Introduction

Bariatric and metabolic surgery is a dynamic and ever-evolving field that has established itself as highly effective therapy for obesity and obesity-related comorbid conditions. Over the past several years, clear trends have emerged. The annualized case volumes between 2015 and 2018 indicate an overall growth rate of 21.9% and account for 760,076 total cases (Table 1) [1]. Significant changes in practice patterns during this time period are demonstrated by a number of well-defined shifts in preferred procedures and approaches. Sleeve gastrectomy (SG) continued to be the most common procedure performed in the USA and grew by 30.4%. This accounts for 62% of all primary bariatric procedures and is performed 2.7 times more frequently than Roux-en-Y gastric bypass (RYGB), the next most common procedure performed. Also, during this time period, placement of laparoscopic adjustable gastric bands declined by 65.9%. While the overall case volume remains low, the rate of duodenal switch (BPD/DS) procedures increased by 115.8%. The vast majority of all bariatric and metabolic surgery cases are performed laparoscopically; however, robotic-assisted surgery, natural orifice transluminal endoscopic surgery (NOTES), and additional alternative approaches are starting to gain popularity (Table 2). In 2018, robotic-assisted procedures accounted for approximately 10% of all cases, nearly twice that of 2015 [1].

Table 1 Frequencies of primary bariatric procedures in the USA, 2015–2018
Full size table
Table 2 Frequencies of bariatric procedures by surgical approach in the USA, 2015–2018
Full size table

This review evaluates contemporary data regarding postoperative weight loss patterns and remission of obesity-related comorbid conditions following bariatric and metabolic surgery, as well as outcomes related to revisional surgery, robotic-assisted surgical platforms, and surgical interventions in adolescent populations. Of particular relevance given the increased frequency of sleeve gastrectomy (SG), recent literature related to postoperative gastroesophageal reflux disease (GERD) and Barrett’s esophagus is also discussed. Except as specifically discussed in the adolescent portion of this review, all publications and data pertain to adult populations.

Weight Loss

As sleeve gastrectomy (SG) has grown in popularity alongside Roux-en-Y gastric bypass (RYGB), several authors have sought to elucidate key differences in the medium and long-term outcomes of both procedures. While observed weight loss at 5 years for both laparoscopic SG and RYGB tends to favor RYGB, differences between the two procedures in this regard do not reliably reach statistical significance [2, 3]. Notably, in a recent randomized controlled trial including 217 patients who underwent SG or RYGB, no statistically significant difference in excess BMI loss at 5 years was identified, although data tended to favor RYGB (61.1% for SG, vs. 68.3% for RYGB, p = 0.22 after adjustment for multiple comparisons) [2•]. Similarly, a clinical equivalence trial between the two procedures including 240 patients who underwent laparoscopic SG or RYGB failed to demonstrate clinical equivalence between the two procedures. While authors cited a slightly higher mean percentage of excess weight loss (%EWL) following RYGB, they were not able to demonstrate a statistically significant benefit to RYGB over SG at 5 years [3•].

While extended data on weight loss outcomes following SG is still lacking, a recent 12-year prospective observational study evaluating RYGB suggests reasonable durability of 2- and 6-year weight loss patterns associated with this procedure [4•]. Among 418 patients in the surgical arm of the study, mean percent change in body weight was found to be − 35% at 2 years, as compared with − 28.0% at 6 years and − 26.9% at 12 years.

For patients with super morbid obesity (BMI > 50) and/or inadequate weight loss after sleeve gastrectomy in the absence of gastroesophageal reflux disease (GERD), biliopancreatic diversion with duodenal switch (BPD/DS) and related procedures remain procedures of choice for maximal weight loss, with %EWL ranging from 70 to 80% at 2 years [5,6,7].

