To determine whether differences in referral reasons explain the higher operative mortality of women in coronary artery bypass surgery. Case series. A tertiary care, private teaching hospital. Consecutive patients who had isolated coronary artery bypass surgery between 1982 and 1987 (total, 2297; 79% male and 21% female). The inhospital mortality rate was significantly higher for women than for men (4.6% compared with 2.6%; P = 0.036; 95% CI for difference in mortality, 0% to 4.0%). Women were older than men (mean, 68.2 and 64.0 years, respectively; P less than 0.001), and a higher percentage of women were referred with unstable angina (P = 0.007), postmyocardial infarction angina (P = 0.029), congestive heart failure (P less than 0.001), and New York Heart Association class IV symptoms (66% compared with 45%, P less than 0.001). More men were referred with a history of an abnormal exercise test (P less than 0.001), and patients referred because of a positive exercise test had a lower mortality (P less than 0.001). Using multivariate analysis, adjustment for the higher preoperative functional class of women and for age accounted for all of the difference in mortality between men and women (odds ratio, 1.04; CI, 0.60 to 1.79; P = 0.89). After correction for functional class alone, there continued to be no significant difference in mortality between men and women (P = 0.40). Differences in functional class and age account for the higher operative mortality of women in coronary bypass surgery. Women are referred for coronary bypass surgery later in the course of their disease than men, and later referral may increase their changes of operative death.