8320
Varicella Vaccine for Immunocompromised Children: Results of Collaborative
Studies in the United States and Canada
Philip LaRussa, Sharon Steinberg, and Anne A. Gershon
Department of Pediatrics, Columbia University College of Physicians
and Surgeons, New York, New York
Varicella vaccine in immunocompromised children was clinically evaluated in 575 US and Canadian children with leukemia in remission by the Varicella Vaccine Collaborative Study. Most children
had chemotherapy stopped 1 week before and 1 week after immunization. Steroids were stopped
for 3 weeks (l week before to 2 weeks after vaccination). Varicella vaccine was safe, immunogenic,
and effective in leukemic children at risk for serious disease or death from chickenpox. The major
side effect was mild rash in 50% ~ 1 month after immunization. About 40% of children who
developed rash were treated with acyclovir. Vaccine efficacy was judged by the degree of protection
after a household exposure to varicella; of 123 exposed children, 17 (14%) developed a mild form
of varicella. The vaccine protected completely against severe varicella. Leukemic vaccinees were
less likely to develop zoster than were comparable children with leukemia who had wild type
varicella. Thus, varicella vaccine, administered carefully with close follow-up, is extremely beneficial
for leukemic children.
Although varicella vaccine was developed in Japan in the
early 1970s [1], it was not until ~6 years later that it began
to be tested in clinical trials in the United States. Since longterm adverse effects could not be predicted at that time, the
most acceptable risk-benefit ratio seemed to be in varicellasusceptible persons at high risk of developing severe or fatal
chickenpox. Thus, among the first US recipients of varicella
vaccine were children with underlying leukemia in remission
but who still received maintenance chemotherapy for leukemia.
In the late 1970s and early 1980s, such children had about a
10% chance of dying of varicella once contracted [2]. Even
today, although antiviral drugs are both useful and available, an
occasional immunocompromised child dies of varicella. Thus,
prevention of varicella continues to be an important goal.
Methods
In total, 575 children with leukemia in remission were immunized in the Varicella Vaccine Collaborative Study sponsored by
the National Institute of Allergy and Infectious Diseases. Many
study details have been published [3-8]. To qualify, children had
to be in continuous remission from acute lymphoblastic leukemia
;:::;: 1 year, have no detectable antibodies to varicella-zoster virus
(VZV), have > 700/mm 3 circulating lymphocytes, and be able to
respond to mitogens in vitro. Two doses of vaccine were usually
given 3 months apart. Most children had chemotherapy stopped
for 1 week before and 1 week after immunization. It was further
recommended that these children not be given steroids for at least
2 weeks after immunization [9].
Reprints or correspondence: Dr. Philip LaRussa, Columbia University College of Physicians & Surgeons, 650 W. 168th St., New York, NY 10032.
The Journal of Infectious Diseases 1996; 174(Suppl 3):S320-3
© 1996 by The University of Chicago. All rights reserved.
0022-1899/96/7483-0014$01.00
Results
Safety. Varicella vaccine was safe, immunogenic, and effective in leukemic children. The first 11 children were no
longer receiving chemotherapy when they were vaccinated;
eventually 66 such children were immunized. Chemotherapy
was suspended for the first dose of vaccine in 509 children.
For most children, chemotherapy was not suspended for the
second vaccine dose, even if there was no seroconversion after
the first dose.
The major adverse effect, varicelliform rash, occurred "'I
month after immunization in 50% of children still receiving
maintenance chemotherapy; in contrast the rate of rash in those
no longer on chemotherapy was 5%. The rash was usually
maculopapular and vesicular and at times resembled a mild
form of chickenpox. About 40% of the vaccinees who developed rash were treated with the antiviral drug acyclovir, usually
by mouth. If the rash progressed despite oral acyclovir therapy
or if systemic symptoms developed, such as moderate-to-high
fever, intravenous acyclovir therapy was instituted. Other unusual but transient adverse effects were headache, upper respiratory tract symptoms, neutropenia, and thrombocytopenia. The
vaccine type VZV is sensitive to acyclovir, but acyclovir did
not appear to interfere with development of the immune response [6].
It was often possible to isolate VZV from the skin rash,
although the virus was never isolated from any other site (e.g.,
the respiratory tract or blood) [10]. Spread of vaccine type
virus to varicella-susceptible contacts occurred only when the
vaccinee had a skin rash; the degree of spread was directly
proportional to the number of skin lesions present [10]. Contact
cases of varicella were extremely mild, with an average of only
38 skin lesions and no systemic signs of illness. There was
only one recognized instance of tertiary spread of virus (i.e.,
from a vaccinee with rash to a susceptible contact who then
lID 1996; 174 (Suppl 3)
Varicella Vaccine in Leukemic Children
developed a rash and subsequently spread vaccine type VZV
to another susceptible contact). These data strongly suggest
that the vaccine virus is attenuated. Another indication of attenuation was the decreased pathogenicity of the vaccine type
virus for healthy children when it was administered by injection
or inhalation [11]. Other differences between vaccine and wild
type VZV are restriction fragment length polymorphism [1214], a lower quantity of the glycoprotein V of VZV in aka
strain [15], and increased temperature sensitivity of aka strain
in cell culture at 39°C compared with wild type VZV [16]. It
is now possible to distinguish between vaccine and wild type
VZV by polymerase chain reaction (Pf.R), circumventing the
need to culture the virus, which is not always possible
[14, 17].
