are not due to mistakes. In addition, certain types of adverse drug events that
once were considered nonpreventable can now be
ameliorated by intensive monitoring and early recognition — or even prevented entirely. For example, the failure to adjust the dose of a nephrotoxic
antibiotic when a patient’s renal function declines
would probably not precipitate an incident report
to the voluntary reporting systems in most hospitals. Yet the number of these errors can be re-
Example of a Potentially Fatal Computerized OrderEntry Error Intercepted by a Pharmacist.
Mix-ups due to large differences in the dosages of conventional and liposomal amphotericin products have
led to deaths.
Courtesy of John Benson, Pharm.D.
book reviews
duced with the help of computer-based decisionsupport systems and electronic health records.
Both of these books provide reviews of methods that are purported to increase drug safety,
such as improved drug naming, labeling, and
packaging; analysis of the root causes of important adverse drug events; failure mode and effects
analysis of the processes of medication use; and
voluntary reporting of important adverse events.
The book by Cohen describes all these methods
in greater depth than the IOM committee does.
It also includes chapters that discuss how to
disclose errors to patients and accounts of some
of the emotional experiences of providers who
have made fatal errors involving medication.
These books with similar titles are authoritative, thorough, and well written. Both aim to improve drug safety and together represent the most
up-to-date and broadest coverage of the subject
now available. Both these books are fact-filled and
visionary, but unfortunately, the benefits of many
of the drug safety measures could have been better assessed and measured in the clinical setting.
Nonetheless, the books might provoke varied responses from readers: frustration or despair because of the enormous gulf between what we do
and what we know, admiration of what has been
accomplished, and hope for the future.
John P. Burke, M.D.
LDS Hospital
Salt Lake City, UT 84103
[email protected]
n engl j med 357;6  www.nejm.org  august 9, 2007
new england journal of medicine
Medical Management
of Vulnerable and Underserved
Patients: Principles, Practice,
and Populations
Edited by Talmadge E. King, Jr., Margaret B. Wheeler,
and Andrew B. Bindman, with four others. 454 pp.,
illustrated. New York, McGraw-Hill, 2007. $41.95.
ISBN 978-0-07-144331-9.
any physicians consider the care of
vulnerable and underserved patients to be
an unsatisfying and often frustrating endeavor,
but this new book offers an optimistic and evidence-based approach to the care of such patients.
The editors define medically vulnerable populations as “those that are wounded by social forces
placing them at a disadvantage for their health.”
The book serves as an important reminder that
many of the issues it raises are relevant to the
broader landscape of American health care. The
authors elicit the voice of the patient to explain
and offer guidance on some of the difficult clinical and societal issues that physicians face in
practice, including behavioral change, noncompliance, health literacy, health disparities, and the
plight of the hidden poor. The book should become an essential primer for teachers, physiciansin-training, and practicing physicians because it
offers important tools for teaching, critical concepts, common pitfalls, core competency concepts,
and discussion questions for each chapter.
A challenge for the editors and authors of this
kind of book is to emphasize the importance of
the physician’s own attitudes and practices toward vulnerable patients without being preachy.
The editors’ decision to use the patient’s voice
and a strong, evidence-based orientation helps
them meet this challenge. The first part of the
book, “Principles,” includes a chapter on the theoretical groundwork and context for understanding
care of complex patients. There are also excellent
reviews of disparities in health and health care
and of the financing and organization of health
care for vulnerable populations. Other chapters
are excellent primers on legal and ethical principles and are written by respected leaders in the
There is a remarkable focus on what health
care providers and systems can actually do to improve care and reduce disparities. The chapters in
the second section of the book, “Practice,” contain
important insights on behavioral change, cross626
cultural communication, health literacy, home
visits, group visits, interactive health technology,
quality improvement, case management, and care
in the context of community. The author of a
superb chapter on cross-cultural communication
emphasizes the importance of assessing one’s own
values in order to improve care and communication. A chapter on behavioral change offers some
practical tools to use with patients who have resistant behaviors in the context of cultural and
financial pressures. The authors of the chapter
on medication adherence make the case that this
problem is not unique to vulnerable populations,
and they offer some common-sense approaches.
