Epidemiologic Study Of Reproductive Outcomes And Environmental Exposures

National Institute Of Environmental Health Sciences

Investigators

Linked publications & trials

Abstract

Summary: In the mid-1980s we carried out a prospective study of early pregnancy in which we enrolled 221 health women who were planning to become pregnant. These women collected daily urine specimens for up to six months. We've assayed these specimens to describe the hormonal events of the menstrual cycle and early pregnancy. 155 women became clinically pregnant during the study, while 44 had pregnancies that ended so early that the pregnancies were detectable only by assay of urinary human chorionic gonadotropin. This unique study has been called a landmark, and continues to provide a rich resource for the description of the earliest stages of pregnancy. (More than 30,000 urine samples are still being stored.) We've published 50 papers from this study over the past two decades, some of which have led to new understanding of the fundamental processes of conception and early pregnancy. In addition, we have continued to make use of large population registries in order to pursue basic questions on pregnancy and maternal and infant health. We have worked especially closely with Norwegian colleagues in the analysis of the Norwegian Medical Birth Registry. Last year's progress. Preterm birth is strongly associated with mortality and morbidity of the baby. Most researchers have reasonably assumed that this risk is due to the immaturity of the newborn. However, it has become more apparent clinically that the causes of preterm delivery might also confer risk independent of the immaturity of the fetus. We have shown in simulations and in analyses of empirical data that half or more of the mortality of preterm babies are attributable to the causes of the early delivery rather than to preterm delivery itself. This insight has repercussions for many aspects of epidemiologic research. For example, statistical adjustments for gestational age are common in etiologic studies of perinatal outcomes. We have shown that such statistical adjustments can severely bias the results, and should not be carried out. This has implications for all studies of environmental effects on newborn outcomes.