Environmental, demographic, and medical factors related to cord blood lead levels.

Blood lead was measured at birth for 11,837 infants in Boston. Extensive maternal demographic, pregnancy, and delivery characteristics were recorded for 4354 of them. For 249 of these, intensive environmental sampling was done. Many medical factors were unrelated to blood lead, including diabetes, venereal diseases, preeclampsia, toxemia, hypertension, age, hematocrit, contraceptive use, presentation, type of delivery, fetal distress, premature rupture of membrane, blood type, gestational age, birthweight, Apgar score, jaundice, and mortality by one month. However, use of tobacco, alcohol, coffee, tea, or marihuana, having had an abortion, receiving welfare and being unmarried, Black, or Hispanic are associated with significantly elevated blood lead. Being college educated, Jewish, younger, and multiparitous are related to lower blood lead levels. Environmental factors covarying with blood lead included dust and soil lead and refinishing activity, but not water, air, or paint lead or traffic density. Many of these predictors are intercorrelated. Simultaneous, step-wise regression models of blood lead ranking these factors are presented. Only 18% of the variance is explainable. Temporal and geographic patterns exist even after taking these maternal and environmental factors into account.