Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Mar 1.
Published in final edited form as: Ann Surg. 2021 Aug 13;277(3):442–448. doi: 10.1097/SLA.0000000000005155

Reduction in Long-term Mortality after Sleeve Gastrectomy and Gastric Bypass Compared to Non-surgical Patients with Severe Obesity

Anita P Courcoulas 1, Eric Johnson 2, David E Arterburn 2, Sebastien Haneuse 3, Lisa J Herrinton 4, David P Fisher 4, Robert A Li 4, Mary Kay Theis 2, Liyan Liu 4, Brianna Taylor 5, Julie Cooper 2, Philip L Chin 5, Gary G Grinberg 4, Anirban Gupta 2, Shireesh Saurabh 2, Scott S Um 5, Panduranga R Yenumula 4, Jorge L Zelada 5, Karen J Coleman 5
PMCID: PMC8840990  NIHMSID: NIHMS1732115  PMID: 34387200

Abstract

Objective:

To separately compare the long-term risk of mortality among bariatric surgical patients undergoing either RYGB or SG to large, matched, population-based cohorts of patients with severe obesity who did not undergo surgery.

Background:

Bariatric surgery has been associated with reduced long-term mortality compared to usual care for severe obesity which is particularly relevant in the COVID-19 era. Most prior studies involved the Roux-en-Y gastric bypass (RYGB) operation and there is less long-term data on the sleeve gastrectomy (SG).

Methods:

In this retrospective, matched cohort study, patients with a body mass index ≥35 kg/m2 who underwent bariatric surgery from January 2005 to September 2015 in three integrated health systems in the United States were matched to nonsurgical patients on site, age, sex, body mass index, diabetes status, insulin use, race/ethnicity, combined Charlson/Elixhauser comorbidity score, and prior health care utilization, with follow-up through September 2015. Each procedure (RYGB, SG) was compared to its own control group and the two surgical procedures were not directly compared to each other. Multivariable-adjusted Cox regression analysis investigated time to all-cause mortality (primary outcome) comparing each of the bariatric procedures to usual care. Secondary outcomes separately examined the incidence of cardiovascular-related death, cancer related-death, and diabetes related-death.

Results:

Among 13,900 SG, 17,258 RYGB, and 87,965 nonsurgical patients, the 5-year follow-up rate was 70.9%, 72.0%, and 64.5%, respectively. RYGB and SG were each associated with a significantly lower risk of all-cause mortality compared to nonsurgical patients at 5-years of follow-up (RYGB: HR=0.43; 95% CI: 0.35,0.54; SG: HR=0.28; 95% CI: 0.13,0.57) Similarly, RYGB was associated with a significantly lower 5-year risk of cardiovascular- (HR=0.27; 95% CI: 0.20, 0.37), cancer- (HR=0.54; 95% CI: 0.39, 0.76), and diabetes-related mortality (HR=0.23; 95% CI:0.15, 0.36). There was not enough follow-up time to assess 5-year cause-specific mortality in SG patients, but at 3-years follow up, there was significantly lower risk of cardiovascular- (HR=0.33; 95% CI:0.19, 0.58), cancer- (HR=0.26; 95% CI:0.11, 0.59), and diabetes-related (HR=0.15; 95% CI:0.04, 0.53) mortality for SG patients.

Conclusion:

This study confirms and extends prior findings of an association with better survival following bariatric surgery in RYGB patients compared to controls and separately demonstrates that the SG operation also appears to be associated with lower mortality compared to matched control patients with severe obesity that received usual care. These results help to inform the trade-offs between long-term benefits and risks of bariatric surgery.

Mini-Abstract

Both gastric bypass and sleeve were each associated with a significantly lower risk of all-cause mortality compared to nonsurgical patients with severe obesity at 5-years of follow-up. These results help to inform the trade-offs between long-term benefits and risks of common bariatric surgical procedures.

Introduction:

Obesity is a leading cause of preventable death in the United States, 1 and it has been cited as a key contributor to the recent decline in US life expectancy.2 That crisis has been exacerbated by the onset of the coronavirus disease 2019 (COVID-19) pandemic, as patients with severe obesity (defined as a body mass index [BMI] ≥ 40 kg/m2) have been associated with a more than two-fold increased risk of death from COVID-19.3 These data clearly illustrate the urgent need to identify interventions that reduce mortality among patients with obesity.

