Background: Bariatric procedures are increasingly performed but their impact on survival is unknown.
Study design: We evaluated short- and longterm mortality rates of patients undergoing gastric bypass on a population level compared with a nonoperated cohort of patients with morbid obesity in a retrospective study, using the Washington State Comprehensive Hospital Abstract Reporting System database and the Vital Statistics database. The study included all patients (age 18 to 65 years) from 1987 to 2001 who underwent gastric bypass with ICD-9 diagnostic codes for obesity. The comparator group included patients of similar age with a diagnosis of obesity or morbid obesity who did not have a bariatric procedure. Survival analysis was used to determine the association of surgeon experience on 30-day mortality and of the procedure on survival while controlling for age, gender, and comorbidity index.
Results: Of the 66,109 obese patients we evaluated, 3,328 had a bariatric procedure. Incidence of the procedure increased from 0.7 to 10.6 per 100,000 from 1987 to 2001, with a 2.5-fold increase in incidence rate of the procedure in the years after 1996 (incidence rate ratio, 2.5; 95% CI, 2.4 to 2.7). Thirty-day mortality was 1.9% and was associated with surgical inexperience. Within the surgeon's first 19 procedures the odds of death within 30 days were 4.7 times higher (95% CI, 1.2 to 18.2) than at later points in a surgeon's case order. At 15 years followup, 16.3% of nonoperated patients had died as compared with 11.8% of patients who had the bariatric procedure. When survival was compared beginning 1 year after the procedure, the adjusted hazard for death was 33% lower than that of nonoperated patients (hazard ratio 0.67; 95% CI, 0.54 to 0.85).
Conclusions: Thirty-day mortality after gastric bypass is higher than previously reported and closely linked to surgeon inexperience. A modest overall survival benefit was associated with the procedure but a marked survival advantage was noted for patients who survive to the first postoperative year.