Acute Insulin Response and β-Cell Compensation in Normal Subjects Treated With Olanzapine or Risperidone for 2 Weeks

A detailed description of this study has been reported previously (2). Briefly, lean (mean BMI ∼24 kg/m²) healthy volunteers were randomized to treatment with placebo (n = 18), olanzapine (10 mg/day; n = 17), or risperidone (4 mg/day; n = 13). Subjects were treated with half-maximal doses for 4 days before dose escalation. Subjects underwent hyperglycemic clamps (4) before randomization and after 15–17 days of treatment. Hyperglycemia was initiated with an intravenous priming dose of 20% (wt/vol) glucose estimated to produce a target level of glycemia (11.1 mmol/l). Half of the priming dose was delivered over the first 5 min and the remainder over the following 10 min. A variable-rate infusion was then used to maintain the target level of glycemia for a total of 240 min.

Sampling for insulin and C-peptide was performed every 2 min from 0 to 10 min and at 15- to 30-min intervals from 10 to 240 min. Baseline values were obtained at 0 and −10 min before glucose infusion. Glucose, insulin, and C-peptide were determined as reported (2). The incremental area under the curve (AUC) for insulin and C-peptide during the first 10 min of hyperglycemia (AUC_0–10) was determined using the trapezoidal rule. The insulin sensitivity index (ISI_clamp) was calculated by dividing the steady-state (180–240 min) average glucose infusion rate by the average insulin concentration over this interval (4). A DI was calculated as the product of insulin secretion and insulin sensitivity (3) using insulin AUC_0–10 and ISI_clamp, respectively. Homeostasis model assessment-1 of insulin resistance (HOMA1-IR) was calculated from mean values for baseline glucose and insulin (5). An alternative calculation of DI used 1/HOMA1-IR in place of ISI_clamp. Parametric and nonparametric statistical comparisons were used, with the Student’s t test and the Wilcoxon Mann-Whitney test for pairwise comparisons. All tests of hypotheses were two sided, with a 5% level of significance.

Glucose concentrations were not significantly different between treatment groups during the first 10 min or during the steady-state phase of the clamp procedure. Insulin levels peaked at 4 min, and the AIR was biphasic and qualitatively similar in all groups at baseline and end point (data not shown). AUC_0–10 for insulin and C-peptide and derived ISI_clamp and DI values are shown in Table 1. AUC_0–10 for insulin increased in both antipsychotic-treated groups, although the change did not reach statistical significance in the risperidone group (P = 0.07). AUC_0–10 for C-peptide also increased in antipsychotic-treated subjects, but the change only reached significance for the risperidone group. Similar decreases in ISI_clamp were observed in olanzapine- and risperidone-treated subjects (−18.9 and −16.6% mean change, respectively). This only reached statistical significance in the olanzapine group, perhaps because of fewer subjects in the risperidone group. No significant between-group differences in mean change ISI_clamp were observed (P > 0.05 for all pairwise comparisons). DI did not change significantly from baseline in any treatment group (P > 0.3), nor were significant between-group differences seen. This was also true if AUC_0–10 for C-peptide was used instead of AUC_0–10 for insulin in the calculation of DI or if 1/HOMA1-IR was used as an independent measure of insulin sensitivity in place of ISI_clamp (not shown).

The current analysis did not show impairment of the AIR to hyperglycemia in normal individuals treated with olanzapine or risperidone for 2 weeks. Instead, the AUC_0–10 for both insulin and C-peptide increased after drug treatment. Likewise, β-cell compensation, as reflected in the DI, was not significantly impacted by these treatments. The latter finding differs from the conclusions of Ader et al. (1), who suggested that olanzapine impaired β-cell compensation in dogs. It is possible that this difference reflects the longer duration of treatment in their study (6 weeks). It should also be noted that the insulin secretory and insulin sensitivity components of DI, which may be interdependent, were derived from the same clamp procedure in the current analysis. However, the results are similar if HOMA1-IR is used to estimate insulin sensitivity. DI did not significantly change with olanzapine treatment in the study by Ader et al. (1). Instead, impaired compensation was inferred because these animals did not demonstrate the increase in DI seen in fat-fed animals reportedly matched for insulin resistance relative to the olanzapine-treated dogs. It is noteworthy that these two groups received different insulin infusion rates in that study and also that the increase in DI seen with high-fat feeding is in direct contrast with the significant decrease seen in this parameter in another study (6) by the same group (62% at 4–6 weeks).

The authors of a 2004 American Diabetes Association Technical Review (7) concluded that olanzapine and risperidone are associated with an increased risk of diabetes. The current results argue against a substantial and generalized impairment of insulin secretion with these agents after short-term treatment; however, the limited size and length of this study preclude us from ruling out deleterious effects in predisposed individuals or after treatment of longer duration. Similarly, we have not seen differential effects on insulin sensitivity using the hyperinsulinemic-euglycemic clamp in normal individuals treated with these medications (8). These studies do not address whether chronic treatment with antipsychotics, especially when accompanied by substantial weight gain, may have detrimental metabolic effects and contribute to the already elevated risk of diabetes in the population of patients that use these med-ications (7). Such information is impor-tant because these drugs are frequently used on a chronic basis. Longer-term prospective studies are needed to betterunderstand potential diabetogenic effects of these agents.