Beyond the type of operation, long-term weight loss following bariatric and metabolic surgery is significantly predicted by short-term changes in energy intake and dietary distribution postoperatively [8]. Specifically, short-term reductions in energy intake (p < 0.001) at 6 months as well as changes in relative proportions of energy from carbohydrates (p < 0.001), fat (p < 0.001), and protein (p < 0.05) were significantly associated with 10-year weight change after bariatric surgery. Men and women with the largest reductions in energy intake at 6 months lost 7.3% and 3.9% more weight, respectively, when compared with study subjects who showed the smallest intake reductions (p < 0.001). Greater weight loss was also observed in patients who favored protein and carbohydrates over fat, and in subjects who favored protein over carbohydrates, than in patients who favored the opposite dietary macronutrient compositions (p < 0.05) [8].

Remission of Obesity-Related Disease

The impact of bariatric and metabolic surgery on obesity-related disease is significant, and for many patients, it is the primary driver for surgical intervention [9] (Table 3). With regard to type 2 diabetes mellitus (T2DM) in particular, remission of disease following surgical intervention is significantly more likely among surgically treated patients than among those who receive no therapy or medical therapy alone [10,11,12]. Although more recent prospective trials and long-term studies question the near-complete remission rates purported by Buchwald’s classic 2009 meta-analysis (95.1% for BPD/DS, 80.3% for RYGB, 79.7% for gastroplasty, and 56.7% for adjustable gastric banding), using contemporary criteria for complete remission, resolution of T2DM does occur in many patients following surgical intervention and is reasonably well-sustained at extended follow-up [3, 4, 7, 12]. A recent meta-analysis examining partial and complete remission of T2DM in patients treated with RYGB reported rates of 52.5% at 5 years, however only 27.5% at 5 years [12]. Among patients undergoing laparoscopic RYGB, baseline disease severity and antidiabetic medical requirement are significantly associated with overall long-term outcomes. In the overall diabetic population, remission of T2DM was observed in 75% of patients at 2 years, as compared with 62% of patients at 6 years and 51% of patients at 12 years. Lowest rates of remission were observed in patients with baseline insulin dependence, followed by those who were reliant on oral antidiabetic medication only, followed by those who did not require medication at baseline (16% at 12 years, versus 56% and 73%, respectively) [4•]. Authors attributed this distribution in outcomes to the variability of pancreatic beta cell function at time of surgery, specifically the greater potential for beta cell recovery in patients not yet requiring insulin. These findings are supported by other recent publications, where duration of T2DM less than 8 years at time of surgical intervention was the main predictor of ultimately achieving satisfactory disease control or remission (HgbA1c% < 6) [11].

Table 3 Obesity-related diseases and comorbid conditions
Full size table

As compared with RYGB, the impact of SG on T2DM remission is less well-studied; however based on recent published evidence, it does not appear significantly inferior to that seen with RYGB at medium-term follow-up [3, 11]. In a recent prospective clinical equivalence trial, complete or partial remission rates of T2DM at 5 years were observed in 37% of patients who underwent SG, as compared with 45% of patients who underwent RYGB (p > 0.99). Similarly, 5-year follow-up results of the STAMPEDE trial found excellent glycemic control (defined as HgbA1C < 6%) of 29% in patients treated with RYGB versus 23% of those treated with SG, without a statistically significant benefit to RYGB observed. Reductions in the rate of insulin use between patients treated with RYGB and SG were also similar (− 35% vs. − 34%, respectively) [11]. With additional long-term data and larger study populations, differences between these two procedures with respect to long-term remission of T2DM may be better understood. BPD/DS is supported in recent literature as the most effective and durable option for treatment of T2DM in surgically treated patients. In their 2015 randomized controlled trial, Mingrone et al. found BPD/DS to be superior to RYGB in sustaining remission of T2DM at 5 years postoperatively—specifically, 53% of patients with baseline T2DM treated with RYGB who experienced remission at 2 years had disease relapse at 5 years, as compared with only 37% of patients treated with BPD/DS [10]. Interestingly, changes in weight were not associated with remission or relapse of T2DM, a finding confirmed by other studies [10, 11]. With regard to cardiovascular disease, bariatric and metabolic surgery has long been associated with improvements in hypertension, hyperlipidemia, and overall cardiovascular health. Particularly in patients with T2DM, several recent publications have highlighted a protective effect of surgical intervention against adverse cardiovascular outcomes. When compared with non-surgically treated controls, surgically treated patients with insulin-dependent T2DM were observed to have significantly higher probability of survival related to non-fatal coronary artery disease at 1-, 5-, and 10-year follow-up (98% vs 89.6%, 92.2% vs 67.6%, and 84% vs 53.1% respectively, log-rank test p < 0.001) [13]. Probability of survival for non-fatal peripheral arterial disease (PAD) was also significantly reduced at 5- and 10-year follow-up (90.5% vs 78.8% and 84.0% vs 53.1% respectively, log-rank test p = 0.007). Outcomes for non-fatal acute myocardial infarction (AMI), stroke, and heart failure (HF) had little to no statistically significant differences between the study populations over the 10 years [13]. These findings were supported in a similar retrospective study, where patients with T2DM who underwent metabolic surgery were significantly less likely to experience a major adverse cardiovascular outcome (MACE) than those managed medically at 8 years (30.8% vs 47.7%, p < 0.001; absolute 8-year risk difference 16.9%; adjusted hazard ratio 0.61). In this analysis, all cause-mortality was also positively impacted (10% vs 17.8%; absolute 8-year risk difference 7.8%; adjusted hazard ratio 0.59) [14].

