The study was performed to reassess the prevalence of gestational diabetes mellitus (GDM) in south Indians and to study the foetal outcome in women with GDM in comparison with normal pregnancies. In 1036 pregnant women, glucose tolerance was tested with 75-g oral glucose load, in the second or third trimester. Those with 2-h plasma glucose of > or = 200 mg/dl were considered as diabetic (WHO Criteria). Those with 2-h values of 140-199 underwent a 3-h glucose tolerance test (GTT) with 100-g glucose load (O'Sullivan and Mahan criteria). GDM was diagnosed in nine women (0.87%) by this criteria. Foetal outcome in 211 GDM referred to the diabetes centre was compared with the outcome in 853 normal pregnancies. There were no cases of aborted pregnancy among the GDM, while six cases were reported among the normal glucose tolerance (NGT). Still birth and the number of premature babies were higher in GDM. Babies with birth weight > or = 3.5 kg were more among the GDM (P < 0.001). There was no difference in the occurrence of congenital anomalies in the two groups. It was noted that congenital abnormalities in the foetus were more common among those born of mothers with higher plasma glucose (9 versus 1.1%). Multiple linear regression analyses in NGT and GDM showed that the birth weight of the baby was dependent on the plasma glucose and the body mass index of the mothers. The results of the present study show that the prevalence of GDM in urban south India was low, when the NDDG criteria was used. It also indicated that the WHO criteria may be more appropriate as the foetal outcome was determined by even a small rise in maternal plasma glucose. With good metabolic control in GDM, the foetal risks are very much reduced.