Clinical risk prediction, coronary computed tomography angiography, and cardiovascular events in new-onset chest pain: the PROMISE and SCOT-HEART trials

BACKGROUND AND AIMS: Whether index testing using coronary computed tomography angiography (CTA) improves outcomes in stable chest pain is debated. The risk factor weighted clinical likelihood (RF-CL) model provides likelihood estimation of obstructive coronary artery disease. This study investigated the prognostic effect of coronary CTA vs. usual care by RF-CL estimates. METHODS: Large-scale studies randomized patients (N = 13 748) with stable chest pain to coronary CTA as part of the initial work-up in addition to or instead of usual care including functional testing. Patients were stratified according to RF-CL estimates [RF-CL: very-low (≤5%), low (>5%-15%), and moderate/high (>15%)]. The primary endpoint was myocardial infarction or death at 3 years. RESULTS: The primary endpoint occurred in 313 (2.3%) patients. Event rates were similar in patients allocated to coronary CTA vs. usual care [risk difference (RD) 0.3%, hazard ratio (HR) 0.84 (95% CI 0.67-1.05)]. Overall, 33%, 44%, and 23% patients had very-low, low, and moderate/high RF-CL. Risk was similar in patients with very low and moderate/high RF-CL allocated to coronary CTA vs. usual care [very low: RD 0.3%, HR 1.27 (0.74-2.16); moderate/high: RD 0.5%, HR 0.88 (0.63-1.23)]. Conversely, patients with low RF-CL undergoing coronary CTA had lower event rates [RD 0.7%, HR 0.67 (95% CI 0.47-0.97)]. The number needed to test using coronary CTA to prevent one event within 3 years was 143. CONCLUSIONS: Despite an overall good prognosis, low RF-CL patients have reduced risk of myocardial infarction or death when allocated to coronary CTA vs. usual care. Risk is similar in patients with very-low and moderate/high likelihood.