Jessica A Zaman, MD, Neil Shah, Glen Leverson, PhD, Jacob Greenberg, MD, EdM, Luke Funk, MD, MPH. University of Wisconsin
Introduction: Bariatric surgery is the most effective treatment for morbidly obese type 2 diabetics. However, guidelines for perioperative glucose control are not well established. We examined optimal glucose control as defined by the American Society for Metabolic and Bariatric Surgery (ASMBS) and its effect on outcomes in bariatric surgery patients with diabetes.
Methods and Procedures: A single-institution, retrospective analysis of 325 patients who underwent laparoscopic gastric bypass or vertical sleeve gastrectomy from 2010 to 2014 was performed. Inpatient finger-stick glucose levels were recorded and defined as optimal if all values were <180 mg/dl. Ninety-day and one-year outcomes, including diabetes resolution, mortality and total costs were compared for patients with and without optimal control.
Results: Of the 155 patients with type 2 diabetes, 124 underwent gastric bypass and 31 had a sleeve gastrectomy. Diabetes resolution at one year was 70.1% (73.4% for gastric bypass and 53.9% for sleeve gastrectomy). Pre-operatively, 112 patients (72%) were on one or more oral antihyperglycemic agents while on discharge only 78 patients (50.3%) required an oral medication. Greater than 50% reduction in long-acting insulin dosage was seen in 93% of gastric bypass patients and 82% of sleeve gastrectomy patients (p = 0.251). Optimal glucose control was achieved in 60 patients (38.7%). Ninety-day and one-year outcomes including total costs were not improved by optimal perioperative glucose control (Table). However, 96.7% of optimally controlled patients experienced diabetes resolution at one year compared to 53.2% in the non-optimally controlled group (p < 0.001).
Conclusion: Bariatric surgery leads to significant type 2 diabetes resolution and a prompt decrease in glucose intolerance in the perioperative period for most patients. Optimal glucose control as defined by the ASMBS did not improve post-operative outcomes in our patient population, but was predictive of long-term diabetes resolution.
Optimal (n=60) |
Non-optimal (n=95) |
p-value | |
Length of stay (median days) | 2.4 | 2.6 | 0.302 |
Emergency department* (%) | 25.4. | 24.2 | 1.000 |
Readmission* (%) | 10.2 | 9.5 | 1.000 |
Any complication* (%) | 11.7 | 8.4 | 0.581 |
One-year mortality (%) | 0 | 1.2 | 1.000 |
Total one-year cost ($) | 18,080 | 17,988 | 0.558 |
*Within 90 days of discharge