Introduction

Obesity is a major health problem in the USA and worldwide. An estimated 66% of American adults are overweight, defined as a BMI of more than 25 kg/m². Half of these are obese (BMI > 30 kg/m²), and nearly 5% have clinically severe obesity (term preferred to morbid obesity, with BMI > 40 kg/m²) ^[1]. Multiple chronic diseases are associated with obesity, including type 2 diabetes mellitus, hypertension, dyslipidemia, metabolic syndrome, coronary heart disease, ischemic stroke, nonalcoholic steatohepatitis (NASH), polycystic ovarian syndrome, respiratory disease (sleep apnea and obesity-hypoventilation syndrome), gallbladder disease, musculoskeletal disease, and certain cancers ^[2]. Furthermore, the risk of death from all causes increases across the range of moderate and severe overweight for men and women in all age groups, a finding especially prominent in the ethnic white population ^[3]. Although there are multiple medical disorders associated with obesity, this review will focus on diabetes.

Treatments for obesity have traditionally consisted of medical nutritional therapy with calorie restriction, exercise, and pharmacologic agents. Lifestyle modification (primarily diet and exercise) can promote successful weight loss, but often fails to achieve weight loss of more than 5–10%, and any weight loss achieved is often not sustained. Likewise, currently available medical and pharmacologic treatments promote limited sustained weight loss. Indeed, participants in the large, recently completed Swedish Obese Subjects (SOS) study who received standard medical treatment for obesity realized an average weight gain of 1.6% over 10 years ^[4••]. It is important to note that even modest weight reduction and increased physical activity can lead to substantial health benefits, especially for individuals at high risk for disease. In the Diabetes Prevention Program trial, only half of overweight participants with impaired glucose tolerance in the lifestyle intervention arm were able to achieve the goal of 7% weight loss by the end of the curriculum (at 24 weeks), and weight loss was not fully sustained over continued follow-up ^[5]. Yet despite the relatively small degree of weight reduction, the risk of development of type 2 diabetes was reduced by 58% as compared with a control group. Comparable risk reduction has been demonstrated in other studies ^[6,7]. In contrast, surgical treatment of obesity, including roux-en-y gastric bypass (RNYGB) surgery and gastric banding procedures, induces impressive absolute weight loss of 30–40 kg (about 60% excess weight or a 10–15 kg/m² reduction in BMI), as shown by Buchwald et al. ^[8] in a meta-analysis examining over 22 000 patients who underwent obesity surgery. With surgical treatment, weight loss may be sustained more than 10–15 years ^[4••].

Surgical weight loss improves the metabolic status of patients

Historically, weight loss via lifestyle interventions and medical therapy has improved glycemic control and mortality in type 2 diabetic patients ^[9,10], although maintenance of weight loss is difficult. Surgical weight loss of 13–27% of initial body weight is sustained for more than 15 years, with greater weight loss realized with RNYGB surgery as compared with other procedures ^[4••]. Thus, it is not surprising that observational studies ^{[11,12,13••]} have shown bariatric surgery leads to a significant reduction in incident type 2 diabetes, as well as substantial improvement or resolution for patients with preexisting disease. Multiple metabolic measures such as fasting plasma glucose and insulin, percentage glycosylated hemoglobin, and use of antidiabetic medications are significantly reduced in diabetic patients treated with obesity surgery, and these effects may be sustained over 5–10 years ^{[11,12,13••]}. A meta-analysis ^[8] of over 130 studies demonstrates resolution or improvement of diabetes in a significant majority of patients. In particular, type 2 diabetes resolved (defined as maintenance of normal blood glucose after discontinuation of all diabetes-related medications) in about 77% of patients, and resolved or improved in 85% ^[8]. Furthermore, dyslipidemias may markedly improve or resolve in 70–95% of surgically treated patients and hypertension in 87–95%. Patients with earlier disease appear to have more complete resolution. In an observational study ^[14••], overall death rates were 40% lower over an average of 7-year follow-up for people who had bariatric surgery compared with those who didn't, and diabetes-related deaths were reduced by 92%. These data strongly suggest additional important health benefits for patients with diabetes. Although most studies were not randomized and thus must be suspect for subject selection bias, no other medical intervention to date realizes this magnitude of success.

