Predictive performance of SCORE2 among people with depression and anxiety disorder

BackgroundIt is not known how well SCORE2, a cardiovascular disease (CVD) risk score commonly used in Europe, performs among people with depression or anxiety disorder. MethodsAmong UK Biobank participants, prevalent depression and anxiety disorder were ascertained through self-report and hospital admission and primary care data. Prediction of 10-year CVD risk using SCORE2 was evaluated among those with and without depression/anxiety disorder by discrimination and calibration measures. Recalibration impact was assessed in terms of the distributions of individuals across predicted risk categories; and potentially preventable cases, estimated by dividing the additional number of individuals in the intermediate or higher predicted risk category by the number needed to treat (NNT) using statins. ResultsAmong 34,228 participants with depression/anxiety disorder, 3,076 (9.0%) had CVD events over a 10-year follow-up period. SCORE2 had a lower C-index (0.695, 95% CI 0.684-0.706) among those with depression/anxiety compared to those without (0.710, 95% CI 0.706-0.713). The calibration plot indicated that predicted risk was lower than observed (E/O ratio 0.782, 95% CI 0.746-0.818; calibration slope 0.899, 95% CI 0.868-0.930) among those with depression/anxiety disorder. After recalibration, 5,529 (38%) of those aged [&ge;]50 years and 1,464 (23%) of those aged <50 years with these mental health conditions were reclassified upward to the intermediate or higher risk category. Dividing these by NNT using statins yielded 191 (11.7%) and 51 (27.4%) additional preventable CVD events among those aged [&ge;]50 and <50 years, respectively. ConclusionsSCORE2 could underestimate 10-year CVD risk among people with depression or anxiety disorder, potentially leading to CVD events that could have been prevented with statin therapy.