Periprosthetic leakage occurs in 1 to 3.5% of patients after aortic valve replacement. Antecedent or active infection increases the risk of perivalvular leakage. Other factors predisposing to perivalvular leakage have not been clearly identified. To determine whether clinical and/or early transthoracic echocardiographic features can identify patients who require reoperation for periprosthetic leakage, we reviewed 1,743 consecutive patients who underwent aortic valve replacement from 1983 through 1991. Thirty patients (1.7%) subsequently required repeat operation for correction of periprosthetic leakage, at a median of 18.0 months (range 0.5-81.0) after aortic valve replacement. Comparison of these 30 patients with periprosthetic leakage to those undergoing primary aortic valve replacement showed no differences in patient age, gender, native valve disease type or etiology, prosthetic valve type, or prosthetic valve size (all p = NS). Of the 30 patients with periprosthetic leaks, 20 had early postoperative transthoracic echocardiography, at a median 5.3 months (range 0.3 - 40.0) after aortic valve replacement. In this early postoperative study, 14 patients (70.0%) had no periprosthetic leak detectable by echocardiography; two patients had trivial leaks, and only four patients had significant periprosthetic leakage by echocardiography. During subsequent operation, aortic valve re-replacement was performed in 26 patients (86.7%), and the periprosthetic leak was repaired in the remaining four patients. Thirty-day and 5-year survival probabilities were 0.90 (70% CI, 0.84-0.96) and 0.73 (70% CI, 0.65-0.82), respectively. It is concluded that in this study, there were no clinical features that distinguished patients undergoing aortic valve replacement who later developed severe perivalvular leakage.(ABSTRACT TRUNCATED AT 250 WORDS)