Plasma haemolysis index and interleukine-6 for the early prediction of cardiac surgery-associated acute kidney injury. A proof-of-concept study.

Abstract

<h4>Introduction</h4>Haemolysis and inflammation contribute to cardiac surgery-associated acute kidney injury (CS-AKI). We aimed to assess the performance of plasma haemolysis index (HI) and interleukine-6 (IL-6) for the prediction of all-stage CS-AKI. We also assessed their ability to predict moderate-to-severe CS-AKI and to discriminate persistent from transient CS-AKI.<h4>Methods</h4>Adult patients undergoing cardiac surgery with cardiopulmonary bypass (CPB) were prospectively included. Haemolysis index and IL-6 were measured immediately after the end of CPB and 6 hours later. Correction for haemodilution relied upon changes in albuminaemia. Persistent CS-AKI was defined as a steady/increasing CS-AKI stage between the 48^th and the 60^th postoperative hour as compared with the worst stage observed within the 48 first hours.<h4>Results</h4>Among 82 patients, CS-AKI occurred in 37 (45%) patients. Postoperative HI and IL-6 were positively correlated to the duration of CPB (r ≤ 0.51, <i>p</i> ≤ 0.0003). Whether we considered isolated measurements of HI or IL-6, their indexation to haemodilution or not, their kinetics and/or their combination, the prediction of all stage CS-AKI was inaccurate (area under the receiver operating characteristic curve [AUC_ROC]≤ 0.68) whereas moderate-to-severe CS-AKI (6 patients only) was predicted with an honourable performance (AUC_ROC = 0.77 [95%CI 0.67;0.86] and 0.87 [95%CI 0.77;0.93] for HI and IL-6, respectively). The persistent/transient nature of CS-AKI was inaccurately predicted (AUC_ROC ≤ 0.68).<h4>Conclusions</h4>In a population in which most CS-AKI cases were mild, although they frequently (41%) persisted >48 hours, CS-AKI was inaccurately predicted by HI and/or IL-6. A better performance for moderate-to-severe CS-AKI prediction is likely. These preliminary findings are yet to be validated.