Bariatric operations have been shown to promote long-term weight loss, improvements in obesity-associated comorbidities, and have been associated with reduced long-term mortality.4 A recent systematic review of 11 observational studies evaluating longer-term (10 or more years) mortality found that bariatric surgery was associated with a 38% lower risk of all- cause mortality (pooled odds ratio 0.62 (95% CI 0.55 to 0.69, p < 0.001) compared to people with severe obesity that did not undergo surgery.4–9 Most patients in these prior studies underwent a Roux-en-Y gastric bypass (RYGB) operation. There is a relative paucity of long-term data addressing the impact of the sleeve gastrectomy (SG) on mortality despite the fact that in 2018, 61% of the estimated 252,000 bariatric procedures performed in the US were SG, followed by RYGB, which accounted for only 17%.10 Observational studies suggest that patients who undergo SG may lose less weight and experience less durable improvement in comorbid health conditions (e.g., diabetes) than patients who undergo RYGB;11–13 however, the risk of reoperation and rehospitalization appears to be lower with SG than RYGB.14, 15 Understanding the effect of SG on mortality is necessary to provide a more complete picture of benefits and risks of RYGB and SG for patients and providers. The current study was designed to separately compare the long-term mortality among bariatric surgical patients undergoing RYGB and SG to large, matched, population-based cohorts of patients with severe obesity who did not undergo those operations. Our primary hypothesis was that bariatric surgery patients, undergoing either RYGB or SG operations, would each be separately associated with lower 5-year mortality when compared to matched non-surgical patients. The secondary aim of this study was to examine the association between bariatric operations and the risks associated with the leading causes of death: cardiovascular-related death, cancer related-death, and diabetes related-death. We were also interested in exploring the heterogeneity of responses to both operations by age, sex, race/ethnicity, and BMI.

Methods:

Design and Setting

We conducted a retrospective observational matched cohort study of adults with severe obesity enrolled in Kaiser Permanente regions in Washington (KPWA), Northern California (KPNC), and Southern California (KPSC). All study procedures were approved by the Institutional Review Board at each site and permitted conduct of the research without explicit consent from participants.

Data Sources

At each study site, electronic medical records, insurance claims, and other data systems were used to extract enrollment, insurance coverage, demographics, blood pressure, height, weight, laboratory values, medications dispensed, deaths, outpatient, inpatient, and emergency department use, and diagnosis and procedure codes of all surgical and non-surgical patients.

Surgical Patients

The bariatric surgery population included adults (19–79 years) with BMI ≥35 kg/m2 who had a primary (first observed, non-revisional) RYGB or SG operation between January 2005 and September 2015. Following our previously published work,13, 15–17 we identified bariatric operations using a combination of patient registries, chart review, and International Classification of Diseases (ICD-9) and Current Procedure Terminology (CPT) procedure codes (available upon request). We excluded patients who had less than one full year of continuous enrollment, a history of cancer (except non-melanoma skin cancer) or who were pregnant.

We identified 32,874 bariatric cases (eFigure 1), including 13,900 SG and 17,258 RYGB patients. Among these, 449 (1.3%) were excluded due to missing pre-operative BMI data, and 1,042 (3.2%) were excluded because their maximum observed preoperative BMI was <35, suggesting they were unlikely to be primary bariatric cases.13 Finally, 25 (0.07%) surgical cases were excluded because they could not be matched to at least one non-surgical patient.

Matched Non-surgical Patients

For each patient who underwent bariatric surgery, we identified up to three matched non- surgical controls from our general medical population via a two-step process. First, among all patients with at least one BMI ≥35 kg/m2 who did not undergo bariatric surgery during the study period (N=1,635,897), we identified a pool of potential controls who were enrolled at the time of the surgery, satisfied the study inclusion/exclusion criteria used for surgical cases, and matched the surgical patients to these controls on study site, sex, baseline age category (19–44, 45–64, 65–69 years), BMI category (35–39.9, 40–49.9, 50.0+ kg/m2), diabetes status (presence/absence based on laboratory, pharmacy, and diagnosis data), race/ethnicity, combined Charlson/Elixhauser comorbidity score18, and insulin use. Second, for each control in the pool we calculated their Mahalanobis distance from the bariatric patient on the basis of age, BMI, Charlson/Elixhauser comorbidity score, and the number of days of health care utilization in the 7–12 months prior to the date of surgery (a marker of comorbidity that is unaffected by utilization related to preparation for bariatric surgery).19 Finally up to three controls were selected based on the shortest Mahalanobis distance, with the restriction that each non-surgical patient could only be used as a control for one surgical patient. These non- surgical patients all received usual medical care, which usually included no specific treatment for obesity since our health systems did not universally provide coverage for obesity pharmacotherapy (leading to very low use overall) and rarely provided intensive lifestyle intervention for weight loss. 20

Analyses

Outcome and censoring definitions

The primary outcome was death from any cause, with death information obtained from a combination of electronic medical records, administrative databases, and state death indices. Secondary outcomes were cardiovascular-related death (coronary artery disease [CAD], heart failure, stroke, and other circulatory diagnoses), cancer related-death, and diabetes related- death. For identifying the secondary outcomes, each death could have had multiple causes listed and we did not restrict our analyses to those listed as the primary cause of death (e.g., a single death could be flagged as having both cardiovascular and diabetes causes). Patients were censored at the first of disenrollment or the end of the study period (9/30/2015).

