Intraoperative epicardial echocardiography or transoesophageal echocardiography in CHD: how much does it matter?

Intraoperative imaging determines the integrity of surgical repairs. Transoesophageal echocardiography represents standard care for intraoperative imaging in CHD. However, some conditions preclude its use, and epicardial echocardiography is used alternatively. Minimal literature exists on the impact of epicardial echocardiography versus transoesophageal echocardiography. We aimed to evaluate accuracy between the two modalities and hypothesised higher imaging error rates for epicardial echocardiography.We retrospectively reviewed all epicardial echocardiograms performed over 16 years and compared them to an age- and procedure-matched, randomly selected transoesophageal echocardiography cohort. We detected un- or misidentified cardiac lesions during the intraoperative imaging and evaluated patient outcomes. Data are presented as a median with a range, or a number with percentages, with comparisons by Wilcoxon two-sample test and Fisher's exact test.Totally, 413 patients comprised the epicardial echocardiography group with 295 transoesophageal echocardiography matches. Rates of imaging discrepancies, re-operation, and incision infection were similar. About 13% of epicardial echocardiography patients had imaging discrepancies versus 16% for transoesophageal (p = 0.2352), the former also had smaller body sizes (p < 0.0001) and more genetic abnormalities (33% versus 19%, p < 0.0001). Death/mechanical support occurred more frequently in epicardial echocardiography patients (16% versus 6%, p < 0.0001), while hospitalisations were longer (25 versus 19 days, p = 0.0003).Diagnostic accuracy was similar between patients undergoing epicardial echocardiography and transoesophageal echocardiography, while rates of death and mechanical support were increased in this inherently higher risk patient population. Epicardial echocardiography provides a reasonable alternative when transoesophageal echocardiography is not feasible.