The chapter on limited health literacy should be
required reading for all medical students and physicians. The role of interactive health technology
in filling in safety-net gaps is emphasized, with
real-life examples of how health information technology has been transformative in the care of
vulnerable patients facing language and cultural
The third section of the book deals with a variety of vulnerable populations that need special
consideration when they seek care. The chapters
on many different population groups stand on
their own and offer important lessons for the care
of patients who often are not mentioned in standard medical textbooks. These groups include patients with a history of incarceration, patients
with limited English proficiency, immigrants,
patients with disabilities, patients who have been
victims of domestic violence, patients with chronic pain, and patients with substance-abuse problems. The information is provided in a nonjudgmental manner on the basis of strong evidence
and with practical guides for improving care. The
chapters on immigrant health and language barriers offer useful guidance, such as advice on how
to work with interpreters, that could help to
bridge divides. The authors of chapter 20, “The
Hidden Poor: Care of the Elderly,” emphasize functional assessment, elder abuse, and the limitations
of prescription-drug coverage for the elderly. Several chapters in the third section serve to remind
us to consider care in the context of family and
community, such as chapter 19, “The Family as
the Context for Care,” and chapter 38, “HIV/AIDS:
Impact on Vulnerable Populations.” Chapter 40,
“Caring for Oneself While Caring for Others,”
offers important insights on reducing clinician
n engl j med 357;6  www.nejm.org  august 9, 2007
book reviews
This book is an ambitious and important contribution to the care of our most wounded patients.
For those of us who regularly care for vulnerable
patients, it provides an excellent resource and
supportive guide. However, it should also become
part of the standard library for all medical students and practicing physicians. All physicians
have much to learn from the practical, evidencebased approaches to the societal issues we all face
in practice. Ultimately, this is a book that could
help all clinicians take better care of all patients,
especially those who may need extra help and
support as they navigate our complex health care
Helen Burstin, M.D., M.P.H.
National Quality Forum
Washington, DC 20005
[email protected]
New and Evolving Infections
of the 21st Century
(Emerging Infectious Diseases of the 21st Century.)
Edited by I.W. Fong and Ken Alibek. 375 pp., illustrated.
New York, Springer, 2007. $129. ISBN 978-0-387-32647-4.
West Nile virus, a striking example of an established viral disease that after being transplanted
to a new continent has adapted and expanded its
dominion in North America. This chapter sets
the stage by illustrating one of the central aspects
of emerging infectious diseases — that most of
them are naturally resident in animal populations
and thus are of equal importance in the field of
veterinary medicine. The events that culminate in
the transition from an often unrecognized infection in a domestic or wild animal to a human
disease are important to understand. Zoonotic
diseases (those transmitted from animals to man)
often have complex dynamics, occasionally leading to cross-species exposures and human infection; in some cases, a disease develops, and in
even rarer cases, the disease is transmitted from
human to human. Understanding the nature of
these events is crucial in the prevention and control of future outbreaks. For instance, the backto-back chapters on hantaviruses and poxviruses
include descriptions of the differences between
genetically promiscuous RNA viruses and viruses
with more conserved DNA genomes. These contrasting chapters serve as examples to illustrate
the complexities that may give rise to clinically
apparent zoonotic infections when they cross the
species barrier into naive human populations.
Numerous chapters in the book, including
those on the severe acute respiratory syndrome
and avian influenza, illustrate the shifting continuum between human medicine and veterinary
medicine. The health problems these infections
cause in both fields are comprehensively handled, and from the perspective of human disease
and diagnosis are well presented.
There are inherent weaknesses in a multiauthored book such as this one. Most chapters
cover the spectrum, from virology and epidemiology to clinical presentation and management,
but they differ greatly in strength and content.
However, these shortcomings are outweighed by
a good number of well-written, high-quality, comprehensive chapters. This book will be a useful
and relatively current resource not only for experts
and researchers in public health and infectious
diseases but also for veterinary and medical professionals as well as graduate students.
t is not a matter of if, but instead
when, the next epidemic will strike. Historically, infectious diseases have inflicted much misery and, in so doing, have shaped humankind as
we know it today. From past plagues, we not only
have vestiges of older immunologic defenses but
also still possess the genetic skeletons of some
viral incursions into our genome. How will the
next epidemic reveal itself, and where will it come
from? This book is well timed, and it places perceived and real viral threats in perspective.
