Treatment of Obesity: Weight Loss and Bariatric Surgery

Abstract

This review focuses on the mechanisms underlying, and indications for, bariatric surgery in the reduction of cardiovascular disease (CVD) as well as other expected benefits of this intervention. The fundamental basis for bariatric surgery for the purpose of accomplishing weight loss is the determination that severe obesity is a disease associated with multiple adverse effects on health which can be reversed or improved by successful weight loss in patients who have been unable to sustain weight loss by non-surgical means. An explanation of possible indications for weight loss surgery as well as specific bariatric surgical procedures is presented, along with review of the safety literature of such procedures. Procedures that are less invasive or those that involve less gastrointestinal rearrangement accomplish considerably less weight loss but have substantially lower perioperative and longer-term risk.

The ultimate benefit of weight reduction relates to the reduction of the co-morbidities, quality of life and all-cause mortality. With weight loss being the underlying justification for bariatric surgery in ameliorating CVD risk, current evidence-based research is discussed concerning body fat distribution, dyslipidemia, hypertension, diabetes, inflammation, obstructive sleep apnea and others.

The rationale for bariatric surgery reducing CVD events is discussed and juxtaposed with impacts on all-cause mortalities. Given the improvement of established obesity-related CVD risk factors following weight loss, it is reasonable to expect a reduction of CVD events and related mortality following weight loss in populations with obesity. The quality of the current evidence is reviewed and future research opportunities and summaries are stated.

Indications for Bariatric Surgery

The fundamental basis for bariatric surgery for the purpose of accomplishing weight loss is the determination that severe obesity is a disease associated with multiple adverse effects on health which can be reversed or improved by successful weight loss in patients who have been unable to sustain weight loss by non-surgical means. The criteria for surgical intervention were established by a NIH consensus panel in 1991¹. Failure of medical treatment to accomplish sustained weight loss is common among persons with severe obesity. The biologic factors involved in the limitations associated with maintaining weight loss are powerful^2,3. Intense lifestyle intervention can produce averages of approximately 10% at 1 year and maintain weight loss at 5.3% over 8 years. The weight loss accomplished is highly variable but is sufficient to accomplish improvement in medical and comorbidity control⁴. Pharmacotherapy may enhance short-term as well as longer-term weight loss⁵. Specific criteria established by the NIH consensus panel indicated that bariatric surgery is appropriate for all patients with BMI (kg/m²) >40 and for patients with BMI 35-40 with associated comorbid conditions. These criteria have held up over the ensuing 24 years to the present, although specific indications for bariatric/metabolic surgical intervention have been identified for persons with less severe obesity, such as persons with BMI 30-35 with type 2 diabetes. The indications for bariatric surgery are evolving rapidly to consider the presence or absence of comorbid conditions as well as the severity of the obesity, as reflected by BMI⁶.

Obesity-related comorbidity is defined as conditions either directly caused by overweight/obesity or known to contribute to the presence or severity of the condition. These comorbid conditions are expected to improve or go into remission in the presence of effective and sustained weight loss.

Obesity-related comorbid conditions are listed in Table 1

Table 1.

Obesity Comorbid Conditions

Premature Mortality
Cardiovascular
- Hypertension
- Atherosclerotic CVD, myocardial infarction, stroke
- Congestive heart failure
- Cardiac arrhythmias
Metabolic
- Type 2 Diabetes, prediabetes
- Dyslipidemia
- Non-alcoholic Fatty Liver Disease (NAFLD)/Steatohepatitis
- Inflammation
Pulmonary
- Obstructive Sleep Apnea
- Asthma
Musculoskeletal
- Degenerative Arthritis
- Immobility
- Pain
Reproductive
- Polycystic Ovarian Syndrome (female)
- Infertility
- Sexual Dysfunction
Genitourinary
- Impaired Renal Function
- Nephrolithiasis
- Stress Urinary Incontinence
Central Nervous System
- Impaired Cognition
- Headache
- Pseudotumor Cerebri
Psychosocial
- Impaired Quality of Life
- Depression
- Other Psychopathology
Cancer

The requirements for patient selection include the BMI criteria described above and failure of medical therapy. Specific criteria regarding designation of the failure of medical therapy have not been formalized but generally include treatment in a variety of medically supervised settings. An understanding or insight into the pathogenesis of obesity and the requirement to reduce energy intake substantially if major weight loss is to be achieved is a requisite⁷. Candidates for bariatric surgery must be assessed for appropriate surgical risk, including the presence of cardiovascular, pulmonary other system disease and control of these comorbid conditions. These principles apply to surgical procedures in general. It is entirely possible, for example, that patients with an exceedingly high risk profile for cardiovascular disease will have experienced end events that indicate that perioperative risk is excessive and the likelihood of reversing cardiovascular disease by improving the risk profile is unlikely to be successful. However, examples of the most severely obese patients whose perioperative risk may be improved by weight loss include patients with congestive heart failure, related anasarca, respiratory failure and inability to ambulate.

