From animal and in vitro studies, it has been suggested that high environmental glucose, ketone, or insulin concentrations and low glucose or insulin concentrations may be etiologic factors for congenital malformations (CMs) in infants of diabetic mothers (IDMs). Transplacental passage of antibody-bound insulin has been demonstrated in humans. Controversy exists regarding the pathophysiology of CMs in human insulin-dependent diabetes mellitus (IDDM) pregnancies. We hypothesized that CMs in IDMs are associated with maternal vasculopathy, poor first-trimester glycemic control (i.e., hyper- and/or hypoglycemia), advanced White class, and high insulin requirements. We studied 165 first pregnancies of women with IDDM from 1978 to 1986. The goals of glucose control were a fasting blood glucose of 1 was measured on entry and every 4 wk to confirm that adequate glycemic control was achieved. Women who enrolled in the project were interviewed during gestation by a geneticist/dysmorphologist who obtained genetic and environmental histories using a standard questionnaire. All live-born infants and stillbirths were examined. Each live-born infant was assessed systematically by two independent examiners, a neonatologist and a geneticist/dysmorphologist; examination with standardized checklists was performed in the newborn nursery as soon after birth as was practical. In first pregnancies in the study, there were 13 IDMs with major CMs (7.9%). By both univariate and multivariate analyses, the following two factors were significantly associated with major CMs: preexisting vasculopathy (retinopathy and/or nephropathy; 54 vs. 25% in the CM and non-CM groups, respectively; P < .05) and increased maternal HbA, at 9 wk gestation (P < .001). CMs were not associated with White class, insulin requirements, and frequency of maternal hypoglycemia as assessed clinically. There was no significant interaction between maternal HbA, and the presence of vasculopathy. We conclude that maternal vasculopathy and poor glycemic control during embryogenesis, but not frequency of maternal clinical hypoglycemia or insulin therapy per se are independent factors associated with major CMs in IDMs. We suggest, based on these data, that in addition to establishment of meticulous glycemic control, IDDM women be assessed for vasculopathy before pregnancy and given appropriate counseling regarding risk of CMs.