Admission BNP as a Dominant Prognostic Signal in Acute Decompensated Heart Failure with Frequent Renal Dysfunction: An Exploratory Multivariable Analysis

Background: Risk stratification for acute decompensated heart failure (ADHF) at admission remains clinically challenging, particularly in patients with concomitant renal dysfunction. Although B-type natriuretic peptide (BNP) is a well-established biomarker in ADHF, its prognostic dominance in prespecified multivariable analyses of cohorts with frequent cardiorenal dysfunction remains incompletely characterized. Methods: We conducted a single-center retrospective cohort study of 220 index admissions for ADHF. A prespecified multivariable logistic regression model was developed to estimate in-hospital mortality using admission variables, including age, sex, heart failure phenotype, systolic blood pressure, estimated glomerular filtration rate, serum albumin, and log-transformed BNP. Model performance was evaluated using discrimination (AUC), bootstrap optimism correction (200 iterations), calibration assessment, and decision curve analysis, in accordance with the TRIPOD guidelines. Results: In-hospital mortality occurred in 7.7% of patients (n = 17). In multivariable analysis, ln(BNP) was the only variable that reached statistical significance as an independent predictor of in-hospital mortality (OR 2.39 per unit increase; 95% CI 1.25–4.59; p = 0.009), representing the dominant prognostic signal in this dataset. Albumin and eGFR showed consistent but non-significant associations. The exploratory model demonstrated apparent discrimination (AUC 0.81), which decreased to 0.73 after optimism correction. A prespecified simplified model including albumin and ln(BNP) yielded an AUC of 0.77. Conclusions: In this exploratory multivariable analysis, admission BNP emerged as the primary prognostic factor for in-hospital mortality in an ADHF cohort with a high prevalence of renal dysfunction. The small number of outcome events limits confidence in model stability and precludes clinical deployment. These findings are hypothesis-generating and support BNP as the strongest prognostic signal observed in this cohort. External validation in larger cohorts is required before any clinical application.