Development and validation of a nomogram for predicting mild cognitive impairment in older patients with chronic heart failure

Chronic heart failure (CHF) and mild cognitive impairment (MCI) are reciprocally associated, yet the determinants that drive cognitive decline in CHF remain incompletely characterized and no bedside-ready prediction tool exists. We quantified the burden of MCI in older adults with CHF, identified its independent correlates and constructed and validated a nomogram for early detection. A total of 375 CHF patients aged ≥ 65 years were consecutively enrolled as the training cohort and dichotomized into MCI ( n = 179) or non-MCI ( n = 196) groups according to comprehensive neuropsychological assessments. Candidate predictors encompassed sociodemographics, clinical indices, biomarkers, lifestyle and psychosocial domains. Multivariable logistic regression with forward conditional was used to select independent predictors. A nomogram was built with the rms package in R. External validation was performed in 135 contemporaneous CHF patients from a second centre. Discrimination (area under the receiver-operating-characteristic curve, AUC), calibration (Hosmer-Lemeshow test), clinical utility (decision-curve analysis) and optimal probability cut-offs (Youden index) were assessed. Eight variables were retained: years of education, diabetes mellitus, NYHA class, left-ventricular ejection fraction (LVEF), NT-proBNP, high-sensitivity cardiac troponin, depressive symptoms, and sleep disorders (all P ≤ 0.01). The nomogram demonstrated strong discrimination in both the training (AUC 0.837, 95% CI 0.796-0.878; sensitivity 78.9%, specificity 76.6%; Youden index 0.555) and validation (AUC 0.788, 95% CI 0.709-0.867; sensitivity 78.3%, specificity 74.7%; Youden index 0.530) sets. Calibration plots aligned well with the forty-five degrees line (Hosmer-Lemeshow P > 0.08 for both). Decision-curve analysis indicated positive net benefit across threshold probabilities of 1-80% (training) and 1-78% (validation). The proposed nomogram provides a well-calibrated, internally and externally validated bedside tool for estimating MCI risk in older CHF patients, enabling clinicians to stratify care and initiate timely multidomain interventions that may help delay cognitive decline and enhance quality of life.