Prognostic value of chronic kidney disease in patients with coronary heart disease: Role of glomerular filtration rate estimating equations

Objective: Chronic kidney disease (CKD) increases risk of coronary heart disease (CHD), and different equations for estimating kidney function on CHD are available. Our aim was to describe the prognostic value of CKD as defined by various creatinine- (Cr-eGFR) and cystatin C-based (Cys-eGFR) estimating equations and their combinations on subsequent cardiovascular disease (CVD) events in patients with CHD. Material and Methods: All patients with CHD aged 30 to 70 years and participating in an in-hospital rehabilitation program between 01/1999 and 05/2000 were enrolled. In all patients, active follow-up was conducted 1, 3, 4.5 and 6 years after discharge from the rehabilitation center. CKD was defined as eGFR <60 mL/min/1.73 m2 (CKD stage 3 to 5) estimated by three Cr-eGFR equations (CG/BSA, MDRD, Chronic Kidney Disease Epidemiology (CKD-EPIcrea) equation) and by two Cys-eGFR equations (Arnal-Dade equation, CKD-EPIcys) and a combination. The primary endpoint of our study were subsequent physician-reported CVD events. Results: 1050 patients with coronary heart disease aged 30 to 70 years at baseline were included in this prospective cohort study (follow-up response 87.1%). During follow-up (mean time 63.4 months) 118 patients (11.2%) experienced CVD events. CKD assessed by the CG/BSA, the MDRD, and CKD-EPIcrea equations showed no statistically significant association with subsequent CVD events after adjustment for multiple covariates (hazard ratio (HR) 1.45 [95% CI, 0.81–2.59], HR 1.47 [95% CI 0.84–2.60], and HR 1.31 [95% CI, 0.72–2.83) respectively). By contrast, the Cys-eGFR equations were much stronger associated with secondary CVD endpoints (Arnal-Dade: HR, 2.01 [95% CI, 1.34–3.04]; CKD-EPIcys HR, 2.22 [95% CI, 1.46–3.37]). The CKD-EPIcys also provided the highest Area Under Curve value. Conclusion: Our study shows that prevalent CKD is an independent risk factor for subsequent CVD in patients with prevalent CHD and implies that Cys-eGFR equations show a better clinical utility.