Statistical models

Cox regression models were used to investigate the association between bariatric surgery vs. usual care (non-surgical controls) and death, comparing RYGB patients to their matched controls and SG patients separately to their matched controls. All patients were followed from the index date (the date of bariatric surgery or, for non-surgical patients, the date of surgery for the patient to whom they had been matched) until death or a censoring event (disenrollment or end of study). For the secondary outcomes, follow-up was also censored at the time of death due to some other cause, so that the interpretation of the results was in terms of cause-specific hazards. Multiple imputation via chained equations was used to impute missing data, with M=10 imputed datasets.21

To permit non-proportional hazards in the effect of bariatric surgery we used a three-knot restricted cubic splines for both procedure-specific analyses.22 Due to the differing average length of follow-up between RYGB and SG, we set procedure-specific knot locations at the tertiles of each group. Cumulative incidence curves for the cause-specific models were constructed using the Nelson-Aalen estimator,23 treating other-cause deaths as competing events.

All analyses were adjusted for a priori identified potential confounders: BMI immediately prior to index date, maximum BMI in prior year, age categories, days of health care and inpatient use use in the 7–12 months before the index date, sex/gender, year of index date, race/ethnicity, insulin use, Charlson/Elixhauser comorbidity score, hypertension diagnosis, systolic and diastolic blood pressures, use of ACE inhibitors, use of ARB medications, use of hypertension medications other than ACE or ARB, insurance category, diabetes status, uncontrolled blood pressure at the index date, dyslipidemia, use of statins, use of non-statin lipid lowering medications, smoking status, diagnosis codes for cerebrovascular disease, neuropathy, coronary artery disease, and mental health status. Models were stratified by site of surgery (KPSC, KPNC, and KPWA). Additional exploratory analyses were conducted to assess for heterogeneity in treatment effects across subgroups defined by BMI, age, sex, diabetes, and race/ethnicity categories.

Finally, we conducted a sensitivity analysis to assess the influence of unmeasured confounding on the 5-year results using the E-value methodology of VanderWeele and Ding.24 The E-value quantifies the minimum strength of association an unmeasured confounder must have with both bariatric surgery and mortality, while simultaneously considering the measured confounders, to negate the observed association between bariatric surgery and mortality in this study 24.

Results:

Among 13,900 SG, 17,258 RYGB, and 87,965 nonsurgical patients, the 5-year follow-up rate was 70.9%, 72.0%, and 64.5%, respectively. The baseline characteristics of both bariatric (RYGB, SG) and matched non-surgical patients are shown in eTable 1. SG cases accounted for 45% of the bariatric surgical operations. In the bariatric surgery group, the average age was 45 years, 82% were female, and 31% of Hispanic ethnicity. One third of both surgical and non- surgical patients had diabetes, with 11% of these patients using insulin.

Table 1 shows the main results of the fully adjusted Cox model comparing both RYGB and SG each to their respective matched non-surgical controls for all-cause mortality. RYGB was associated with a significantly lower risk of all-cause mortality at both 5- (HR=0.43; 95% CI: 0.35,0.54) and 7-years (HR=0.45; 95% CI: 0.35,0.59) follow-up. The cumulative unadjusted incidence of mortality for RYGB was 1.04% at 5-years and 1.78% at 7-years compared to 2.88% and 4.46% respectively for the matched non-surgical patients. SG was also associated with a significantly lower risk of all-cause mortality at 5-years (HR=0.28; 95% CI: 0.13,0.57), with an unadjusted cumulative incidence or mortality of 1.02% for SG and 2.74% for matched non- surgical patients. There was not enough follow-up time to assess 7-year all-cause mortality in SG patients. Figure 1A shows a plot of the adjusted cumulative incidence of all-cause mortality for both RYGB and SG each with their respective matched non-surgical controls. In this figure both surgical groups were matched to their control non-surgical patients and any imbalances in group characteristics in eTable 1 were adjusted for in the Cox model. The E-values (relative risk) for the point estimate and upper confidence bound for incident mortality after RYGB at 5 years were 0.54 and 3.11, respectively, and for SG at 5-years were 0.57 and 2.90, respectively (eTable3).

Table 1.

