Integrated biomarkers and renal Doppler for early acute kidney injury diagnosis in hepatitis C virus cirrhosis.

BACKGROUND

Acute kidney injury (AKI) complicates up to 65% of hospitalizations for decompensated cirrhosis and is associated with high short-term mortality. Traditional markers such as serum creatinine are unreliable in cirrhosis, while novel biomarkers like neutrophil gelatinase-associated lipocalin (NGAL), cystatin C, and Doppler renal resistive index (RRI) show promise but have not been systematically validated in hepatitis C virus-endemic, resource-limited settings such as Egypt.

AIM: To evaluate the diagnostic and prognostic utility of NGAL, cystatin C, and RRI for early AKI detection, subtyping [pre-renal, hepatorenal syndrome (HRS), acute tubular necrosis (ATN)], and mortality risk stratification in cirrhotic patients.

METHODS

In this prospective cohort study, 120 cirrhotic patients with ascites were enrolled and categorized into three groups: Non-AKI (n = 42), pre-renal AKI (n = 43), and intrinsic AKI (n = 35), the latter subdivided into HRS (n = 20) and ATN (n = 15). Serum NGAL and cystatin C were quantified using commercially available enzyme-linked immunosorbent assay kits; RRI was obtained via standardized Doppler ultrasonography. AKI was staged per International Club of Ascites-Kidney Disease Improving Global Outcomes guidelines. To reflect real-world presentation, the cohort included both prevalent (present at admission) and incident (developing postadmission) AKI cases. Diagnostic performance was assessed using receiver operating characteristic analysis, and Cox regression models evaluated 30-day mortality. An NGAL-RRI-based machine-learning model (XGBoost) was developed for AKI subtyping, and a decision-tree model was constructed. Point-of-care ultrasound was associated with the reclassification of 42% of diagnostically ambiguous AKI cases.

RESULTS

AKI was present in 65% (78/120) of patients. Of these, 52 (66.7%) had AKI at admission, and 26 (33.3%) developed AKI during hospitalization. NGAL distinguished AKI with area under the curve (AUC) of 0.91 (95%CI: 0.87-0.95) [cutoff > 150 ng/mL, sensitivity of 90% (85%-94%), specificity of 85% (78%-90%)]. RRI > 0.74 identified HRS with AUC of 0.89 (95%CI: 0.83-0.94) with sensitivity of 82%, specificity of 88%. Cystatin C showed moderate diagnostic value [AUC of 0.86 (95%CI: 0.81-0.91)]. Urinary granular casts were specific for ATN (100%), while renin and aldosterone levels were highest in HRS. NGAL > 150 ng/mL was independently associated with higher 30-day mortality (hazard ratio = 3.1, 95%CI: 1.7-5.5), while RRI > 0.74 was associated with poor renal recovery (hazard ratio = 2.2, P = 0.003). The internally validated NGAL-RRI XGBoost model achieved a cross-validated AUC of 0.92 for AKI subtyping. Excluding borderline fractional excretion of sodium cases improved HRS AUC from 0.91 to 0.93. Implementation of an NGAL-guided triage protocol corresponded to a 28% lower rate of intensive care unit transfers in a simulated decision framework (14/50 avoided).

CONCLUSION

NGAL and RRI outperform creatinine in early AKI detection, subtyping, and prognostication in cirrhotic patients, while cystatin C adds diagnostic value, particularly in ambiguous cases. An NGAL-RRI-based algorithm, enhanced by cystatin C when needed, represents a practical, resource-efficient tool for AKI triage in resource-limited settings.