Although chemotherapy for leukemia was temporarily discontinued during the vaccination period, there was no observed
difference in the relapse rate of vaccinees and matched control
leukemic children who had experienced natural varicella. In
both groups, the relapse rate was ~25% [6].
Immunogenicity and efficacy. A seroconversion to VZV,
measured by the fluorescent antibody to membrane antigen
assay, occurred in 82% of vaccinees with acute lymphoblastic
leukemia after 1 dose and in 95% after 2 doses [3- 5]. Vaccine
efficacy was judged by the degree of protection after a household exposure to varicella; after such an exposure, > 80% of
unvaccinated susceptible contacts will develop clinical varicella [18]. Of 123 leukemic vaccinees whose siblings developed varicella, 106 (86%) were completely protected from
chickenpox. No child developed severe varicella; the 14% who
developed clinical varicella almost always had modified infections. An average child who is otherwise healthy develops
250-500 skin lesions during chickenpox [18]. The leukemic
vaccinees who had breakthrough varicella after household contact developed on average < 100 skin lesions. The child with
the most severe illness after household contact was still receiving chemotherapy when she became ill; she developed 640
skin lesions, but therapy with acyclovir was not given.
Over a follow-up of up to 11 years, 13% of vaccinees who
originally seroconverted to VZV became seronegative. However, there is no evidence that either the incidence or the severity of breakthrough varicella has increased with time. Thus,
any suggestion of waning immunity based on antibody titers
is counterbalanced by the lack of evidence of any change in
the clinical protection afforded by the vaccine.
The role of the VZV-specific cell-mediated immune response to vaccination was investigated in a limited manner
at the beginning of the US-Canadian collaborative study
(table 1). Although the number of patients studied was small,
this provides evidence that the cell-mediated response may
provide protection to vaccinees after household exposure to
varicella [19].
Zoster in leukemic vaccinees. Zoster develops when latent
virus reactivates from dorsal root ganglia infected during primary varicella. A live attenuated vaccine could potentially pre-
S321
Table 1. Attack rate after household exposure to varicella zoster
virus (VZV).
VZV antibody/
CMI* status
Positive/positive
Negative/positive
Positive/negative
Negative/negative
No. of breakthroughs/
no. exposed
3/27
Attack rate (%)
11
0/3
o
2/6
3/3
33
100
NOTE. Data are from [19] (with permission).
* VZV antibody/cell-mediated immunity (CMI) at time of exposure.
vent zoster by blocking or limiting primary infection of dorsal
root ganglia with wild type virus upon subsequent exposure to
varicella. For this strategy to work, vaccine virus would have
to show less tendency to establish latent infection or be less
likely to reactivate than wild type virus. In addition, once infection occurs and latency is established with vaccine strain (after
immunization) or wild type virus (after primary varicella), reexposure to virus (either vaccine or wild type) may potentiate
the VZV-specific immune response, further decreasing the risk
of developing zoster by preventing reactivation of latent virus.
To investigate the role of these potentially protective mechanisms, the risk of developing zoster was studied in subgroups
of the cohort of vaccinated children with leukemia. The risk
of developing zoster was compared in 96 vaccinated leukemic
children and an equal number of matched children with leukemia who had wild type varicella. The crude incidence rates
were 0.80 and 2.46 per 100 person-years, respectively. A
Kaplan-Meier life table analysis showed that the risk of developing zoster was significantly lower in the vaccinated group
[7]. This finding validates our first hypothesis that vaccine virus
may have less of a tendency to establish latent infection or to
reactivate than wild type virus. A subsequent analysis also
showed the risk of zoster was lower in leukemic vaccinees
who received two doses of vaccine compared with those who
received one dose (P < .01, Cox proportional hazard model).
This same analysis revealed that the risk of zoster was also
lower in vaccinees who, subsequent to vaccination and seroconversion, had household exposure to wild type varicella compared with vaccinees without household exposure [20]. These
studies validate our second hypothesis that there is a protective
boosting effect from either reexposure to the attenuated vaccine
type virus or to wild type virus after a primary immune response.