Pre-operative psychological assessment is commonly done to identify patients who require preoperative intervention or disqualification altogether. Active substance abuse is a standard contraindication to surgery. Although a requirement for mandatory preoperative weight loss among all patients is not justified by published literature, individual patients deemed to be at exceedingly high risk due to the severity of obesity and its comorbid conditions are appropriate in selected cases. The literature surrounding psychological evaluation and its likelihood to predict success is evolving⁸. Psychological assessment prior to bariatric surgery may identify patients with psychopathology such as major depression, binge eating disorder, substance abuse, among others that may impact the decision to proceed with surgery or indicate referral for further preoperative assessment and intervention⁹. In addition, psychological assessment may contribute to predicting postoperative weight loss^9-11.

Specific Bariatric Surgical Procedures

Surgical procedures in the past have been considered to function as restrictive in which the size of the gastric pouch is greatly reduced, malabsorptive in which malabsorption of nutrients contributes to weight loss, and a combination of restrictive and malabsorptive components. It is now clear that this construct is an oversimplification and, to some extent, inaccurate. There is ample evidence that neural and endocrine signaling pathways affecting eating behaviors, reduction of appetite, satiety, energy intake, and possibly physical activity are all operative to a variable extent.

Roux-en-Y Gastric Bypass

Roux-en-Y gastric bypass was developed by Mason in the 1970's in response to unacceptable complication rates that followed ileojejunal intestinal bypass, a procedure which resulted in malabsorption, diminished food intake, and substantial weight loss with its associated benefits but unacceptable complication rates¹². In this procedure, the stomach is transected creating a gastric pouch of approximately one ounce capacity. A Roux-en-Y gastrojejunostomy is done, thus diverting ingested nutrients from the body of the stomach, duodenum, and proximal jejunum. The vagal trunks are not disturbed but a variable number of branches to the body of the stomach are divided in the process of dividing the stomach. Associated endocrine changes are described below. While malabsorption of energy-containing nutrients is minimal, if any, malabsorption of calcium, iron and vitamin B12 and possibly other micronutrients occurs.

Sleeve Gastrectomy

In this procedure, approximately 80% of the body of the stomach is resected, creating a tubular stomach based on the lesser curvature of the stomach. No gastrointestinal to small intestine anastomosis is required. Although some restriction on food intake may occur, gastric emptying is accelerated.

Biliopancreatic diversion with duodenal switch

This is a more complex procedure in which a sleeve gastrectomy is done. An anastomosis between the proximal duodenum and bypassed intestine is accomplished, thereby creating a degree of malabsorption of nutrients. This procedure is infrequently performed due to higher incidence of short- and long-term complications.

Implantation of Devices

Adjustable Gastric Banding

An adjustable gastric band is placed about the proximal stomach to constrict the size of the gastric pouch and outlet. The rate of gastric emptying can be adjusted by a balloon connected to a subcutaneous port.

Intermittent vagal blockade

In this procedure, leads are placed about the vagal trunks at the diaphragm to produce intermittent vagal blockade. Weight loss occurs by reduction of appetite and establishment of early satiety. The intermittent blockade is hypothesized to avoid neural adaptation as occurred in the past with truncal vagatomy. A device for this purpose has been approved by the Food and Drug Administration¹³.

Gastrointestinal Endoscopic Devices

While several endoscopically placed devices or suturing procedures are under development, placement of gastric balloon(s) has recently been approved by the Food and Drug Administration¹⁴.

Bariatric Surgery Safety

While the benefits of weight loss among individuals with severe obesity, particularly those with comorbid conditions, are unquestioned, these benefits must be considered in the context of surgical complications. In the past, complications including perioperative mortality were as much as tenfold more frequent than occur at the present time. For example, a population based study by Flum and Dellinger reported 2% mortality following gastric bypass, considerably higher than 0.5% commonly reported by those surgeons who chose to report their outcomes¹⁵. In response, the bariatric surgical community enacted a number of changes to result in this improved safety record. Included was the identification of the importance of surgeon and center experience, the establishment of pathways, care protocols, and quality initiatives and incorporation of all of these aspects of care into an accreditation of centers program, presently administered by the American Society for Metabolic and Bariatric Surgery and the American College of Surgeons¹⁶. The transition to laparoscopic methodology occurred during the same time period and also contributed to the improved safety.

Incomplete retention or follow-up in reported clinical series has been a limitation to interpretation of registry-based reports on bariatric surgical safety¹⁷. The multi-center bariatric surgery research consortium funded by NIH, known as LABS, however achieved 100% 30-day follow-up among 2,458 participants¹⁸. The LABS Consortium reported 30-day mortality was 0.3% in all patients, 0.2% for laparoscopic gastric bypass (RYGB), 2.1% for open gastric bypass. There was no mortality among LAGB patients. A serious complication occurred in 4.1% of all patients, 4.8% laparoscopic gastric bypass, 7.8% open gastric bypass, 1.0% LAGB. Patient factors in this study that predicted major complication include extremes of BMI, obstructive sleep apnea, inability to walk 200 feet and a history of deep vein thrombosis (DVT). Factors in other studies include age, gender (male), other comorbidities, and smoking¹⁹. Provider factors predicting complications include surgeon and center experience²⁰. Registries report slightly lower perioperative mortality as well as data regarding less severe complications such as wound infection or dehydration, although among patients with 80-85% 30-day follow-up^17,20. These mortality and complication rates compare very favorably to multiple commonly performed surgical procedures such as coronary bypass graft, arthroplasty, cholecystectomy and hysterectomy^21,22.

