Background: Monitoring of renal function becomes increasingly important in the aging population of HIV-1 infected patients. We compared Cockroft & Gault (C&G), Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), Modification of Diet in Renal Disease (MDRD), Cystatin C- and 24 h urine-based estimated GFR (eGFR) with the gold standard, measured GFR (mGFR) using [125I]-iothalamate.
Methods: Substudy within a randomized, multinational trial comparing continuing zidovudine/ lamivudine with switching to tenofovir/ emtricitabine in patients with suppressed HIV-1 infection. Accuracy (defined as the mean difference between eGFR and mGFR) and precision (defined as standard deviation (SD) of the mean difference between eGFR and mGFR) of the eGFRs were calculated using linear regression and Bland & Altman analysis.
Results: We included 19 patients, 18 men, 15 Caucasian, mean (SD) age 46.0 y (± 8.9) and BMI 23.9 kg/m2 (± 3.0). Mean (SD) mGFR was 102 ml/min/1.73 m2 (± 19), 4 patients had mild renal dysfunction. All eGFRs tended to underestimate true GFR, with best accuracy for C&G (-1 ml/min/1.73 m2), CKD-EPI (-1 ml/min/1.73 m2), 24 hcreatinine clearance (-2 ml/min/1.73 m2) and MDRD-6 (0 ml/min/1.73 m2), and worst for cystatin C-based (-9 ml/min/1.73 m2) and MDRD-4 estimations (-10 ml/min/1.73 m2). Accuracy worsened at higher mGFR, but was not significantly influenced by age. C&G tended to overestimate at higher BMI. Precision was comparable for all GFR estimations.
Conclusions: In this limited number of patients with preserved renal function and suppressed HIV-infection C&G and CKD-EPI appeared to be the best reflection of real GFR and most practical tool for monitoring GFR.