Together these findings suggest obesity surgery may be indicated as primary treatment for type 2 diabetes ^[15], especially as diabetes becomes more prevalent in younger adults. Recently, a randomized controlled trial of laparoscopic gastric banding versus optimal conventional medical therapy addressed this question. Dixon et al. ^[13••] randomized recently diagnosed type 2 diabetics (diagnosis of ≤2 years) and BMI of 30–40 kg/m², inclusive of some patients who's weight falls below current guideline, to laparoscopic adjustable gastric banding or intensive lifestyle modification with medical management, and demonstrated that surgically treated patients were more likely to achieve resolution of their diabetes (73 versus 13%, P < 0.001), a 5.5 relative risk for remission in the surgically treated group. Further investigation and longer follow-up data are needed, but early surgical intervention for type 2 diabetes will likely become increasingly utilized in the future. If surgery is to be recommended earlier, it is important to understand the current indications, benefits, and risks associated with the procedures [Fig. 1].

Obesity surgery background

In 1991, a National Institutes of Health expert consensus panel recommended that surgical treatment of obesity should be considered for informed and motivated patients in whom operative risks are acceptable. Candidates should have a BMI of more than 40 kg/m² or between 35 and 40 with high-risk comorbid conditions such as cardiopulmonary problems (sleep apnea, pickwickian syndrome, and cardiomyopathy), diabetes mellitus, or physical problems that interfere with lifestyle ^[16]. The common types of surgical procedures performed in the USA for obesity include RNYGB surgery, gastric banding (including adjustable and nonadjustable bands), and to a lesser degree gastroplasty (principally vertical-banded gastroplasty) ^[17]. Newer procedures, including gastric sleeve are under evaluation. Banding procedures and gastroplasty induce weight loss via gastric restriction and early satiety; gastric bypass has a component of malabsorption in addition to restriction. In the USA, the number of surgical procedures for obesity has increased dramatically, over 10-fold in a decade, with 13 000 procedures performed in 1998, 70 000 in 2002, and over 130 000 in 2005 ^[18,19] with continued annual increases and estimates over 200 000 in 2007. Increasing medical and public awareness of sustained weight loss, increased ease with newer laparoscopic procedures, low complication rates, and increased population with obesity are all likely to contribute to the soaring number of procedures.

Hormonal mechanisms of weight loss with obesity surgery

The mechanisms by which obesity surgery induces sustained weight loss, and therefore improvement in the medical comorbidities, including diabetes mellitus are likely related to more than simple restriction of intake or malabsorption of nutrients. Evidence has emerged that alterations in gut hormones such as peptide YY (PYY), ghrelin, glucagon-like peptide-1 (GLP-1), and leptin are likely playing a critical role, and that the type of procedure performed may influence these hormones and the magnitude of weight reduction. Ghrelin, a hormone produced in the stomach fundus, which increases hunger and food intake, is decreased after RNYGB surgery in weight-stable patients ^[20], and therefore may be contributing to the greater success for sustained weight loss after surgery. This effect is not seen following laparoscopic-adjustable gastric banding techniques ^[21]. Satiety hormones such as PYY and GLP-1, which are secreted in the distal small bowel, have been shown to be increased in patients who have undergone RNYGB surgery ^[22,23], possibly contributing to decreased appetite and weight loss, despite concurrent decrease in plasma leptin level, another hormone associated with satiety ^[24]. Meal stimulated PYY and GLP-1 concentrations are increased as early as 4–6 weeks after RNYGB surgery, suggesting that the altered anatomy and rapid delivery of nutrients to the distal intestine promote their secretion ^[22,23]. Increases in GLP-1 may also mediate improved insulin secretion in response to meals and participate beyond weight loss to the substantial metabolic improvements in patients with type 2 diabetes who have undergone RNYGB surgery. Additional investigations are needed to fully understand the changes in gut hormone secretion and their role in appetite, satiety, and weight loss after obesity surgery. Additional gut hormones likely contribute to the magnitude and sustained weight loss, although to date, they have not been as well characterized.