Additional investigation into the long-term risk of cardiovascular events and need for coronary revascularization in patients with and without T2DM suggests bariatric surgery offers a significant protective effect. In a recent propensity-matched retrospective cohort analysis with a median follow-up of over 6 years, bariatric surgery was associated with a significantly reduced rates of myocardial infarct (1.8% vs 10.0%, RR 0.18), coronary catheterization (1.9% vs 8.8%, RR 0.22), percutaneous coronary intervention (0.4% vs 7.8%, RR 0.05), and coronary artery bypass grafting (0.6% vs 2.3%, RR 0.26) [15].

Bariatric and metabolic surgery presents a uniquely efficacious therapy for patients with obesity and HF. The relative safety and efficacy of bariatric and metabolic surgical intervention in patients with severe HF has been well-documented in the literature by several authors in small patient populations. Surgical interventions both decrease hospital readmission rates significantly and improve overall cardiac function [16, 17]. More recent data suggests patients with a history of bariatric surgery subsequently admitted with HF have a nearly 50% reduction in in-hospital mortality and significantly shorter LOS than patients with morbid obesity and no history of bariatric surgery, as well a propensity score-matched patients with no history of bariatric surgery (in-hospital mortality 0.96% vs 1.86%, p = 0.0013, and 0.96% vs 1.86%, p = 0.0011, respectively; LOS 4.8 ± 4.4 versus 5.7 ± 5.7 days, p < .001, and 4.8 ± 4.4 versus 5.4 ± 6.3 days, p < .001, respectively) [18]. Furthermore, when used as primary prevention, bariatric surgery may significantly reduce the incidence of new HF diagnoses in patients with obesity. As found by a large-scale recent retrospective Swedish cohort study, patients with obesity who undergo RYGB are only half as likely to develop HF as similarly obese patients who undergo intensive lifestyle modification at a median follow-up of 4.1 years (hazard ratio 0.54, 95% CI 0.26–0.82) [19].

Obesity is a well-established risk factor for the development of certain malignancies, including breast, colon, rectum, corpus uteri, esophageal, gallbladder, gastric cardia, renal, liver, meningioma, multiple myeloma, ovary, pancreas, and thyroid [20] (Table 4). Bariatric and metabolic surgery is associated with a reduced risk of subsequent malignancy overall in controlled studies of patients with morbid obesity [21,22,23]. While bariatric surgery alone does not appear to confer any protective effect against malignancy, weight loss at 1 year following bariatric surgery is significantly associated with reduced risk of any cancer in both adjusted and unadjusted models (hazard ratio 0.897, p = 0.005 for every 10% weight loss in adjusted models; hazard ratio 0.876, p ≤ 0.001 for every 10% weight loss in unadjusted models) [22]. This protective benefit appears to increase as time progresses. At extended follow-up (mean 3.5 years), surgically treated patients with severe obesity had a 33% lower hazard of developing any malignancy when compared to those who did not receive bariatric surgery (hazard ratio 0.67, p < 0.001). This effect was more pronounced when analysis was limited to obesity-associated malignancies including postmenopausal breast (hazard ratio 0.58, p < 0.001), colon (hazard ratio 0.59, p = 0.04), endometrial (hazard ratio 0.50, p < 0.001), and pancreatic cancer (hazard ratio 0.46, p = 0.04) [23].

