Bariatric surgery: Mechanisms, indications and outcomes


Professor Paul O'Brien, Centre for Obesity Research and Education, Monash Medical School, The Alfred Hospital, Melbourne,
Vic. 3004, Australia. Email:


The rising problem of obesity is causing major health problems, reduced quality of life and reduced life expectancy. It now generates approximately 10% of all health costs. The progression of the problem indicates preventive measures have been unsuccessful so far. Only bariatric surgical treatments have been able to achieve substantial and durable weight loss. Gastric banding and gastric bypass are used in more than 90% of bariatric operations. The proportion of each varies from greater than 95% bands in Australia, about 50/50 in Europe and USA and nearly 100% bypass in South America. The availability of follow up is a prime determinant of choice. Understanding the mechanisms of effect for the bariatric procedures is central to optimizing their effect. The traditional narrow concepts of restrictive (blocking the transit of food) and malabsorptive (preventing the absorption of food) should be discarded and the importance of induction of satiety, change of taste, diversion of chyme, neural and hormonal mediation and the effects of aversion need to be included. The primary mechanism of effect for gastric banding is the generation of a background of satiety and early post-prandial satiation via specifically structured vagal afferents at the level of the band. At five years after banding or bypass, there is typically a loss of 30–35 kg representing 50–60% of excess weight. This weight loss has been shown to be associated with major improvement or complete resolution of multiple common and serious health problems plus improvement in quality of life and in survival. Level 1 evidence supports the use of the gastric band over optimal lifestyle therapy. Randomized controlled trials has shown gastric banding to achieve better weight loss, health and quality of life than optimal lifestyle therapies for adults above a BMI of 30 and adolescents above a BMI of 35. In adults with mild to severe obesity and type 2 diabetes gastric banding leads to remission in three out of four individuals. Perioperative risk is significant with gastric bypass and late revisional procedures can be required after both procedures. Gastric banding is indicated in any adult who has a BMI over 30, has problems with their obesity and has made substantial effort to reduce their weight by lifestyle methods. Gastric bypass or biliopancreatic diversion should be considered in those with BMI greater than 35 if banding is contraindicated or has been unsuccessful.

The evolution of obesity

Obesity is likely to be the disease of the 21st century. The growth of obesity is worldwide, a pandemic, with the World Health Organization (WHO) estimating more than 1.6 billion people are currently overweight and 400 million obese.1 The rapid increase has been best demonstrated by the careful population measurements by the Centers for Disease Control and Prevention (CDC) in the United States.2 They show the prevalence of obesity has accelerated after 1980. Between 1960 and 1980 the percentage of adults who were obese increased only marginally, from 13.4% to 14.4%. In the next 20 years it doubled to 30.4%.2 It has continued to increase during the present decade. It had risen to 33.8% in 20083 and projected figures for 2010 and 2020 are 37.4% and 44.2%, respectively.4 Currently, more than 77 million adults in the USA are suffering the disease of obesity. A similar rate of growth is occurring in Australia with a current prevalence of approximately 22% or 2.9 million adults.5

Obesity and its treatment is becoming an important part of gastroenterology

Acknowledgment of the problem of obesity is the first step to dealing with it. This has come slowly as the traditional view that weight control is a personal responsibility is strongly embedded. Increasingly, governments, healthcare managers and physicians have accepted that we now live in an obesigenic environment, that the obesity is causing much illness, premature death and high healthcare costs and that the medical community should seek to address it as a disease. This attitude is growing strongly in gastroenterology. Obesity is a disease of overnutrition. It is driven in part by gut hormones and liver-derived insulin resistance. It generates diseases such as non-alcoholic steatohepatitis, several common GI cancers and gastro-oesophageal reflux disease. Endoscopic and surgical procedures on the gut are leading the attempts to generate substantial weight loss in the obese. Leading journals are encouraging the GI physician and surgeon to be knowledgeable and skilled in managing this disease.6

Increased recognition of the co-morbidities and costs

Obesity is a one of the most virulent of pathogens. It causes or exacerbates numerous common and serious diseases. Type 2 diabetes is the paradigm of an obesity-derived disease. There are now an estimated 285 million people with type 2 diabetes world-wide.7 Increased weight and diabetes has been directly and strongly linked through two major epidemiological studies, the Nurses' Health study of 112 000 women8 and the Male Health Professionals study of 51 000 males.9 From baseline BMI values in the low 20s, they have shown a direct and substantial rise in diabetes with weight increase. For the Nurse's Health study, from a baseline of BMI of 22, there is 5 times the risk at BMI 25, 27 times the risk at BMI 30 and 93 times at BMI 35.