Other studies. Studies in Japan, where immunization of
leukemic children was first done, indicated a lower rate of
adverse events than found in US studies. In leukemic children
with chemotherapy suspended, the rate of vaccine-associated
rash was <25% [21]. Some Japanese children were then immunized without stopping their chemotherapy, but this led to an
unacceptably high incidence and severity of rashes [21-24].
In general, it seems likely that Japanese children were receiving
S322
LaRussa et al.
less immunosuppressive chemotherapy than the American children. When daily 6-mercaptopurine therapy was continued in
small US studies, high fevers and hepatitis were observed in
2 patients [25]. These children had <700/mm 3 lymphocytes
on the day of immunization. Although the number of leukemic
children in remission is small, it seems prudent to immunize
them after withholding their leukemia chemotherapy 1 week
before and 1 week after vaccination, particularly since chemotherapy for leukemia in the 1990s is likely to be more immunosuppressive than that used in the early 1980s.
Japanese and US studies show that children with solid tumors
who have been immunized seem to fare better than leukemic
children with regard to tolerance of varicella vaccine [6, 26].
Yet, small numbers of children with lymphoma who have been
immunized have had unacceptably high rates of rash and
fever [27].
Natural varicella may be severe in patients who have undergone organ transplantation. The only studies of varicella vaccine in such patients were small and assessed children about
to receive a renal transplant. Vaccination of 23 uremic children
~2 months before transplantation was tolerated very well and
resulted in an 87% seroconversion rate after a single dose
of vaccine [28]. Efficacy data are not yet available, but the
investigators had the impression that the incidence of both
varicella and zoster had decreased in the population of uremic
children after use of the vaccine.
Future Studies in Immunocompromised Patients
There is ample need to define the safety and efficacy of
varicella vaccine in additional immunocompromised children
for whom natural varicella continues to pose a danger. These
include children with solid tumors and children about to have
renal transplantation, for whom data are promising but scant.
Children about to undergo liver transplantation and those in
the early phase of human immunodeficiency virus infection
also might benefit from varicella vaccine. Given the problems
that have occurred in lymphoma patients, however, it may be
that these patients are too immunocompromised to tolerate the
vaccine. Children who have had bone marrow transplants are
another group in whom varicella vaccine should be studied.
The availability of modem diagnostic methods (e.g., PCR) and
antiviral drugs that are effective against VZV are significant
advances that make it possible to study the vaccine further in
high-risk populations. If varicella vaccine is routinely used in
infants, however, there may be less of a need to immunize
immunocompromised children in the future, as the amount of
circulating wild virus will decrease.
Conclusions
On the basis of studies in US and Canadian children with
acute lymphoblastic leukemia in remission, it appears that vaccination of such children against varicella is safe, although
JID 1996; 174 (Suppl 3)
antiviral therapy may be necessary-particularly to treat some
vaccine-associated rashes. Two doses of vaccine must be given
to assure a seroconversion rate of ~90%, but chemotherapy is
temporarily withheld only for the first dose. High-risk children
who are vaccinated should have laboratory tests for development of antibodies to VZV, since some vaccine failures may
occur, and the child may not be protected if seroconversion
has not occurred.
Vaccinated leukemic children can attend school and live
more normal lives, and their parents no longer need to fear
that they might contract natural varicella. In addition, the incidence of clinical reactivation (zoster) in vaccinated children is
lower than in unvaccinated leukemic children. Thus, varicella
vaccine, administered carefully with close follow-up, is extremely beneficial to leukemic children. These benefits ought to
be carefully extended to other immunocompromised children,
particularly until the varicella vaccine is used routinely in all
infants.
References
1. Takahashi M, Otsuka T, Okuno Y, Asano Y, Yazaki T, Isomura S. Live
vaccine used to prevent the spread of varicella in children in hospital.
Lancet 1974;2:1288-90.
2. Feldman S, Hughes W, Daniel C. Varicella in children with cancer: 77
cases. Pediatrics 1975; 80:388-97.
3. Gershon A. Immunoprophylaxis of varicella-zoster infection. Am J Med
1984; 76:672-7.
4. Gershon AA, Steinberg S, Gelb L. Live attenuated varicella vaccine: use
in immunocompromised children and adults. Pediatrics 1986; 78(suppl):
757-62.
5. Gershon AA, Steinberg SP. Persistence of immunity to varicella in children
with leukemia immunized with live attenuated varicella vaccine. N Engl
J Med 1989;320:892-7.
6. Gershon AA, LaRussa P, Steinberg S. Live attenuated varicella vaccine:
current status and future uses. Semin Pediatr Infect Dis 1991; 2: 171-8.
7. Hardy IB, Gershon A, Steinberg S, LaRussa P, et al. The incidence of
zoster after immunization with live attenuated varicella vaccine. A study
in children with leukemia. N Engl J Med 1991;325:1545-50.
