Abstract
BACKGROUND: Chest pain is responsible for millions of visits to the emergency department (ED) annually. Cardiac ultrasound can detect ischemic changes, but varying accuracy estimates have been reported in previous studies. We synthetized the available evidence to yield more precise estimates of the accuracy of cardiac ultrasound for acute myocardial ischemia in patients with chest pain in the ED and to assess the effect of different clinical characteristics on test accuracy. METHODS: A systematic search for studies assessing the diagnostic accuracy of cardiac ultrasound for myocardial ischemia in the ED was conducted in MEDLINE, EMBASE, CENTRAL, CINAHL, LILACS, Web of Science, two trial registries and supplementary methods, from inception to December 6th, 2022. Prospective cohort, cross-sectional, case-control studies and randomized controlled trials (RCTs) that included data on diagnostic accuracy were included. Risk of bias was assessed with the QUADAS-2 tool and a bivariate hierarchical model was used for meta-analysis with paired Forest and SROC plots used to present the results. Subgroup analyses was conducted on clinically relevant factors. RESULTS: Twenty-nine studies were included, with 5043 patients. The overall summary sensitivity was 79.3% (95%CI 69.0-86.8%) and specificity was 87.3% (95%CI 79.9-92.2%), with substantial heterogeneity. Subgroup analyses showed increased sensitivity in studies where ultrasound was conducted at ED admission and increased specificity in studies that excluded patients with previous heart disease, when the target condition was acute coronary syndrome, or when final chart review was used as the reference standard. There was very low certainty in the results based on serious risk of bias and indirectness in most studies. CONCLUSIONS: Cardiac ultrasound may have a potential role in the diagnostic pathway of myocardial ischemia in the ED; however, a pooled accuracy must be interpreted cautiously given substantial heterogeneity and that important patient and test characteristics affect its diagnostic performance. PROTOCOL REGISTRATION: PROSPERO (CRD42023392058).