Mid- and longer-term complications have been well described although determination of their incidence is limited by progressively greater numbers of patients lost to follow-up²³. These include intestinal obstruction, marginal ulcer, ventral hernia, and gallstones. Metabolic complications reported include nephrolithiasis and hypoglycemia. Mineral and vitamin deficiencies as well as weight regain are reported in variable numbers of patients. Reports of micronutrient deficiencies vary substantially as follows: Iron, 33-55%; calcium/vitamin D, 24-60%; vitamin B12, 24-70%; copper, 10-15%; thiamine, <5%²⁴. Established guidelines recommend routine nutrient supplementation to include multivitamins, iron, minerals, calcium, and vitamin D²⁵.

Complications specific to LAGB placement continue to occur in the longer-term at approximately 2% per year. These long-term complications include erosion of the gastric wall by the band, slippage or herniation of the body of the stomach thereby creating obstruction within the band, and complications of the port, including infection. Combined with disappointing long-term weight loss (see below), application of LAGB in the US and Europe has diminished dramatically recently²⁶.

Regarding gastrointestinal endoscopic devices, the literature looks promising but does not have long-term data¹³. Weight loss is modest while the device safety is good. The safety record of the FDA-approved device is excellent but the sustainability of the long-term weight loss following the approved 6 month intervention remains to be determined²⁷.

In summary, both perioperative and long-term complications occur following all bariatric surgical procedures. Multiple steps have been taken in the recent years to reduce perioperative mortality to the presently reported minimum comparable to other commonly performed surgical procedures. Longer-term complications requiring reoperation or micronutrient deficiencies require careful surveillance and prompt intervention. These complications are generally judged to occur with sufficiently low frequency and severity so as to not constitute a contraindication to the performance of bariatric surgery in general.

Weight loss following bariatric surgery

Weight loss following bariatric surgery has been studied and reported both short- and longer-term following all surgical procedures undertaken, as weight loss is the primary objective of bariatric surgery. Mean weight loss is uniformly reported. It is important to recognize, however, the high variability of weight loss following apparently standardized operative procedures such as RYGB or LAGB²⁸. Following RYGB, the LABS consortium reported similar and rapid weight loss 6 months following surgery by stratifying of weight loss into five separate trajectories ranging from 12% total body weight (TBW) loss to 45% TBW 3 years following surgery. Similarly, for LAGB, trajectories are identified for most but not all patients 1 year following surgery. Factors involved in the high degree of variation of weight loss have been examined and reported but do not fully explain the extent of the variability. Predictors of weight loss vary among several reports and include both patient and provider factors. These factors include but are not limited to the presence of specific comorbid conditions such as diabetes, gender, age and behavioral variables, including physical activity and eating behaviors¹⁰. Weight loss following RYGB at years 1, 2, and 3 reported in the 30-35% TBW range²⁹. Initial reports of weight loss following LAGB in Australia suggested weight loss was similar to that seen following RYGB. Data from the US as well as Europe, however, have not confirmed comparable weight loss following LAGB, closer to 15.9% TBW at 3 years²¹. As noted earlier, this lesser weight loss, compared to RYGB, has led to a substantial reduction in the application of LAGB as treatment for severe obesity. The weight loss following biliopancreatic diversion/duodenal switch tends to be slightly greater than that following RYGB while weight loss following sleeve gastrectomy is comparable or is slightly less than RYGB in several reports^30-33. Those studies with non-surgical comparator groups, primarily the Swedish Obese Subjects trial and a prospective clinical trial with a population base comparator from Utah, indicate that the non-surgical patients do not experience long-term weight loss. This is not unexpected, given the requirement that patients selected for surgery undergo and fail medical treatment prior to selection for surgical intervention. Longer-term follow-up has been reported by Pories as well as the Swedish and Utah studies. All show rapid weight loss during the first 12 months following RYGB followed by modest regain of weight until approximately Year 3-5. Following Year 3-5, weight loss tends to be maintained in the 30% TBW range^34-36. Thus, it is well established that maintenance of weight loss following RYGB at 10-20 years is maintained.

In general, procedures less invasive or those that involve less gastrointestinal rearrangement such as LAGB, vagal blocking, and endoscopic procedures such as balloon placement accomplish considerably less weight loss but have substantially lower perioperative and longer-term risk. A research need is a more effective determination of the likely weight loss that will be achieved following any of these interventions including lifestyle intervention and medication, and the amount of weight loss needed to achieve a specific response such as improved control or remission of a specific comorbid condition. At such time, as a more accurate identification of the weight loss required to achieve a specific clinical outcome and the relative risk involved is determined, it will be possible to more accurately identify appropriate candidates for specific procedures, taking into account the expected weight loss and risk profiles.