Operative morbidity and mortality

Observational studies suggest that long-term survival for patients with surgical weight loss is improved ^[11]. Although early procedures such as the purely malabsorptive jejunoileal bypass surgery were associated with serious complications, including hepatic failure, septic arthritis, oxalate stones, bypass enteritis, protein malnutrition, and vitamin B12 and D deficiencies, current procedures do not carry similar risk. Current procedures have operative risk similar to cholecystectomy. Estimates of early operative mortality, defined as mortality at 30 days or less, vary but in general are low at 0.1–0.33% ^[18,19,25], but up to 0.5% for gastric bypass surgery ^[8], higher in the elderly ^[26] and with less-experienced surgeons ^[27]. Complications of obesity surgery can occur in up to 10% of patients ^[28]. Reoperation during same admission occur in 6–9%, technical complications, including obstruction, anastomotic, hemorrhagic, wound, and splenic injury in 1–2%, and systemic complications in 3–7%, which most commonly involve the pulmonary system ^[18]. In addition, the rate of overall hospital admission in the year after RNYGB surgery may be increased two-fold, with most admissions for gastrointestinal or surgical-related complications ^[25]. Thus, the immediate risk compared with long-term benefit must be carefully considered for each patient, particularly the elderly who have higher immediate risk and relatively fewer years to realize long-term benefit.

Reduced long-term mortality following bariatric surgery

Obesity is a chronic illness, which shortens life expectancy ^[3]. Several studies suggest a significant mortality benefit in obese patients who undergo surgery for treatment of their obesity. However, it is important to note that though carefully performed, these studies were not randomized, thus are subject to selection bias.

Bariatric patients have improved survival over extended 15-year follow-up (11.8 versus 16.3% mortality, representing hazard ratio of 0.67 for death in surgical cohort) compared with patients of similar age with a diagnosis of obesity or morbid obesity who did not have a bariatric procedure ^[27]. Similarly, the SOS study ^[4••] found an adjusted hazard ratio of 0.71 for death in surgical group compared with medical control groups with a mean follow-up of 10.9 years. The most common causes of death in surgical and medical groups remain cardiovascular events and cancer, both of which have been linked with obesity and are leading causes for death in the general population. Most of the surgical patients in the SOS study underwent vertical-banded gastroplasty, which is less popular in the USA and generally results in less weight loss. To isolate the effect of RNYGB surgery on mortality, Adams et al. ^[14••] examined mortality over 18 years of follow-up (mean 7.1 years) in nearly 8000 patients who had undergone RNYGB surgery compared with obese controls, and found the adjusted rate of death from any cause was 40% lower in the surgical group (hazard ratio 0.60). Furthermore, surgical patients had multiple lower disease-specific death rates, including diabetes (92% lower), coronary artery disease (59% lower), and cancer (60% lower) ^[14••]. Other groups have supported similar results ^[29].

Dumping syndrome

After obesity surgery, patients must modify their diet. One of the common adverse effects from restrictive and malabsorptive procedures is the dumping syndrome, which can occur in a majority of patients if dietary changes are not made. Symptoms include watery diarrhea, nausea, bloating, and abdominal cramping. This is the result of rapid delivery of simple sugars into the small bowel, increasing the osmotic load. The resultant volume shifts can result in adrenergic stimulation causing vasomotor symptoms such as palpitations, confusion, or syncope. Alterations in gut peptides contribute to the dumping syndrome ^[30]. Dumping syndrome can usually be treated with dietary modifications, including limiting simple sugars, and eating frequent small meals, which include protein and fiber ^[31].

Micronutrient deficiencies

Micronutrient deficiencies are common following obesity surgery, more so with gastric bypass than with gastric banding procedures because of the malabsorption, which results from bypassing the duodenum and proximal small bowel. Serum iron deficiency may occur in 51% of women and 22% of men, as well as deficiencies of vitamin B12 (37%) and folate (35%) in patients who underwent RNYGB surgery ^[32]. Anemia may occur in half of patients ^[32,33•]. Routine use of an oral multivitamin does not always prevent deficiencies ^[33•]. Fat-soluble vitamin (A, D, E, and K) deficiency is less common, with vitamin A deficiency reported in 11% of patients and vitamin D₂₅OH deficiency in 7% ^[34]. However, the defined cutoff for vitamin D₂₅OH deficiency is controversial. A lower cutoff for normal vitamin D levels than many would consider clinically important (of 9 ng/ml) was used in this study so results may underestimate the scope of the problem. This concept is supported by the finding that 58% of patients with normal vitamin D₂₅OH (defined as >8.9 ng/ml) had elevated parathyroid hormone (PTH) ^[35,36•]. At higher cutoff for vitamin D₂₅OH deficiency (20 ng/ml), 23% of RNYGB surgery patients were demonstrated to have deficiency at 1 year following surgery. Although there was less hyperparathyroidism (25%) than in other studies, this could be due to either classification or shorter duration of follow-up ^[36•]. The implications of low vitamin D include accelerated bone turnover and risk for osteopenia and osteoporosis.

