The prevalence of type 2 diabetes is increasing worldwide.1 More than 20 million people in the US were suffering from diabetes in 2005, with approximately 30% being undiagnosed.2 It is estimated that by 2050 over 48 million Americans will have diabetes.3 In the US more than $132 billion was spent in 2002 on type 2 diabetes and its related complications.
The prevalence of type 2 diabetes is increasing worldwide.1 More than 20 million people in the US were suffering from diabetes in 2005, with approximately 30% being undiagnosed.2 It is estimated that by 2050 over 48 million Americans will have diabetes.3 In the US more than $132 billion was spent in 2002 on type 2 diabetes and its related complications. Nearly $92 billion was spent on direct medical expenditures and $40 billion to cover indirect costs resulting from lost productivity secondary to disability and early mortality.4 The costs are anticipated to grow to $192 billion by 2020.4
Evidence suggests that bariatric surgery may represent a valid and costeffective treatment for diabetes, concomitantly inducing a consistent survival advantage.5–8 However, surgical procedures are not yet considered a standard of care for diabetes management. This review focuses on the metabolic effects of bariatric procedures and possible cultural barriers that may affect the acceptance of surgery as a treatment for patients with type 2 diabetes.
Conventional Treatment for Type 2 Diabetes
Currently, according to the American Diabetes Association (ADA), diabetes is considered “a chronic illness that requires continuing medical care and patient self-management education to prevent acute complications and to reduce the risk of long-term complications.” 9 While the risk of microvascular complications can be decreased with intensive medical treatment and adequate glycemic control,10 reduction of macrovascular disease and cardiovascular mortality remain difficult to obtain with conventional therapeutic strategies. The Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Release Controlled Evaluation (ADVANCE) study showed that intensive medical treatment significantly reduces microvascular events but causes no significant effects in major macrovascular events.11,12
While these large clinical trials have reported positive results, several studies show the difficulty of patients reaching glycemic goals. The UK Prospective Diabetes Study (UKPDS) examined the percentage of patients who attained the target successful glycemic control in three different treatment modalities. Glyated hemoglobin (HbA1c) levels <7% were achieved after nine years of monotherapy with diet, insulin, or sulfonylurea by only 9, 28, and 24% of patients, respectively.12 Data from the National Health and Nutrition Examination Survey (NHANES) 1999–2000 have recently showed that only 37% of adults with previously diagnosed diabetes achieved the target HbA1c goal of <7%.13
Diabetes management includes the control of related comorbidities.The UKPDS group investigated the effects of tight control of blood pressure compared with a less tight antihypertensive treatment. There was a significantly decreased risk in the tight control group of diabetes-related deaths, strokes, and microvascular end-points.14 In the Scandinavian Simvastatin Survival Study (4S) the authors reported that patients with diabetes treated with a statin had a significant relative reduction of major coronary heart disease (CHD) event risk (CHD death or non-fatal myocardial infarction [MI]) compared with patients treated with placebo.15 In spite of clear evidence of the benefit of blood pressure and cholesterol control in patients with type 2 diabetes, the available data show goals for blood pressure and cholesterol are not being met. NHANES 1999–2000 found that the goals of blood pressure and total cholesterol were attained only by 35.8 and 48.2% of subjects with type 2 diabetes respectively.13 Only 7.3% of adults in this study had achieved all three recommended goals of HbA1c level <7%, blood pressure <130/80mmHg, and total cholesterol level <200mg/dl.13 Data from major studies indicate that conventional medical treatment of most patients with diabetes generally does not produce satisfactory control.12–13
‘Metabolic’ Surgery
Available evidence has established that obese patients can frequently develop normal glycemia and HbA1c levels after bariatric procedures. Resolution (or remission) was defined as discontinuation of all diabetesrelated medications and maintenance of blood glucose levels within the normal range. In a recent meta-analysis by Buchwald and colleagues, a complete remission of diabetes was found in 78.1% of patients who underwent bariatric surgery. Taking into account either complete remission or the improvement of diabetes, the percentage increases to 86.6% of patients.5 The rate of diabetes resolution was as high as 80.3–95.1% when only gastrointestinal bypass procedures were included.5 In type 2 diabetes patients following gastrointestinal bypass procedures, Polyzogopoulou et al. found improved insulin sensitivity, as well as improved beta-cell acute insulin response to glucose stimulation.16 In an earlier meta-analysis, Buchwald and co-workersfound a rate of hyperlipidemia improvement of 70%.17 Considering the gastrointestinal bypass procedures alone, 96.9–99.1% of patients had improvement of hyperlipidemia. Hypertension resolved in 61.7% of patients and it resolved or improved in 78.5% of patients.17
Given the defined benefits of tight metabolic control, it is not surprising that surgical interventions result in significant reductions of cardiovascular risk and improved survival.18 A retrospective cohort study investigating long-term mortality among nearly 10,000 Roux-en-Y gastric bypass (RYGB) patients compared with severely obese controls found a 40% decrease in adjusted long-term mortality from any cause in the surgery group compared with the obese control group. The decrease in diabetes-related mortality in the surgery group was 92%.7
In spite of its dramatic outcomes, bariatric/metabolic surgery is chosen by only 1% of the morbidly obese19 and has not yet become an accepted treatment for diabetes. Perhaps this may be related to the common perception of bariatric surgery, which is often seen as ‘extreme,’ to be reserved as a treatment of last resort. Several other factors, however, may limit surgical referral. Cultural barriers and widespread misconceptions about the regulation of bodyweight exist. In fact, severe obesity is rarely viewed as a genuine disease and obese patients are too often victimized with social stigma based on the assumption that weight can be controlled by deciding to eat less and exercise more.