Expected benefit on CVD risk factors

The ultimate benefit of weight reduction, whether medical or surgical, relates to the reduction of the co-morbidities, quality of life and all-cause mortality. Despite the importance of assessing these risks and taking steps to implement effective medical management with variable success³⁷, surgery has proven to be more effective^35,36. To be covered in this section are the following: body fat distribution, dyslipidemia defined as hypertriglyceridemia and/or low HDL cholesterol (HDL-C) and variably increases in LDL cholesterol (LDL-C), hypertension and prehypertension, insulin resistance, diabetes and prediabetes, non-alcoholic fatty liver disease (NAFLD), inflammation [high sensitivity C-reactive protein (hsCRP), interleukin-6 (IL-6), white blood cell count (WBC), oxidized LDL, intracellular adhesion molecule-1 (ICAM-1) and adiponectin], vascular reactivity and obstructive sleep apnea.

Body fat distribution

The relationship of obesity to CVD events relate in part to alterations in body fat distribution, i.e. increased central/visceral vs. subcutaneous/peripheral, the so-called metabolic syndrome phenotype^38-40. This distribution of excess body fat relates to an excess delivery of free fatty acids (FFA) to the liver wherein defects in insulin action result with subsequent impact on other components of the metabolic syndrome, i.e. dyslipidemia, glucose intolerance, NAFLD and inflammation among others⁴¹. With medical weight loss, percentage reductions in visceral adipose tissue are similar to or exceed other adipose tissue depots but this relative benefit is somewhat reduced with more weight loss⁴².

In general, in patients without diabetes the relative amounts of loss of visceral adipose tissue from 3 to 12 months post-bariatric surgery are similar or greater than the percentage loss of total or subcutaneous adipose tissue^43-45, but at 24 months were variably greater in the visceral depot^46,47. Following LABG, a preferential mobilization of visceral fat was observed at 2 and 6 months, as compared with total and subcutaneous AT; but this outcome was reserved only for patients with excessive amounts of visceral adipose tissue before surgery and this preferential visceral fat reduction occurs only in those⁴⁷. When changes in body composition were compared after malabsorptive biliointestinal bypass (BIBP) and restrictive LAGB during a 4-year follow-up, the effects of BIBP were greater on total fat loss and trunk fat⁴⁹. When omentectomy accompanies a laparoscopic Roux-en-Y gastric bypass (RYGB) procedure, changes in glucose homeostasis, lipid levels, and adipokine profile at 90 days postoperatively have been variably reported^50,51.

Dyslipidemia

The dyslipidemia of obesity reflects mostly the insulin resistant metabolic environment that accompanies excess body fat. This includes hypertriglyceridemia, lower levels of HDL-C, variable increases in apolipoprotein B and VLDL-C, and small dense LDL and HDL⁵². Although LDL-C can be increased in moderately to severely obese patients, this is not nearly as prevalent as the aforementioned lipid and lipoprotein abnormalities. In a meta-analysis of 75 papers in which follow-up lipids were measured up to 4 years post RYGB, baseline and follow-up levels of LDL-C were reported in 48 studies, and baseline LDL-C was123± 7 mg/dL⁵³. Although heterogeneity among studies for LDL-C and all other lipids was high, subgroup analyses revealed reductions in LDL-C by intervals of 1 month up until 4 years [standard mean difference (SMD) −1.31 to −0.52, 95^th% confidence intervals (CI), p<0.00001]. HDL-C levels were assessed in 47 studies. Herein, a time dependent trend was noted. At an interval up to 6 months, no significant change in HDL-C was seen however by 12 months an increase was seen (SMD +1.10, +0.57 to +1.63 95 % CI, p<0.0001), an effect maintained through all subsequent time points assessed including at 4 years. Plasma triglycerides were examined in 55 studies with no change at 1 month but highly significant effects of RYGB on triglycerides were seen up to 4 years (SMD –0.57, 95 % CI, –0.76 to –0.37,p<0.00001).

From another large series that included 73 studies in which patients were examined for nearly 4 years, metabolic surgery produced a decrease in LDL-C from 116 to 90mg/dL and triglycerides from 188 to 127mg/dl, and an increase in HDL-C from 46 to 55mg/dL⁵⁴. When data from the LABS-2 study were examined the prevalence of dyslipidemia improved at 3 years post RYGB in 62% of patients²⁸ and fasting hypertriglyceridemia (>200 mg/dl) remitted in 86% patients, while low HDL-C (<40mg/dl) in 86%²⁴. Important to consider is that all of these studies did not report data about the use of lipid-altering medications post-surgery⁵⁵.

