American Journal of Epidemiology
Copyright © 1998 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol. 148, No. 4
Printed in U.S.A.
Counterpoint: Responding to Inadequate Critique of Birth Defects Paper
David N. Cowan, 1 Gregory C. Gray,2 and Robert F. DeFraites3
Before responding to Haley's critique (1) of the
three papers recently published in the New England
Journal of Medicine (2-4), we would like to place our
research in perspective. We conducted an exploratory
comparison of the available medical data for two large
cohorts to see if there was an association between Gulf
War service and the risk of birth defects. As pointed
out by Wegman et al., there is a "serious lack of
information about risk factors" (5, p. 704), and so our
study was conducted without specific environmental
exposure hypotheses. It was an initial, rather than a
final, evaluation of reproductive outcomes among
Gulf War veterans.
Haley goes into great detail about why he believes
the findings of our report are not valid, including
developing and arguing a complex scenario of "given," "if," and "then" statements. A review of Haley's
argument regarding the birth defects study is in order.
The first " i f statement ("The wartime exposures
(e.g., chemical or biologic toxins) in deployed veterans caused both the symptoms of the Gulf War syndrome and birth defects" (1, p. 320)) is offered with no
consideration for biologic plausibility or even speculation regarding which exposures might possibly cause
both a "Gulf War syndrome" and birth defects. Future
research efforts would benefit if Haley could offer a
hypothesis regarding a Gulf War exposure that might
cause both a newly described chronic condition and
paternally mediated teratogenesis. We hope he will be
forthcoming, particularly since the role of and mechanism for male-mediated teratogenesis remain unclear.
Haley's argument also has errors in logic. In this "if
statement, he lists the "Gulf War syndrome" separately from birth defects, while in the following paragraph he redefines them as one and the same ("... birth
defects are part of the Gulf War syndrome ..." (1, p.
320)), which forces the " i f statement to be true. This
has implications for his next " i f statement as well, in
that it implies, with no supporting data, that persons
with an increased risk of birth defects left the military
at a higher rate than did persons without increased
risk.
The second "if statement ("The symptoms of the
Gulf War syndrome caused many affected veterans in
the deployed group to be separated from active duty
prematurely during the postwar period of follow-up"
(1, p. 320)) deserves further examination. Haley uses
separation data from the hospitalization study (2) to
criticize our birth defects study (4). This is inappropriate because the two studies' cohorts differed in
composition and had different start and stop dates for
subject follow-up. Haley apparently failed to note that
the birth defects paper (4) reported loss from active
duty at two dates (September 1992 and September
1993) and not the date reported in the hospitalization
paper (December 1993) (2). Haley (1) attempts to
validate the second "if statement utilizing the December 1993 data. At the two September dates we reported, there was little difference in attrition; that is,
the relative risks for male Gulf War veterans were 1.01
and 1.04, respectively, and among women the relative
risks were 0.93 and 0.99. Our study evaluated births
only until the end of September 1993; attrition after
that date is irrelevant. Finally, in spite of Haley's
convoluted logic regarding the reasons for leaving
military service, Gray et al. (6) provide detailed information indicating that undiagnosed Gulf War-related
illnesses are not likely to be causing substantial attrition among deployed veterans.
Because the first "if statement is purely speculative
and the second "if statement is false, Haley's two
"then" statements ("Failure to include records of births
and birth defects from nonmilitary hospitals in the
study preferentially excluded the birth defects caused
by the war" (1, p. 320) and "Finding no difference in
the rates of postwar birth defects in the deployed and
nondeployed groups would not rule out an excess of
birth defects caused by exposures in the war" (1, p.
320)) are not true.
Haley apparently also misunderstands our comments regarding births identified but not taking place
Received for publication January 28, 1998, and accepted for
publication March 30, 1998.
1
Exponent Health Group, assigned to the Gulf War Health Studies office, Department of Defense, Falls Church, VA.
2
Naval Health Research Center, San Diego, CA.
3
Office of the Army Surgeon General, Falls Church, VA.
Reprint requests to Dr. David Cowan, Exponent Health Group,
8201 Corporate Drive, Suite 680, Landover, MD 20785.
326
Counterpoint: Reply by Cowan et al.
in military hospitals. There was concern that births
occurring during the follow-up period but occurring in
civilian hospitals may have been at high risk for defects and that, if these births were differentially associated with deployment, then they could have been a
source of bias. However, we found that births to male
Gulf War veterans were slightly less likely to have
occurred in civilian facilities, and essentially all births
to female veterans, both deployed and not, occurred in
military facilities (4).