8. Lawrence R, Gershon A, Holzman R, Steinberg S, NIAID Varicella Vaccine Collaborative Study Group. The risk of zoster after varicella vaccination in children with leukemia. N Engl J Med 1988;318:543-8.
9. Lydick E, Kuter BJ, Zajac BA, Guess HA. Association of steroid therapy
with vaccine-associated rashes in children with acute lymphocytic leukaemia who received Oka/Merck varicella vaccine. NIAID Varicella
Study Group. Vaccine 1989; 7:549-53.
10. Tsolia M, Gershon A, Steinberg S, Gelb L. Live attenuated varicella
vaccine: evidence that the virus is attenuated and the importance of skin
lesions in transmission of varicella-zoster virus. J Pediatr 1990; 116:
184-9.
11. Bogger-Goren S, Bernstein JM, Gershon A, Ogra PL. Mucosal cell mediated immunity to varicella zoster virus: role in protection against disease.
J Pediatr 1984; 105:195-9.
12. Martin JH, Dohner D, WellinghoffWJ, Gelb LD. Restriction endonuclease
analysis of varicella-zoster vaccine virus and wild type DNAs. J Med
Virol 1982; 9:69- 76.
13. Brunell P, Geiser C, Novelli V, Lipton S, Narkewicz S. Varicella-like
illness caused by live varicella vaccine in children with acute lymphocytic leukemia. Pediatrics 1987; 79:922-7.
14. LaRussa P, Lungu 0, Hardy I, Gershon A, Steinberg S, Silverstein S.
Restriction fragment length polymorphism of polymerase chain reaction
JID 1996; 174 (Suppl 3)
15.
16.
17.
18.
19.
20.
Varicella Vaccine in Leukemic Children
products from vaccine and wild-type varicella-zoster virus isolates. 1
Virol 1992;66:1016-20.
Kinchington PR, Remenick 1, Ostrove 1, Straus S, Ruyechan W, Hay 1.
Putative glycoprotein gene of varicella-zoster virus with variable copy
numbers of a 42 base pair repeat sequence has homology to herpes
simplex virus glycoprotein C. 1 Virol 1986;59:660-8.
Hayakawa Y, Yamamoto T, Yamanishi K, Takahashi M. Analysis of
varicella zoster virus (VZV) DNAs of clinical isolates by endonuclease
HpaI. 1 Gen ViroI1986;67:1817-29.
Hughes P, LaRussa PS, Pearce 1M, Lepow ML, Steinberg SP, Gershon A.
Transmission of varicella-zoster virus from a vaccinee with underlying
leukemia, demonstrated by polymerase chain reaction. J Pediatr 1994;
124:932-5.
Ross AH, Lencher E, Reitman G. Modification of chickenpox in family
contacts by administration of gamma globulin. N Engl 1 Med 1962;
267:369-76.
Gershon A, LaRussa P, Steinberg S. Clinical trials in immunocompromised
individuals. Pediatr Clin North Am 1996 (in press).
Gershon AA, LaRussa P, Steinberg S, Mervish N, Lo SH, Meier P. The
protective effect of immunologic boosting against zoster: an analysis in
leukemic children who were vaccinated against chickenpox. 1 Infect
Dis 1996; 173:450-3.
8323
21. Sato Y, Miyano T, Kawauchi K, Yokoyama M. Use of live varicella
vaccine in acute leukemia and malignant lymphoma. Biken 11984;27:
111-3.
22. Ha K, Baba K, Ikeda T, Nishida M, Yabuuchi H, Takahashi M. Application
of a live varicella vaccine in children with acute leukemia or other
malignancies without suspension of anticancer therapy. Pediatrics 1980;
65:346-50.
23. Konno T, Yamaguchi Y, Minegishi M, Goto Y, Tsuchiyama S. A clinical
trial of live attenuated varicella vaccine (Biken) in children with malignant diseases. Biken 1 1984;27:73-5.
24. Nunoue T. Clinical observations on varicella-zoster virus vaccinees treated
with immunosuppressants for a malignancy. Biken 1 1984;27:115-8.
25. Arbeter A, Granowetter L, Starr S, Lange B, Wimmer R, Plotkin S. Immunization of children with acute lymphoblastic leukemia with live attenuated varicella vaccine without complete suspension of chemotherapy.
Pediatrics 1990; 85:338-44.
26. Takahashi M, Asano Y, Kamiya H, Baba K, Yamanishi K. Active immunization for varicella zoster virus. In: Nahmais AI, Dowdle WR, Schnazi
RE, eds. The human herpesviruses. New York: Elsevier, 1981:414-31.
27. Takahashi M. Clinical overview of varicella vaccine: development and
early studies. Pediatrics 1986; 78 S:736-41.
28. Broyer M, Boudailliez B. Varicella vaccine in children with chronic renal
insufficiency. Postgrad Med 1 1985; 6 1(suppl 4):103-6.