22,748 pregnant women originally recruited 1984-1987 for a study evaluating risk factors for neural-tube defects and enrolled mainly between weeks 15-20 of pregnancy (at time of plasma AFP measurement or amniocentesis).
Data on diet, dietary retinol intake and vitamin supplements were collected by telephone questionnaire, data on pregnancy outcome by questionnaires sent to obstetricians (77%) and mothers (23%) after delivery. For this study, defects were classified as: cranial-neural-crest (craniofacial, CNS except neural tube, thymic and heart), neural tube, musculoskeletal/urogenital, and other (GI and non-GI).
Results
Among the 22,748 women, there were 339 births with birth defects, 121 of cranial-neural-crest origin. The prevalence of cranial-neural-crest defects among the 500 women who consumed >15,000 IU of retinol daily (dietary plus supplements) was 1.8%, compared with 0.51% among the 6410 women who consumed <5,000 IU (relative risk of 3.5). Looking at retinol from supplements alone, the prevalence of cranial-neural-crest defects among the 317 women consuming >10,000 IU daily was 2.21% vs. 0.46% among the 11,083 consuming < 5,000 IU daily (relative risk 4.8).
Dose-response curves suggested a threshold value of about 10,000 IU daily. Subgroup analysis suggests main effect occuring before week 7 of pregnancy. Consumption of over 10,000 IU of vitamin A daily would be expected to lead to a birth defect in one of every 57 births.