Pre-test probability of obstructive coronary artery disease in the new guidelines: too much, too little or just enough?

Abstract

Abstract Background Previous 2013 ESC guidelines recommended the use of the Modified Diamond-Forrester method to assess the pre-test probability (PTP) of obstructive coronary artery disease (CAD). The 2019 ESC Chronic Coronary Syndrome guidelines updated this recommendation with a major downgrade in PTP. The aim of this study was to compare the performance of these two methods in patients with stable chest pain undergoing coronary computed tomography angiography (CCTA) for suspected CAD. Methods We performed a retrospective analysis on prospectively collected data from a cohort of consecutive patients undergoing CCTA for suspected CAD from October 2016 to 2019. Key exclusion criteria were age <30 years-old, known CAD, suspected acute coronary syndrome or symptoms other than chest pain. Obstructive CAD was defined as any luminal stenosis ≥50% on CCTA. Whenever invasive coronary angiography (ICA) was subsequently performed, patients were reclassified if luminal stenosis was <50%. The two PTP prediction models were assessed for calibration and discrimination. Results A total of 320 patients (median age 63 years [IQR 53–70], 59% women) were included. Chest pain characteristics were: 48% atypical angina, 38% non-anginal chest pain, 14% typical angina. The observed prevalence of obstructive CAD was 16.3% (n=52). Patients with obstructive CAD were more often male, were significantly older and had a higher prevalence of typical angina and cardiovascular risk factors (except for family history of CAD). On average, individual PTP was 22.1% lower in the new guidelines. The 2013 prediction model significantly overestimated the likelihood of obstructive CAD (mean PTP 37.3% vs 16.3%; relative overestimation of 130%, p-value for miscalibration 0.005). The updated 2019 method showed good calibration for predicting the likelihood of obstructive CAD (mean PTP 15.2% vs 16.3%; relative underestimation of 6.5%, p-value for miscalibration 0.712). The two approaches showed similar discriminative power, with a C-statistics of 0.730 and 0.735 for the 2013 and 2019 methods, respectively (p-value for comparison 0.933). Stratification by gender produced similar results. Conclusions In patients with stable chest pain undergoing CCTA, the updated 2019 prediction model allows for a more precise estimation of pre-test probabilities of obstructive CAD than the previous model. Adoption of this new score may improve disease prediction and change the downstream diagnostic pathway in a significant proportion of cases. Graph 1 Funding Acknowledgement Type of funding source: None