Résumé
ABSTRACT
Background
Cardiovascular diseases are the leading cause of death worldwide, particularly in older adults. While the Systematic Coronary Risk Evaluation 2‐Older Persons (SCORE2‐OP) model estimates 10‐year cardiovascular risk in this population, its validation in general European cohorts remains limited. Given growing relationships between sleep disturbances and cardiovascular risk, we aimed to validate SCORE2‐OP in a French cohort and assess the incremental value of excessive daytime sleepiness (EDS).
Methods
We included 4626 participants aged 70+ from the Three‐City cohort, without cardiovascular disease and dementia at baseline. The SCORE2‐OP model calibrated for low‐cardiovascular risk regions was used to estimate 10‐year cardiovascular risk. Calibration was assessed using observed‐to‐expected cumulative incidence ratios and calibration curves; discrimination using the area under the curve (AUC). Cox proportional hazards models examined associations between sleep symptoms (poor sleep quality, EDS, insomnia symptoms, sleep apnea [proxy]) and cardiovascular events. Incremental predictive value of significant symptoms was quantified by ΔAUC and net reclassification improvement (NRI).
Results
Over 10 years, the observed cumulative incidence of fatal and non‐fatal coronary heart disease or stroke was 10.55%; 95% confidence intervals (CI) = (9.62; 11.48), while the SCORE2‐OP predicted risk was 14.04%; 95% CI = (13.85; 14.24), yielding an observed‐to‐expected ratio of 0.75; 95% CI = (0.69; 0.80). Discrimination was moderate (AUC = 61.71%, 95% CI = [58.64; 64.78]). EDS was the only sleep symptom independently associated with cardiovascular events (adjusted hazard‐ratio = 1.32, 95% CI = [1.05; 1.65]). Adding EDS to SCORE2‐OP did not improve overall discrimination (ΔAUC = +0.72%, 95% CI = [−0.05; 1.50]) or reclassification (NRI = +1.33%, 95% CI = [−3.27; 5.93]). Sex‐stratified analyses showed significant improvement in discrimination in men (ΔAUC = +2.27%, 95% CI = [0.54; 4.00]), but not in women. Moreover, EDS improved reclassification in low (< 7.5%) and intermediate (7.5%–15%) cardiovascular risk groups (NRI = +12%, 95% CI = [3.56; 20.43] and NRI = +12.44%, 95% CI = [7.23; 17.65], respectively).
Conclusions
In this French cohort, SCORE2‐OP overestimated cardiovascular risk and showed moderate discrimination. EDS improved SCORE2‐OP performance in intermediate cardiovascular risk groups where treatments are uncertain, highlighting its clinical relevance; although implications for prevention strategies require further study.
External validation of SCORE2‐OP in older French adults and incremental value of excessive daytime sleepiness for cardiovascular risk prediction: SCORE2‐OP showed moderate discrimination and overestimated risk by 25%. Adding EDS did not improve overall model performance but enhanced classification in low‐ and intermediate‐risk groups.