On Adjusting for the Outcome of Previous Pregnancies in Epidemiologic Reproductive Studies

If mothers experience an abnormal outcome in one or several pregnancies, it is not certain whether this mishap reflects constitutional defects in the mothers, genetic defects in the children, or continuous exposure of the mothers to an unfavorable environment.1 Thus, it is not clear whether adverse pregnancy outcome can be regarded as a marker of, for example, unmeasured genetic risk factors that can confound an observed relation be? tween an environmental exposure and the outcome of the mothers' subsequent pregnancies. A recent article by Weinberg2 concluded that bias can result when adjustment is made for any variable that both is caused in part by the exposure under study and is correlated with the outcome under study. In the article, spontaneous abortion history was used as an example of a variable that should not be adjusted for when one studies the relation between loss of a pregnancy and long-term exposure also connected with earlier pregnan? cies. When applying Weinberg's reproductive example, one should note, however, that the results of the article concern situations in which exposure associated with past pregnancies is etiologically relevant also to out? comes of subsequent pregnancies.