Vitamin D has been shown to increase insulin release from pancreatic islet cells in vitro, and to improve insulin secretion in vitamin D-deficient animals. Few attempts have been made to evaluate this issue directly in humans. We studied 35 otherwise healthy diabetic subjects in the early spring at the seasonal nadir of 25-hydroxyvitamin D [25(OH)D] concentrations (mean 35 +/- 7 nmol/L). Fasting glucose, insulin, C-peptide, and glucagon concentrations, and their responses to Sustacal stimulation were not related to indexes of mineral metabolism. In 20 subjects, a double-blinded, placebo-controlled, crossover trials of 1,25-dihydroxyvitamin D[1,25](OH)2D] treatment (1 micrograms/d for 4 d) had no effect on fasting or stimulated glucose, insulin, C-peptide, or glucagon concentrations. However, insulin and C-peptide responses to Sustacal after 1,25(OH)2D treatment were related to duration of diabetes (r2 = 0.28, P = 0.052 and r2 = 0.25, P = 0.002, respectively) in that short duration correlated with improvement after 1,25(OH)2D treatment. Hence, vitamin D nutrition, or 1,25(OH)2D therapy, had no major effect on glucose homeostasis in non-insulin-dependent diabetes mellitus.