The reduction in LDL-C up to 2 years post metabolic surgery, however, only appears to occur for operations with more weight reduction, i.e. RYGB or biliopancreatic diversion vs. sleeve gastrectomy or LAGB although increases in HDL-C and reductions in triglycerides can occur with all^56-59. Plasma levels of the pro-atherogenic lipoprotein, lipoprotein (a), are not changed after metabolic surgery^60,61.

Hypertension

Obesity is often associated with hypertension (BP>140/90) and in the Edmonton Obesity Staging System, 98% of 5787 obese patients had at least one comorbidity and hypertension was present in 76%⁶². Although difficult to assess from many publications that cite population statistics, the prevalence of hypertension in the severely obese population is ~65%⁶³, not that different from the prevalence in Edmonton. However, from the most recent data from NHANES (2010), 52% of American subjects with a BMI ≥35 kg/m² had treated or untreated hypertension vs. 43% with a BMI ≥30 but ≤35 kg/m^{2 64}. Now, how effective is metabolic surgery in correcting this common co-morbidity of excess body fat?

The effects of metabolic surgery on the prevalence of hypertension are variable, procedure-related and time-dependent. During the active weight loss phase blood pressure decreases and anti-hypertensive drugs are often discontinued⁶⁵. However, after weight stabilization the results are less clear, perhaps related to the duration of hypertension pre-operatively. In a systematic review and meta-analysis of 21 studies using a variety of surgical approaches reduced the relative risk of hypertension at intervals between 24-50 months by 46±8% and hypertension risk reached a nadir when BMI was reduced by 10 kg/m^2,66. Data from LABS-2 demonstrated that cohort persistent remediation from hypertension at 3 and 6 years was nearly 40%²⁸ and the Utah-Obesity study demonstrated a 2 and 6 year relative risk of remission of hypertension of 8.2 and 2.90, respectively⁶⁷. However, the Swedish Obesity Study revealed recidivism of hypertension at 6-8 years of follow-up with no significant difference from baseline⁶⁸. Whether or not this relates to permanent changes in the arterial wall based on years of hypertension pre-operatively remains unclear.

Diabetes and prediabetes

The last decade has been one to not only document the benefit of metabolic surgery in patients with type 2 diabetes and glucose tolerance, but produce sufficient data from randomized controlled trials to render metabolic surgery as an option for the treatment of type 2 diabetes. The most recent update and convincing study is from the Rubino group that carried out an open-label, randomized controlled trial to compare medical or surgery by RYGB or biliopancreatic diversion in 60 patients aged 30-60 years with a BMI of 35 kg/m² or more and a history of type 2 diabetes of at least 5 years, and 53 completed a 5 year follow-up⁶⁹. Diabetes remission was defined at 2 years, as an HbA1c concentration of ≤6·5% and a fasting plasma glucose of ≤5·6 mmol/L without pharmacological treatment for 1 year. Overall, 50% of the 38 surgical patients sustained a diabetes remission at 5 years, compared with none of the 15 medically treated patients (p=0·0007). A similar medical vs. surgical trial for the treatment of type 2 diabetes was carried out in 150 patients with uncontrolled type 2 diabetes (HbA1c – 9.3±1.5%) by Schauer et al. with a 91% follow-up at 3 years⁷⁰. The primary end point of an HbA1C of ≤6.0% was met by only 5% of the medical group vs. 38% of RYGB patients and 24% with sleeve gastrectomy, all in the setting of much less use of glucose-lowering medications in the surgical groups than in the medical-therapy group. As expected the amount of weight reduction was only ~4.0% in the medical group vs. 22-24% in the surgical groups. In 4 randomized trials wherein RYGB was further compared with sleeve gastrectomy, there was no significant difference between procedures with the reduction in HbA1c or fasting plasma glucose or the in change in weight, BMI or the number or type of drugs used to treat type 2 diabetes^30,31. Overall, the amounts of weight loss have been a definite predictor of diabetes remission. When LABG vs. RYGB was controlled for weight loss, RYGB was clearly superior to LABG in the induction of remission. It has, therefore, been demonstrated that both weight loss and RYGB contribute to the superior remission of diabetes following gastric bypass compared to LABG⁷¹. Recurrence of diabetes following induction of remission may occur in as many as 58% of patients 15 years after bariatric surgery, predominantly in those who had gastric banding⁷². Further research is needed to determine the long-term benefit of gastric bypass and sleeve gastrectomy in patients with type 2 diabetes.

A related glucose-centric topic is the prevention of type 2 diabetes in severely obese patients by metabolic surgery. Using a systematic review and meta-analysis wherein medical vs. surgical approaches were examined in patients with impaired fasting glucose or impaired glucose tolerance, non-surgical approaches reduced new onset type 2 diabetes by 14-56% using a number of different interventions whereas bariatric surgery was 90% effective⁷³. The factors that were associated with effectiveness were weight loss, young age and fasting insulin levels. In a systematic review and meta-analysis that extended the outcome to an admixture of patients with normal glucose tolerance, impaired fasting glucose with or without impaired glucose tolerance at baseline, medical strategies reduced new onset type 2 diabetes from 15-63% based on the various types of intervention whereas metabolic surgery was 84% successful⁷⁴.