Obesity contributes to ischaemic heart disease, stroke and the diseases that are linked to the metabolic syndrome, such as hypertension, the dyslipidaemias, obstructive sleep apnoea, non-alcoholic steatohepatitis and polycystic ovary syndrome. Cancer risk is markedly increased, particularly for colorectal cancer10 but also oesophageal, pancreatic, renal, endometrial, breast and gall bladder cancers.11 In addition, back pain, degenerative diseases of the hips and knees and depression are common. Mortality risk is markedly increased with obesity, now competing with smoking as the most prevalent preventable cause of death.12

Costs of the obesity pandemic are large. The direct healthcare costs in the United States for obesity during 2010 are estimated to be US$194 billion4 and the Americans themselves are spending US$59 billion on all the options offered to fight fatness. The total annual direct health care cost of overweight and obesity for Australia for the year 2005 has been estimated to be $21 billion.13

Evolution of bariatric surgery

Bariatric surgery began in the 1950s with jejuno-ileal bypass. It was superseded in the 1970s by gastric stapling procedures, such as Roux Y gastric bypass (RYGB), various forms of gastroplasty, and by the malabsorptive procedure of biliopancreatic diversion (BPD). All procedures were able to achieve substantial weight loss in the morbidly obese and yet bariatric surgery during this time really had no impact at a community level as it failed to attract even 1% of those suffering the problem. Factors such invasiveness, risks of death or complications and unknown long term effects were discouraging this approach.

In the last 15 years, the application of a laparoscopic approach to complex abdominal surgery including bariatric surgery, the development of the laparoscopic adjustable gastric banding procedure (gastric banding), improved safety and the better documentation of clinical effectiveness has led a surge of interest in bariatric surgery. In Australia, there were less than 400 bariatric procedures in 1993. In 2008 there were more than 14 000. Worldwide, it is estimated that a total of 344 000 procedures were performed in that year.14 Roux Y gastric bypass, both laparoscopic and open, was the most common (47%) followed by gastric banding (42%), sleeve gastrectomy (5%) and BPD (2%).14 The trend in Europe is for RYGB to increase, although gastric banding remains dominant. The opposite trend is occurring in the USA with gastric banding now moving ahead of RYGB in frequency. In Australia, gastric banding is the method of choice in more than 95% of bariatric procedures.

The major differences in surgical approaches across the world can be seen to be derived directly from the funding of the procedures and their follow up requirements. Gastric banding is unique among surgical procedures in that the placement of the band is simply the first step of a process of care that continues permanently. The adjustment of the band to achieve control of appetite is the central component of effect and therefore the follow up program is crucial. The uptake of gastric banding is directly related to the healthcare support for this follow up process. In Australia, follow up is covered under Medicare and banding has become the preferred approach. In Europe, there is generally little funding for follow up and banding struggles. In South America, there is no aftercare funding and almost no banding.

The central question is not which operation is best but when bariatric surgery becomes indicated. All current procedures achieve substantial weight loss. The type of procedure is the detail. Given the prevalence and pathogenicity of obesity, weight loss is arguably the most powerful treatment in medicine today. Substantial weight loss has major benefits to health, survival and quality of life. Type 2 diabetes is the paradigm of an obesity related disease. With sufficient weight loss many patients will have remission of their disease.15 Most will have improvement in control of blood sugar levels. Hypertension, obstructive sleeps apnoea, dyslipidaemia, non-alcoholic steatohepatitis, polycystic ovary syndrome, metabolic syndrome, gastro-oesophageal reflux disease, asthma and depression have been shown to improve or disappear.16 Obese people who lose weight live longer than a comparative group of obese without weight loss. Comparison of the long-term survival of patients after gastric banding with a community group who were obese showed a 72% reduction in the relative risk of dying.17