Results of matched, fully-adjusted Cox proportional hazards model comparing the risk of incident, all-cause mortality in gastric bypass and sleeve gastrectomy patients compared to matched non-surgical patients

Roux-en-Y gastric bypass
1 Year after Index Date 3 Years after Index Date 5 Years after Index Date 7 Years after Index Date
Number of RYGB patients still at risk at this time point 14725 10597 6530 3251
Hazard Ratios (95% CI) comparing gastric bypass vs matched non-surgical patients 0.31 (0.24, 0.40) 0.30 (0.24, 0.37) 0.43 (0.35, 0.54) 0.45 (0.35, 0.59)
Cumulative Incidence for gastric bypass 0.28% 0.55% 1.04% 1.78%
Non-surgical patients matched to gastric bypass patients
Number of matched patients still at risk at this time point 39076 26732 16078 8026
Cumulative Incidence for matched non-surgical patients 0.46% 1.55% 2.88% 4.46%
Sleeve gastrectomy
1 Year after Index Date 3 Years after Index Date 5 Years after Index Date 7 Years after Index Date
Number of sleeve patients still at risk at this time point 10473 5302 1121 83
Hazard Ratios (95% CI) comparing sleeve gastrectomy vs matched non-surgical patients 0.29 (0.19, 0.45) 0.35 (0.23, 0.53) 0.28 (0.13, 0.57) NA
Cumulative Incidence for sleeve gastrectomy 0.28% 0.56% 1.02% NA
Non-surgical patients matched to sleeve gastrectomy patients
Number of matched patients still at risk at this time point 28128 13268 2521 170
Cumulative Incidence for matched non-surgical patients 0.46% 1.55% 2.74% NA
Open in a new tab

Figure 1.

Figure 1.

Figure 1.

Figure 1.

Figure 1.

Open in a new tab

Estimates of the cumulative incidence of all-cause mortality following Roux-en-Y gastric bypass and sleeve gastrectomy versus matched non-surgical control patients.*

Figure 1A. All-Cause Mortality

*Separate estimates for all-cause mortality are derived from Cox models for all-cause mortality (A), and from Kaplan-Meier plots for cardiovascular-related (B), cancer-related (C) and diabetes-related mortality (D).

Figure 1B. Cardiovascular-Associated Mortality

Figure 1C. Cancer-Associated Mortality

Figure 1D. Diabetes-Associated Mortality

Table 2 shows the secondary outcome results with the matched, fully adjusted Cox model comparing the risk of incident, cause-specific mortality for both RYGB and SG to their matched non-surgical controls. RYGB was associated with a significantly lower risk of cardiovascular- (HR=0.27; 95% CI: 0.20, 0.37), cancer- (HR=0.54; 95% CI: 0.39, 0.76), and diabetes-related (HR=0.23; 95% CI:0.15, 0.36) mortality through 5-years. There was not enough follow-up time to assess 7-year cause-specific mortality in SG patients. At 3-years follow up for SG, there were significantly lower risk of cardiovascular- (HR=0.33; 95% CI:0.19, 0.58), cancer- (HR=0.26; 95% CI:0.11, 0.59), and diabetes-related (HR=0.15; 95% CI:0.04, 0.53) mortality.

Table 2.

Results of matched, fully-adjusted Cox proportional hazards model comparing the risk of incident, cause-specific mortality in gastric bypass and sleeve gastrectomy patients compared to matched non-surgical patients

Hazard ratios comparing gastric bypass vs. matched non-surgical patients
Hazard Ratio (95% CI) at 1 Year after Index Date Hazard Ratio (95% CI) at 3 Years after Index Date Hazard Ratio (95% CI) at 5 Years after Index Date Hazard Ratio (95% CI) at 7 Years after Index Date
Cardiovascular-related mortalitya 0.46 (0.39, 0.54) 0.18 (0.13, 0.25) 0.27 (0.20, 0.37) 0.76 (0.53, 1.09)
Cancer-related mortalitya 0.69 (0.57, 0.82) 0.44 (0.29, 0.64) 0.54 (0.39, 0.76) 0.91 (0.56, 1.48)
Diabetes-related mortalitya 0.49 (0.39, 0.61) 0.19 (0.12, 0.31) 0.23 (0.15, 0.36) 0.45 (0.25, 0.80)
Hazard ratios comparing sleeve gastrectomy vs. matched non-surgical patients
Hazard Ratio (95% CI) at 1 Year after Index Date Hazard Ratio (95% CI) at 3 Years after Index Date Hazard Ratio (95% CI) at 5 Years after Index Date Hazard Ratio (95% CI) at 7 Years after Index Date
Cardiovascular-related mortalitya 0.46 (0.30, 0.71) 0.33 (0.19, 0.58) 0.57 (0.19, 1.71) NA
Cancer-related mortalitya 0.42 (0.21, 0.86) 0.26 (0.11, 0.59) 0.39 (0.06, 2.29) NA
Diabetes-related mortalitya 0.23 (0.10, 0.53) 0.15 (0.04, 0.53) 0.69 (0.15, 3.24) NA
Open in a new tab
a

For cause-specific mortality analyses, the cause of death could be listed as “primary” or “contributing”. Due to small sample sizes, estimates for sleeve gastrectomy could not be reliably calculated 7 years after the index date.

