Development of predisposition, injury, response, organ failure model for predicting acute kidney injury in acute on chronic liver failure

Background and Aim There is limited data on predictors of acute kidney injuryin acuteon chronic liver failure. We developed a PIRO model (Predisposition, Injury, Response, Organ failure) for predicting acute kidney injuryin a multicentric cohort of acuteon chronic liver failurepatients. Patients and Methods Data of 2360 patients from APASL-ACLF Research Consortium (AARC) was analysed. Multivariate logistic regression model (PIRO score) was developed from a derivation cohort (n=1363) which was validated in another prospective multicentric cohort of acuteon chronic liver failurepatients (n=997). Results Factors significant for P component were serum creatinine[(≥2 mg/dL)OR 4.52, 95% CI (3.67-5.30)], bilirubin [(<12 mg/dL,OR 1) vs (12-30 mg/dL,OR 1.45, 95% 1.1-2.63) vs (≥30 mg/dL,OR 2.6, 95% CI 1.3-5.2)], serum potassium [(<3 mmol/LOR-1) vs (3-4.9 mmol/L,OR 2.7, 95% CI 1.05-1.97) vs (≥5 mmol/L,OR 4.34, 95% CI 1.67-11.3)] and blood urea (OR 3.73, 95% CI 2.5-5.5); for I component nephrotoxic medications (OR-9.86, 95% CI 3.2-30.8); for R component, Systemic Inflammatory Response Syndrome,(OR-2.14, 95% CI 1.4-3.3); for O component, Circulatory failure (OR-3.5, 95% CI 2.2-5.5). The PIRO score predicted acute kidney injurywith C-index of 0.95 and 0.96 in the derivation and validation cohort. The increasing PIRO score was also associated with mortality (P<.001) in both the derivation and validation cohorts. Conclusions The PIRO model identifies and stratifies acuteon chronic liver failurepatients at risk of developing acute kidney injury. It reliably predicts mortality in these patients, underscoring the prognostic significance of acute kidney injuryin patients with acuteon chronic liver failure.