Serum uric acid significantly improves the accuracy of cardiovascular risk score models

Abstract

Abstract Aims This study evaluated the impact of serum uric acid (sUA) on the accuracy of pooled cohort equations (PCE) model, Systematic COronary Risk Evaluation 2 (SCORE2), and SCORE2-older persons. Methods and results We evaluated 19 769 asymptomatic self-referred adults aged 40–79 years free of cardiovascular disease and diabetes who were screened annually in a preventive healthcare setting. sUA levels were expressed as a continuous as well as a dichotomous variable (upper sex-specific tertiles defined as high sUA). The primary endpoint was the composite of death, acute coronary syndrome, or stroke, after excluding subjects diagnosed with metastatic cancer during follow-up. Mean age was 50 ± 8 years and 69% were men. During the median follow-up of 6 years, 1658 (8%) subjects reached the study endpoint. PCE, SCORE2, and high sUA were independently associated with the study endpoint in a multivariable model (P < 0.001 for all). Continuous net reclassification improvement analysis showed a 13% improvement in the accuracy of classification when high sUA was added to either PCE or SCORE2 model (P < 0.001 for both). sUA remained independently associated with the study endpoint among normal-weight subjects in the SCORE2 model (HR 1.3, 95% CI 1.1–1.6) but not among overweight individuals (P for interaction = 0.01). Subgroup analysis resulted in a significant 16–20% improvement in the model performance among normal-weight and low-risk subjects (P < 0.001 for PCE; P = 0.026 and P < 0.001 for SCORE2, respectively). Conclusion sUA significantly improves the classification accuracy of PCE and SCORE2 models. This effect is especially pronounced among normal-weight and low-risk subjects.