Figures 1B, C, D show the unadjusted Kaplan Meir plots of cardiovascular-, cancer-, and diabetes-related mortality for RYGB, SG, and their matched controls. The time varying hazards ratios comparing the risk of incident mortality following RYGB and SG versus matched controls for all-cause mortality, cardiovascular, cancer, and diabetes deaths are shown in Figure 2A, B, C, and D and show a sharp decline in the HR beginning immediately after the index date and continuing through 3-to-4 years after which the HR begins to increase again.

Figure 2.

Figure 2.

Figure 2.

Figure 2.

Figure 2.

Open in a new tab

Time-varying hazard ratios comparing the risk of incident mortality following Roux-en-Y gastric bypass and sleeve gastrectomy versus matched non-surgical patients. Separate estimates are provided for all-cause mortality (A), cardiovascular-related (B), cancer-related (C) and diabetes-related mortality (D).

Figure 2A. All-Cause Mortality Time-Varying Hazard Ratio

Figure 2B. Cardiovascular Associated Mortality Time-Varying Hazard Ratio

Figure 2C. Cancer-Associated Mortality Time-Varying Hazard Ratio

Figure 2D. Diabetes-Associated Mortality Time-Varying Hazard Ratio

Analyses of heterogeneity of treatment effects are shown in eFigures 2A-E. There were no significant effects of pre-operative BMI ≥ 50 kg/m2, diabetes status, age ≥ 65 years, sex, or race/ethnicity on mortality when comparing RYGB or SG patients to control patients.

Discussion:

This large, contemporary, multisite observational study identified that bariatric surgical operations were associated with a significantly lower risk of mortality through 5 years follow-up compared to the nonsurgical control patients. This study also specifically included large numbers of patients that underwent the most common current bariatric operation, the SG, identifying a clear association with lower mortality following SG when compared to nonsurgical patients that received usual care. For the SG operation, this is a key finding that is consistent with several other observational studies involving mostly RYGB patients.4 This study also extends the prior literature by examining cause-specific mortality. We found significant reductions in cardiovascular-, cancer-, and diabetes-related mortality at 5-years for RYGB and at 3-years for SG. We found no significant differential effects of SG and RYGB operations on mortality across subgroups defined by pre surgical BMI, diabetes status and insulin use, age, sex, or race/ethnicity.

Our findings are consistent with other major long-term observational studies in the field, including the Swedish Obese Subjects study, the Utah Obesity Study, and a nationwide study involving Veteran’s Administration patients.5–7, 9, 25 In a more recent study from Canada, a population-based matched cohort study of over 13,000 patients who underwent bariatric surgery (87% had RYGB) were compared to nonsurgical patients where the primary outcome was all-cause mortality, with cause-specific mortality as the secondary outcome.26 At a median follow-up of 4.9 years, the overall mortality rate was 1.4%) in the surgery group and 2.5% in the non-surgery group, indicating a lower risk of all-cause mortality (adjusted HR, 0.68 [95% CI, 0.57 to 0.81]). Bariatric surgery was also associated with lower cardiovascular and cancer mortality. The lowered observed all-cause mortality of surgery was significant across most subgroups with the largest absolute effects for men and patients aged 55 years or older.26 Our study found similar improvements in all-cause and cause specific (cardiovascular-, cancer-, and diabetes- related) mortality, while also including a much larger proportion of people undergoing SG and a sample that can be used to reflect bariatric practices in the U.S. We also found that the mortality benefits were similar across all subgroups defined by pre surgical BMI, diabetes status and insulin use, age, sex, or race/ethnicity. The clinical implication is that the association between bariatric procedures and improved mortality is not restricted to any specific clinical subgroup of patients.

In this study the cumulative mortality rates for both RYGB and SG were quite low – just 1% at 5 years for both procedures and 1.78% at 7 years for RYGB. These rates are somewhat lower than other published observational studies. For example, in New York state, from 1999 to 2005, the mean bariatric surgery mortality rate was 2.5% with 8–14 years of follow-up and did not differ by operation type (57% of total cases were RYGB).27 In a long-term Veteran’s Administration study, which included 74% RYGB mostly male and higher medical risk patients, estimated mortality rates were 2.4% at 1 year, 6.4% at 5 years, and 13.8% at 10 years for surgical patients and 1.7% at 1 year, 10.4% at 5 years, and 23.9% at 10 years for matched control patients.9 One study among 33,540 patients in Israel did report 5-year mortality rates similar to ours: 1.3% for RYGB and 0.8% for SG from 2005 to 2014.28 It is possible that the lower rates of mortality we report for RYGB and SG may be reflective of the commercially-insured population being studied along with contemporary bariatric care where nearly all cases are performed by the minimally invasive/laparoscopic approach.29 Our findings might not apply to small community-based bariatric practices or operations performed in populations with a large number of Medicaid patients.

