Thirty-day mortality after valve replacement for aortic stenosis over the last 22 years. A multivariate risk stratification

Over the last 22 years (1965-86), the 30-day mortality rate (30-DMR) after valve replacement for primary or predominant aortic stenosis (AS, N = 690) fell from 20.0% for the first 100 consecutive patients, via 7.8% for the next 490, to 2.0% for the final 100 (P less than 0.0001). There was, however, a wide scatter in 30-DMR in the 11 consecutive 2-year periods, as well as during later years. The variation in 30-DMR was paralleled by changes in a high risk prognostic index (derived from a logistic regression model) and in a high coronary artery disease (CAD) score. In patients with a high CAD score who died within 30 days, 93% (N = 25) had CAD at autopsy compared with 37% (N = 11) of those with a low CAD score (P less than 0.0001). In 205 patients evaluated by coronary arteriography, the 30-DMR was 4.1% in 122 without CAD, 3.6% (3.8% for triple vessel/left main stem) for 55 with CAD who underwent bypass grafting, and 17.9% for 28 with CAD who did not have bypass grafting (P less than 0.0001). Left ventricular failure (LVF; episodes of pulmonary oedema and/or stasis), age, pronounced hypertrophy/strain in the ECG, and a high CAD score were independent incremental risk factors for 30-DMR. Quantitatively, LVF increased the risk 10 times more than pronounced hypertrophy/strain and a high CAD score. LVF also neutralized the influence of age. Modifying (symptom-masking) digitalis and/or diuretic treatment in functional class II patients (N = 189) increased the 30-DMR from 0.9% to 9.1% (P less than 0.01). The scatter of operative year-specific 30-DMR was related to changes in preoperative prognostic patient profiles and to unrevascularized CAD. Operative intervention in AS patients, even with discrete symptoms, and consistent revascularization of significant CAD, should be strongly advocated.