New approaches to the diagnosis and treatment of infective endocarditis: review of 100 consecutive cases.

To assess the effect of our new, more aggressive approach to treating infective endocarditis, we retrospectively reviewed our recent experience with this disease. Between 1983 and 1989, we treated 100 patients with endocarditis, in 94 of whom the diagnosis was confirmed. Fifty-four (57%) of the 94 patients had native valve endocarditis, and the other 40 patients (43%) had prosthetic valve endocarditis. The patients' mean ages were 50 years for native valve disease and 58 years for prosthetic valve disease (p < 0.05). The male-to-female ratio was 4:1. Among the patients with confirmed endocarditis, 87 (93%) had significant underlying risk factors for endocarditis. Upon physical examination of the 94 patients, 16 (17%) were afebrile, 15 (16%) had negative blood cultures, and 26 (28%) had no cardiovascular symptoms or immunologic findings. Echocardiography was of limited value: its sensitivity was 56% for native valve endocarditis and 33% for prosthetic valve endocarditis. The ratio of affected valves was 5 aortic:4 mitral: 1 tricuspid in both forms of the disease. Viridans streptococcus was isolated in 23 (25%) of the confirmed cases, Enterococcus faecalis in 17 (18%), Staphylococcus aureus in 13 (14%), and coagulase-negative staphylococcus in 14 (15%). Gram-negative, anaerobic, and fungal organisms accounted for only 13 (14%) of the confirmed cases. The mean duration of intravenous therapy was 29 days. Twenty (37%) of the native valve patients and 16 (40%) of the prosthetic valve patients received antibiotics on an outpatient basis. Vancomycin was used in 44 (47%) of the cases, nafcillin or ampicillin in 40 (44%), and other beta-lactam agents in 9 (10%). The mean hospital stay was significantly longer for prosthetic valve endocarditis patients than for those with native valve disease (29 versus 23 days; p < 0.01). Cardiac catheterization was performed in 9 native valve patients (17%) and 6 prosthetic valve patients (15%). Valve surgery was performed in 33 native valve patients (61%) and 22 prosthetic valve patients (55%). Failure, defined as in-hospital death or recurrent endocarditis, accounted for 14 (28%) of the native valve cases (17% death and 11% relapse) and 8 (20%) of the prosthetic valve cases (10% death and 10% relapse), for an overall failure rate of 24%. The rate of failure was independent of the infecting pathogen or the type of antimicrobial therapy applied. Our experience verified that, in many patients with significant underlying risk factors, the diagnosis of endocarditis may be made on an empiric basis.