Association between the prognostic nutritional index and cognitive impairment in patients with heart failure: a cross-sectional study

Background The prognostic nutritional index (PNI), calculated from serum albumin and peripheral lymphocyte count, is commonly interpreted as a nutritional–inflammatory marker. This study investigated the association between the PNI and cognitive impairment (CI) among patients with heart failure (HF). Methods This cross-sectional study included 153 patients with HF. Before cognitive status classification and statistical analysis, one point was added to the original Montreal Cognitive Assessment (MoCA) total score for participants with fewer than 12 years of education. Participants were classified into the CI group (education-corrected MoCA <26) or the non-CI group (education-corrected MoCA ≥26). Results CI was identified in 93 patients (60.8%). The median PNI was lower in the CI group than in the non-CI group [44.10 (40.55–47.00) vs. 47.08 (44.62–50.28), p < 0.001]. After adjusting for age, sex, and educational attainment, each five-point increase in the PNI was associated with lower odds of prevalent CI (odds ratio (OR): 0.562, 95% confidence interval (95% CI): 0.348–0.910; p = 0.019). The base clinical model had an area under the curve (AUC) of 0.868, and the PNI-extended model had an AUC of 0.882; the increment was small and non-significant (ΔAUC: 0.014, 95% CI: −0.008 to 0.036; DeLong p = 0.204). In a sensitivity analysis, albumin was independently associated with CI (OR: 0.362 per 5-g/L increase, 95% CI: 0.172–0.759; p = 0.007), and the PNI-extended model did not outperform the albumin-extended model (AUC: 0.882 vs. 0.887; ΔAUC: −0.005, 95% CI: −0.022 to 0.013; DeLong p = 0.588). Conclusion Lower PNI was independently associated with higher odds of prevalent CI in patients with HF. However, its incremental discrimination beyond age, sex, and educational attainment was small and statistically non-significant, and the PNI-extended model did not outperform the albumin-extended model in the sensitivity analysis. These findings support an association but do not establish incremental clinical utility; the decision-curve analysis is exploratory, and external validation is required.