Abstract

Background: Stress echocardiography (SE) is a noninvasive, functional method to assess presence of flow-limiting coronary artery disease (CAD). There is a prevailing consensus that core laboratory interpretations are more accurate than clinical site results. We sought to determine concordance and discordance between site and core laboratory interpretation of stress echocardiograms in this substudy of the PROMISE trial and identify factors that may influence disagreement.
Methods: Site and core lab readings for all 986 PROMISE patients with interpretable SE were compared for discordance (any rest or stress abnormality vs none). Patient, echocardiographic, and site-related factors affecting disagreement were examined and 90-day and one year outcomes were assessed.
Results: 66 (6.7%) of 986 studies had discordant results (kappa=0.46; 95% CI 0.35-0.57): 51 abnormal by site but normal by core lab and 15 normal by site and abnormal by core lab. Echo contrast was used in 199/986 studies (20.2%). There was less disagreement on rest wall motion interpretation(34 discordant; disagreement rate 3.4%) than stress (61 discordant studies; disagreement rate 6.2%). Of discordant stress studies, 53/61 were abnormal by site vs 8/61 abnormal by core lab. Predictors of discordance were [...]. Discordant SE results were associated with greater need for a second noninvasive test (OR=2.34 [1.13-4.82]) and referral to catheterization (OR=18.31 [9.55-35.12]).
Conclusion: In patients undergoing SE, disagreement in site vs core lab interpretations was infrequent and occurred with both rest and stress assessments. Discordance was most commonly due to an abnormal site read when normal by core lab, and led to additional noninvasive and invasive testing. Attention to identifiable patient and imaging characteristicsmay reduce interpretive variation and reduce downstream care.