Surgical Treatment of Diabetes: Making a Case for a Pragmatic Approach

The development of type 2 diabetes mellitus (T2DM) is one of the most serious consequences of weight gain and obesity. Despite an understanding of this association by patients and their physicians, weight loss through diet and exercise, the typical lifestyle modifications recommended to break the yoke of diabetes that hangs on obese individuals, requires intensive interventions that can be difficult to maintain (1, 2). Moreover, lifestyle measures are generally inadequate in patients with severe obesity who are at particularly high risk of developing T2DM and its complications. Although developed primarily to facilitate weight loss, bariatric surgery has been noted to cause marked improvement in diabetes control (3). Emerging data have provided convincing evidence that bariatric surgery can produce rapid, effective, and sustainable remission of T2DM in addition to weight loss and is superior to nonsurgical interventions (3–7). Both the American Diabetes Association (8) and the American Association of Clinical Endocrinologists, The Obesity Society, and American Society for Metabolic & Bariatric Surgery (9) have now expanded their treatment recommendations to include bariatric surgery as an option for individuals with morbid obesity, body mass index (BMI) ≥ 40 kg/m² or ≥ 35 kg/m² with comorbid conditions such as T2DM.

Recently, the US Food and Drug Administration approved laparoscopic adjustable gastric banding (LAGB) for individuals with BMI of 30–34 kg/m² and comorbidities. LAGB is a relatively simple bariatric procedure in which a saline-filled silicon band is fitted around the stomach near the esophageal junction. The level of gastric restriction imposed by the band can be adjusted by infusing saline through a sc port maintaining the desired level of gastric restriction. LAGB is the least invasive of the commonly used bariatric surgical procedures, with a perioperative mortality rate of approximately 0.05% (10). The safety, ease, and greater effectiveness for weight loss than what is usually obtained with lifestyle measures make the rationale for use of LAGB at lower levels of BMI, and raise the possibility of applying this procedure more widely in the treatment of T2DM. The outcomes of two previous randomized controlled studies have demonstrated that subjects receiving LAGB have superior weight loss and improvement in diabetes control relative to those getting medical or lifestyle management (4, 11).

In this issue of the JCEM, Ding et al (12) add another study comparing the effectiveness of LAGB to medical/lifestyle measures for the treatment of T2DM. They conducted a 12-month randomized, controlled, single-center trial comparing 23 diabetic subjects assigned LAGB with 22 subjects given intensive medical and weight management (IMWM) in a structured program based on previous successful regimens. Notably, they recruited diabetic patients without regard to ongoing treatment or duration of disease and with a broad range of diabetic control. In this sense, the study was arguably more pragmatic than some previous trials in that it involved a population reflective of the average diabetes practice. The average BMI for the study group was 36.5 ± 3.7 kg/m²; they had been diagnosed with diabetes for 9 ± 5 years (10 ± 6 in the surgery group and 8 ± 4 in the control), and their glycated hemoglobin (HbA1c) was 8.2 ± 1.2%. About 40% of the patients were on insulin (72% in the surgery group and 18% in the IMWM group). After 1 year, the primary endpoint (HbA1c < 6.5% and fasting glucose < 7.0 mmol/L) was achieved in 33% of the LAGB group and 23% of the IMWM group, a difference that did not differ statistically. HbA1c reduction over the course of the trial was similar between groups (∼1%), but weight loss was greater in the surgically treated subjects (−13.5 ± 1.7 vs −8.5 ± 1.6 kg; P = .027). There was a slightly greater decrease in systolic blood pressure in IMWM than LAGB in this mostly normotensive cohort, whereas changes in other measures of cardiovascular risks and general health status improved similarly between groups. No participants experienced severe hypoglycemia during the trial. The authors conclude that a multidisciplinary IMWN produces similar benefits to LAGB for diabetes control and cardiometabolic risk over 1 year. In addition, they explain the difference between their results and previous trials that showed a relative benefit of LAGB as being due primarily to the older, more diabetic cohort they enrolled. This latter conclusion, leaning on the pragmatic nature of their recruitment, is the major new information added by this study to what is known about surgical treatment of diabetes, and it bears some scrutiny.

The recent successes of bariatric surgery to remedy dysglycemia and other metabolic abnormalities of diabetes (3–7) have been tempered by emerging evidence that results are not as robust in patients with advanced disease. Thus, higher baseline C-peptide, lower blood glucose, absence of insulin therapy, and better-preserved β-cell function before surgery have all been associated with more successful T2DM resolution after one or another bariatric procedures (13–15). Consistent with this, patients with longer duration of diabetes and worse β-cell function had smaller declines in HbA1c than those with a shorter time from diagnosis, despite comparable weight loss (16).

Therefore the argument advanced by Ding et al (12) for their results has surface validity. Certainly the duration of diabetes, the baseline HbA1c, and the number of patients in their cohort taking insulin was higher than the group studied by Dixon et al (4), who recruited only patients with a T2DM diagnosis of 2 years or less. However, the patients in the present paper were only marginally different than those reported by Courcoulas et al (11), where in fact a benefit from surgery was demonstrated. It is clear that T2DM is a heterogeneous condition, perhaps as much syndrome as disease, such that it is difficult to distinguish clear differences between two cohorts, especially ones limited to 20–30 subjects, unless specific criteria are used in their selection.

Although the disproportionately high percentage of LAGB subjects taking insulin in the Ding et al (12) study, ie, 72%, may explain the relatively small effects of surgery on diabetes control, what also sets the results here apart is the superior effect of the IMWM program. Nearly a quarter of the subjects randomized to lifestyle/medical treatment met the primary outcome, substantially more than the 0% and 13% of similarly treated subjects in the studies of Courcoulas et al (11) and Dixon et al (4), respectively. This response is a testament to the efficacy of multidisciplinary approaches to lifestyle treatments (1, 2) and the utility to subjecting these sorts of programs to systematic investigation and modification, as the authors have done. However, there appears to be some effect of random chance in this result as well. To wit, a similar group of diabetic subjects treated with an identical regimen by this group in a parallel trial of gastric bypass had only 13% diabetes remission (17). The variability in the response to IMWM delivered at a single center to obese T2DM subjects provides a tangible example of the need to enhance the power of comparative effectiveness studies with larger samples.

The work of Ding et al (12) raises questions about the comparative effectiveness of LAGB in the treatment of the broad swath of T2DM patients, particularly those with advanced disease. This level of doubt is not present for bariatric procedures like gastric bypass that have larger weight loss effects than LAGB and have been demonstrated in randomized controlled trials to have substantially greater effects on diabetes control (4–6, 11). To be sure, the randomized controlled trials comparing surgery and medical treatment of diabetes are challenging endeavors, but they have been very influential in shaping the landscape around diabetes treatment while also driving research into a potentially fruitful area of physiology. In the sense that Ding et al (12) studied diabetic patients without regard to their treatment modalities or severity of disease, they have added a level of “pragmatism” to this field of research. On the other hand, like most of the other randomized controlled trials of bariatric surgery in diabetes, theirs was performed at an elite academic medical center. A next step in this line of investigation would be to extend the generalizability of surgery for diabetes outside the ivory tower. Although certainly a daunting task, the string of superb clinical studies in this area over the past decade suggests that it is a reasonable goal.