Non-alcoholic fatty liver disease (NAFLD)

NAFLD is diagnosed by either imaging or histology with no alternative explanation for fatty liver including alcoholic liver disease. Of interest many times liver transaminases are normal with biopsy-proven NAFLD⁷⁵. The prevalence varies around the world but is typically more common in Western nations (20-40%) and using proton magnetic resonance spectroscopy in 2,287 subjects from a multiethnic, population-based sample (32% white, 48% black, and 18% Hispanic) the prevalence was highest in Hispanics (45%) with 33% in whites and 24% in blacks⁷⁶. This high prevalence is similar to that of the metabolic syndrome and reflects the insulin resistance related to both. NAFLD is present in ~90% of patients who qualify for metabolic surgery and approximately 33% of these have biopsy proven non-alcoholic steatohepatitis (NASH)⁷⁷, a precursor of more serious liver disease including cirrhosis and need for transplantation. In fact patients with NASH by biopsy have an increased risk of death within a median follow-up of 10.2 years after bariatric surgery⁷⁸. Although medical management including weight loss, pioglitazone, vitamin E, pentoxifylline, ursodeoxycholic acid and most recently liraglutide has had variable effect, bariatric surgery has proven more effective⁷⁹.

Data from the Swedish Obesity Study, a non-randomized study of 3,570 obese subjects which compared several types of bariatric operations including RYGB and gastric banding to medical management for up to 10 years, revealed a reduction in ALT at 2 years that was maintained at 10 years vs. the non-surgical control group⁸⁰. Retrospective or cohort studies in general have demonstrated that improvement of NAFLD is more likely after RYGB than after other interventions, however, data at present are insufficient to indicate reductions in liver-specific mortality, liver transplantation, or quality of life⁸¹. One study examined the impact of metabolic surgery on NAFLD in 381 patients at baseline with second and third biopsies at 1 and 5 years, post-operatively⁸². At both 1 and 5 years, major reductions in steatosis and ballooning degeneration ensued with no change in inflammation. And in the 27% of patients diagnosed with NASH, steatosis and ballooning also improved after 5 years, but fibrosis and inflammation did not. In some patients fibrosis actually increased, an outcome that was associated with more severe obesity and insulin resistance. Perhaps the largest study to examine the benefit of bariatric surgery was carried out in 1236 obese patients (BMI - 48.4±7.6 kg/m²) wherein RYGB (n=681) was compared to adjustable gastric banding (n=555)⁸³. At baseline, NAFLD was present in 86%, and as severe in 22% patients. In general RYGB patients had a higher BMI and more severe NAFLD. All NAFLD parameters improved after surgery (P<0.001) but improved more after RYGB than after gastric banding, and the amount of weight loss related to this benefit. Overall, metabolic surgery appears to be the best treatment for NAFLD in patients who qualify for surgery.

Inflammation

Systemic inflammation is routinely assessed by non-specific metrics such as the erythrocyte sedimentation rate (ESR) or high sensitivity C-reactive protein (hsCRP) but other biomarkers can also be utilized, i.e. interleukin-6 (IL-6), white blood cell count (WBC), oxidized LDL, intracellular adhesion molecule-1 (ICAM-1) and adiponectin. Following medically-managed weight reduction, in general the fall in hsCRP relates to the amount of weight reduction^84,85. Adiponectin is an adipokine that is not pro-inflammatory but anti-inflammatory and also is associated with insulin sensitivity. A systematic review that examined weight loss using low calorie diets and exercise reported an 18-48% increase in adiponectin⁸⁶. Of note, the one study that reported an increase in adiponectin by 48% was a Mediterranean diet that resulted in a 15% weight reduction, suggesting this change is also related to the amount of weight loss⁸⁷.

Subjects undergoing metabolic surgery that lost 33% of their original weight had a highly significant median hsCRP reduction from 0.83 to 0.18 mg/d ensued⁸⁸. In another study wherein gastric stapling was compared to gastric banding and patients were defined at baseline as low vs. high CVD risk based on a hsCRP of <1.0 mg/mL vs. >3.0 mg/mL, the mean reduction in hsCRP for high CVD risk patients was greater for gastric stapling vs. gastric banding, −1.10 ± 0.94 mg/L vs −0.67 ± 0.82 mg/L, respectively⁸⁹. Following sleeve gastrectomy levels of hsCRP and IL-6 decreased and adiponectin increased, however, these measurements were made at 1 year, a time at which nadir weight and weight stability may have not been assured⁹⁰. In another one year analysis following metabolic surgery, there was a significant decrease in levels of IL-6 (p < 0.001), hsCRP (p < 0.001), and increase in plasma levels of adiponectin (p < 0.001), and levels of IL-6 and hsCRP correlated with BMI⁹¹. However, in another study, 1 year post gastric banding, hsCRP levels decreased from 1.33+/−1.21 mg/dl to 0.40+/−0.61 mg/dl but. IL-6 and TNF-α levels did not change⁹². Additional data also support the variability of metabolic surgery to decrease levels of IL-6^92-94. In a retrospective study in which 62 subjects that underwent a RYGB and a median follow-up of 15 months, there was a greater reduction in hsCRP with more surgical weight loss⁹⁵. This relationship between change in BMI/weight post-operatively and reduced inflammation also relates to adiponectin⁹⁶. Of interest when patients with or without remission of type 2 diabetes post RYGB were compared up to 24 months post-operatively, the non-remission group had a higher number of leukocytes (6,867 vs. 5,424), and hsCRP (0.27 vs. 0.12 mg/dL), MCP-1 (118 vs. 64 ng/mL), and lower adiponectin (9.4 vs. 15.4 ng/mL) than the remission group⁹⁷. Additional benefits of gastric bypass on inflammation and CVD risk may also relate to reduced levels of oxidized LDL and lipoprotein-associated phospholipase A2⁹⁸.

