Epidemiological methods to assess the correlation between industrial contaminants and rates of congenital anomalies

The present knowledge of epidemiological methods, applied to assess the correlation between industrial contaminants and rates of congenital anomalies is reviewed. The concept of congenital anomalies may be extended to include other adverse reproductive outcomes, such as malformations, infant mortality, stillbirths, spontaneous abortions, intrauterine growth retardation, ectopic pregnancies, multiple births, altered secondary sex ratio, and parental sub-fertility. The review of occupational exposures associated with congenital anomalies indicated: (1) inconsistency of the reported associations; (2) more positive than negative associations; (3) solvents are the best studied, and the most frequently reported teratogenic chemicals; (4) common congenital anomalies are the most frequently studied diagnostic categories, while other defects are grouped into larger categories, with little biological meaning. The review of environmental exposures indicated that: (1) single-site studies outnumber multi-site ones; (2) results are heterogeneous; (3) congenital anomalies are, in general, unspecific, and grouped into large categories, such as those defined by anatomic systems. Recent developments in molecular biology anticipate the possibility to measure exposures directly, instead of by different "proxies", as well as to analyze the genetic predisposition for the teratogenic response to given environmental agents. The strategy of building up large banks of biological materials has already started in several birth defects registries. The following procedural guidelines to assess the teratogenicity of a pollutant are recommended: (1) strength of the association; (2) consistency of findings in different studies; (3) specificity of the association; (4) time-exposure relationship; (5) existence of a dose-response gradient between exposure and disease occurrence; (6) biological plausibility; (7) coherence of the evidence with natural history of the disease; (8) experimental (or quasi-experimental) evidence and (9) reasoning by analogy.