Our cause-specific mortality findings also compliment other recent, large observational studies on the effect of RYGB and SG operation for diabetes remission, micro- and macro- vascular complications from diabetes, and weight loss up to 8-years of follow-up.16, 17, 30 In these studies, surgery was associated with a lower risk for incident microvascular disease (neuropathy, nephropathy, and retinopathy) at 5 years compared to control, and surgery was associated with a significantly lower incidence of macrovascular disease events and all-cause mortality at both 5 and at 8 years. Furthermore, bariatric surgery is known to result in durable weight loss and improvements or remission of obesity-related comorbid conditions, such as type 2 diabetes, lipids, and other conditions.31

This study has several limitations including loss to follow-up of 29%−35% at 5-years. In addition, our analyses did not directly compare RYGB to SG as we compared each bariatric surgical procedure to its own matched control group. Therefore, we were unable to make any inference about differences in hazards ratios for SG versus RYGB. We also used state mortality indices which do not include patients who died in another state as we did not use national death index data for this study. The observational design of the study also precludes causal inference, and unmeasured confounding may have persisted despite model adjustment for many major risk factors for mortality (e.g., age, smoking status, diabetes status). However, our sensitivity analyses using the E-value methodology indicate that each follow-up time point had an E-value greater than 2.5 for the upper bound of the CI; suggesting that the results could only be explained by an unmeasured confounder that was associated with both receipt of bariatric surgery and risk for mortality by a risk ratio of 2.5. Such an unmeasured confounder, however, seems unrealistic. For example, in eTable 1, the strongest measured confounder in our analysis was age, and the HR difference comparing between age 18–44 vs 65+ groups was 2.74. Thus, any unmeasured confounders would need to exceed the effect of age on mortality risk to invalidate our results. Currently there are no long-term randomized studies investigating the impact of bariatric operations on mortality. Such level I evidence would be important to have more confidence that it is the bariatric operation lowering mortality as suggested by our study, or if the effects we see in our work and other observational studies are biased by potential unmeasured confounders such as diet, exercise, motivation, and healthy lifestyle habits. Until such time, large observational studies will continue to be used to estimate the effect of bariatric surgery on mortality.

There are also several strengths of the current study which include a high proportion and large number of SG operations, representing current bariatric operative trends for which survival/mortality data have been lacking. In addition, we used rigorous statistical methods to carefully match bariatric patients to their non-surgical controls and control for factors that might account for the differences in mortality between these groups. In addition, we performed heterogeneity of treatment effects analyses to look for differences in mortality outcomes in important subsets of patients based on clinical factors such as age, BMI, and race/ethnicity.

In summary, these data appear to both confirm and extend prior findings of an association between lower mortality following bariatric surgery, including among patients undergoing the SG procedure, as it appears that SG and RYGB are each separately associated with lower mortality with respect to matched controlled non-surgical patients. All of these findings can assist patients and their health care providers in understanding the trade-offs between surgical risks of reoperation and re-interventions and long-term weight loss, diabetes and other metabolic improvements including improved survival.14 Finally, in light of the current COVID-19 pandemic, bariatric surgery appears to have the potential to modify COVID-19-related mortality as recent studies have demonstrated increased risk for requiring ventilation and increased mortality from the virus in those with obesity as well as possible protection from more serious disease in those who have undergone bariatric surgery.3, 32, 33

Supplementary Material

Supplemental Data File

ACKNOWLEDGEMENTS

Conflicts of Interest Statement:

AC had a research grant from Allurion Inc; DA has grants from NIH and PCORI and travel paid by IFSO Latin America Chapter and World Congress on Interventional Therapy for Diabetes; KC has funding for research from NIDDK, NHLBI, NIMH, and FDA and is paid a stipend for reviewing grants for NIH. (outside of the submitted work)

No authors have spouses, partners, or children that have financial relationships that may be relevant to the submitted work.

SH, MKT, DF, JC, RL, PC, SS, GG, PY, AG, EJ, LH, JZ, SU, and BT have nothing to declare.

Funding/Support and Role of Funder/Sponsor:

The study was funded by NIH/NIDDK R01DK105960-01. The sponsor did not play a role in the collection, management, analysis or interpretation of data, the preparation, review, or approval of the manuscript, or the decision to submit the manuscript for publication.