Vascular reactivity

Endothelial dysfunction, a nitric oxide-dependent component of reduced vascular reactivity, is defined as an imbalance between relaxing and contractile endothelial factors that is common in patients with severe obesity^99,100. Following weight loss, there is evidence that improvements in endothelial function can occur, but this outcome is inconsistently apparent and convincing evidence for normalization is lacking¹⁰¹.

At a very early interval after RYGB, weight loss led to significant improvements in brachial artery diameter and endothelial independent vasodilation¹⁰². A similar reduction in flow mediated dilatation was noted at 6 months S/P gastric banding¹⁰³. When subjects lost an average of 33% of their original weight, flow-mediated dilation (FMD) showed significant improvements after surgery from 7.4 % to 18.9 % (p < 0.001)⁸⁸. A recent systematic review included 8 studies on FMD (9 data sets; 269 patients) and 4 on nitrate-mediated dilation (NMD) (4 data sets; 149 patients). After 3-24 months following 4 different types of bariatric surgery (mostly RYGB), there was a significant improvement in FMD (MD: 5.65%; 95% CI: 2.87, 8.03; P<0.001), whereas NMD did not change (MD: 2.173%; 95% CI: −0.796, 5.142; P=0.151), and the percentage change in BMI was associated with changes in FMD (Z=−4.26, P<0.001) and NMD (Z=−3.81, P<0.001)¹⁰⁴. Most of these observations of vascular reactivity were performed at relatively brief intervals after surgery; thus, as for hypertension, the duration of reduced vascular reactivity associated with obesity may contribute to the response post metabolic surgery.

Obstructive sleep apnea

Obstructive sleep apnea (OSA) is common in patients with severe obesity, ~35-45% of patients at the time of bariatric surgery^105,106. With non-surgical approaches to weight reduction and OSA, the benefit typically relates to the amount of weight reduction and the severity of the OSA and in general metabolic surgery is more successful than non-surgical approaches¹⁰⁶. A systematic review of the metabolic surgical literature examined 69 studies in which 13,900 patients were included and RYGB, sleeve gastrectomy, gastric banding and BPD were compared¹⁰⁷. Although BPD was associated with the best results and gastric banding with the least beneficial outcome on OSA, all operations realized major reductions with >75% of patients experiencing resolution or at least some improvement.

When a systematic review of the surgical vs. non-surgical approach to modifying obesity-related sleep disordered breathing (not OSA) was inspected, 19 surgical (n=525) and 20 non-surgical (n=825) studies reporting primary endpoints of change in BMI and apnea hypopnea index (AHI) were examined¹⁰⁸. Unfortunately, the surgical vs. non-surgical groups were not matched in terms of BMI or the amount of weight reduction, 51.3 vs. 38.3 kg/m² and −11.9 vs. −3.1 kg/m² BMI units, respectively. However, both groups experienced a benefit in AHI, 29/hour vs. 11/hour, respectively. Despite AHI being substantially reduced, clinical OSA indicating continued CPAP use was common¹⁰⁹.

Mechanisms for cardiometabolic benefit of metabolic surgery

Gut hormones, changes in bile acid metabolism and the microbiome all relate to the benefits of metabolic surgery on cardiometabolic risk¹¹⁰. Changes in gut hormones relate to changes in energy balance and include increases in peptides that increase satiety, i.e. GLP-1, GIP, PYY_3–36, oxyntomodulin and gastrin) and those that reduce hunger-promoting factors, i.e. ghrelin¹¹¹. Moreover, although studies done at intervals up to 12 months post-metabolic surgery have demonstrated reductions in energy expenditure when expressed per fat free mass, modest increases were identified when data were expressed per body weight¹¹². Thus, some contribution from both sides of the energy balance equation may be operational in maintaining reduced weight after surgery.