The hierarchy of obesity therapies

Obesity is a chronic disease. Resolution of the disease of obesity requires substantial and durable weight loss. The therapeutic options available are listed in Table 1 in order of their risk, side effects, invasiveness and costs. We should always begin with the simplest and safest and work down the list. Lifestyle therapies (diet, exercise, behavioral change) should always be the first line of management. Multiple randomized controlled trials (RCTs) have shown that a modest weight loss of between 2 and 5 kg can be achieved at 12 months.18–21 This level of weight loss is associated with a clinically valuable reduction of metabolic risk22–24 but generally will not solve the problems of obesity. Lifestyle therapies should be applied optimally and sought to be maintained permanently. If, however, they fail to resolve the obese patient's problems, the next level of therapy should be considered. Current drug therapies add little further benefit.25 Very low energy diets can be effective if taken correctly but are inevitably short term. The recent versions of the intragastric balloon have yet to show effectiveness by RCT and remain short term options. In spite of vigorous research effort, no additional endoscopic approaches are yet available which can provide even medium term benefit.

Table 1.  Weight loss approaches and their relative risks, side effects, invasiveness and costs
Ranking by riskApproachRisk score
1.Lifestyle changes—eat less, more activity and exercise, modify behaviour.1
2.Drugs and very low energy diets2
3.Endoscopic approaches—intragastric balloon et al.4
4.Gastric Banding5
5.Sleeve gastrectomy7
6.Roux en Y gastric bypass (RYGB)8
7.Open biliopancreatic diversion (BPD)9
8.Laparoscopic biliopancreatic diversion10

The bariatric surgical procedure of gastric banding becomes the next option to consider. It should be preferred ahead of other bariatric procedures on the hierarchy of risk (Table 1) for three reasons:

  • 1Safety profile.26,27 Mortality after gastric banding is rare. In systematic reviews, mortality has occurred in 1 in 200026 or 1 in 3000,28 ten to fifteen times less than for RYGB in the same reviews. At the Centre for Bariatric Surgery (CBS) in Melbourne we have now treated over 5700 patients with gastric banding without mortality.
  • 2Minimal invasiveness. It is truly minimally invasive, not just because of the laparoscopic approach but also the minimal dissection needed to place the band. Outpatient placement is now standard practice for many groups in the USA and Canada with the mean length of post operative stay of 2 h.29 Thus it can be placed with less risk and less time in hospital than a liver biopsy.
  • 3Complete and easy reversibility. There is no intention of reversal of gastric banding. It has occurred in only 128 of 5710 (2.2%) of CBS patients. However obesity is not a curable disease. Better treatments are inevitable and the option of reversibility allows the potential to change over to newer therapies in the future.

The other bariatric procedures are available for consideration should gastric banding be contraindicated or has failed. Less than 1% of our patients at CBS have required conversion.

How does bariatric surgery work?

A key requirement, not yet fulfilled for all bariatric procedures, is to understand better how the therapeutic effect is achieved. Table 2 lists some of the likely mechanisms. There has been a major increase in our understanding of mechanisms for gastric banding and RYGB in recent years.

Table 2.  How do bariatric procedures work?
Bariatric Surgery—Options for Weight Loss Effect
1. Reduce appetite, induce satiety
2. Alter the taste of food
3. Restrict intake
4. Divert nutrients from duodenum
5. Malabsorption of nutrients
6. Increase energy expenditure
7. Aversion effect—dumping, steatorrhoea, vomiting

The induction of satiety and satiation after gastric banding has been demonstrated in a prospective blinded crossover study as being a key factor.30 Satiety is the background feeling of not being hungry, without relationship to eating. Satiation is the feeling of satisfaction after eating. Both are induced by optimal setting of the band adjustment.30 As a result the gastric banding patient can follow the eating guidelines which center on eating a small amount of good food slowly. They eat the amount of food that could be compressed into half a cup (125 gm) three or less times per day. Food must be chewed well. Swallowed food is squeezed across the area of the band by oesphageal peristalsis. Between two and six peristaltic waves are needed to clear a single small bite of food. Feelings of satiation are generated with each squeeze.31–33 Eating for a maximum period of 20 min and taking one bite per minute generates enough signals to achieve satiation. Figure 1 shows the position of the band at the cardia of the stomach. Note that there is no ‘pouch’ of stomach above the band.

