Renal function assessment in patients with systemic lupus erythematosus

Few studies have evaluated the glomerular filtration rate (GFR) in patients with systemic lupus erythematosus (SLE). Even though the National Kidney Foundation (NKF) suggests using the equations to estimate GFR, rheumatologists continue using creatinine clearance (CCl). The main objective of our study was the assessment of different equations to estimate GFR in patients with SLE: Simplified MDRD study equation (sMDRD), CCl, Cockcroft Gault (CG), CG calculated with ideal weight (CGi), Mayo Clinic Quadratic (MCQ), and Chronic Kidney Disease Epidemiology Collaboration Equation (CKD-EPI). CKD-EPI was considered as the reference standard, and it was compared with the other equations to evaluate bias, correlation (r), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), percentage of measurement of GFR between 70–130% of GFR measured through CKD-EPI (P30) and to compute the ROC curves. Adequacy of the 24-h urine collection was evaluated. To classify patients into GFR < 60 ml/min/1.73 m2, the best sensitivity and NVP were obtained with sMDRD: the best PPV and specificity with MCQ. P30 was 99.3% with sMDRD, 77.5% CCl, 91.7% CG, 96.7% CGi, and 77.2% with MCQ. The lowest bias was for sMDRD and the highest for CCl. Only 159 (52.6%) urine collections were considered adequate, and when these patients were re-evaluated, the statistical results improved for CCl. CGi was better in general than CG. CCl should not be considered as an adequate GFR estimation. Ideal weight is better than real weight to calculate GFR through CG in patients with SLE.