Reoperation is not an independent predictor of mortality during aortic valve surgery

Reoperations on aortic valves are associated with increased mortality, which may affect valve prosthesis selection at the time of initial aortic valve replacement. We analyzed our experience to determine whether reoperation itself independently predicts mortality during aortic valve surgery.

Demographic, intraoperative, and outcome data were collected prospectively on patients undergoing primary or redo aortic valve replacement or Bentall procedures after previous aortic valve replacement with or without concomitant coronary bypass grafting at a single institution from 1990 through 2002. Logistic regression analyses validated by means of bootstrap methodology identified the predictors of hospital mortality and the independent effect of reoperation.

Of 2673 patients undergoing aortic valve surgery, 2375 were primary operations, 216 were reoperations, and 82 were Bentall-after-aortic valve replacement procedures. Of 298 reoperations, 32 were third and 5 were fourth procedures. Mortality was 2.3% for primary operations, 4.6% for redo aortic valve replacement, and 2.4% for Bentall-after-aortic valve replacement procedures. Most patients underwent elective procedures, with mortalities of 1.6%, 1.7%, and 2.5%, respectively. Hospital mortality was independently predicted by peripheral vascular disease (odds ratio, 3.6), active endocarditis (odds ratio, 2.9), worsening New York Heart Association class (odds ratio, 2.3), and need for annular enlargement (odds ratio, 2.1). Reoperation itself did not predict hospital mortality.

The risk of mortality during aortic valve surgery is due mostly to active endocarditis, New York Heart Association class, and comorbidity. We failed to find a significant effect of reoperation on perioperative mortality. Mechanical valves, with their attendant anticoagulation-related morbidity, should not be implanted solely because of anticipated high mortality associated with bioprosthetic rereplacement.