, Volume 59, Issue 5, pp 945-953

First online:

Gastric bypass surgery vs intensive lifestyle and medical intervention for type 2 diabetes: the CROSSROADS randomised controlled trial

  • David E. CummingsAffiliated withDepartment of Medicine, University of Washington Email author 
  • , David E. ArterburnAffiliated withGroup Health Research Institute
  • , Emily O. WestbrookAffiliated withGroup Health Research Institute
  • , Jessica N. KuzmaAffiliated withFred Hutchinson Cancer Research Center
  • , Skye D. StewartAffiliated withDepartment of Surgery, University of Washington
  • , Chun P. ChanAffiliated withDepartment of Surgery, University of Washington
  • , Steven N. BockAffiliated withDepartment of Surgery, University of New Mexico
  • , Jeffrey T. LandersAffiliated withGroup Health Physicians
  • , Mario KratzAffiliated withFred Hutchinson Cancer Research Center
    • , Karen E. Foster-SchubertAffiliated withDepartment of Medicine, University of Washington
    • , David R. FlumAffiliated withDepartment of Surgery, University of Washington



Mounting evidence indicates that Roux-en-Y gastric bypass (RYGB) ameliorates type 2 diabetes, but randomised trials comparing surgical vs nonsurgical care are needed. With a parallel-group randomised controlled trial (RCT), we compared RYGB vs an intensive lifestyle and medical intervention (ILMI) for type 2 diabetes, including among patients with a BMI <35 kg/m2.


By use of a shared decision-making recruitment strategy targeting the entire at-risk population within an integrated community healthcare system, we screened 1,808 adults meeting inclusion criteria (age 25–64, with type 2 diabetes and a BMI 30–45 kg/m2). Of these, 43 were allocated via concealed, computer-generated random assignment in a 1:1 ratio to RYGB or ILMI. The latter involved ≥45 min of aerobic exercise 5 days per week, a dietitian-directed weight- and glucose-lowering diet, and optimal diabetes medical treatment for 1 year. Although treatment allocation could not be blinded, outcomes were determined by a blinded adjudicator. The primary outcome was diabetes remission at 1 year (HbA1c <6.0% [<42.1 mmol/mol], off all diabetes medicines).


Twenty-three volunteers were assigned to RYGB and 20 to ILMI. Of these, 11 withdrew before receiving any intervention. Hence 15 in the RYGB group and 17 in the IMLI group were analysed throughout 1 year. The groups were equivalent regarding all baseline characteristics, except that the RYGB cohort had a longer diabetes duration (11.4 ± 4.8 vs 6.8 ± 5.2 years, p = 0.009). Weight loss at 1 year was 25.8 ± 14.5% vs 6.4 ± 5.8% after RYGB vs ILMI, respectively (p < 0.001). The ILMI exercise programme yielded a 22 ± 11% increase in \( \overset{\cdot }{V}{\mathrm{O}}_{2 \max } \) (p<0.0001), whereas \( \overset{\cdot }{V}{\mathrm{O}}_{2 \max } \) after RYGB was unchanged. Diabetes remission at 1 year was 60.0% with RYGB vs 5.9% with ILMI (p = 0.002). The HbA1c decline over 1 year was only modestly more after RYGB than ILMI: from 7.7 ± 1.0% (60.7 mmol/mol) to 6.4 ± 1.6% (46.4 mmol/mol) vs 7.3 ± 0.9% (56.3 mmol/mol) to 6.9 ± 1.3% (51.9 mmol/mol), respectively (p = 0.04); however, this drop occurred with significantly fewer or no diabetes medications after RYGB. No life-threatening complications occurred.


Compared with the most rigorous ILMI yet tested against surgery in a randomised trial, RYGB yielded greater type 2 diabetes remission in mild-to-moderately obese patients recruited from a well-informed, population-based sample.

Trial registration: NCT01295229


Bariatric surgery Diabetes Intensive lifestyle Metabolic surgery Randomised controlled trial