Urinary biomarkers may provide prognostic information for subclinical acute kidney injury after cardiac surgery

Objective: This study aimed to determine the biomarker-specific outcome patterns and short-and long-term prognosis of cardiac surgery–asoociated acute kidney injury (AKI) identified by standard criteria and/or urinary kidney biomarkers. Methods: Patients enrolled (N = 200), originated a German multicenter study (NCT00672334). Standard risk injury, failure, loss, and end-stage renal disease classification (RIFLE) criteria (including serum creatinine and urine output) and urinary kidney biomarker test result (neutrophil gelatinase-associated lipocalin, midkine, interleukin 6, and proteinuria) were used for diagnosis of postoperative AKI. Primary end point was acute renal replacement therapy or in-hospital mortality. Long-term end points among others included 5-year mortality. Patients with single-biomarker-positive subclinical AKI (RIFLE negative) were identified. We controlled for systemic inflammation using C-reactive protein test. Results: Urinary biomarkers (neutrophil gelatinase-associated lipocalin, midkine, and interleukin 6) were identified as independent predictors of the primary end point. Neutrophil gelatinase-associated lipocalin, midkine, or interleukin 6 positivity or de novo/worsening proteinuria identified 21.1%, 16.9%, 30.5%, and 48.0% more cases, respectively, with likely subclinical AKI (biomarker positive/RIFLE negative) additionally to cases with RIFLE positivity alone. Patients with likely subclinical AKI (neutrophil gelatinase-associated lipocalin or interleukin 6 positive) had increased risk of primary end point (adjusted hazard ratio, 7.18; 95% confidence interval, 1.52-33.93 [P =.013] and hazard ratio, 6.27; 95% confidence interval, 1.12-35.21 [P =.037]), respectively. Compared with biomarker-negative/RIFLE-positive patients, neutrophil gelatinase-associated lipocalin positive/RIFLE-positive or midkine-positive/RIFLE-positive patients had increased risk of primary end point (odds ratio, 9.6; 95% confidence interval, 1.4-67.3 [P =.033] and odds ratio, 14.7; 95% confidence interval, 2.0-109.2 [P =.011], respectively). Three percent to 11% of patients appear to be influenced by single-biomarker-positive subclinical AKI. During follow-up, kidney biomarker-defined short-term outcomes appeared to translate into long-term outcomes. Conclusions: Urinary kidney biomarkers identified RIFLE-negative patients with high-risk subclinical AKI as well as a higher risk subgroup of patients among RIFLE-AKI-positive patients. These findings support the concept that urinary biomarkers define subclinical AKI and higher risk subpopulations with worse long-term prognosis among standard patients with AKI.