Background: Cognitive impairment (CI) is common in patients with heart failure (HF) and impacts treatment adherence and other aspects of patient life in HF. Recognition of CI in patients with HF is therefore important. We aimed to develop a risk model with easily available patient characteristics, to identify patients with HF who are at high risk to be cognitively impaired and in need for further cognitive investigation.
Methods & results: The risk model was developed in 611 patients ≥ 60 years with HF from the TIME-CHF trial. Fifty-six (9 %) patients had CI (defined as Hodkinson Abbreviated Mental Test ≤ 7). We assessed the association between potential predictors and CI with least-absolute-shrinkage-and-selection-operator (LASSO) regression analysis. The selected predictors were: older age, female sex, NYHA class III or IV, Charlson comorbidity index ≥ 6, anemia, heart rate ≥ 70 bpm and systolic blood pressure ≥ 145 mmHg. A model that combined these variables had a c-statistic of 0.70 (0.63-0.78). The model was validated in 155 patients ≥ 60 years with HF from the ECHO study. In the validation cohort 51 (33 %) patients had CI (defined as a Mini Mental State Exam ≤ 24). External validation showed an AUC of 0.56 (0.46-0.66).
Conclusions: This risk model with easily available patient characteristics has poor predictive performance in external validation, which may be due to case-mix variation. These findings underscore the need for active screening and standardized assessment for CI in patients with HF.
Keywords: Cognitive impairment; Heart failure; Risk model.
© 2022 The Authors.