From 2016 to 2020, over 900 000 MBS procedures were performed in inpatient and hospital-owned ambulatory surgical centers in the USA. For adults with or without diabetes, women and persons with class III obesity had a significantly higher incidence of MBS. Adults with obesity and diabetes had nearly threefold higher age- and sex-adjusted incidence of MBS than those with obesity but without diabetes. Overall, SG was the most common MBS surgical approach; however, RYGB was performed more often among adults with than without diabetes. Although the South had the largest numbers of MBS procedures, the highest incidence occurred in the Northeast region compared with other regions.

To our knowledge, this is the first report documenting the population-based incidence of MBS among US adults with obesity. Previous studies reported that the number of MBS procedures increased and then plateaued during 2010−2015, followed by a 22.1% decline in 2019 and 2020 in the USA.7 21 We found that the incidence of MBS among adults with obesity increased about 6% annually from 2016 to 2019 with similar changes over time regardless of diabetes status. Despite no sex differences in efficacy and complications risk,22 patients having MBS are predominantly women.21 In our study, 80% of patients who had MBS were women, those with diabetes having a nearly twofold higher incidence and those without diabetes a nearly threefold higher incidence than men. Given the magnitudes of the IRs (with vs without diabetes) by sex, our data suggest that the relative association of diabetes with MBS was stronger for men than for women.

Consistent with other studies, the highest number of patients with obesity and MBS procedures was in the South region.21 23 However, the Northeast region had the highest incidence of MBS, suggesting geographic differences in utilization and/or access to surgery. A higher proportion of MBS was paid for by private insurance.23 However, the age-, sex-, and obesity level-adjusted incidence of MBS among adults with obesity and public insurance (ie, Medicaid, Medicare) as the expected primary payer was similar to the incidence among adults with obesity who had MBS procedures with private insurance as the expected primary payer. Medicaid coverage of MBS has expanded substantially in the last decade.24 Previous studies suggested that individual state policies, insurance coverage, and economic status may contribute to geographic variation in the use of MBS.25 26 Enhancing bariatric surgery facilities and expanding insurance coverage could help reduce these regional disparities.

Between 2016 and 2020, 5–7% of MBS procedures for patients with diabetes were done in ambulatory settings, which were lower than among patients without diabetes (7–11%). Percentages were lower for patients with diabetes, likely due to the comorbidity profile and the type of MBS procedure performed. SG and laparoscopic gastric banding are now often done in the ambulatory setting, but gastric bypasses are still most commonly done in the inpatient setting, with at least an overnight hospital stay.27 Options for weight loss procedures are expanding, especially with the advent of endoscopic sutured gastroplasty, which has less risk than surgery.28 The advancements in minimally invasive MBS procedures allow for faster recovery and discharge and may lead to higher proportions of MBS in ambulatory settings.29 NASS only includes hospital-owned facilities which comprise only a small fraction of the procedures done in ambulatory settings. Physician and/or corporate-owned ambulatory surgery centers comprise over 70% of the ambulatory surgery market, so our estimates likely under-represent the ambulatory setting MBS procedures.30

A study in the early 2000s found a declining proportion of patients having MBS with no comorbidities, with a proportion of 64% in 2002.13 That trend appears to have continued. Overall, we saw half of patients with no comorbidities, likely due to increased availability of facilities and physicians to perform procedures, especially among those with a lower risk profile.31 In contrast to the previous study using NIS patients with obesity data,13 we see an even distribution of MBS procedures across median household income quartiles. This is encouraging that lower income patients have more access to MBS than previously noted. Although only 3% of MBS cases were performed in hospitals located in non-metro areas, the incidence of MBS in metro and non-metro areas, at least among adults with diabetes, was not significantly different, suggesting that adults in non-metro areas had access to metro hospitals. The discrepancies between proportions and incidences of MBS among adult populations highlight the importance of assessing incidences at population level rather than proportions at patient level when comparing differences across subpopulations.

Previous studies have demonstrated that MBS produces improved glycemic control and leads to higher rates of diabetes remission compared with lifestyle and medical management alone.4 However, the use of MBS in patients with obesity and diabetes did not increase after the ADA lowered the recommended BMI level.32 In the USA, about 42% of adults have obesity and 9% of adults have class III obesity,33 but according to ASMBS, only 1% of eligible patients have MBS.34 Glucagon-like peptide-1 agonists have now been approved for weight management35; this and other medications used for weight loss may also impact MBS rates now and in the future.36

This study is subject to several limitations. First, the information from NHIS for the denominator in incidence calculations is self-reported and subject to recall bias; an under-report of diabetes could overestimate the incidence of MBS in adults with diabetes, and therefore overestimate the association of diabetes with MBS. Second, NHIS respondents could include a small portion of adults with previous MBS, which would result in a slight underestimate of MBS incidence in the US adult population with obesity. Third, federal hospitals are not included in NIS, which excludes a small segment of the US population. Fourth, NASS only includes ambulatory surgeries in hospital-owned facilities, which represent a small fraction of the ambulatory surgery centers performing MBS. This underestimates our total numbers of MBS procedures and the proportion occurring in ambulatory settings. Fourth, the NHIS only has self-reported BMI, which can be lower than measured BMI.37 However, after calibrating self-reported BMI using mean differences of measured and self-reported BMI, we did not detect any substantial change of estimates (online supplemental tables 3 and 4). Fifth, since we did not have information on the expected primary payer from the NHIS,15 the primary payer was derived from insurance status, age, retired status, and disabled status.16 We only classified adults without insurance as self-pay, which could underestimate the number who would have self-paid for an MBS procedure since insurance does not always cover MBS. This classification could overestimate the incidence of MBS among adults who would have self-paid. Finally, because this is a cross-sectional observational study, we are unable to delve into the underlying reasons for the observed trends; further longitudinal studies are warranted, especially those addressing clinical effectiveness of MBS procedures to improve diabetes outcomes or overall health.

Notwithstanding these limitations, this is the first nationally representative study to assess MBS incidence among adults with obesity by diagnosed diabetes status. Our study contributes to understanding the current patterns and recent trends in MBS utilization among US adults with obesity.