Objective.We performed a systematic review of the literature to describe current knowledge of cardiovascular (CV) risk prediction algorithms in rheumatic diseases.Methods.A systematic search of MEDLINE, EMBASE, and Cochrane Central databases was performed. The search was restricted to original publications in English, had to include clinical CV events as study outcomes, assess the predictive properties of at least 1 CV risk prediction algorithm, and include patients with rheumatoid arthritis (RA), ankylosing spondylitis (AS), systemic lupus erythematosus (SLE), psoriatic arthritis (PsA), or psoriasis. By design, only cohort studies that followed participants for CV events were selected.Results.Eleven of 146 identified manuscripts were included. Studies evaluated the predictive performance of the Framingham Risk Score, QRISK2, Systematic Coronary Risk Evaluation (SCORE), Reynolds Risk Score, American College of Cardiology/American Heart Association Pooled Cohort Equations (PCE), Expanded Cardiovascular Risk Prediction Score for Rheumatoid Arthritis (ERS-RA), and the Italian Progetto CUORE score. Approaches to improve predictive performance of general risk algorithms in patients with RA included the use of multipliers, biomarkers, disease-specific variables, or a combination of these to modify or develop an algorithm. In both SLE and PsA patients, multipliers were applied to general risk algorithms. In studies of RA and SLE patients, efforts to include nontraditional risk factors, disease-related variables, multipliers, and biomarkers largely failed to substantially improve risk estimates.Conclusion.Our study confirmed that general risk algorithms mostly underestimate and at times overestimate CV risk in rheumatic patients. We did not find studies that evaluated models for psoriasis or AS, which further demonstrates a need for research in these populations.
Effect of competing mortality risks on predictive performance of the QRISK3 cardiovascular risk prediction tool in older people and those with comorbidity: external validation population cohort study
