Glimepiride treatment and IGF-I in adolescents with type 1 diabetes: a prospective, randomized, double-blind, placebo-controlled study.

S erum IGF-I is reduced in adolescents with type 1 diabetes, and injections of IGF-I improve glycemic control (1). The fact that sulfonylureas can increase IGF-I directly and independent of insulin has not been included in standard literature (2). The first observation of a stimulatory effect on serum IGF-I was made in hypophysectomized rats (3). In in vitro experiments, glibenclamide stimulated growth of human chondrocytes via IGF-I and independent of insulin (4). Glibenclamide and glimepiride had dose-dependent stimulatory effects on IGF-I transcription and production in human liver cells (HuH7) (5). We recruited 40 pubertal patients with type 1 diabetes of a duration of 1 year (negative for C-peptide) at Ulm (n 20) and Bern (n 20). They were randomly allocated at the start of treatment and each participant underwent a 6-week course of either glimepiride (one daily dose of 8.2 mol 4 mg; n 20) or placebo (n 20) in addition to the multiple injection intensive insulin therapy (Table 1). One patient receiving glimepiride was withdrawn because of viral encephalitis. The primary end point in our study had been defined as the increment of IGF-I between start of treatment and 6–8 weeks thereafter. Assuming a SD of 200 ng/ml, we estimated that in a twosided statistical test with an level of 0.05 and a power of 80%, sample sizes of 17 patients per group would be sufficient to attain a significant result, if a true rise in IGF-I from 300 ng/ml (5th percentile) to 500 ng/ml (50th percentile) occurred. The study protocol was approved by the local ethics committees at both centers. At the time of allocation, both groups were not relevantly different regarding age, sex, weight, height, blood pressure, insulin dose, fasting serum glucose, hypoglycemic events, IGF-I, IGF binding protein-3 (IGFBP-3), HbA1c, or serum lipids. No remarkable changes (Mann-Whitney U test) in IGF-I or IGFBP-3 could be observed during glimepiride treatment (Table 1). When compared with the placebo group, no differences could be found. Glimepiride did not influence weight, blood pressure, insulin dosage, fasting serum glucose, rate of hypoglycemias, HbA1c, or serum lipids. In adolescents with type 1 diabetes, the peripheral mode of application of insulin is likely to lead to IGF-I insufficiency, consecutively to growth hormone hypersecretion and an insulin-resistant state. In case oral sulfonylureas could effectively increase IGF-I, they could present a suitable therapeutic option because they are inexpensive, easy to administer, and do not endanger patients by hypoglycemias. An increase of IGF-I to the upper normal range would be desirable and would not likely be associated with severe side effects (6). For safety reasons, glimepiride, which exhibited a higher stimulatory effect on IGF-I than glibenclamide (5), was given at a usual dose. We anticipated that a treatment duration of 6 weeks should be sufficient to induce a change in IGF-I. The reason why IGF-I did not increase significantly probably lies in the low serum concentrations of glimepiride (median 0.16 mol/l) achieved with our protocol. Glimepiride levels were up to four times higher in the cell culture experiments (5). The authors consider it appropriate to suggest further studies using higher concentrations of sulfonylureas.