In order to determine whether there is any relationship between pancreatic reserve and the insulin dose required for achieving a good metabolic control in type 2 diabetic patients with secondary failure to oral hypoglycaemic agents, fasting and post-glucagon C-peptide were determined in thirty-nine type 2 diabetic patients with secondary failure to sulphonylureas and hyperglycaemia < 250 mg/dl who attended an outpatient clinic. M-value was calculated in patients performing self-monitoring of blood glucose. Otherwise, pre- and post-prandial glycaemias were measured bi-weekly as outpatients. HbA1c and fructosamine were assessed monthly. A patient was considered well controlled when he or she fulfilled all the requirements of the European NIDDM Policy Group and the insulin dose necessary for these goals was correlated to the pancreatic reserve. There were two drop-outs. Thirty-five out of the thirty-seven patients complied with the objectives in an average time of 3.14 +/- 1.93 months. At the beginning of the study mean HbA1c was 8.01 +/- 1.40% and fructosamine 343.81 +/- 59.05 micromol/l, whereas at the end of the study the values were 6.91 +/- 0.94% and 291.89 +/- 38.59 micromol/l, respectively (both p < 0.001). Body weight increased from 68.95 +/- 12.40 to 69.44 +/- 12.54 kg (n.s.), while hypoglycaemic events decreased from 1.70 +/- 2.37 to 0.88 +/- 1.33 events/week (p < 0.05). To attain all the objectives, 19.03 +/- 5.98 i.u. (0.28 +/- 0.08 i.u./kg) of insulin were required. Basal and post-glucagon C-peptide were 1.97 +/- 1.24 and 3.29 +/- 1.85 ng/ml, respectively, with an increase of 1.32 +/- 0.78 ng/ml. All these values inversely correlated with insulin dose, especially the increase during the test (r = -0.652 with i.u./kg and r = -0.599 with i.u., both p < 0.01). In conclusion, C-peptide test is a good indicator of the insulin dose required for achieving the aims of metabolic control in type 2 diabetic patients.