AIM: Although higher adiponectin levels predict a low risk of type 2 diabetes, elevated adiponectin levels predict higher mortality in older persons. METHODS: We examined the associations of adiponectin with anthropometric, metabolic and hematological variables and renal function in 361 community-dwelling elderly women aged 76±8. Renal function was assessed using the estimated glomerular filtration rate (eGFR). RESULTS: By univariate analysis, hemoglobin (r=-0.307, p<0.001) and creatinine-based eGFR (r=-0.121, p<0.05) were inversely associated with adiponectin. After adjustment for percentage body fat, HDL cholesterol and serum leptin, hemoglobin and creatinine-based eGFR remained independent predictors of adiponectin (standardized β coefficient=-0.248, p<0.0001, and -0.101, p=0.03, respectively). A similar but stronger relationship was observed between adiponectin and cystatin C-based eGFR (standardized β coefficient=-0.180, p=0.02). Elderly women with anemia had higher serum adiponectin than those without anemia (17.7±9.1 vs. 14.1±7.0 µg/mL, p=0.001). Women with creatinine-based eGFR <45 mL/min/1.73 m2 (19.2±10.6 µg/mL) had higher adiponectin than those with eGFR ≥45 -< 60 mL/min/1.73 m(2) (14.4±6.9 µg/mL) and those with eGFR ≥60 mL/min/1.73 m(2) (14.5±7.4 µg/mL, p<0.05). CONCLUSIONS: We conclude that hemoglobin and eGFR are independently associated with serum adiponectin in community-dwelling elderly women. Anemia and reduced renal function may contribute to elevated adiponectin levels in the elderly and may represent important confounders of the relationship between elevated adiponectin and mortality in this population.
AIM: Although higher adiponectin levels predict a low risk of type 2 diabetes, elevated adiponectin levels predict higher mortality in older persons. METHODS: We examined the associations of adiponectin with anthropometric, metabolic and hematological variables and renal function in 361 community-dwelling elderly women aged 76±8. Renal function was assessed using the estimated glomerular filtration rate (eGFR). RESULTS: By univariate analysis, hemoglobin (r=-0.307, p<0.001) and creatinine-based eGFR (r=-0.121, p<0.05) were inversely associated with adiponectin. After adjustment for percentage body fat, HDL cholesterol and serum leptin, hemoglobin and creatinine-based eGFR remained independent predictors of adiponectin (standardized β coefficient=-0.248, p<0.0001, and -0.101, p=0.03, respectively). A similar but stronger relationship was observed between adiponectin and cystatin C-based eGFR (standardized β coefficient=-0.180, p=0.02). Elderly women with anemia had higher serum adiponectin than those without anemia (17.7±9.1 vs. 14.1±7.0 µg/mL, p=0.001). Women with creatinine-based eGFR <45 mL/min/1.73 m2 (19.2±10.6 µg/mL) had higher adiponectin than those with eGFR ≥45 -< 60 mL/min/1.73 m(2) (14.4±6.9 µg/mL) and those with eGFR ≥60 mL/min/1.73 m(2) (14.5±7.4 µg/mL, p<0.05). CONCLUSIONS: We conclude that hemoglobin and eGFR are independently associated with serum adiponectin in community-dwelling elderly women. Anemia and reduced renal function may contribute to elevated adiponectin levels in the elderly and may represent important confounders of the relationship between elevated adiponectin and mortality in this population.