Table 4 Relative risk of obesity-related malignancies in highest body mass index (BMI) patients versus normal BMI
Full size table

Postoperative Sequelae: Reflux, Barrett’s Esophagus, and Feeding Intolerance

A comprehensive summary of common late postoperative complications following bariatric surgical intervention is provided in Tables 5 and 6 [1, 24,25,26,27,28,29,30]. Particularly with the increased prevalence of SG as compared with RYGB and other bariatric surgical interventions, postoperative gastroesophageal reflux disease (GERD) is of increasing concern. Recent long-term retrospective data suggests new-onset GERD symptoms occur in 47.8% of patients at an average follow-up of 8.48 years (range 6.1–10.3 years), despite an associated %EWL of 60% [28]. While these authors reported use of a 34-french bougie for all patients examined, which has since been discouraged, these results still merit attention. Along with worsening GERD symptoms, proton-pump inhibitor (PPI) use increased significantly among the study population: 52% of patients were reliant on PPI therapy at the end of the study period, as compared with only 15% preoperatively (relative risk 2.933, p < 0.0001). Among the 7 patients included in this study who underwent secondary RYGB for GERD, only 57.1% experienced complete resolution of their symptoms following revisional surgery [28].These findings were duplicated in a similar recent retrospective review, where all patients underwent SG completed with a 48-French bougie [29••]. At an average 58-month follow-up, incidence of GERD symptoms, visual analogue scales (VAS) mean score, and PPI use all increased significantly when compared with preoperative values (68.1% vs 33.6%, p < 0.0001; 3 vs 1.8, p = 0.018; 57.2% vs 19.1%, p < 0.0001). Of even more concern, upward migration of the Z line and biliary-like esophageal reflux was present in 73.6% and 74.5% of patients on upper endoscopy, respectively. While no significant correlations were found between reflux symptoms and endoscopic findings, Barrett’s esophagus (BE) was newly diagnosed in 17.2% of patients postoperatively, and a significant increase in the incidence and severity of erosive esophagitis (EE) was observed, up to 59.8% at an average of 66 months postoperatively [29, 31]. Particularly given the poor correlation between visualized esophageal pathology and GERD symptoms, these authors suggest routine endoscopic surveillance of all SG patients may be indicated postoperatively.

Table 5 Postoperative long-term complications and deleterious effects following bariatric and metabolic surgery
Full size table
Table 6 Nutritional deficiencies following bariatric surgery
Full size table

In contrast to the above studies, long-term feeding tolerance in patients undergoing RYGB, biliopancreatic diversion (BPD), and duodenal switch (DS) appears fairly robust [32]. In a cross-sectional study of 196 patients who underwent one of these three bariatric procedures, mean food tolerance score (FTS) was 24.6, 24.0, and 23.7 for RYGB, BPD, and DS at mean follow-up of 87.9 months (scale of 1–27, 1 being horrible and 27 perfect). Alimentary satisfaction was observed in 73.3% of patients overall. No significant differences were found between the 3 procedures for FTS, alimentary satisfaction, or frequency of vomiting [32].

