Albuminuria and polyvascular disease improve multivariate predictive models after an acute cardiovascular event. The AIRVAG cohort

Abstract

Background and aims: There is no consensus regarding risk stratification tools for secondary prevention in atherosclerotic cardiovascular disease. Our aim was to compare the discriminative performance of the Framingham, REGICOR, SCORE, and REACH risk functions and the Bohula-TIMI and SMART risk scores, as well as to assess the potential added value of other clinical variables for the prediction of recurrent events in patients with established vascular disease.

Methods: A cohort of 269 patients with established vascular disease (52.8% coronary, 32% cerebrovascular, 15.2% PAD) was included. The survival functions of risk groups (low/medium/high) according to commonly used cutoff points for each function/score were compared, and hazard ratios for each were estimated using Cox regression. We calculated Δ Harrell's C statistic, cat-NRI, and cNRI after adding new predictors to a base model including age, sex, total cholesterol, current smoking status, hypertension, and diabetes.

Results: After six years of follow-up (median 4.82 years), 61 events occurred (23%). High-risk groups had a higher risk of recurrent event: SMART (HR: 3.17 [1.55-6.5]), Framingham (HR: 3.08 [1.65-5.75]), REGICOR (HR: 2.71 [1.39-5.27]), SCORE (HR: 2.14 [1.01-4.5], REACH (HR: 5.74 [2.83-11.7]), B-TIMI (HR: 3.68 [0.88-15.3]). Polyvascular disease (three territories HR: 5.6 [2.2-14.25]), albuminuria (HR: 3.55 [2.06-6.11]), and heart failure (HR: 3.11 [1.34-7.25]) also increased risk. Discrimination (Harrell's C) was low but improved after adding albuminuria and polyvascular disease. Both variables also improved the performance of the base model (cNRI.326 [.036; .607]).

Conclusions: The Framingham, REGICOR, SCORE, and REACH functions and the B-TIMI and SMART scores showed low yet similar performance in secondary prevention. Albuminuria and polyvascular disease improved the predictive performance of major classical cardiovascular risk factors.

Keywords: Análisis de supervivencia; Cardiovascular risk; Cohort studies; Estudios de cohortes; Modelos de riesgos proporcionales; Prevención secundaria; Proportional hazard models; Riesgo cardiovascular; Secondary prevention; Survival analysis.