The concept that gestational subclinical hypothyroidism could have deleterious effects on the intellectual outcome of progeny was championed more than three decades ago by Evelyn Man in a series of publications. Her studies lay fallow until the Spanish group directed by Morreale de Escobar and the Dutch group headed by Vulsma provided the rationale for her results. Although the findings of the Spanish and Dutch groups elucidated the pathophysiologic basis for Man's conclusions, questions remained regarding the reliability of her biochemical measurements and possible bias in patient selection. In view of the uncertainty surrounding the validity of Man's work, we decided to try to confirm her findings. Our initial goal was to obtain an estimate of the prevalence of subclinical hypothyroidism in an unselected population living in New England. We accomplished this with two separate prospective studies involving 12 000 pregnant women residing in Maine. We found that 2.3% had TSH concentrations of >6.0 mU/l and 0.3% had TSH values of >12 mU/l at 17 weeks' gestation. We next did a retrospective study, utilizing sera that had been stored at -20 degrees C for 8 years, obtained in week 17 of gestation from 25 000 women. We identified 62 women with subclinical hypothyroidism and 124 matched controls. Fourteen of the hypothyroid mothers had been diagnosed and treated before and during pregnancy on a dosage of thyroxine that was unchanged. WISC IQs of the offspring of the 124 control and 62 hypothyroid mothers were determined at 8+/-0.5 years. The mean and s.d. of IQs of the children of the 124 control and of the 14 treated hypothyroid mothers were significantly higher than those of the children of the 48 untreated hypothyroid women. More than twice as many children of the untreated mothers had IQs of >1 s.d. below the control mean, and four times as many of the children had IQs 2 s.d. below the control mean, as did the children of the controls. A comparison of the mean hormonal values of the treated and untreated mothers at 17 weeks showed no significant difference in any of the biochemical markers. We surmise that the circulating level of thyroxine was normal in the treated mothers at a critical time before 17 weeks' gestation, but by 17 weeks it was insufficient to meet the growing demands of pregnancy. Treatment should begin as early as possible in pregnancy with the goal of maintaining free thyroxine (FT(4)) in the upper half of the normal reference range and TSH in the lower half of the normal reference range. In view of these data, we believe that all pregnant women should be screened for hypothyroidism as early in pregnancy as possible (or even before conception). To be cost-effective, screening programs should be based on those designed for congenital hypothyroidism, in which filter-paper blood specimens are forwarded to regional laboratories for thyroid function determinations.