There were three objectives of this study: to investigate possible specificity in the association between specific cardiac defects and chromosomal anomalies; to evaluate ways of categorizing cardiac defects into larger groups with epidemiological similarities that could indicate similarities in etiology or pathogenesis; and to analyze the relationship between specific cardiac defects and diabetes. We pooled data on infants (aged 1 year or younger) with congenital cardiovascular defects from three large birth defect registries in California, Sweden, and France. The registries in Sweden and France obtained data through reporting from various sources; in California, medical records were reviewed. For severe congenital heart defects, the percentage of infants with identified chromosomal anomalies varied between 0.9% for d-TGV to 68.4% for ECD. In general, specific cardiac conditions have different risk factors. For example, conotruncal defects have been traditionally grouped, but the data presented in this paper indicates more differences for risk factors for the components of conotruncal defects: tetralogy of Fallot, d-TGV, common truncus, and DORV. In general, we suggest the strategy of "splitting" rather than "lumping" when searching for specific genetic factors and/or teratogens. Adequate analysis thus requires large registries or collaboration among registries. The findings did not support constellations between mothers' diabetes and specific defects.