Figure 1.

The adjustable gastric band placed at the very top of the stomach with no significant gastric pouch above. It is proposed that compression of vagal afferents within this area of gastric cardia mediate the satiety effect.

At optimal adjustment, the band compresses the cardia of the stomach to generate a sense of satiety, of reduced appetite throughout the day. With eating, multiple additional signals are generated. Neural and hormonal mediators for this effect have been sought. A number of gut peptides are known to influence appetite. Ghrelin is the only hormone that increases appetite (orexigenic). Glucagon-like pepetide-1 (GLP-1), peptide YY (PYY), pancreatic polypeptide (PP), gastric inhibitory peptide (GIP) and cholecystokinin (CCK) are known to decrease appetite (anti-orexigenic). None of these appear to change in either basal or post-prandial circulating level with the gastric band.

The vagal afferents are most likely the key mediators of this effect through activation of intraganglionic lamellar endings (IGLEs).34,35 These are specifically structured vagal afferent tension/distension receptors attached to the connective tissue sheath of the myenteric ganglia. They are sensitive to distension and distortion and are slowly adapting. Additional possible vagal afferent mediators include the intramuscular arrays (IMAs)34 which lie within the circular and longitudinal muscle, are long straight varicose axons running parallel to the muscle fibres. They are particular concentrated in the fundus of the stomach, have close proximity to the interstitial cells of Cajal, the pacemaker cells of the smooth muscle. Morphologically, they appear capable of detecting distension or distortion of the gastric wall but electrophysiological study does not support this role.36

The mechanisms involved in RYGB (Fig. 2) are less defined and probably multiple. The traditional concept of distension of the small gastric pouch and delayed emptying through a narrow stoma has been considered a central effect. There is an enhanced post-prandial rise of the satiety-inducing hormone, PYY, an effect that is sustained for at least 24 months.37 However, this same cross-sectional and prospective study did not confirm the previously reported postprandial rise of the incretin, GLP-138 which brings into question the proposed role of incretins in the effect of RYGB on type 2 diabetes.39 Diet-induced thermogenesis is noted in the rat model of RYGB40 but remains to be established as an effect in humans.

Figure 2.

The RYGB involved complete separation of a small section of upper stomach from the body of stomach, creation of a roux en Y length of proximal jejunum which is anastomosed to the proximal gastric pouch. Intake is restricted by the small pouch, gastric emptying is restricted by the narrow gastrojejunostomy and food is diverted from the duodenum.

Comparison of bariatric surgery with non-surgical therapies

We have performed three RCTs in which we have compared gastric banding with optimal non-surgical programs. The initial RCT was of mild to moderately obese adults (BMI 30–35). We compared optimal non-surgical therapy, including lifestyle measures, drug therapy and very low energy diets with the gastric band and showed significantly better weight loss, health and quality of life for the banding group.41 Adverse events were similar between groups. The gastric band patients had lost 86% of their excess weight (%EWL) compared to 21% EWL in the non-surgical group (Fig. 3). This substantial weight loss effect has remained at 6 years follow up. The gastric banding participants showed almost complete resolution of the metabolic syndrome, from 38% to 3% (vs 38% to 24%; a non-significant change) and markedly improved quality of life as measured by the SF-36. The second study was of obese adults (BMI 30–40) with type 2 diabetes. There was 73% remission rate of diabetes in the gastric band group and 13% in the lifestyle group.42 Again, the metabolic syndrome was significantly improved in the banding group alone. The third RCT was of obese adolescents (BMI > 35; age 14–18 years).43 The gastric band group lost 79% of their excess weight and showed a significant improvement on the metabolic syndrome which reduced from 36% to zero. There was also an improved quality of life.

Figure 3.