Robotic Surgery

Robotic-assisted surgical platforms are an increasingly utilized adjunct to traditional laparoscopy in bariatric and metabolic interventions worldwide. While surgeon comfort, improved technical dexterity, and greater ease of use in patients with high body mass index (BMI) are purported as selling points of the robotic platform, the majority of outcomes research to date has failed to demonstrate a consistent benefit of robotic-assisted bariatric and metabolic surgery over traditional laparoscopy. A recent extensive retrospective review of 77,991 patients who underwent primary RYGB (7.5% robotic-assisted) and 189,503 patients who underwent primary SG (6.8% robotic assisted) found that on the whole, aggregate bleeding complications and transfusion requirements were less in the robotic-assisted cohorts (for example 1.0% vs 0.5%, p = 0.0005 for transfusion and 0.8% vs 0.4%, p = 0.03 for aggregate bleeding, for RYGB in 1:3 matched patient factor analysis) [33]. Outcomes were otherwise similar for both SG and RYGB, with several exceptions. For RYGB, lower mortality rates (p = 0.05) and surgical site infections (SSIs) (p = 0.0006) were observed, despite longer operative times using the robotic platform (151.9 min vs 114.6 min, p < 0.0001). For SG, a longer length-of-stay (LOS) was observed (p < 0.0001) along with higher rates of conversion (p > 0.0001), 30-day intervention (p = 0.01), operative drain placement (p < 0.0001), sepsis (p = 0.01), and organ space SSI (p = 0.0002) in the robotic-assisted cohort. Despite statistical significance, the majority of these differences were fairly minute, leading study authors to conclude potential benefits of robotic-assisted RYGB and SG found most likely represented a complex interplay between surgeon experience and the surgical platform used [33].

With regard to surgical education, another retrospective review focused on robotic-assisted SG utilizing residents for tissue exposure concluded that while significantly increased cost supply was associated with the robotic platform as compared with its laparoscopic equivalent, preliminary morbidity rates and operative times were similar to historic laparoscopic controls. No differences in operative times between BMI cohorts 50–59 kg/m2 and 30–49 kg/m2 were observed, suggesting a robotic platform may be beneficial in patients with higher BMIs [34].

Revisional Surgery

As reasons for weight recidivism and other common postoperative sequelae such as GERD become better understood with time, patients who experience these after effects following a primary bariatric surgical intervention frequently present for revisional surgical consideration. Large-scale prospective data in this area is generally lacking. However, in a single-center, retrospective review of 534 patients presenting for revisional evaluation, 64% sought evaluation for weight regain (true of all procedures except BPD/DS), 26% for dysphagia, and 21% for lap band-related complications [35]. Seventy-eight and 88% of patients presenting with band slippage or insufficient weight loss, respectively, were medical tourists, leading study authors to conclude poor follow-up played a substantial role in these postoperative issues. Patients with uncontrolled psychosocial and medical issues were also much more likely to struggle with weight recidivism despite an anatomically successful surgery. Patients who underwent revisional surgery experienced higher rates of postoperative complications than those undergoing primary surgery (41% vs 15%, p < 0.0001), most commonly wound infections (24%). Revisional surgeries took on average 0.4 h longer than primary surgeries, and patients remained in the hospital for a median of 2 days longer, both statistically significant differences. Revisional surgery (most commonly RYGB) was successful in assisting weight loss, comparable to that seen in the primary surgery group and significantly more than in patients managed with lifestyle modification alone (%EWL 61.2%, p < 0.0001) [35].BPD/DS may present a valuable revisional option for selective patients with super morbid obesity and/or weight regain following SG in the absence of reflux symptoms [5, 36]. A recent expert-consensus paper examining revisional BPD/DS cited a vast majority of experts (96.7%) as supporting this position, specifically for a 2-stage operation. Ninety percent of authors also favored RYGB over BPD/DS for patients with weight regain after SG with associated reflux, and/or enlarged fundus [5]. While BPD/DS remains limited in general applicability by technical difficulty and operative duration, several simplified versions of the procedure such as single-anastomosis duodenoileostomy (SADI) may lead to similar postoperative outcomes with increased ease of completion [6, 36]. Indeed, this concept is well-supported in a recent retrospective cohort study where patients with failed SG underwent SADI or RYGB, primarily based on the absence or presence of reflux symptoms, as well as desire for additional weight loss. Patients who underwent SADI experienced 8.7%, 12.4%, and 19.4% more weight loss compared to those who underwent RYGB at 6, 12, and 24 months postoperatively (all p < 0.001), while all patients treated with revisional RYGB for GERD or dysphagia all experienced improvement in these symptoms. Similar rates of complications and nutritional deficiencies were observed in both groups [36].

