Jianchao Liu, Hongtao Cai, Junbo Feng
Background Patients hospitalized with acute decompensated heart failure (ADHF) remain at high risk for early readmission and death after discharge. Static biomarkers may not fully capture the evolving inflammatory, nutritional, and renal status during hospitalization. We aimed to identify in-hospital inflammatory-nutritional-renal trajectory phenotypes and assess their associations with 90-day post-discharge outcomes. Methods This single-center retrospective cohort study included 633 adults hospitalized with ADHF between January 2018 and December 2025 who survived to discharge and had serial measurements of neutrophil-to-lymphocyte ratio (NLR), serum albumin, and estimated glomerular filtration rate (eGFR) at admission, 48–72 h after admission, and before discharge. Multivariate latent class mixed modelling was used to identify trajectory phenotypes. The primary outcome was all-cause readmission or all-cause mortality within 90 days after discharge. Associations were evaluated using multivariable logistic regression, with supplementary time-to-event analyses. Results Overall, 172 patients (27.2%) experienced the primary outcome. Three phenotypes were identified: Phenotype 1, low inflammation-stable nutrition-stable renal function ( n = 348, 55.0%); Phenotype 2, moderate inflammation-declining albumin-stable renal function ( n = 196, 31.0%); and Phenotype 3, high inflammation-low albumin-worsening renal function ( n = 89, 14.0%). Event rates were 14.9%, 33.2%, and 61.8%, respectively. In the fully adjusted model, Phenotype 2 (odds ratio [OR], 2.12; 95% confidence interval [CI], 1.34–3.35) and Phenotype 3 (OR, 4.85; 95% CI, 2.58–9.12) were associated with the primary outcome. Additional adjustment for baseline NLR, albumin, and eGFR attenuated but did not eliminate these associations. The area under the receiver operating characteristic curve increased from 0.732 for the clinical model to 0.754 after adding baseline biomarkers and to 0.768 after further adding trajectory phenotype. Conclusions In-hospital inflammatory-nutritional-renal trajectory phenotypes were associated with graded 90-day risks after ADHF discharge and provided limited additional prognostic information beyond baseline clinical and biomarker data. Prospective external validation is required before clinical implementation.