Accordingly, bariatric surgery is considered as an extreme way to enforce healthier lifestyle habits and a form of treatment that works through mechanical changes rather than through alterations in the underlying pathophysiology of the disease.
Scientific data related to bodyweight regulation and the mechanisms of action of bariatric procedures clearly refute the common misconceptions. Bodyweight is maintained within a relatively narrow range by a precise and powerful biologic system and voluntary efforts to reduce weight are resisted by compensatory biologic responses. On the other hand, recent evidence shows that gastrointestinal surgery does not ‘treat’ diabetes solely by restriction or malabsorption. Although the mechanism is incompletely understood, RYGB does alter gut hormone secretion,20–22 which may contribute to sustained reductions in food intake,20 improved insulin sensitivity,23–26 and enhanced insulin secretion.16,24,26–28
The term ‘bariatric’ surgery, derived from the Greek word baros for weight,29 defines a branch of general surgery that deals with the control of weight. Successful outcomes of bariatric surgery are measured as percent of excess bodyweight loss. However, weight loss is only one of the outcomes of such surgery. The truly significant benefits of these procedures include the resolution/improvement of obesity-related comorbidities, such as diabetes, hypertension, and hyperlipidemia. The argument explained above represents a strong rationale for a conceptual shift that defines conventional weight-loss surgery and novel gastrointestinal procedures as ‘metabolic surgery’ as opposed to simply bariatric surgery.
Risk–Benefit Ratio
For the development of a meaningful risk–benefit ratio, it is important to evaluate mortality data from conventional medical treatment. Among UKPDS patients (individuals newly diagnosed with type 2 diabetes at the time of study entry), over the 10-year follow-up mortality was 44%, with cardiovascular disease being the leading cause of death (51.5%).11
Further support for the significant effects of type 2 diabetes on mortality comes from a 12-year follow-up of men who reported using medication for diabetes, who had an absolute risk of cardiovascular disease death that was three times higher than male controls.31 As already discussed, the decrease in diabetes-related mortality after RYGB is substantial, and in one study was up to 92%.7 This potential survival advantage of surgery needs to be balanced with the risk of death from the surgical procedure itself. For this comparison we can evaluate mortality rates of bariatric operations with those of other surgical procedures. A US average mortality rate for seven operations was calculated in 2000.32 These ranged from 0.3% for hip replacement to 10.7% for craniotomy. A metaanalysis35 suggests that mortality rates from bariatric surgical procedures compare very favorably with those of other surgical procedures. Another more recent meta-analysis agreed, reporting a total mortality at ≤30 days of 0.30%.35–36
With any surgical procedure for any type of disease, the indication for surgical treatment is usually determined by the assessment of the balance between risks from the disease and risk from surgery itself. Diabetes should be no exception.
An informative example comes from treatment of coronary artery disease. A population-based study demonstrated that the seven-year incidence rates of MI in subjects with diabetes without prior MI at baseline (20.2%) were essentially the same as the incidence rates of MI in subjects without diabetes but with prior MI (18.8%). The hazard ratio for CHD deaths between these two groups was not significantly different from 1.0.33 Guidelines from the American College of Cardiology (ACC)/American Heart Association (AHA) agree that coronary artery bypass graft (CABG) surgery should be performed in patients with stable angina in certain settings.34 In 2000, the US national average mortality rate for CABG was 3.5%.32 Bariatric surgical procedures, by comparison, demonstrate a mortality rate that is merely one-tenth as high as that for CABG35 and offer long-term resolution of diabetes, yet are rarely considered in the algorithm of diabetes management.
A Return on Investment
Cremieux and co-workers8 have recently compared bariatric surgery patients with obese control patients who did not have surgery. The improvements in morbidity and mortality contributed to decreases in healthcare utilization.18,37,38 All costs for bariatric surgery are recouped within two years for laparoscopic patients and within four years for open surgeries.8
This return on investment should not be a surprise considering the welldocumented immediate and long-lasting reductions on myriad obesityrelated comorbid conditions.6,7,16,32,35,39–47
Surgical Treatment of Type 2 Diabetes— Patient Selection
Current National Institutes of Health (NIH) guidelines define eligibility for surgical management of morbid obesity using criteria primarily based on body mass index (BMI). Specifically, patients with a BMI >40kg/m² or with a BMI >35kg/m² with comorbidities (including diabetes) are eligible for bariatric surgery.48 BMI has been adopted worldwide as the predominant measure to guide classification of obesity and to determine risk of morbidity and mortality due to obesity.49 Due to these criteria and classifications, BMI is frequently thought of as a primary measure of risk associated with obesity and is one of the most important reasons why a surgical operation is warranted.