Furthermore, Haley appears to have ignored our
dose-response (risk by time-in-theater) analyses (4,
table 3 and text). If there were some teratogen present
in the Gulf War environment, regardless of whether it
also caused symptoms of the "Gulf War syndrome,"
there would likely be some dose-response association;
for example, longer or higher cumulative levels of
exposure, or greater opportunity to experience some
threshold level of exposure, might result in more birth
defects. We found no association with time-in-theater
among deployed service members, no matter how we
evaluated the data for two definitions of birth defect.
These analyses, which include an internal control of
exposure, should relieve any concerns regarding comparisons between deployed and nondeployed service
members.
Haley calls for future studies to include all birth
defects. All recorded birth defects were captured and
included in our report. Furthermore, although we included tables (4, tables 5 and 6) with birth defects
comprising 5 percent or more of all serious defects
(and found no significant increases among Gulf War
veterans), we also evaluated but did not publish all
specific groups of defects occurring among our population. Even among the rare events, there were no significant or substantial associations with Gulf War service.
Finally, Haley calls for additional studies that capture all births occurring among deployed and nondeployed members, including those among persons leaving active duty. Logistically, it would be very difficult
to conduct such a historical longitudinal study, particularly in the absence of a specific, testable exposure or
biologically plausible defect hypothesis. Anticipated
difficulties in subject tracing, locating, and contacting,
identifying births, obtaining informed consent to contact hospitals and physicians, and ascertaining defects
from data recorded in a myriad of different fashions
may well introduce serious biases. However, other
studies utilizing other methodologies have been completed or are underway. These include the recently
published paper evaluating Goldenhar's syndrome (7);
an examination of births among both separated and
active duty personnel in both military and civilian
facilities occurring in seven states that maintain active
Am J Epidemiol
Vol. 148, No. 4, 1998
327
case ascertainment programs for birth defects (8); a
health study underway by the Department of Veterans
Affairs that will capture reproductive outcomes in a
nationwide sample of veterans (Han Kang, Environmental Epidemiology Service, Department of Veterans Affairs, personal communication, 1997); a study
of the health of female veterans (9); and a study
underway in London evaluating birth defects among
children of British Gulf War veterans (10).
We acknowledge that it has not been "proven" that
children of Gulf War veterans are not at higher risk of
birth defects, even though none of the peer-reviewed
studies completed to date has identified a significant
association (4, 6, 11). Furthermore, it is possible that
there are some rare defects that may be associated with
some specific exposures occurring in the war. However, if the ongoing studies fail to detect such an
association, one must question whether it is the most
appropriate or best use of limited resources to continue
to pursue this area of research, particularly in the
absence of a clearly defined hypothesis regarding measurable exposures and specific birth defects.
REFERENCES
1. Haley R. Point: bias from the "healthy-warrior effect" and
unequal follow-up in three government studies of health effects of the Gulf War. Am J Epidemiol 1998; 148:315-23.
2. Gray GC, Coate BC, Anderson CM, et al. The postwar hospitalization experience of US veterans of the Persian Gulf
War. N Engl J Med 1996;335:1505-13.
3. Kang H, Bullman T. Mortality among U.S. veterans of the
Persian Gulf War. N Engl J Med 1996;355:1498-504.
4. Cowan D, Gray G, DeFraites R, et al. Birth defects among
children of Persian Gulf War veterans. N Engl J Med 1997;
337:1175-6.
5. Wegman DH, Woods NF, Bailar JC. Invited commentary:
how would we know a Gulf War syndrome if we saw one?
Am J Epidemiol 1997;146:704-ll.
6. Gray GC, Knoke JD, Berg SW, et al. Counterpoint: responding to suppositions and misunderstandings. Am J Epidemiol
1998,148:328-33.
7. Araneta MRG, Moore CA, Olney RS, et al. Goldenhar syndrome among infants born in military hospitals to Gulf War
veterans. Teratology 1997;56:244-51.
8. Department of Veterans Affairs. Annual report to Congress:
federally sponsored research on Persian Gulf veterans' illnesses. Washington, DC: The Research Working Group of the
Persian Gulf Veterans Coordinating Board, Department of
Veterans Affairs, 1997.
9. Klemm R, Rhodes L, Ugwu L, et al. The Persian Gulf Women's Health Linkage Study: a longitudinal study designed to
assess short-term and long-term effects of Gulf War service.
Presented at session 2110 (board 2) of the 125th Annual
Meeting of the American Public Health Association, Indianapolis, Indiana, November 11, 1997.
10. Doyle P, Roman E, Maconochie N. Birth defects among
children of Persian Gulf War veterans. (Letter). N Engl J Med
1997;337:1175.
11. Penman AD, Currier MM, Tarver RS. No evidence of increase
in birth defects and health problems among children bom to
Persian Gulf War veterans in Mississippi. Milit Med 1996;
161:1-6.