Estimating glomerular filtration rate in patients with acute kidney injury: a prospective multicenter study of diagnostic accuracy

Abstract

Abstract Background Estimating glomerular filtration rate (GFR) in acute kidney injury (AKI) is challenging, with limited data comparing estimated and gold standard methods to assess GFR. The objective of our study was to assess the performance of the kinetic estimated GFR (KeGFR) and Jelliffe equations to estimate GFR in AKI, using a radioisotopic method (technetium-diethylenetriaminepentaacetic acid) as a reference measure. Methods We conducted a prospective multicenter observational study in hospitalized patients with AKI. We computed the Jelliffe and KeGFR equations to estimate GFR and compared these estimations to measured GFR (mGFR) by a radioisotopic method. The performances were assessed by correlation, Bland–Altman plots and smoothed and linear regressions. We conducted stratified analyses by age and chronic kidney disease (CKD). Results The study included 119 patients with AKI, mostly from the intensive care unit (63%) and with Stage 1 AKI (71%). The eGFR obtained from the Jelliffe and KeGFR equations showed a good correlation with mGFR (r = 0.73 and 0.68, respectively). The median eGFR by the Jelliffe and KeGFR equations was less than the median mGFR, indicating that these equations underestimated the mGFR. On Bland–Altman plots, the Jelliffe and KeGFR equations displayed a considerable lack of agreement with mGFR, with limits of agreement >40 mL/min/1.73 m2. Both equations performed better in CKD and the KeGFR performed better in older patients. Results were similar across AKI stages. Conclusions In our study, the Jelliffe and KeGFR equations had good correlations with mGFR; however, they had wide limits of agreement. Further studies are needed to optimize the prediction of mGFR with estimatation equations.