Funding: NIH-NIDDK R01DK105960-01

Footnotes

Data Access, Responsibility, and Analysis:

Eric Johnson, David Arterburn, and Sebastien Haneuse had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

REFERENCES

  • 1.Masters RK, Reither EN, Powers DA, Yang YC, Burger AE, Link BG. The impact of obesity on US mortality levels: the importance of age and cohort factors in population estimates. Am J Public Health. Oct 2013;103(10):1895–901. doi: 10.2105/AJPH.2013.301379 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Woolf SH, Schoomaker H. Life Expectancy and Mortality Rates in the United States, 1959–2017. JAMA. Nov 26 2019;322(20):1996–2016. doi: 10.1001/jama.2019.16932 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Tartof SY, Qian L, Hong V, et al. Obesity and Mortality Among Patients Diagnosed With COVID- 19: Results From an Integrated Health Care Organization. Ann Intern Med. Aug 12 2020;doi: 10.7326/M20-3742 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Wiggins T, Guidozzi N, Welbourn R, Ahmed AR, Markar SR. Association of bariatric surgery with all-cause mortality and incidence of obesity-related disease at a population level: A systematic review and meta-analysis. PLoS Med. Jul 2020;17(7):e1003206. doi: 10.1371/journal.pmed.1003206 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Sjostrom L, Narbro K, Sjostrom CD, et al. Effects of bariatric surgery on mortality in Swedish obese subjects. N Engl J Med. Aug 23 2007;357(8):741–52. doi: 10.1056/NEJMoa066254 [DOI] [PubMed] [Google Scholar]
  • 6.Adams TD, Gress RE, Smith SC, et al. Long-term mortality after gastric bypass surgery. N Engl J Med. Aug 23 2007;357(8):753–61. doi: 10.1056/NEJMoa066603 [DOI] [PubMed] [Google Scholar]
  • 7.Perry CD, Hutter MM, Smith DB, Newhouse JP, McNeil BJ. Survival and changes in comorbidities after bariatric surgery. Ann Surg. Jan 2008;247(1):21–7. doi: 10.1097/SLA.0b013e318142cb4b [DOI] [PubMed] [Google Scholar]
  • 8.Maciejewski ML, Livingston EH, Smith VA, et al. Survival among high-risk patients after bariatric surgery. JAMA. Jun 15 2011;305(23):2419–26. doi:jama.2011.817 [pii] 10.1001/jama.2011.817 [DOI] [PubMed] [Google Scholar]
  • 9.Arterburn DE, Olsen MK, Smith VA, et al. Association between bariatric surgery and long-term survival. JAMA. Jan 06 2015;313(1):62–70. doi: 10.1001/jama.2014.16968 [DOI] [PubMed] [Google Scholar]
  • 10.Estimate of Bariatric Surgery Numbers -, November 21). Retrieved October 22, 2020, from https://asmbs.org/resources/estimate-of-bariatric-surgery-numbers.
  • 11.McTigue KM, Wellman R, Nauman E, et al. Comparing the 5-Year Diabetes Outcomes of Sleeve Gastrectomy and Gastric Bypass: The National Patient-Centered Clinical Research Network (PCORNet) Bariatric Study. JAMA Surg. May 1 2020;155(5):e200087. doi: 10.1001/jamasurg.2020.0087 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Arterburn D, Wellman R, Emiliano A, et al. Comparative Effectiveness and Safety of Bariatric Procedures for Weight Loss: A PCORnet Cohort Study. Ann Intern Med. Dec 4 2018;169(11):741–750. doi: 10.7326/M17-2786 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Arterburn DE, Johnson E, Coleman KJ, et al. Weight Outcomes of Sleeve Gastrectomy and Gastric Bypass Compared to Nonsurgical Treatment. Ann Surg. Mar 13 2020;doi: 10.1097/SLA.0000000000003826 [DOI] [PubMed] [Google Scholar]
  • 14.Courcoulas A, Coley RY, Clark JM, et al. Interventions and Operations 5 Years After Bariatric Surgery in a Cohort From the US National Patient-Centered Clinical Research Network Bariatric Study. JAMA surgery. Jan 15 2020; doi: 10.1001/jamasurg.2019.5470 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Li RA, Liu L, Arterburn D, et al. Five-year Longitudinal Cohort Study of Reinterventions After Sleeve Gastrectomy and Roux-en-Y Gastric Bypass. Ann Surg. Jun 7 2019;doi: 10.1097/SLA.0000000000003401 [DOI] [PubMed] [Google Scholar]