Bile acid metabolism clearly changes post-metabolic surgery and mechanisms to implicate these alterations to benefit include their beneficial effects on satiety, gut hormones, incretins, energy metabolism and the gut microbiome, with the majority of these effects mediated via the bile acid receptors FXR and TGR5¹¹¹. Elevation of bile acids is commonly seen post-metabolic surgery¹¹³, and in murine models of atherogenesis activation of FXR and TGR5 reduced the expression of pro-inflammatory cytokines and chemokines within the arterial wall and atherosclerotic plaque volume ¹¹⁴. Finally, the gut microbiome is modified following metabolic surgery and this change seems to play an important role in the metabolic benefits gained from bariatric surgery. Two types of surgeries, RYGB and VBG, result in similar changes in the microbiome, an effect that can be maintained for at least a decade. Moreover, when microbiota from bariatric surgery patients are transferred into germ-free mice, decreases in fat mass ensue¹¹⁵. All of these mechanisms may be closely related and reflect changes in glucose, lipid/lipoprotein metabolism and inflammation that ultimately may be the mediators of reduced CVD risk. This is clearly an incredibly important area of research that may be applicable far beyond metabolic surgery and related weight reduction.

Current evidence for reduction in CVD events vs. impact on all-cause mortality

Effect of bariatric surgery on long-term survival

It has not been possible to conduct a prospective randomized clinical trial of bariatric surgery vs. continued non-surgical treatment (usual care) of severe obesity adequately powered and of sufficient duration to assess the impact on CVD events and longevity. Given the improvement of established obesity-related CVD risk factors following weight loss, it is reasonable to expect a reduction of CVD events and related mortality following weight loss in populations with obesity. The Look AHEAD trial randomized 5,145 individuals with obesity and with type 2 diabetes to intensive lifestyle intervention (ILI) or usual care¹¹⁶. Following median follow-up of 9.6 years, weight loss was 6.0% in the ILI group versus 3.5% in the usual care participants. Despite improved CVD risk factor status for all metrics except LDL-C, a reduction of CVD events or mortality was not demonstrated. In contrast, the Swedish Obese Subjects (SOS) study reported weight loss following a variety of bariatric surgical procedures to be 17% 5 years following surgery, 16% in 15 years, and 18% in 20 years post-surgery¹¹⁷. Weight was essentially unchanged in the usual care group matched for multiple clinical parameters. This weight loss in the surgical group was associated with a reduction of CVD events: adjusted HR 0.67; 95% CI 0.54-0.83, p <.001. In addition, mortality was reduced: HR 0.71, p = .001¹¹⁸. More recently, a systematic review and meta-analysis of 14 studies included 29,208 patients who underwent bariatric surgery, with a follow-up 2-14.7 years¹¹⁹. These studies took place in the US, Canada, Italy, Australia, and Sweden. Although not analyzed or reported by Kwok et al, weight loss among the studies varied from 15-30% or more depending primarily on the specific bariatric surgical procedure performed. Overall mortality was reduced more than 50% (OR 0.48). The incidence of myocardial infarction (OR 0.46) and stroke (OR 0.49) was also reduced¹²⁰.

Quality of Evidence

The great majority of published literature regarding bariatric surgery consists of observational data^29,35. A limited number of these observational trials with several-year follow-up have constructed comparator groups, matched from non-surgical populations. Medical and ethical considerations have prevented conduct of an adequately powered randomized control trial (RCT) to test the hypothesis that bariatric surgery is superior to usual care^54,67. However, recently, several RCTs have successfully been conducted evaluating medical vs. surgical intervention as primary treatment for type 2 diabetes^69,117.

Summary and Conclusions

It is reasonable to hypothesize that the greater improved CVD and mortality following bariatric surgery compared to lifestyle intervention is a function of the substantially greater weight loss that follows surgery, although neuroendocrine factors following gastrointestinal modification may also contribute. The survival benefit occurs primarily as the result of reduced CVD death although reduced death due to all types of cancers also contributes substantially to this survival benefit^121-123.

Figure: 1.

Diagram of Surgical Options. Image credit: Walter Pories, M.D. FACS.

Nonstandard Abbreviations and Acronyms

NIH

National Institutes of Health

BMI

Body Mass Index

CVD

Cardiovascular Disease

LABS

Longitudinal Assessment of Bariatric Surgery

LAGB

Laparoscopic Adjustable Gastric Banding

TBW

Total Body Weight

US

United States

VLDL-C

Very Low Density Lipoprotein-Cholesterol

BP

Blood Pressure

NHANES

National Health and Nutrition Examination Survey

HbA1c

Hemoglobin A1c

ALT

Alanine Aminotransferase

TNF-α

Tumor necrosis factor alpha

CPAP

Continuous Positive Airway Pressure

GLP-1

Glucagon-like peptide-1

GIP

Gastric inhibitory polypeptide

PYY_3–36

peptide tyrosine tyrosine or pancreatic peptide YY_3-36

FXR

Farnesoid X Receptor

TGR5

Transmembrane Bile Acid Receptor 5

Look AHEAD

Look AHEAD (Action for Health in Diabetes)