Vitamin D deficiency reduces insulin secretion and still occurs in East London Asians in whom the prevalence of diabetes mellitus is at least four times that of Caucasians. Vitamin D status was assessed in 44 of 65 non-diabetic subjects 'at risk' of diabetes (spot blood glucose level >6.0 mmol/l <2 h post cibum, or>4.6 mmol/l >2 h post cibum on two separate occasions) and in 15 of 60 age and sex-matched 'low-risk' control subjects who attended for oral glucose tolerance test (OGTT) after screening of 877 omnivorous subjects not known to have diabetes. It was found that 95% of at-risk and 80% of low-risk subjects were vitamin D deficient (serum 25-hydroxy-vitamin D <11 ng/ml). Diabetes was present in 16, impaired glucose tolerance in 12 and normoglycaemia in 19 at-risk subjects, imparied glucose tolerance in 2, and normoglycaemia in 13 low-risk subjects. Correlation of 30-min OGTT blood glucose, specific insulin and C-peptide levels with 25-hydroxy-vitamin D concentrations in 44 at-risk subjects were -0.31 (p=0.04), 0.59 (p=0.0001) and 0.44 (p=0.006). In 15 'not-at-risk' subjects 30-min OGTT specific insulin and C-peptide levels correlated with 25-hydroxy-vitamin D, r=0.39 (p=0.04) and 0.16 (p=0.43), respectively. Serum alkaline phosphatase concentration was higher in at-risk than not-at-risk subjects (59.6 vs 46.5 IU/l, p=0.012); corrected calcium concentrations were comparable (2.38 vs 2.39 mmol/l, p=0.7). Following treatment with 100,000 IU vitamin D by i.m. injection, specific insulin, C-peptide [30 min on OGTT] and 25-hydroxy-vitamin D concentrations had risen 8-12 weeks later [mean +/- DS] from 57 +/- 62 to 96.2 +/- 82.4 mU/l [p=0.0017], 1.0 +/- 0.4 to 1.7 +/- 0.8 pmol/ml [p=0.001] and 3.6 +/- 1.8 to 13.5 +/- 7.4 ng/ml [p=0.0001], (but not to low-risk group values of 179 +/- mU/l, 2.7 +/- 1.14 pmol/ml and 8.16 +/- 6.4 ng/ml), respectively. Both total serum alkaline phosphatase and corrected calcium concentrations rose following vitamin D treatment in the at-risk subjects by 11.1 +/- 8.22 (from 44 to 55 IU/l) and 0.15 +/- 0.18, (2.43 to 2.57 mmol/l), respectively (p=0.004). Abnormal glucose tolerance was unchanged by vitamin D treatment. The value of early and sustained repletion with vitamin D in diabetes prophylaxis should be examined in communities where vitamin D depletion is common.