1/ Wow! What intervention showed a 56% REDUCTION IN CANCER MORTALITY? MUST READ in @AnnalsofIM. Bariatric Surgery and All-Cause Mortality: A Population-Based Matched Cohort Study acpjournals.org/doi/10.7326/M1… 
2/ Overall, #bariatric surgery was associated with...
✅47% reduction in cardiovascular mortality
✅56% reduction in #cancer mortality
✅36% reduction in other medical mortality at 4.9 years of follow-up.
Lets dive into the study in latest @AnnalsofIM
acpjournals.org/doi/10.7326/M1…
✅47% reduction in cardiovascular mortality
✅56% reduction in #cancer mortality
✅36% reduction in other medical mortality at 4.9 years of follow-up.
Lets dive into the study in latest @AnnalsofIM
acpjournals.org/doi/10.7326/M1…
3/ The mission was simple (👏👏to @McMasterU &
@STJOESHAMILTON): Determine the association bw bariatric surgery & all-cause mortality. Esp since other studies of mortality post bariatric surg were limited by cohort selection bias, completeness of follow-up, & confounders.
@STJOESHAMILTON): Determine the association bw bariatric surgery & all-cause mortality. Esp since other studies of mortality post bariatric surg were limited by cohort selection bias, completeness of follow-up, & confounders.
4/ How did they do it?
Population-based matched cohort study in a universal health care system in Ontario, @Canada. 13,679 patients who underwent bariatric surgery from Jan 2010 to Dec 2016 and 13,679 matched nonsurgical patients.
Full Methods here: acpjournals.org/doi/10.7326/M1…
Population-based matched cohort study in a universal health care system in Ontario, @Canada. 13,679 patients who underwent bariatric surgery from Jan 2010 to Dec 2016 and 13,679 matched nonsurgical patients.
Full Methods here: acpjournals.org/doi/10.7326/M1…
5/✅Primary outcome: all-cause mortality (BEST one!)
✅Secondary outcome: cause-specific mortality.
✅Pts matched by age, sex, BMI, diabetes duration.
✅Cause of death classified: cardiovascular, oncologic, other med (resp, GI, infxn, etc), or external (trauma, suicide, accident)
✅Secondary outcome: cause-specific mortality.
✅Pts matched by age, sex, BMI, diabetes duration.
✅Cause of death classified: cardiovascular, oncologic, other med (resp, GI, infxn, etc), or external (trauma, suicide, accident)
6/ So what did they find? HINT: It's all good news for patients. Even JT agrees!
👏Overall mortality LOWER: 1.4% in surgery group vs 2.5% in nonsurgery group
👏Adjusted HR of all-cause mortality: 0.68 [95% CI 0.57-0.81] (i.e. good to be in the surgery group)
👏Overall mortality LOWER: 1.4% in surgery group vs 2.5% in nonsurgery group
👏Adjusted HR of all-cause mortality: 0.68 [95% CI 0.57-0.81] (i.e. good to be in the surgery group)
7/ Pts in the bariatric surgery group had a lower risk of dying, period! Plus lower risk of dying of cardiovascular dz or cancer!
Looking at cause specific mortality:
👍Lower cardiovascular mortality. HR 0.53 CI 0.34-0.84
👍My fave: ‼️Lower cancer mortality‼️ HR 0.54 CI 0.36-0.80
Looking at cause specific mortality:
👍Lower cardiovascular mortality. HR 0.53 CI 0.34-0.84
👍My fave: ‼️Lower cancer mortality‼️ HR 0.54 CI 0.36-0.80
8/ If you’re older, it gets better (although it hurts to hear that 55+ is 'old').
✅TWICE as many pts aged 55y+ in nonsurgery group died vs surgery group! (n=188 vs 87)
✅55y+ had an absolute risk reduction of 3.3%, CI 2.3-4.3 w/ HR of mortality in surgery grp 0.53, CI 0.41-0.69
✅TWICE as many pts aged 55y+ in nonsurgery group died vs surgery group! (n=188 vs 87)
✅55y+ had an absolute risk reduction of 3.3%, CI 2.3-4.3 w/ HR of mortality in surgery grp 0.53, CI 0.41-0.69
9/ The pipe-smoking monocle-wearing skeptic may ask about the confounders. and yes residual confounding by unmeasured factors is always possible in observational research...but authors used data from multiple linked administrative databases and...(next tweet)
10/ To determine‼️the most comprehensive association bw bariatric surgery and mortality to date‼️, they adjusted for confounders incldng
➡️comorbidities
➡️previous procedures
➡️socioeconomic status
➡️smoking status
➡️substance use disorder
➡️cancer screening
➡️psychiatric history
➡️comorbidities
➡️previous procedures
➡️socioeconomic status
➡️smoking status
➡️substance use disorder
➡️cancer screening
➡️psychiatric history
11/ Paging my #healthdisparities peers! Their study is the first to include #SDOH confounders of mortality such as
👏neighborhood marginalization
👏rural residence
👏psychiatric history
👏cancer screening
👏previous health care utilization, together within the same study.
👏neighborhood marginalization
👏rural residence
👏psychiatric history
👏cancer screening
👏previous health care utilization, together within the same study.
12/ Limits: yes, residual confounding by unmeasured factors is always possible in observational research and the observational design limits causal inference…but I mean, come on! Cancer prevention efforts must include consideration for bariatric surgery. Again, bravi, team! /fin