Weight loss, expressed as % of excess weight lost (%EWL) in a RCT of gastric banding versus optimal non-surgical therapy in mild to moderately obese patients (data derived from continued follow up of gastric banding patients from the RCT41). The gastric banding group show durability of the weight loss at 6 years follow up.

Comparison between bariatric surgical procedures

Table 3 lists a range of comparators and the position of each current option against these comparators. The key outcome comparators between procedures are weight loss, health benefit and adverse events at the medium term (3–10 years). Short term data (<3 years) are largely irrelevant and long term data (>10 years) would be preferable but are almost totally absent. Systematic review of the medium term weight loss outcomes for gastric banding26,28,44 have shown no difference between RYGB and gastric banding but suggest there is a better weight loss with biliopancreatic diversion. Insufficient studies of the latter procedure preclude definitive conclusion. Figure 4 shows the relative % EWL for the three principal procedures.

Table 3.  Comparison of attributes of the principal bariatric procedures
AttributeGastric bandRYGBSleeve gastrectomyBPD +/− DS
Side effects++++++
Reversible easilyYesNoNoNo
Minimally invasive++++++++
Low revision rate++?+
Requires follow up++++++++
Figure 4.

Systematic review of the weight loss achieved in the medium term (3–10 years) by the principal bariatric surgical procedures of gastric banding, gastric bypass (RYGB) and biliopancreatic diversion (BPD) (Modified from Reference 44).

The effects of bariatric surgery on type 2 diabetes have been subject of two systematic reviews.15,28 Buchwald reported 86% of 1835 patients from multiple case studies showed remission or improved control.15 Maggard et al.28 reviewed 21 case series and reported a range of 64–100% showing remission or improvement. Numerous methodological problems existed in most of the published reports. RCTs constituted less than 5% of 618 studies, the definition of diabetes and its remission were not adequately described and there was extensive but poorly reported loss to follow up. Nevertheless, there was a trend towards better outcomes for the RYGB patients than the gastric band patients with complete remission reported in 70.1% (95% CI 59–83) at more than 2 years after RYGB and 58.3% (95% CI 42–74) after gastric banding. There was a direct relationship between weight loss and remission across the studies.

There have been multiple observational studies of the various comorbidities of obesity and generally these have reported either total remission or major improvement. Of particular relevance to the gastroenterologist are the studies of gastro-esophageal reflux disease45 and non-alcoholic steatohepatitis.46

What are reasonable expectations of risks?

Death is worst thing that can happen to a bariatric surgical patient. Mortality rates associated with gastric bypass of up to 2% were common47–49 but appear to have improved in recent years. The National Institutes of Health (NIH) in the USA has funded a longitudinal assessment of bariatric surgery, known as the LABS study, to provide careful evaluation of the outcomes from bariatric surgery. The perioperative safety data have now been published.27 A total of 4776 patients were enrolled and treated at 10 sites, specifically selected for their bariatric surgical expertise. RYGB and gastric banding were used. There were 15 deaths in the RYGB patients (0.44%) and no deaths in the gastric banding group, a highly significant difference. These findings are consistent with the systematic reviews.26,28

Gastric banding

This is my preferred option for the primary treatment of severe obesity. When applied properly, it is effective, safe and gentle. Adjustability permits maintenance of effect as long as the band is in place. Reversibility permits access to other therapeutic options that may be developed in the future. There is a need for long term skilled aftercare and there is a maintenance requirement, with approximately 10% needing some revisional procedure in 10 years. Removal and replacement of the band for abnormal proximal gastric enlargement is effective and revised patients have a weight loss equal to the total group.

Roux en Y gastric bypass

Roux en Y gastric bypass (RYGB) has been known to be the most effective of the stomach stapling procedures since the 1980s.50 It achieves good weight loss, particularly in the short term. However it carries significant risk, it is non-adjustable and essentially not reversible. Its effectiveness tends to fade with time. In the first 12 months after RYGB, a weight loss of 60–70% EWL can be expected. This effect is maintained for 12 months and then begins a gentle fade to average 50% EWL for those still attending follow up at 5 years. It is very effective in resolving type 2 diabetes, possibly through a post-prandial increase of the incretins GLP-1 and GIP. It is now commonly, but not universally, performed laparoscopically. The significant mortality of 0.44% seen in the LABS study involved expert bariatric surgeons.27 The community mortality rate was reported as 1.9% for the 15 years between 1990 and 2004 and is probably still at or above 1.0%.49 In the LABS study, 109 of the patients had further abdominal surgery.

