Event Status of the Starr-Edwards Aortic Valve to 20 Years: A Benchmark for Comparison

Background. Considerable effort and expense has been invested in the evolutionary development of cardiac valvular prostheses with the goal of reducing or minimizing specific events related to these prostheses. It is important to measure any improvement achieved with such development by comparison against a historic standard. The Starr-Edwards caged-ball prosthesis model 1260 has been used for 30 years as the predominant or sole model of its kind for aortic valve replacement. This historic opportunity provides a benchmark for subsequent improvement and comparison of current prostheses. Methods. Between 1969 and 1991, 1,100 patients (median age, 57 years; 838 men and 194 women) underwent aortic valve replacement with or without coronary artery bypass grafting (aortic valve replacement, 964; aortic valve replacement plus coronary artery bypass grafting, 136) with the 1260 Starr-Edwards caged-ball prosthesis. Results. Operative mortality was 6.2% (68 patients). Univariate patient characteristics predictive of early mortality were female sex (p = 0.003), age (>56 years; p = 0.002), recent operative interval (1985 to 1991 versus 1969 to 1976 or 1977 to 1984; p = 0.002), presence of atrial fibrillation (p = 0.001), and small valve size (7A to 8A = 19 to 21 mm; p < 0.001). Follow-up extended to 11,293 patient-years (mean, 24.8 years) and was 96.9% complete. Survival at 5, 10, 15, and 20 years for all patients including operative mortality was 76.6%, 59.6%, 44.9%, and 31.2%, respectively. Operative variables predictive of poor late survival were advanced New York Heart Association class (III or IV); (p = 0.0001), older age (>56 years; p = 0.0001), and lower (<0.56) ejection fraction (p = 0.0001). Freedom from thromboemboli and anticoagulant-related bleeding at 5 years was 90.8% and 98.7%, respectively. Univariate model for greater risk of late thromboemboli identified female sex (p = 0.04), older age (>56 years; p = 0.0002), and New York Heart Association class III or IV (p = 0.0058), as risk factors. Multivariate analysis for thromboemboli demonstrated older age (p = 0.0007) and New York Heart Association class III or IV (p = 0.0041) as significant. Alternatively, univariate analysis for late bleeding found only the most recent operative interval (p = 0.009) as significant, and the rarity of events prevented a multivariate query. There were no valve failures. Conclusions. The late results of survival and freedom from late anticoagulant-related bleeding or thromboemboli are excellent, especially in larger (9A and above) sizes, and with the long implant record comparable with more recent prostheses, the Starr-Edwards valve provides an excellent, safe, and durable alternative in the aortic position and provides a benchmark against which to compare other prostheses.