RESULTS
A total of 1,502 patients who underwent isolated surgical aortic valve replacement with a bioprosthetic valve were identified, 1,090 (72.6%) of whom received a bovine prosthesis and 412 (27.4%) of whom received a porcine prosthesis. Patients who received bovine valves were significantly more likely to have chronic lung disease and peripheral vascular disease when compared to patients who received porcine valves (Table 1). Those who received bovine valves also had significantly higher STS Predicted Risk of Mortality scores than those who received porcine valves. Rate of elective vs urgent vs emergent surgical status differed significantly between the two groups. Implanted valve size was significantly smaller in the bovine group than in the porcine group.
To account for differences in baseline and operative characteristics, propensity score matching was utilized to create similar groups. Matching yielded 412 risk-adjusted pairs, with an adequate balance indicated by SMDs <0.1 for 23 out of the 24 variables that were matched on (Table 2).
Postoperative outcomes of the matched study population are reported in Table 3. There were no significant differences in operative mortality, length of stay, blood product transfusion, reoperation for bleeding, duration of intubation, pneumonia, stroke, renal failure requiring dialysis, 30-day readmission rates, or aortic valve reintervention rates in the matched cohorts (p>0.05). The mean interval from surgery to the time of postoperative echocardiogram was 0.69±1.47 years in the bovine group and 0.58±1.33 years in the porcine group. There were no significant differences in postoperative mean aortic valve gradient or ejection fraction in the matched cohorts.
Kaplan–Meier estimates demonstrated comparable survival in propensity-matched cohorts of patients who received bovine versus porcine aortic valves (Figure 1, p=0.99, log-rank). On multivariable Cox proportional-hazards regression, valve type was not significantly associated with long-term mortality (HR 1.02, 95% CI: 0.74, 1.40, p=0.924, Table 4). Variables which were significantly associated with long-term mortality after bioprosthetic SAVR included age, chronic dialysis, history of cerebrovascular accident, prior cardiovascular intervention, atrial fibrillation, and urgent surgical status (Table 4).
There were no significant differences in competing-risk cumulative incidence estimates for all-cause readmissions (p=0.68, Figure 2) or aortic valve reinterventions (p=0.25, Figure 3) in the matched cohorts.