Abstract
Cancer survivors are at increased risk of type 2 diabetes, which usually develops from obesity and insulin resistance. Whether diabetes susceptibility is due to shared risk factors for cancer and insulin resistance or directly related to cancer and its treatment is unknown. We investigated effect modification between malignancy and body mass index (BMI) as determinants of insulin sensitivity in patients with hematologic malignancies and controls without cancer. In a cross-sectional study of 43 individuals without diabetes (20 patients with treated hematologic malignancies; 23 controls without malignancies), we measured insulin-stimulated whole-body glucose use (M) by hyperinsulinemic euglycemic clamp. Insulin sensitivity index (ISI) was calculated by dividing M over steady-state plasma insulin. Inflammatory cytokines were measured in plasma. Controls were more obese and included more non-White individuals and women vs patients with hematologic malignancies. Patients with cancer exhibited greater insulin sensitivity (median ISI, 42.4 mg/kg/min/[μU/mL]; interquartile range [IQR], 33.9-67.2 vs 23.4 mg/kg/min/[μU/mL]; IQR, 12.9-29.2; P < .001) and higher interleukin-6 (IL-6) and monocyte chemoattractant protein-1 (MCP-1) concentrations vs controls. Patients with cancer demonstrated greater reduction in ISI with increasing BMI vs controls, which remained significant after adjustment for sex and race (β = -2.6 units; 95% confidence interval, -4.8 to -0.4; P interaction = .024). This interaction also remained significant after adjusting for log IL-6 (P interaction = .048) and log MCP-1 (P interaction = .021). Cancer survivors had disproportionately greater insulin resistance with increasing BMI vs controls without malignancies. Effect modification between cancer and BMI in determining insulin sensitivity implicated cancer-specific etiologies in glucose dysregulation and could partially explain excess diabetes diagnoses among oncology patients.