CARDIO-PAL score predicts 6-month mortality in heart failure inpatients with validation-cohort AUC 0.80 and 54% mortality in the high-risk group
Synopsis
Using heart failure inpatients admitted to St. Boniface Hospital in Winnipeg, Canada, between 2019 and 2024 and randomly split 80%/20% into development (n = 2036) and validation (n = 508) cohorts, the authors applied penalized LASSO regression and a multivariable logistic model to derive the CARDIO-PAL score, which in the holdout validation cohort achieved discrimination with an AUC of 0.80 (95% CI 0.75–0.84) and calibration slope 0.84 with intercept -0.04, with 6-month mortality of 2%, 15%, and 54% across low-, intermediate-, and high-risk groups.
Interpretation
The CARDIO-PAL score predicts 6-month mortality in heart failure inpatients, with a validation-cohort AUC of 0.80 (95% CI 0.75–0.84) and calibration slope 0.84 with intercept -0.04. Available prognostic tools commonly predict survival beyond 1 year, whereas this work moves the prediction window to 6 months, addressing short-term prognosis. Single-center retrospective cohort with a development cohort of n = 2036 and a validation cohort of n = 508, using holdout validation to assess discrimination and calibration.
Eight candidate variables were retained: age (aOR 1.01), PPSv2 (aOR 2.08), systolic blood pressure (aOR 1.14), discontinued ACE inhibitor or angiotensin II receptor blocker (aOR 2.58), beta-blocker (aOR 1.59), mineralocorticoid receptor antagonist (aOR 2.67), right ventricular systolic pressure (aOR 2.06), and urea (aOR 2.13). The variable set spans functional status (PPSv2), imaging parameters, laboratory values, and medication use rather than relying on a single dimension. Variables were selected by penalized LASSO regression in the development cohort and then entered a multivariable logistic regression, with adjusted odds ratios reported.
Six-month mortality across low-, intermediate-, and high-risk groups was 2%, 15%, and 54%, respectively. This stratification translates the score into communicable absolute-risk bands that can support goals-of-care discussions and palliative care referral. Group mortality came from score-based risk groups in the validation cohort, while 6-month mortality was 29% and 31% in the development and validation cohorts.
Perspective
The score targets heart failure inpatients for short-term 6-month mortality prognostication, in settings where goals-of-care discussions and palliative care referral decisions are made; the authors propose using it collaboratively with other prognostic scores and serial clinical assessments rather than as a replacement for clinical judgment.
Only the abstract is available here, so the specific point values for each variable, risk-group cutoffs, handling of missing data, and external validation remain open questions that affect whether the score can be used directly.