Background: This study was aimed at defining clinical and anatomic patterns in cases of surgical endocarditis (SE). Methods: SE cases done between 1981 and 1997 at our metropolitan county hospital were retrospectively analyzed. Results: A total of 106 consecutive episodes of SE involving 125 valves in 100 patients were studied. SE included 71 aortic, 42 mitral, and 12 tricuspid valves. The etiologies included intravenous drug abuse (IVDA) in 48 (45%) and dental source in 30 (28%). A congenitally deformed valve was present in 19 (18%). Compared to non-IVDA, IVDA episodes of SE were more often superimposed on previously normal valves (38/48 [79%] vs. 30/58 [52%])**, S. aureus infections (17/43 [40%] vs. 9/54 [17%])*, active endocarditis (38/48 [79%] vs. 32/58 [55%])*, and surgically treated on an urgent basis (10/48 [21%] vs. 4/58 [7%])*. Overall, macroemboli occurred in 53 (50%) of SE and was associated with pseudoaneurysm*, preoperative neurologic dysfunction,** and operative death.** The operative mortality (defined by Society of Thoracic Surgeons) for SE was 5/106 (4.7%). Macroembolism,** aortoventricular discontinuity,** abscesses,* pseudoaneurysm,** and preoperative renal failure* were associated with mortality. Prosthetic valve endocarditis was present in 10 of 106 episodes of SE (9.4%). *p ≤ 0.05; **p ≤ 0.01. Conclusion: (1) The aortic valve is most commonly associated with SE, (2) SE of a previously normal valve is more likely to occur with IVDA than other etiologies, (3) macroemboli occur in half of SE and is associated with an increased operative mortality.