Adolescent Bariatric and Metabolic Surgery

Since the hallmark publication by Inge et al. in 2014 examining outcomes from the Teen-Longitudinal Assessment of Bariatric Surgery (Teen-LABS) study which cited similar outcomes and postoperative complication rates in adolescent and adult populations undergoing bariatric and metabolic surgery, several subsequent publications have further delineated specific and significant beneficial effects towards youth-onset T2DM [37, 38]. Indeed, youth-onset T2DM now accounts for a substantial percentage of new pediatric diabetes cases annually and appears to be significantly more aggressive than its adult counterpart. Nearly 50% of teens with T2DM progress to insulin dependence after a median of 11 months, suggesting rapid loss of pancreatic beta-cell function which outpaces that typically seen in adult disease [38•]. In a secondary analysis of the Teen-LABS data released in 2018, in which surgical outcomes related to diabetes were compared with a population of adolescent patients with T2DM managed with lifestyle modification and/or medical therapy, surgical therapy was vastly superior. While patients who underwent metabolic surgery experienced greater weight loss (BMI − 29.0% vs + 3.7%) and improvement in HgbA1C (6.8 to 5.5%, vs 6.4 to 7.8%) at 2 years than those managed medically, adolescents in the surgical arm of the study also had a greater than expected improvement in T2DM (95% with HgbA1C < 6.5%) when compared to adult populations, despite similar operations, weight loss, and definitions of disease [38•].

With regard to other obesity-related comorbid conditions, bariatric and metabolic surgery in adolescents is also associated with significant improvements in blood pressure, dyslipidemia, serum inflammatory markers, and abnormal kidney function not seen in medically treated counterparts, as well as functional mobility and musculoskeletal pain [38,39,40]. Younger adolescent patients appear more likely than older adolescent patients to resolve dyslipidemia, while girls appear more likely than boys to resolve elevated blood pressure [39].

In reflection of these findings, a recent policy statement released by the American Academy of Pediatrics (AAP) advocated for increased access for pediatric patients with severe obesity (defined as BMI ≥ 35 or ≥ 120% of the 95th percentile for age and sex, whichever is lower) to multidisciplinary programs with high-quality pediatric metabolic and bariatric surgical capabilities. While the majority of pediatric patients undergo bariatric surgery as adolescents (defined as age 13 to 18 years), available data does not support a lower age limit for weight-loss surgery in youth. As such, the AAP supports individualized patient-centered care and avoidance of “unsubstantiated lower age limits” when considering surgical referral [41].

Conclusion

Bariatric and metabolic surgical intervention remains a highly effective therapy for obesity and obesity-related comorbid conditions. While SG and RYGB produce similar weight loss patterns and remission of obesity-related disease at medium-term follow-up, highly prevalent rates of GERD and dysphagia following SG may ultimately limit the long-term success of this procedure. Routine surveillance endoscopy may also be indicated in SG, given high rates of esophagitis and Barrett’s esophagus found in these patients postoperatively. When utilized as therapy for T2DM, bariatric and metabolic surgery is most effective in adolescent patients and adult patients who are non-insulin dependent. Surgery is also effective for the prevention of coronary artery disease and related interventions in both diabetic and nondiabetic populations. Robotic-assisted laparoscopic platforms have generally equivalent outcomes to laparoscopic intervention, with a potential benefit in high BMI patients. Revisional surgery is commonly performed for inadequate weight loss or reflux/dysphagia following primary surgery, with outcomes characterized by higher complication rates and longer inpatient LOS. Bariatric and metabolic surgical intervention in adolescent patients with morbid obesity and obesity-related comorbidities is safe and effective, with overall similar outcomes to those seen in adult populations. With regard to youth-onset T2DM in particular, outcomes related to metabolic surgical intervention outpace those seen in adult populations.