Preoperative risk evaluation and stratification of long-term survival after valve replacement for aortic stenosis. Reasons for earlier operative intervention.

Abstract

Predictability of prognosis was analyzed in 630 patients who were alive 30 days after valve replacement (1965-1986) for aortic stenosis. Follow-up totaled 4,072 patient-years. During the operative periods of 1965-1971 (n = 62), 1972-1976 (n = 164), 1977-1981 (n = 158), and 1982-1986 (n = 246), decreases in cardiothoracic index (0.56 +/- 0.06, 0.53 +/- 0.06, 0.52 +/- 0.06, and 0.51 +/- 0.06; p less than 0.0001), in the prevalence of functional classes III and IV (87%, 76%, 68%, and 62%; p less than 0.0001), and of left ventricular (50%, 39%, 36%, and 30%; p less than 0.05) were accompanied by improved survival (5 year/10 year +/- SE: 73 +/- 6%/53 +/- 6%, 82 +/- 3%/65 +/- 4%, 89 +/- 3%/79 +/- 4%, and 87 +/- 3%/-; p = 0.002) despite increasing age (56 +/- 9, 57 +/- 11, 59 +/- 10, and 61 +/- 11 years; p less than 0.01). A Cox regression analysis identified peak-to-peak systolic gradient (p = 0.0001; inversely related to death rate), cardiothoracic index (p = 0.0003), left ventricular failure (p = 0.0005), prosthetic orifice diameter of 15 mm or less (p = 0.001), age (p = 0.003), ventricular ectopic beats (p = 0.004), male gender (p = 0.03), and antianginal/antiarrhythmic treatment (p = 0.03) as independent risk factors. A prognostic index calculated from the final Cox model stratified the patients into eight risk groups, having observed +/- SE/predicted/expected (matched background population) 10-year survival rates of 90 +/- 7% (n = 29)/94%/91%, 84 +/- 6% (n = 61)/89%/87%, 86 +/- 3% (n = 105)/83%/81%, 75 +/- 4% (n = 165)/75%/77%, 62 +/- 6% (n = 128)/63%/74%, 51 +/- 8% (n = 84)/47%/71%, 29 +/- 9% (n = 40)/31%/67%, and 16 +/- 9% (n = 18)/14%/54% (p less than 0.000001/-/-). Excess mortality relative to the background populations prevailed predominantly in risk groups 5-8 and was mainly caused by congestive heart failure. Thus, improved long-term survival during the 22-year operative period was related to improved preoperative patient status. Earlier operation (= low prognostic index) inferred a survival rate comparable to that of a matched background population. The prognostic index was probably predominantly related to preoperative myocardial damage that caused late predictable death from congestive heart failure.