Prognostic factors for successful insulin therapy in subjects with type 2 diabetes

Objective: To assess which factors influence or predict the efficacy of ins ulin therapy in subjects with type 2 diabetes, who were poorly controlled d espite maximal doses of oral glucose lowering agents. Research design and methods: Seventy-five patients with type 2 diabetes par ticipated (mean age (CSD), 67 +/- 8 years; body mass index, 25.8 +/- 5.0 kg /m(2); median time since diagnosis of diabetes, 8 years (range 1-36); 27 ma les and 48 females). They were transferred to insulin therapy, in which cas e either insulin alone, or a combination of insulin and glibenclamide was e mployed. The importance of baseline parameters (glycaemic control, beta-cel l function, measures of insulin resistance) was assessed by comparing good and poor responders (defined as achieved HbA(1c) < 8.0 or > 9.0%) to insuli n therapy, and by multiple logistic regression analysis of these baseline p arameters and achieved metabolic control. Results: During insulin therapy, HbA(1c) levels decreased from 10.9 +/- 1.3 to 8.2 +/- 1.1% (p < 0.001), and fasting blood glucose levels decreased fr om 14.0 +/- 2.3 to 8.2 +/- 2.1 mmol/l (p < 0.001). Thirty patients reached HbA(1c) levels < 8.0%, 21 of them even < 7.5%. The mean increase in body we ight was 4.5 kg. HbA(1c) after 6 months was 7.0 +/- 0.6% in the good respon ders, and 9.8 +/- 0.6% in the poor responders (p < 0.001), despite a compar able insulin dose. Baseline metabolic control was similar in both groups. A lso, glucagon-stimulated and calculated insulin secretion, as well as param eters of insulin resistance, such as fasting serum insulin levels, free fat ty acids, and serum triglycerides, were not different between both groups, and certainly not higher in the poor responders. Also previous metformin us e was not different. However, poor responders were more obese than good res ponders, and had significantly longer known duration of diabetes. Multiple logistic regression confirmed that only duration of diabetes and body mass index were independent predictors of response to insulin therapy. Conclusions: We conclude that in elderly patients with type 2 diabetes impr ovement of glycaemic control can be achieved at the expense of some weight gain. Measurement of residual insulin secretion prior to institution of ins ulin treatment does not discriminate between good and poor responders to th is mode of therapy. Especially in obese patients with longer duration of di abetes more attention is needed in order to achieve optimal glycaemic contr ol. Combination of insulin with newer drugs, like thiazolidinediones, may p erhaps achieve this. (C) 1999 Elsevier Science B.V. All rights reserved.