The title of the book, New and Evolving Infections
of the 21st Century, may seem prophetic, given that
the first decade of this century has yet to run its
course. Nevertheless, the book will be a valuable
contribution to education and preparedness. Its
editors have called on experts in infectious diseases of the very recent past to document and
compile current and recent threats, while also
providing examples of evolving insidious infections that in some cases are relatively common
in our population. The book is divided into three
sections: “Evolving Infectious Diseases,” “Newly
Jonathan L. Heeney, D.V.M., D.V.Sc.(Path.), Ph.D.
Recognized Human Viruses,” and “Emerging ViUniversity of Cambridge
ruses in Asia.”
Cambridge CB3 0ES, United Kingdom
The first section begins with a chapter on [email protected]
n engl j med 357;6  www.nejm.org  august 9, 2007
Book Reviews
Vaccine: The Controversial Story
of Medicine’s Greatest Lifesaver
By Arthur Allen. 523 pp., illustrated. New York, W.W. Norton,
2007. $27.95. ISBN 978-0-393-05911-3.
ver the past several decades, most
vaccine-preventable diseases in the United
States and other Western nations have been reduced to rare occurrences to such a degree that
physicians, nurses, health care workers, and patients are unfamiliar with them. In contrast,
some of these same infections remain the major
causes of illness and death in resource-poor nations. A number of books about specific infections (such as polio and smallpox) and the vaccines that prevent them have appeared in recent
years, but Arthur Allen has chosen to highlight
most of the vaccines that have evolved in the two
centuries since Edward Jenner introduced smallpox vaccination in 1796. Allen’s themes are the
research and development that led to the creation
of these vaccines, the personalities of the people
who were involved in this progress, and the concomitant rise of the antivaccine movement in the
United States.
Allen is a freelance writer who has published
articles in the Washington Post, the New York Times
Magazine, the New Republic, Mother Jones, Redbook,
and other publications. In writing this book, his
first, he has attempted to interview both those
who are involved in vaccine development and policy and those who have become active and vocal
in the antivaccine movement. As a journalist, he
provides thoughtful insights into the personalities and activities of all who have been available
to him for interviews or less formal conversations.
As a result, he is able to present much information that is unknown to many of us who trumpet vaccines as exerting the greatest impact on
infectious diseases — second only to clean water
Allen begins with the history of Jenner and
the controversies that surrounded the adoption
of the smallpox vaccine during the 150 years
after the publication of Jenner’s seminal paper.
After further commentary on Louis Pasteur and
the development of rabies vaccines in France, Al-
len devotes the rest of his book to events, products, and people in the United States. He concentrates most energetically on the personalities
— Jonas Salk and Albert Sabin, among others.
Fortunately, he was able to interview Maurice Hilleman, a power in the vaccine industry, before
his death in 2005. He recounts the stories of the
contamination of yellow fever vaccine with hepatitis B virus during World War II, the paralytic
polio that was transmitted by Cutter’s Salk vaccine, and the swine flu vaccine that caused Guillain–Barré syndrome in 1976.
In the last third of the book, Allen discusses
the controversies surrounding the diphtheria, pertussis, and tetanus vaccine; the measles, mumps,
and rubella vaccine; and the use of thimerosal
(ethyl mercury). He again describes proponents
and opponents of the vaccines from a journalist’s
perspective. Although some might disagree, I
found his writing well balanced, and he includes
more information about opponents of vaccines
than one usually finds in similar sources. The audience for this book should be broader than just
vaccinologists. Physicians who administer vaccines in their practices — including pediatricians,
family physicians, internists, and geriatricians
— will find much information that affects their
patients and their interactions with those patients.
Allen has cast a different light on much of the
contention that has arisen in recent years, and
he challenges those of us who are committed to
its resolution and to the continued prevention of
Samuel L. Katz, M.D.
Duke University Medical School
Durham, NC 27710
[email protected]
Dr. Katz reports receiving consulting fees from the Merck
Vaccine Advisory Board.
Book Reviews Copyright В© 2007 Massachusetts Medical Society.
the journal’s web and e-mail addresses:
For letters to the Editor: authors.nejm.org
For information about the status of a submitted manuscript:
To submit a meeting notice: [email protected]
The Journal’s Web pages: www.nejm.org
n engl j med 357;6  www.nejm.org  august 9, 2007