Agent Orange and Spina Bifida

Veterans and Agent Orange: Update 1996, National Academy of Sciences

In the continuing saga of Agent Orange, the National Institute of Medicine's Committee to Review the Health Effects in Vietnam Veterans of Exposure to Herbicides has added yet another wrinkle to this tragedy of science, that is. The Committee has determined that there is "limited/suggestive evidence of an association between" exposure to Agent Orange and the birth defect, spina bifida.

This association was based on a 1989 Centers for Disease Control (CDC) study that reported a relative risk of 1.7 with a 95 percent confidence interval of 0.6 to 5.0 for spina bifida in children of Vietnam veterans versus non-Vietnam veterans. For the Vietnam veterans group with the greatest opportunity for exposure to Agent Orange, the reported association was 2.7 (95% C.I. 1.2 - 4.5).

However:

Weak association. For the data overall, the odds ratio was 1.7 - a weak association. For a weak association to hold up, the data must be very good.


Data dredging. For the data overall, no statistically significant association was identified as the lower end of the 95 percent confidence interval extends below 1.0 to 0.6. So they data dredged, and, voila!, a significant association was identified among those children of Vietnam veterans with the greatest opportunity for exposure.


Exposure misclassification. Actual exposures to Agent Orange are not known. As a surrogate "guesstimate" of exposure, "opportunity for exposure" was used. This raises the specter of misclassification of exposure.


Recall bias. CDC failed to verify children reported free of birth defects indeed had no birth defects. As the NIM Committee put it, "... the issue of recall bias is a major concern with this study." Recall bias would have biased the results toward reporting an increased risk.


Biological plausibility. What's the alleged biological mechanism, showing biological plausibility, through which Agent Orange causes spina bifida? No word on that!


Negative data. The Committee seems to have discounted a 1984 CDC study that used the general population of Atlanta as the controls (rather than non-Vietnam vets), reported an odds ratio of 1.1 (95% C.I. 0.6-1.7.) In this study, CDC reported no association.

As the data is of questionable quality, the weak association is certainly suspect to the point where data dredging was relied upon to "establish" the association. As usual, negative data were discounted.

I think it's great the federal government wants to compensate the children of Vietnam veterans. Exactly what makes torturing epidemiology a prerequisite?

For more on the NIM report, click to see the guest commentary of Dr. Michael Gough, Director of Science and Risk Studies at the Cato Institute.

Material presented on this home page constitutes opinion of the author.

Copyright © 1996 Steven J. Milloy. All rights reserved. Site developed and hosted by WestLake Solutions, Inc.