BACKGROUND: Cardiac conduction abnormalities occur in endocarditis and have been associated with infection extension and increased mortality. There have been no prospective studies of electrocardiographic (ECG) conduction changes in endocarditis. We examined the incidence of ECG changes in a large prospective cohort with suspected endocarditis and correlated changes with echocardiographic evidence of invasive infection and mortality. METHODS: One hundred thirty-seven of 1396 (10%) suspected cases of endocarditis were classified as "definite" or "possible" by the Duke criteria and had an interpretable ECG. ECG conduction changes were classified as old (pre-existing hospitalization), new (evident on admission or developed during hospitalization), or indeterminate. New or indeterminate abnormalities were considered "ECG conduction changes." Echocardiogram results were reviewed to identify infected valves and invasive infection. RESULTS: ECG conduction changes were present in 36 of 137 (26%) patients. Patients with ECG conduction changes were more often male (69% vs 46%, P =.005) and had prosthetic valves (47% vs 23%, P <.001). There were no significant differences in microbiology results or treatment with cardiac surgery. In 76 (55%) patients, at least one infected valve was identified by echocardiography; 15 of 76 (20%) patients were determined to have evidence of invasive infection. Eight of 15 (53%) invasive infections exhibited ECG conduction changes compared with 16 of 61 (26%) isolated valve infections (P =.046). Eleven of 36 (31%) patients with ECG conduction changes died during hospitalization compared with 15 of 101 (15%) patients without changes (P =.039). CONCLUSIONS: ECG conduction changes commonly occur in endocarditis despite more sensitive diagnostic criteria and are associated with increased mortality and invasive infection.
BACKGROUND:Cardiac conduction abnormalities occur in endocarditis and have been associated with infection extension and increased mortality. There have been no prospective studies of electrocardiographic (ECG) conduction changes in endocarditis. We examined the incidence of ECG changes in a large prospective cohort with suspected endocarditis and correlated changes with echocardiographic evidence of invasive infection and mortality. METHODS: One hundred thirty-seven of 1396 (10%) suspected cases of endocarditis were classified as "definite" or "possible" by the Duke criteria and had an interpretable ECG. ECG conduction changes were classified as old (pre-existing hospitalization), new (evident on admission or developed during hospitalization), or indeterminate. New or indeterminate abnormalities were considered "ECG conduction changes." Echocardiogram results were reviewed to identify infected valves and invasive infection. RESULTS: ECG conduction changes were present in 36 of 137 (26%) patients. Patients with ECG conduction changes were more often male (69% vs 46%, P =.005) and had prosthetic valves (47% vs 23%, P <.001). There were no significant differences in microbiology results or treatment with cardiac surgery. In 76 (55%) patients, at least one infected valve was identified by echocardiography; 15 of 76 (20%) patients were determined to have evidence of invasive infection. Eight of 15 (53%) invasive infections exhibited ECG conduction changes compared with 16 of 61 (26%) isolated valve infections (P =.046). Eleven of 36 (31%) patients with ECG conduction changes died during hospitalization compared with 15 of 101 (15%) patients without changes (P =.039). CONCLUSIONS: ECG conduction changes commonly occur in endocarditis despite more sensitive diagnostic criteria and are associated with increased mortality and invasive infection.