Waist-hip ratio (WHR) defined obesity phenotype and risk of diabetes in cancer survivor women from Sister Study

Introduction : Cancer survivors are at elevated risk of developing diabetes and often experience treatment-related changes in body composition and insulin sensitivity. However, few studies have evaluated whether general and central obesity, independently and jointly, are associated with diabetes risk in long-term cancer survivors. We examined whether obesity phenotype (general, central, or combined) is associated with incident diabetes among women cancer survivors. Methods : We included 1,924 women (aged 35 to 74 years) from the Sister Study (enrolled 2003-2009) who reported a history of cancer other than breast or non-melanoma skin cancer and had no diabetes at baseline. We excluded participants diagnosed <1 year before enrollment (median time since diagnosis: 11.6 years). Participants were followed through September 2021. General obesity was defined as examiner-measured body mass index (BMI)≥30 kg/m 2 , and central obesity as waist-hip ratio (WHR)≥0.85. Participants were categorized as: no obesity (BMI<30, WHR<0.85), central-only (WHR≥ 0.85, BMI<30), general-only (BMI ≥30, WHR<0.85), and combined obesity (BMI≥30 and WHR ≥0.85). Incident diabetes was identified by self-reported new physician diagnosis during follow-up. Multivariable Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for incident diabetes across obesity phenotypes. Results : At baseline 16.5% of participants had central-only obesity, 14.0% had general-only obesity, and 13.3% had combined obesity. Over a median follow-up of 12.9 years, 157 women (8.2%) developed incident diabetes. After adjusting for potential confounders, central obesity alone was associated with more than 2-fold higher diabetes incidence (HR 2.14, 95% CI 1.25-3.68), and general obesity alone showed a similar >2-fold association (HR 2.33, 95% CI 1.38-3.92), each compared with no obesity. The combined obesity phenotype was associated with the highest incidence, showing a more than 6-fold risk (HR 6.48, 95% CI 4.13-10.19) with a significant multiplicative interaction between general and central obesity (p for interaction <0.0001). Alternative central obesity parameters (waist circumference ≥88 cm and waist-height ratio ≥0.50) showed similar associations. Associations were generally consistent across categories of time since cancer diagnosis (<5 years, 5-<15 years, and ≥15 years). These associations also were seen among women with prevalent hypertension, dyslipidemia, or cardiovascular disease. Conclusion : Our findings suggest that central and general obesity, especially in combination, are strongly positively associated with diabetes incidence in cancer survivors. Incorporating WHR-based obesity assessment into survivorship care may help identify women cancer survivors who would benefit from targeted diabetes prevention and monitoring.