PATHOPHYSIOLOGY AND NATURAL HISTORY ENDOCARDITIS Prosthetic valve endocarditis

Fifty-three (3.6%; actuarially 4.1% at 48 months) of 1465 consecutive in-hospital survivors of valve replacement from 1975 to July 1979 (aortic, mitral, or aortic and mitral, only one untraced) developed prosthetic valve endocarditis (PVE). Incremental risk factors for developing PVE were native valve endocarditis (p < .0001), black race (p = .0001), mechanical prosthesis (vs bioprosthesis) (p = .005), male sex (p = .04), and longer cardiopulmonary bypass time (p = .09). In general, the hazard function for developing PVE was greatest at 3 weeks after valve replacement. Patients with native valve endocarditis had a tendency to develop PVE early after valve replacement, as did patients in whom mechanical prostheses were used. PVE associated with Staphylococcus epidermidis tended to appear within 6 months of valve replacement, whereas streptococcal PVE tended to appear later after valve replacement. PVE took an atypical form in some patients, but patients with possible PVE (n = 6) had the same findings as those with certain PVE (n = 47). In 11 patients bacteriologic confirmation of PVE was not obtained. The typical prosthetic and periprosthetic characteristics ofPVE were present in 30 of the 40 cases in which observations were possible. PVE is a serious condition; 34 (64%) of our 53 patients died. Most deaths occurred within 3 months of the first evidence of PVE. Recovery of some patients is possible with appropriate medical and surgical treatment, but more intense preventive measures are indicated. Circulation 69, No. 2, 223-232, 1984 PROSTHETIC VALVE ENDOCARDITIS (PVE), a serious complication of cardiac valve replacement, is reported to occur in from 1% to 9.4% of patients. "-'9 In many studies the true incidence is probably underestimated because it is calculated from retrospective study of patients who have developed endocarditisl-12 or it is estimated by nonactuarial methods.'3"'Underestimation of incidence may also result from restrictive criteria for PVE, which omit patients in whom no positive cultures were obtained.' 10 16-18 The present study assesses the incidence and time course of PVE, describes the fate of patients with it, and identifies incremental risk factors for developing PVE and for dying from it. From these associations, improved understanding, prevention, and treatment may result. Materials and methods Study group. Primary aortic, mitral, or concomitant aortic and mitral valve replacement for valvular heart disease was From the Departments of Surgery and Medicine, The University of Alabama in Birmingham School of Medicine and Medical Center, and the Birmingham Veterans Administration Hospital, Birmingham. Address for correspondence: John W. Kirklin, M.D., University of Alabama in Birmingham, Department of Surgery, University Station, Birmingham, AL 35294. Received June 28, 1983; revision accepted Oct. 20, 1983. *Present address: the Thoracic Surgical Clinic, Karolinska Hospital, Stockholm, Sweden. Vol. 69, No. 2, February 1984 performed in 1533 patients from January 1, 1975, to July 1, 1979. The study group consisted of the 1465 hospital survivors, 818 of whom had undergone isolated aortic valve replacement, 452 of whom had undergone isolated mitral valve replacement, and 195 of whom had undergone concomitant aortic and mitral valve replacement (three of these also having tricuspid valve replacement). Valve replacement was accompanied by coronary artery bypass grafting in 385 patients and in 117 patients by other procedures such as repair of ascending aortic aneurysm, left ventricular aneurysmectomy, reconstruction of ventricular outflow obstruction, or tricuspid annuloplasty. Mechanical prostheses (Bjork-Shiley or Starr-Edwards) were used in 1036 patients, and bioprostheses (Hancock or Carpentier-Edwards types) were used in 429. All patients received antimicrobial agents for a period (4 to 10 days) after operation. Different protocols were used because of surgeon-to-surgeon differences, patients' drug allergies, and changing ideas as to optimal prophylaxis. Commonly used protocols included nafcillin alone for 2 or 3 postoperative days and oral cephalexin for 4 to 7 more days, nafcillin followed by oral cephradine or cloxacillin, and intravenous cephapirin as the only drug. Dicumarol was started on the second postoperative day and use continued for only 6 to 12 weeks in most patients with bioprostheses. However, dicumarol was prescribed in a lifelong regimen in all patients with mechanical prostheses. Definitions. Each patient was classified as having certain PVE, possible PVE, or no PVE during the follow-up period. PVE was classified as certain in patients (1) with an illness clinically consistent with infectious endocarditis, positive blood cultures for the same organism in at least two samples, and no apparent source of bacteremia other than the heart; (2) with positive blood cultures as above, and with typical or suggestive findings (see below for definition) at repeat operation or autopsy; (3) with organisms cultured or stained from prosthetic or 223 by gest on A ril 6, 2017 http://ciajournals.org/ D ow nladed from