Candidate surrogate end points for ESRD after AKI

Abstract

AKI, a frequently transient condition, is not accepted by the US Food and Drug Association as an end point for drug registration trials. We assessed whether an intermediate-term change in EGFR after AKI has a sufficiently strong relationship with subsequent ESRD to serve as an alternative end point in trials of AKI prevention and/or treatment. Among 161,185 United States veterans undergoingmajor surgery between 2004 and 2011,we characterized in-hospitalAKI by KidneyDisease ImprovingGlobalOutcomes creatinine criteria and decline in EGFR from prehospitalization to postdischarge time points and quantified associations of these values with ESRD and mortality over a median of 3.8 years. An EGFR decline of30%at 30, 60, and 90 days after discharge occurred in 3.1%, 2.5%, and 2.6%, of survivors without AKI and 15.9%, 12.2%, and 11.7%, of survivors with AKI. For patientswith in-hospital AKI comparedwith those with no AKI and stable EGFR, a 30% decline in EGFR at 30, 60, and 90 days after discharge demonstrated adjusted hazard ratios (95% confidence intervals) of ESRD of 5.60 (4.06 to 7.71), 6.42 (4.76 to 8.65), and 7.27 (5.14 to 10.27), with corresponding estimates for 40% decline in EGFR of 6.98 (5.21 to 9.35), 8.03 (6.11 to 10.56), and 10.95 (8.10 to 14.82). Risks for mortalitywere smaller but consistent in direction. A 30%-40%decline in EGFR after AKI could be a surrogate end point for ESRD in trials of AKI prevention and/or treatment, but additional trial evidence is needed.