The effects of combined insulin and sulfonylurea therapy on glycaemic control and B-cell function was studied in 15 Type 2 (non-insulin-dependent) diabetic patients who had failed on treatment with oral hypoglycaemic agents. The patients were first treated with insulin alone for four months. Five patients were given two daily insulin doses and ten patients one dose. During insulin treatment the fasting plasma glucose fell from 14.5 +/- 0.8 to 8.8 +/- 0.4 mmol/l and the HbA1 concentration from 12.6 +/- 0.4 to 9.2 +/- 0.2%. This improvement of glycaemic control was associated with a suppression of basal (from 0.31 +/- 0.04 to 0.10 +/- 0.02 nmol/l) and glucagon-stimulated (from 0.50 +/- 0.08 to 0.19 +/- 0.04 nmol/l) C-peptide concentrations. Four months after starting insulin therapy the patients were randomised to a four-month double-blind cross-over treatment with insulin combined with either 15 mg glibenclamide per day or with placebo. Addition of glibenclamide to insulin resulted in a further reduction of the fasting plasma glucose (7.9 +/- 0.5 mmol/l) and HbA1 (8.3 +/- 0.2%) concentration whereas the basal (0.21 +/- 0.03 nmol/l) and glucagon-stimulated C-peptide concentrations (0.34 +/- 0.06 nmol/l) increased again. Addition of placebo to insulin had no effect. The daily insulin dose could be reduced by 25% after addition of glibenclamide to insulin, while it remained unchanged when insulin was combined with placebo. The fasting free insulin concentration did not differ between the glibenclamide and placebo periods (28 +/- 6 vs 30 +/- 5 mmol/l).(ABSTRACT TRUNCATED AT 250 WORDS)