Comparison of secondary failure rate between three second generation sulphonylureas.

Sulphonylureas are effective hypoglycaemic agents but have been reported to have a high failure rate with time. The introduction of a second generation of sulphonylureas may have modified this effect, but little information is available. Two hundred and forty-eight non-insulin dependent diabetic patients were prospectively randomly treated with one of three second generation sulphonylureas--gliclazide (86), glibenclamide (84), and glipizide (78), and were followed for 5 yr to assess the secondary failure rate. Six out of 86 (7%) failed with gliclazide, 15 out of 84 (17.9%) failed with glibenclamide, and 20 out of 78 (25.6%) failed with glipizide. Gliclazide was significantly better than glipizide (p less than 0.005), but not other differences were significant. With all three drugs, in the patients who failed on treatment, body mass index (BMI) at diagnosis was significantly lower than in those who responded to treatment (p less than 0.001). There are differences in secondary failure rate between second generation sulphonylureas. Measurement of the BMI at diagnosis may be useful in predicting those patients most likely to fail to respond.