Acute kidney injury requiring renal replacement therapy (AKI-RRT) is strongly associated with mortality after cardiac surgery; however, options for early identification of patients at high risk for AKI-RRT are extremely limited. Early after cardiac surgery, the predictive ability for AKI-RRT even of one of the most extensively evaluated novel urinary biomarkers, neutrophil gelatinase–associated lipocalin (NGAL), appears to be only moderate. We aimed to determine whether the NGAL/hepcidin-25 ratio (urinary concentrations of NGAL divided by that of hepcidin-25) early after surgery may compare favorably to NGAL for identification of high-risk patients after cardiac surgery.
This is a prospective substudy of the BICARBONATE trial, a multicenter parallel-randomized controlled trial comparing perioperative bicarbonate infusion for AKI prevention to usual patient care. At a tertiary referral center, 198 patients at increased kidney risk undergoing cardiac surgery with cardiopulmonary bypass were included into the present study. The primary outcome measure was defined as AKI-RRT. Secondary outcomes were in-hospital mortality and long-term mortality. We compared area under the curve of the receiver operating characteristic (AUC-ROC) of urinary NGAL with that of the urinary NGAL/hepcidin-25 ratio within 60 minutes after end of surgery. We compared adjusted AUC and performed cross-validated reclassification statistics of the (logarithmic) urinary NGAL/hepcidin-25 ratio adjusted to Cleveland risk score/EuroScore, cross-clamp time, age, volume of packed red blood cells, and (logarithmic) urinary NGAL concentration. The association of the NGAL/hepcidin-25 ratio with long-term patient survival was assessed using Cox proportional hazard regression analysis adjusting for EuroScore, aortic cross-clamp time, packed red blood cells and urinary NGAL.
Patients with AKI-RRT (n = 13) had 13.7-times higher NGAL and 3.3-times lower hepcidin-25 concentrations resulting in 46.9-times higher NGAL/hepcidin-25 ratio early after surgery compared to patients without AKI-RRT. The NGAL/hepcidin-25 ratio had higher AUC-ROC compared with NGAL for risk of AKI-RRT and in-hospital mortality (unadjusted AUC-ROC difference 0.087, 95% confidence interval [CI], 0.036–0.138, P < .001; 0.082, 95% CI, 0.018–0.146, P = .012). For AKI-RRT, the NGAL/hepcidin-25 ratio increased adjusted category-free net reclassification improvement (cfNRI; 0.952, 95% CI, 0.437–1.468; P < .001) and integrated discrimination improvement (IDI; 0.040, 95% CI, 0.008–0.073; P = .016) but not AUC difference. For in-hospital mortality, the ratio improved AUC of the reference model (AUC difference 0.056, 95% CI, 0.003–0.108; P = .037) and cfNRI but not IDI. The urinary NGAL/hepcidin-25 ratio remained significantly associated with long-term mortality after adjusting for the model covariates.
The urinary NGAL/hepcidin-25 ratio appears to early identify high-risk patients and outperform NGAL after cardiac surgery. Confirmation of our findings in other cardiac surgery centers is now needed.