Should Bariatric Surgery be a Treatment Option for Type 2 Diabetes?   (March 28, 2012))

Bariatric surgery to "reverse" diabetes is again in the news. On March 26, the New England Journal of Medicine published the results of a surgery study, Bariatric Surgery versus Intensive Medical Therapy in Obese Patients with Diabetes.

This study, with as usual a cute acronym, the STAMPEDE (Surgical Therapy And Medications Potentially Eradicate Diabetes Efficiently) trial, is without a doubt the best study yet published to evaluate if bariatric surgery might be a treatment option for treatment of type 2 diabetes - but it's not perfect (see below). STAMPEDE was designed as a prospective trial (looking ahead, not looking back at data from the past), and randomized patients into surgery vs. state-of-the-art medical therapy, while evaluating A1C and other end-points over a one-year period of time, in patients with T2DM and moderately-high Body Mass Index (BMI) levels -- one of the first times researchers have studied doing diabetes weight loss surgery in patients that aren't severely obese. (Usually, bariatric surgery is restricted to patients with BMI levels over 35.)  The study's primary endpoint was the percent of treatment group achieving an A1C less than or at 6% at 1 year. It's planned to continue to follow the patients for a total of five years.

STAMPEDE involved 150 men and women with a clinical diagnosis of type 2 diabetes between the ages of 20 and 60 years old with a BMI of 27-43. 50 patients were treated with advanced medical therapy (AMT), 50 with Laparoscopic Sleeve Gastrectomy and AMT, and 50 with Roux-en-Y Gastric Bypass and AMT. There were eight dropouts from the study: seven withdrew immediately after randomization from the AMT group; one patient did not undergo sleeve gastrectomy. 

There were no deaths during the year of followup; a listing of adverse events are available on-line and included four reoperations in the surgically treated groups.

A press release indicated that "This study was funded by Ethicon Endo-Surgery, Inc., a subsidiary of Johnson & Johnson, which is a company that designs and manufactures medical devices and surgical instruments. Dr. Schauer is a paid advisory board member of Ethicon, and is in complete compliance with all Cleveland Clinic COI policies."

The results of STAMPEDE? Not surprisingly, patients who underwent surgery lost more weight and were more successful at controlling their A1C. The target A1C level  (6.0% or less at 12 months) occurred in 5 of 41 patients (12%) in the AMT group, as compared with 21 of 50 (42%) in the gastric-bypass group and 18 of 49 (37%) in the sleeve-gastrectomy group. Note: these are extremely tight targets for A1C, and according to at least one recent study (ACCORD), these targets might put the patients at increased risk of death.

Should this study be considered definitive? No. As I indicated above, the study design is not perfect. And the follow-up thus far is way too short (one year). And a cost-benefit analysis should be done.

Does this change my prior opinion that bariatric surgery to "cure" diabetes should be considered experimental? No, not at all. My main concern with STAMPEDE or other such trials is that patients randomized to surgery are treated differently than those randomized away from surgery. The study design was not "double-blinded" -- both the patients and the staff knew whether the patient had been operated upon. This introduces what is called "investigator bias" (which, as one author defines it, is "bias resulting from a conflict of interest arising from passionate beliefs held by the investigators"): a huge issue when the lead researcher is being paid by the trial's funding agent.

Could a similar study have been designed that would be double-blind? Yes! A study with "sham surgery"  being done in one arm of the study is, IMHO, easy to design and in my thinking, very ethical. In such a trial, patients would agree as part of the informed consent that all participants in the trial would have surgery, but that some would receive sham surgery. That is, some of the patients would be wheeled into the operating room, and have anesthesia and an incision in ther belly identical to that if receiving "real" surgery, but the surgeon would not change anything internally. The surgeon doing the surgery would then bow out of the study, and not disclose to any study staffers or to the patient whether real or sham surgery had been done (unless there was an emergency requiring "unblinding" of what had been done).

Why do I stress sham surgery? Because in the 1950's, there was a popular surgical procedure to treat angina pectoris (cardiac chest pain) in which the internal mammary artery was tied off. But then, a surgeon named H.K. Beecher described the placebo effect of such surgery in a classic paper following a randomized trial of internal mammary artery ligation versus a sham operation. Thereafter, this surgical procedure was no longer done: it was the placebo effect, not the tying off of an artery, that resulted in the improvement.

To paraphrase what I've said before, bariatric surgery for diabetes control is experimental surgery, even if short-term improvement of diabetes is likely, and even though Medicare apparently will be reimbursing the procedure. And as such, it remains my opinion that this surgery should only rarely be undertaken, and when done, it should be done as part of a clinical trial, with a highly qualified team, and appropriate safeguards, including informed consent from the patient that they are taking part in an experiment.