Differential Effect of Baseline Adiponectin on All-Cause Mortality in Hemodialysis Patients Depending on Initial Body Mass Index. Long-Term Follow-Up Data of 4.5 Years

Abstract

We sought to investigate the interaction of adiponectin levels and body mass index (BMI) for predicting all-cause mortality in a cohort of hemodialysis (HD) patients. Longitudinal, observational cohort study. HD unit. Sixty patients (mean age: 64 ± 13 years, 39 men) with end-stage renal disease on maintenance HD followed up for 4.5 years represented the prospective study cohort. Associations between baseline plasma adiponectin levels and initial BMI with all-cause mortality were assessed taking into account the assumption of nonlinear correlations. The association between adiponectin, BMI, and serum levels of interleukin-10 (IL-10) and interleukin-6 (IL-6) with survival was determined cross-sectionally. All-cause mortality. Nonlinear survival modeling showed that there was a U-shaped association of BMI with all-cause mortality, whereas there was an inverse U-shaped association for plasma adiponectin levels. Using a BMI of 24 kg/m(2) as a cutoff, an interaction effect of BMI on the association between adiponectin and mortality was observed (P = .045). In participants with BMI ≥ 24 kg/m(2), each 15 μg/mL increase in plasma adiponectin levels was associated with a decreased hazard of death (hazard ratio: 0.57, 95% CI: 0.32 to 0.99) in unadjusted analysis. In HD patients with BMI < 24 kg/m(2), no significant association was observed between adiponectin and mortality (P = .989). Cross-sectional analysis showed that in the subgroup of patients in whom the protective effect of adiponectin was observed (BMI ≥ 24 kg/m(2)), a positive linear association existed between adiponectin and IL-10 levels (r = 0.345, P = .027) as well as a negative association with IL-6 levels (r = -0.322, P = .040). No association was observed in patients with BMI < 24 kg/m(2), neither with IL-10 nor with IL-6. Obesity possibly modifies the effect of adiponectin on all-cause mortality in HD patients, thus explaining the published conflicting results in recent literature regarding the association of plasma adiponectin levels and mortality in chronic kidney disease patients.