  • 16.Fisher DP, Johnson E, Haneuse S, et al. Association Between Bariatric Surgery and Macrovascular Disease Outcomes in Patients With Type 2 Diabetes and Severe Obesity. JAMA. Oct 16 2018;320(15):1570–1582. doi: 10.1001/jama.2018.14619 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.O’Brien R, Johnson E, Haneuse S, et al. Microvascular Outcomes in Patients With Diabetes After Bariatric Surgery Versus Usual Care: A Matched Cohort Study. Ann Intern Med. Sep 4 2018;169(5):300–310. doi: 10.7326/M17-2383 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Gagne JJ, Glynn RJ, Avorn J, Levin R, Schneeweiss S. A combined comorbidity score predicted mortality in elderly patients better than existing scores. J Clin Epidemiol. Jul 2011;64(7):749–59. doi: 10.1016/j.jclinepi.2010.10.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Mahalanobis PC. On the generalized distance in statistics. Proc Natl Acad Sci India. 1936;2(1):49–55. [Google Scholar]
  • 20.Saxon DR, Iwamoto SJ, Mettenbrink CJ, et al. Antiobesity Medication Use in 2.2 Million Adults Across Eight Large Health Care Organizations: 2009–2015. Obesity. 2019;27(12):1975–1981. doi: 10.1002/oby.22581 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Raghunathan TE, Lepkowski JM, Van Hoewyk J, Solenberger P. A Multivariate Technique for Multiply Imputing Missing Values Using a Sequence of Regression Models. Survey Methodology 2001;27:85–95. [Google Scholar]
  • 22.Harrell FE. Regression Modeling Strategies. Secaucus, NJ: Springer-Verlag New York, Inc.; 2006. [Google Scholar]
  • 23.Kalbfleisch John D., and Prentice Ross L.. The Statistical Analysis of Failure Time Data. Vol. 360. John Wiley & Sons, 2011. [Google Scholar]
  • 24.Haneuse S, VanderWeele TJ, Arterburn D. Using the E-Value to Assess the Potential Effect of Unmeasured Confounding in Observational Studies. JAMA. Feb 12 2019;321(6):602–603. doi: 10.1001/jama.2018.21554 [DOI] [PubMed] [Google Scholar]
  • 25.Carlsson LMS, Sjoholm K, Jacobson P, et al. Life Expectancy after Bariatric Surgery in the Swedish Obese Subjects Study. N Engl J Med. Oct 15 2020;383(16):1535–1543. doi: 10.1056/NEJMoa2002449 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Doumouras AG, Hong D, Lee Y, Tarride JE, Paterson JM, Anvari M. Association Between Bariatric Surgery and All-Cause Mortality: A Population-Based Matched Cohort Study in a Universal Health Care System. Ann Intern Med. Aug 18 2020;doi: 10.7326/M19-3925 [DOI] [PubMed] [Google Scholar]
  • 27.Telem DA, Talamini M, Shroyer AL, et al. Long-term mortality rates (>8-year) improve as compared to the general and obese population following bariatric surgery. Surg Endosc. Mar 2015;29(3):529–36. doi: 10.1007/s00464-014-3714-4 [DOI] [PubMed] [Google Scholar]
  • 28.Reges O, Greenland P, Dicker D, et al. Association of Bariatric Surgery Using Laparoscopic Banding, Roux-en-Y Gastric Bypass, or Laparoscopic Sleeve Gastrectomy vs Usual Care Obesity Management With All-Cause Mortality. JAMA. Jan 16 2018;319(3):279–290. doi: 10.1001/jama.2017.20513 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Encinosa WE, Bernard DM, Chen CC, Steiner CA. Healthcare utilization and outcomes after bariatric surgery. Med Care. Aug 2006;44(8):706–12. doi: 10.1097/01.mlr.0000220833.89050.ed [DOI] [PubMed] [Google Scholar]
  • 30.Aminian A, Zajichek A, Arterburn DE, et al. Association of Metabolic Surgery With Major Adverse Cardiovascular Outcomes in Patients With Type 2 Diabetes and Obesity. JAMA. Sep 2 2019;doi: 10.1001/jama.2019.14231 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Arterburn DE, Telem DA, Kushner RF, Courcoulas AP. Benefits and Risks of Bariatric Surgery in Adults: A Review. JAMA. Sep 1 2020;324(9):879–887. doi: 10.1001/jama.2020.12567 [DOI] [PubMed] [Google Scholar]
  • 32.Kim T, Roslin M, Wang JJ, et al. Body Mass Index as a Risk Factor for Clinical Outcomes in Patients Hospitalized with COVID-19 in New York. Obesity (Silver Spring). Oct 31 2020;doi: 10.1002/oby.23076 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Uccelli M, Cesana GC, De Carli SM, et al. COVID-19 and Obesity: Is Bariatric Surgery Protective? Retrospective Analysis on 2145 Patients Undergone Bariatric-Metabolic Surgery from High Volume Center in Italy (Lombardy). Obes Surg. Oct 31 2020;doi: 10.1007/s11695-020-05085-z [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Data File
NIHMS1732115-supplement-Supplemental_Data_File.pdf (501.8KB, pdf)

RESOURCES