Clinical usefulness of novel biomarkers for the detection of acute kidney injury following elective cardiac surgery.

Background/aims: Acute kidney injury (AKI) is common following cardiac surgery and predicts a poor outcome. However, the early detection of AKI has proved elusive and most cases are diagnosed only following a significant rise in serum creatinine (SCr). We compared a panel of early biomarkers of AKI for the detection of AKI in patients undergoing heart surgery. This study included serum cystatin C (CyC) and urinary levels of neutrophil gelatinase-associated lipocalin (NGAL), interleukin-18 (IL-18), retinol-binding protein (RBP) and N-acetyl-beta-D-glucosaminidase (NAG).

Methods: We retrospectively identified 15 patients undergoing open cardiac surgery who developed AKI within 72 h postoperatively. For these, we identified 15 matched controls also having undergone surgery but without AKI. Serial serum and urine samples had prospectively been postoperatively obtained from all patients at 0, 2, 4, 6, 10, 24, 48 and 72 h after admission to the intensive care unit. AKI was defined as a >50% increase in SCr. CyC was measured by nephelometry, while NGAL, IL-18, and RBP were measured by ELISA and NAG was measured by spectrophotometry. The urinary biomarkers were normalized to urinary creatinine (UCr) concentration. Each marker was assessed at each time point for its predictive value using receiver operating characteristic curves to predict AKI.

Results: Following the exclusion of 1 case due to a urinary tract infection, the final cohort consisted of 29 patients aged 62.9 +/- 13.7 years with baseline SCr of 73.2 +/- 11.9 micromol/l. While there were no differences in the demographics between cases and controls, the aortic clamp time was predictably higher in AKI cases than in controls (60.6 +/- 13.9 vs. 43.0 +/- 9.2 min, p < 0.05). Each biomarker differed significantly between cases and controls for at least one time point. The optimal area under the curve (AUC) was for CyC at 10 h (sensitivity 0.71, specificity 0.92, cutoff 1.31 mg/l), NGAL at 0 h (sensitivity 0.84, specificity 0.80, cutoff 49.15 microg/g UCr), IL-18 at 2 h (sensitivity 0.85, specificity 0.73, cutoff 285.65 ng/g UCr), RBP at 0 h (sensitivity 0.75, specificity 0.67, cutoff 2,934.65 microg/g UCr) and NAG at 4 h (sensitivity 0.86, specificity 0.67, cutoff 37.05 U/mg UCr). Using a combination of all 5 biomarkers analyzed at the optimal time point as above, we were able to obtain an AUC of 0.98 (0.93-1.02, p < 0.001) in this limited sample.

Conclusion: The use of serum and urinary biomarkers for the prediction of AKI in patients undergoing cardiac surgery is highly dependent on the sampling time. Of the evaluated markers urinary NGAL had the best predictive profile. The previously unstudied marker of urinary RBP showed similar predictive power as more established markers. By combining all 5 studied biomarkers we were able to predict significantly more cases, suggesting that the use of more than one marker may be beneficial clinically.