Biliopancreatic diversion

It is the most metabolically severe of the current options and therefore hasn't proved to be popular with patients or surgeons in spite of favorable published outcomes. Biliopancreatic diversion (BPD) has been available for 30 years51 and yet remains a very minor part of bariatric surgery. Worldwide, it constitutes less than 2% of bariatric surgery.14 However, it does generate good weight loss and should be considered on occasions as a second line bariatric surgical option.

Sleeve gastrectomy

The sleeve is the first element of the duodenal switch procedure, a variant of the BPD. It has lately become popular as a single procedure because of ease of surgery, relative effectiveness and perceived lack of need for close follow up. A systematic review of the 36 studies available to mid-2009 showed 55% EWL at 3 years.52 There has been only one medium term study which reports a 40% weight regain by 5 years.53 There is a general expectation that the sleeve will fail to maintain acceptable levels of weight loss in the medium term as the tube of residual stomach inevitably expands. Continuation to completion of the duodenal switch would then need to be considered. Leaks are relatively frequent (>1%) and tend to persist for months generating morbidity, anxiety and costs.

Who should be considered and who should not?

There is level 1 evidence supporting a better outcome for using gastric banding in the mild to moderately obese (BMI 30–35) when compared with lifestyle therapy.41,42 This approach is cost-effective.54,55 When the two treatment paths are modeled over time, the gastric banding approach is dominant i.e. it provides increased number of quality-adjusted life years at a lower cost than the existing option of non-surgical therapies. Any person who is obese (BMI > 30), is suffering from the medical, physical or psychosocial consequences of the obesity and has diligently sought a solution through a range of lifestyle options over time, should be considered for gastric banding. Because the stapling group of surgical options lack level 1 data, are of greater risk, and are not controllable or reversible, maintenance of existing cutoff of BMI > 40 or BMI > 35 with major comorbidities should remain for these procedures.

Gastric banding is unsuitable for those who are mentally defective or otherwise unable to engage in the ‘partnership’ needed for optimal outcome. Other contraindications include portal hypertension and remote living which could preclude adequate follow up.

Needs and challenges

Bariatric surgery is never a quick fix. It is a process of care that begins with a careful initial clinical evaluation and detailed patient education and it continues beyond the operative procedure through a permanent follow up. All procedures have the potential for perioperative complications and death. Revisional surgery is relatively common as maintenance of the correct anatomy is intrinsic to effectiveness. But bariatric surgery can provide a solution to the problem of obesity. It achieves substantial weight loss, improved health and quality of life and a longer life. We need to optimize these benefits and minimize the risks and the costs. The following are some of the areas for further research and development:

  • 1A better understanding of the mechanisms of action of each procedure is required to enable optimum surgery and follow up.
  • 2We need careful data management for all patients. Bariatric surgical procedures should be incorporated into national clinical registries to enable objective assessment of the risks and benefits across the community.
  • 3There is a need for more randomised controlled trials to define the benefits of weight loss on various comorbidities of obesity. More study is needed in particular for the patients with metabolic diseases—type 2 diabetes, metabolic syndrome, non-alcoholic steatohepatitis, the dyslipdidaemias, polycystic ovary syndrome and obstructive sleep apnoea.
  • 4We need to know more about who should be offered bariatric surgery, and define the most safe and efficient pathways for assessment, surgery and aftercare.
  • 5We need better cost-effectiveness evaluation of the bariatric surgical approach to disease management in comparison with existing options.

Bariatric surgery has the potential to be one of the most important and powerful treatment approaches in medicine. High quality of clinical care, good science and comprehensive data management will allow optimal application of this approach to be realized.

Disclosures: CORE receives unrestricted research grants from Allergan and Applied Medical.