Aims: The most suitable method for predicting the glomerular filtration rate (GFR) in obesity is currently debated. Therefore, multiple GFR/creatinine clearance prediction methods were applied to (morbidly) obese and nonobese patients ranging from moderate renal impairment to glomerular hyperfiltration and their predictions were rated based on observed fosfomycin pharmacokinetics, as this model drug is exclusively eliminated via glomerular filtration.
Methods: The GFR/creatinine clearance predictions via the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), Modification of Diet in Renal Disease (MDRD; indexed and de-indexed by body surface area) and creatinine clearance via the Cockcroft–Gault formula (CLCRCG) using different body size descriptors were compared to the fosfomycin clearance (CLFOF) from 30 surgical patients (body mass index = 20.1–52.0 kg m−2), receiving 8000 mg as intravenous infusion.
Results: The concordance between CLFOF and creatinine clearance predictions was highest for CLCRCG employing either ideal body weight or adjusted body weight (if body mass >1.3 ideal body weight; CLCRCG_ABW-Schwartz, concordance-correlation coefficient [95% confidence interval] = 0.474 [0.156; 0.703], CCC) and GFR predictions via the de-indexed MDRD equation (concordance-correlation coefficient = 0.452 [0.137; 0.685]). The proportion of predicted GFR values within ±30% of the observed CLFOF (P30 = 72.3–76.7%) was only marginally lower than the reported P30 in the original CKD-EPI and MDRD publications (P30 = 84.1–90.0%).
Conclusion: This analysis represents a successful proof-of-concept for evaluating GFR/creatinine clearance prediction methods: Across all body mass index classes CLCRCG_ABW-Schwartz or the de-indexed MDRD were most suitable for predicting creatinine clearance/GFR also in (morbidly) obese, CKD stage <3B individuals in therapeutic use. Their application is proposed in optimising doses for vital therapies in obese patients requiring monitoring of renal function (e.g. methotrexate dosing).