Abstract 9515: Heterogeneity of Benefit on Quality of Life With an Invasive versus Conservative Strategy for Chronic Coronary Disease. Results From the ISCHEMIA Trial

Abstract

Background: The ISCHEMIA trial recently showed that, on average, patients with chronic coronary disease (CCD) and at least moderate ischemia randomized to an invasive strategy (INV) reported less angina and better quality of life (QOL) than those treated conservatively (CON)—an effect sustained through follow-up. To better support shared decision-making for individual patients, we explored the heterogeneity of treatment benefit of the INV vs CON strategy within the ISCHEMIA trial. Methods: Disease-specific QOL was assessed with the Seattle Angina Questionnaire (SAQ; scores range 0-100, higher=less angina/better QOL). An ordinal (proportional odds) model for 1-year SAQ summary score (SAQ SS) produced estimates of the odds of better QOL with randomization to INV vs CON, with potential heterogeneity of treatment benefit explored via interactions between patient factors and treatment. The final model included both significant interactions (accounting for multiple testing) and main effects. Results: Among 4617 patients with CCD from 37 countries randomized to INV vs CON treatment (mean age 64 ± 10 y, 23% women, 40% diabetes), mean SAQ SS was 74.1 ± 18.9 at baseline and 85.7 ± 15.6 at 1 year (p<0.001). Patients with lower baseline SAQ SS (i.e., more angina) and younger patients derived greater symptomatic treatment benefit from INV management (model-based interaction p=0.005 and p=0.004, respectively; Figure). No other patient factor was associated with a differential treatment effect. In the final model, female sex and diabetes were also associated with worse 1-year QOL, independent of treatment. Conclusion: Using data from a large multinational randomized trial, we found that younger patients and those with more angina at baseline derived greater QOL benefit from INV management. As the primary benefit of revascularization in CCD is symptom improvement, this model could be used to better inform individual patient’s decision-making for initial treatment.