Abstract
Type 1 hepatorenal syndrome (T1HRS) is the most rapidly progressive form of functional renal failure in decompensated cirrhosis, with 30-day mortality exceeding 50%. We aimed to develop and internally validate a bedside nomogram to predict 30-day death in intensive care unit (ICU) patients with T1HRS. Using the Medical Information Mart for Intensive Care IV database, adults who met diagnostic criteria for T1HRS within 48 hours of ICU admission (2008-2019) were randomly split 7:3 into development (n = 340) and internal validation (n = 146) sets. Predictors collected within 24 hours of diagnosis were selected using least absolute shrinkage and selection operator-Cox; a multivariable Cox model was constructed and internally validated. Performance was assessed by area under the curve, calibration slope, and decision-curve analysis. Overall 30-day mortality was 47.1% (229/486). Seven variables were retained: model for end-stage liver disease-sodium, total bilirubin, international normalized ratio, serum creatinine, mean arterial pressure, urine Na+, and terlipressin use. The model achieved the area under the curves of 0.82 (development) and 0.81 (validation) at 30 days, with a calibration slope of 0.99. The findings were robust to competing-risk, terlipressin exclusion, and Child-Pugh substitution. A 7-variable nomogram provides accurate, well-calibrated, and clinically actionable prediction of 30-day mortality in ICU patients with T1HRS, ready for prospective evaluation and trial stratification.