The widely used MDRD formula underestimates kidney function in obese patients with diabetes mellitus. Therefore we aimed to evaluate the predictive value of estimated glomerular filtration rate (eGFR) for cardiovascular events in a typical cohort of patients with diabetes mellitus.
A total of 988 patients were analyzed. Cox regression models including the variables HbA1c, age, duration of diabetes, eGFR and urinary albumin to creatinine ratio (UACR) were run. First the whole collective was analyzed, in a second step the cohort was split into four different groups according to body mass index (BMI) and eGFR (Group 1, 475 Pts: eGFR > 60 ml/min; BMI < 30 kg/m2, Group 2, 274 Pts: eGFR > 60 ml/min;BMI > 30 kg/m2, Group 3, 110 Pts,: eGFR < 60 ml/min; BMI > 30 kg/m2 and Group 4, 129 Pts.: eGFR < 60 ml/min;BMI < 30 kg/m2). eGFR was calculated using MDRD, Cockroft–Gault, and CKD-EPI formula. The endpoint was defined as unplanned cardiovascular hospitalization.
Patients (571 male, 417 female) were 61 ± 22 years of age, mean duration of diabetes was 14.3 ± 12.3 years. After a median follow-up of 29 months 95 (9.6 %) patients reached the defined endpoint. The first model, including all patients showed that UACR (HR 1.001, p < 0.001) and eGFR (HR 0.957, p < 0.001) were significant predictors of the composite endpoint. In obese patients eGFR completely lost its predictive value for cardiovascular events.
The prevalence of normoalbuminuria in patients with an eGFR below 60 ml/min was 59.4 %.
In obese patients eGFR is not predictive for cardiovascular events.