The role of sulphonylurea in combination therapy assessed in a trial of sulphonylurea withdrawal

Aims To evaluate the effect of adding insulin to sulphonylurea (SU) and the effect of SU withdrawal on glycaemic control in Type 2 diabetic patients who failed on treatment with SU alone. Method One hundred and seventy-five patients were included in a placebo-controlled multicentre study. During phase I (4 months), premixed insulin was added to glibenclamide therapy; during phase II (1-4 months, depending on response) the insulin dose was fixed, while placebo or glibenclamide replaced the open SU therapy. Insulin sensitivity (KITT), beta-cell function (C-peptide) and metabolic control (HbA(1c)) were monitored. Results HbA(1c) improved from 9.65% to 7.23% (P < 0.0001) during phase I. A high HbA(1c) value (P < 0.0001) and a high KITT-value (P = 0.045) at baseline were associated with a beneficial response to combination treatment. During phase II, glycaemic control was unchanged in the control (glibenclamide) group. In the placebo group, after SU withdrawal, fasting blood glucose (FBG) increased by 10% or more within 4 weeks in 79% of the patients. Patients (67 of 112) with an FBG increase greater than or equal to 40% during phase II were defined as 'SU responders' by protocol. In a multivariate analysis only a long duration of diabetes was associated with SU response. There were more GAD-antibody positive patients among non-responders (18% vs. 4%, P = 0.0263). Conclusions Poor glycaemic control in combination with preserved insulin sensitivity and lack of GAD antibodies predicts a beneficial response to combination therapy, which can be achieved in 75 % of patients with SU failure.