This study compared treatment with a prandial glucose regulator (repaglinide) and a sulphonylurea (glibenclamide) in Muslim Type 2 diabetic patients who practice Ramadan fasting. Two hundred and thirty-five patients, previously treated with a sulphonylurea, were randomised to receive either repaglinide (n=116, preprandially three-times daily) or glibenclamide (n=119, preprandially once- or twice-daily) 6 weeks before Ramadan. During Ramadan, patients changed their eating pattern to two meals daily, and the daily dose of repaglinide was redistributed to two preprandial doses. After Ramadan, patients resumed their regular meal pattern and treatment dosage for 4 weeks. During Ramadan, a statistically significant reduction in mean serum fructosamine concentration from baseline was observed in the repaglinide group (-16.9+/-4.9 micromol/l, -3.8%, P<0.05) but not the glibenclamide group (-6.9+/-4.8 micromol/l, -0.8%). Difference in change in HbA(1c) from baseline was not statistically significant between groups. The number of hypoglycaemic events with midday blood glucose <4.5 mmol/l was significantly lower in the repaglinide group (2.8%) than the glibenclamide group (7.9%) (P=0.001). Apart from hypoglycaemia, both treatments were equally well tolerated. Type 2 diabetic Muslims using prandial repaglinide showed a trend towards better glycaemic control and had a lower frequency of hypoglycaemia than patients using glibenclamide during Ramadan.