While BMI does represent one conveniently calculated parameter of obesity, it is far from a fully inclusive measure of all risk. For instance, ethnicity plays a large role in obesity risk. One meta-analysis found that the relationship between percent body fat and BMI differed among ethnic groups such that, for a given level of body fat, age, and gender, different ethnic groups have different BMI levels.50 Razak et al. demonstrated that South Asians and Chinese have distributions of elevated glucose and lipid factors similar to Europeans at significantly lower BMI values.51 This supports that BMI-centric definitions of obesity and associated risk are applicable primarily to Caucasians.
Waist circumference should also be considered in an obesity risk profile. NHANES reported that patients who had a high waist circumference were significantly more likely to have hypertension, diabetes, dyslipidemia, and the metabolic syndrome compared with those having a normal waist circumferences. This was the case across different BMI categories.52 Also, while BMI alone is a good predictor of the risk of developing diabetes and/or metabolic syndrome, it is less adequate as a measure of the overall risk of morbidity and mortality in patients with established diabetes.
There is therefore no evidence that a specific BMI cut-off could predict diabetes-specific risks, and recent investigations of metabolic surgery in less obese patients53–56 show that no specific level of BMI can provide a distinction between patients who would benefit from surgery and those who would not with regard to resolution/improvement of diabetes.57 Using current BMI cut-offs may also delay a potentially life-saving option for patients with lower BMI but at similar risk from diabetes.
Furthermore, BMI-based policies might be perceived as discriminatory since the BMI-related risk is influenced by gender and ethnic differences. For example, an individual of South Asian descent with a large waist circumference, several comorbidities (including diabetes), and a BMI of 34kg/m² would have a substantially different risk–benefit profile for surgical treatment of diabetes than an individual of European descent with a normal waist circumference, one comorbidity, and a BMI of 36kg/m². However, based on our current guidelines, the latter would be offered surgery while the former would not. The growing evidence indicating that the antidiabetic mechanisms of some gastrointestinal operations cannot be explained by changes in food intake and bodyweight alone58,59 suggests that it might be appropriate to expand the indications for surgical treatment of diabetes to patients who do not meet existing BMI-based criteria.
Several small studies and case reports have examined the efficacy of surgical treatment of type 2 diabetes in overweight and mildly obese patients. Cohen et al. performed RYGB in 37 type 2 diabetes patients with BMIs between 32 and 35kg/m². Diabetes resolved in all patients post-operatively, with HbA1c levels of <6% off any oral antidiabetic agents.53 Lee and co-workers reported excellent long-term control of diabetes in patients with BMIs between 27 and 35 who underwent RYGB.54 Cohen’s group also reported post-operative resolution of diabetes (with HbA1c levels <6% off all antidiabetic medications) in two patients with BMI between 29 and 31kg/m² who had undergone laparoscopic duodenal–jejunal bypass (DJB).42 Chiellini et al.23 studied patients with type 2 diabetes and BMI 26–33kg/m² who underwent biliopancreatic diversion. Post-surgery, all patients had normalized blood glucose and dramatic improvement in all measures of glucose homeostasis. A recent study looked at the effect of laparoscopic DJB on 20 type 2 diabetes patients with BMI between 20 and 30kg/m². There was a significant reduction in fasting glycemia and HbA1c over six months post-operatively, and a significant increase in C-peptide between the third and sixth months post-operatively.55 DePaula et al. have also reported dramatic improvement of diabetes in non-obese patients undergoing laparoscopic ileal interposition.56
Future Research
Sample sizes and duration of post-operative follow-up are insufficient to assess long-term efficacy and the safety of surgery in non-obese patients. Large-scale randomized clinical trials with optimal medical and lifestyle therapy of diabetes as the comparator will be critical to advance the field, to define the role of surgery, and to implement surgery as a mainstay for type 2 diabetes therapy. These trials may indeed help to obtain critical information to identify and define more appropriate criteria for surgical indication, especially metrics of diabetes (such as pancreatic insulin reserve or insulin resistance) and metabolic disease, in agreement with the modern concept of ‘metabolic surgery.’
Conclusions
Bariatric surgery and gastrointestinal surgical procedures in particular have been shown to dramatically improve type 2 diabetes. This evidence challenges long-held medical theories that consider diabetes as a chronic, progressive, and irreversible illness. This surgery—also defined as metabolic surgery—used for diabetes treatment involves a low risk of mortality in the immediate post-operative period and a significant survival advantage over the long term compared with the mortality associated with the diagnosis of diabetes. BMI is actually thought of as the most important parameter to evaluate obesity risk and indications for gastrointestinal bypass procedures. In our opinion, it would be more accurate to consider various factors in patient selection, including ethnicity, waist circumference, metrics of diabetes, and metabolic disease. Currently, we cannot argue that every patient with diabetes is a candidate for gastrointestinal bypass procedures, but in the ‘appropriate’ patients surgery should be considered as an available treatment option, especially when diabetes is difficult to control by conventional medical therapies.