Postprandial hyperinsulinemic hypoglycemia after gastric bypass surgery

Hyperinsulinemic hypoglycemia is a newly recognized complication of RNYGB surgery ^[37–39]. Symptoms occur postprandially (1–4 h after a meal) and include palpitations, tremor, sweating, anxiety, and hunger, and may include more severe symptoms of neuroglycopenia such as confusion, seizure, or loss of consciousness. During episodes, low blood glucose with elevated insulin and c-peptide are documented. Symptoms are similar to those experienced with the late dumping syndrome, but present typically more than 1–2 years after surgery, usually after dietary and weight stability has been achieved. Initial treatment, which leads to clinical improvement for many patients who experience postprandial symptoms, includes dietary modification with frequent small meals low in carbohydrate and glycemic index and higher in protein and unsaturated fats. Unfortunately, a subset of patients remains refractory to medical nutritional therapy. Pharmacologic therapy with acarbose, somatostatin analogs, and diazoxide has success for many of these patients. However, some patients who have not responded to these pharmacologic interventions have undergone partial pancreatectomy to manage severe neuroglycopenic symptoms ^[38,39]. Review of surgical specimens from resected pancreata shows diffuse islet hyperplasia, with abnormally large numbers and irregular size and shape of islets, without evidence of insulinoma ^[38,39]. Understanding the implications of these histology results is difficult, as pancreatic islet histology of severely obese patients after massive weight loss has not been well characterized.

Currently cases are rare and are less common in patients with preexisting diabetes, but the incidence may be expected to increase in number as more gastric bypass surgeries are performed and with longer duration of follow-up. Anecdotally, some patients with postprandial hyperinsulinemic hypoglycemia report symptoms consistent with hypoglycemia predating their surgical procedure. It may be of value to query patients for these symptoms for medical evaluation prior to surgical referral to avoid this complication.

The cause of the hyperinsulinemic hypoglycemia remains incompletely understood, but may be related to elevated GLP-1 or other gut hormones ^[40•]. GLP-1 is secreted from the L cells of the distal small bowel and increases insulin secretion in response to a meal. Patients with hyperinsulinemic hypoglycemia after gastric bypass surgery have significantly higher fasting GLP-1 concentrations, as well as GLP-1 response to a liquid mixed meal, when compared with overweight controls with similar BMI or asymptomatic postgastric bypass patients ^[40•]. The role of GLP-1 in this syndrome could be related to its action in response to a meal, or to an effect on pancreatic islet cells over time, as it has been known to promote islet growth and inhibit apoptosis in Zucker diabetic rats ^[41]. Further investigations are ongoing to better understand the role of GLP-1 in hyperinsulinemic hypoglycemia after gastric bypass or to identify other causes of the syndrome.

Conclusion

The epidemic of obesity is spreading in the USA and worldwide. Surgical treatment of obesity has been shown to be effective in producing significant and sustained weight loss. Although there are several surgical approaches for weight management currently in use, improvements in diabetes including achievement of near normal glycemia without medication or reduced medications are realized in many patients. Additionally, dyslipidemia and hypertension are also improved. Increasing evidence supports improved survival with bariatric surgery compared with medical and nutritional treatments for the overweight patient, and may be anticipated specifically for patients with diabetes although this has been incompletely characterized. Early surgical intervention for overweight type 2 diabetes may be clinically appropriate in patients for whom operative risks are acceptable. Understanding beneficial and detrimental consequences of bariatric surgery as strategy to treat obesity is essential for patients and physicians to make informed decisions about treatment options, and consideration of bariatric surgery is especially important for patients with type 2 diabetes as there are substantial metabolic benefits for this group.

Acknowledgement

There are no conflicts of interest.