Effect of comorbid and fatal coexistent conditions on sex and race differences in vascular surgical mortality.

Many previous studies of vascular procedures have found sex and race differences in surgical mortality that were attributed to differential prevalence of comorbidity. Adjustment for selected comorbid conditions does not entirely remove bias. In addition to adjustments for other covariates, surgical mortality ratios in this study were adjusted for coexistent conditions that caused postoperative death but were unrelated to the procedure. The adjusted mortality was, therefore, attributable to the procedure. Medicare administrative and death certificate data on beneficiaries aged 65-99 years who resided in Indiana and Kentucky and who had 6,016 major vascular procedures in 1994-1997 were used. In Cox proportional hazard models, male-to-female and nonwhite-to-white surgical mortality ratios were adjusted for age, sex, or race; weighted Charlson comorbidity score; length of hospital stay; and fatal coexisting conditions (FCCs). Altogether, 3,333 patients died within 30 postoperative days. There were sex and/or race differences in mortality caused by aortic aneurysm, stroke, and diabetes (P < 0.05). Unadjusted, all-cause 30-day mortality was higher in women and nonwhite patients than in men and white patients following coronary artery bypass graft (CABG) procedure (P < 0.03). Mortality following all non-CABG procedures combined was lower in women than in men (P < 0.02). In multivariate analyses, 30-day mortality following CABG, adjusted for covariates, was lower in men than in women (hazard ratio [HR] = 0.88, 95% confidence interval [CI] 0.79-0.98), but there was no sex difference after adjustment for only FCC (HR = 0.94, 95% CI 0.85-1.05). Mortality following all non-CABG procedures combined was higher in men than in women, but this difference was insignificant after adjustment for comorbidity and/or FCC (HR = 1.05, 95% CI 0.93-1.17). Age- and sex-adjusted 30-day mortality following CABG was higher in nonwhite patients than in white patients (HR = 1.37, 95% CI 1.08-1.74), and this race difference persisted after further adjustments. There were no significant sex or race differences in surgical mortality following carotid endarterectomy, non-CABG thoracoabdominal procedures, or procedures in the limbs. Adjustments for covariates did not alter race difference in post-CABG surgical mortality. Adjustment for comorbid conditions slightly affected sex differences in mortality following CABG and all non-CABG procedures combined, but adjustment for FCC reduced these differences to insignificant levels.