Skip to main content

Association between bariatric surgery and all-cause mortality: a population-based matched cohort study in a universal health care system

Doumouras AG, Hong D, Lee Y, Tarride J, Paterson JM, Anvari M. Ann Intern Med. 2020; 173(9):694-703. Epub 2020 Aug 18. DOI:

Background — Mortality after bariatric surgery has been previously studied, but cohort selection bias, completeness of follow-up, and collection of confounders have limited the inference of results.

Objective — To determine the association between bariatric surgery and all-cause mortality.

Design — Population-based matched cohort study.

Setting — Ontario, Canada.

Participants — 13 679 patients who underwent bariatric surgery from January 2010 to December 2016 and 13 679 matched nonsurgical patients.

Intervention — Bariatric surgery.

Measurements — The primary outcome was all-cause mortality, with cause-specific mortality as the secondary outcome. Patients were matched according to age, sex, body mass index, and diabetes duration.

Results — 13 679 patients who underwent bariatric surgery were matched to 13 679 nonsurgical patients. After a median follow-up of 4.9 years, the overall mortality rate was 1.4% (n = 197) in the surgery group and 2.5% (n = 340) in the nonsurgery group, with a lower adjusted hazard ratio (HR) of overall all-cause mortality (HR, 0.68 [95% CI, 0.57 to 0.81]). Patients aged 55 years or older had an absolute risk reduction of 3.3% (CI, 2.3% to 4.3%), with a lower HR of mortality in the surgery group (HR, 0.53 [CI, 0.41 to 0.69]). Observed relative effects were similar across sex; however, the observed association in absolute terms was greater in men. Surgery also was associated with lower cardiovascular mortality (HR, 0.53 [CI, 0.34 to 0.84]) and lower cancer mortality (HR, 0.54 [CI, 0.36 to 0.80]).

Limitation — The observational design limits causal inference.

Conclusion — Bariatric surgery was associated with substantially lower all-cause, cardiovascular, and cancer mortality. The lowered observed mortality of surgery was significant across most subgroups. The largest absolute effects were for men and patients aged 55 years or older.