Dynamic renal and hepatic trajectories during hospitalization and in-hospital mortality in acute heart failure

Background In acute heart failure (AHF), organ dysfunction evolves during hospitalization and reflects dynamic hemodynamic stress. While admission abnormalities are recognized risk markers, the prognostic significance of renal and hepatic in-hospital trajectories remains unclear. We aimed to evaluate whether dynamic renal and hepatic trajectories identify distinct in-hospital mortality phenotypes beyond baseline measurements. Methods We retrospectively analyzed consecutive patients hospitalized for AHF between January 2019 and January 2026. Worsening renal function (WRF) was defined according to KDIGO criteria. Hepatic injury was defined as total bilirubin ≥1.5 mg/dL and/or AST/ALT ≥2× the upper limit of normal. Renal trajectories were classified as: preserved renal function without WRF, baseline dysfunction without WRF, de novo WRF, and superimposed WRF. Hepatic trajectories were defined using admission, peak, and last in-hospital laboratory values as: (1) stable normal (no injury at admission and no injury at peak), (2) de novo injury (no injury at admission but injury at peak), (3) persistent injury (injury at admission and injury at last measurement), and (4) resolved injury (injury at admission but no injury at last measurement). The primary endpoint was in-hospital mortality. Multivariable logistic regression adjusted for age, sex, admission systolic blood pressure, heart rate, and log-transformed NT-proBNP. Results Among 2,261 patients, overall in-hospital mortality was 4.3% (97/2261). Renal trajectories were associated with graded mortality (p<0.001): 0.94% (7/746) in preserved renal function without WRF, 3.50% (34/971) in baseline dysfunction without WRF, 9.74% (19/195) in de novo WRF, and 12.32% (43/349) in superimposed WRF. Compared with preserved renal function without WRF, adjusted ORs were 5.33 (95% CI 1.11–25.71; p=0.037), 13.95 (2.67–72.95; p=0.002), and 17.21 (3.60–82.32; p<0.001), respectively. Hepatic trajectories were also associated with graded mortality (p<0.001). Mortality was: 2.04% (37/1812) in stable normal patients, 3.33% (3/90) in resolved injury, 14.29% (39/273) in persistent injury, and 27.91% (24/86) in de novo injury. Compared with stable normal hepatic function, adjusted ORs were 1.35 (95% CI 0.16–11.11; p=0.782) for resolved injury, 6.78 (2.66–17.29; p<0.001) for persistent injury, and 22.28 (7.53–65.94; p<0.001) for de novo injury. Patients who developed both de novo WRF and de novo hepatic injury exhibited the highest mortality (25.0%, 7/28). Compared with patients without dynamic organ deterioration (2.20%, 37/1685), the adjusted OR was 12.48 (95% CI 4.81–32.39; p<0.001). Conclusions Dynamic renal and hepatic trajectories during AHF hospitalization identify graded risk phenotypes beyond static admission measurements. De novo multiorgan deterioration defines the highest-risk subgroup and may provide incremental value for early in-hospital risk stratification.