Bariatric surgery and prevention of type 2 diabetes in Swedish obese subjects.

Carlsson LM, Peltonen M, Ahlin S, et al. Bariatric surgery and prevention of type 2 diabetes in Swedish obese subjects. N Engl J Med. 2012;367:695-704.  

The Swedish Obese Subjects (SOS) study compared the long-term effects of bariatric surgery with the effects of usual care for type 2 diabetes prevention among 3,429 subjects with obesity.

  • Bariatric surgery: banding, vertical banded gastroplasty, or gastric bypass
  • Usual care: weight loss with professional guidance or no professional guidance

Subjects in the current analysis were aged 37-60 years, had body mass index (BMI) ≥34 kg/m2 in men and BMI ≥38 kg/m2 in women, and had did not have diabetes at baseline. Study participants underwent bariatric surgery (n=1,658) or were members of the control group (n=1,771), which was matched to the surgery group on a variety of variables; the surgery group had higher mean body weight and more severe risk factors vs the control group at baseline. All participants entered the study with the intention to lose weight. Examinations occurred at baseline and after 6 months and 1, 2, 3, 4, 6, 8, 10, and 15 years. Due to high rate of subjects lost to follow-up, the current analysis includes data from 15-year follow-up, as well as sensitivity analyses.

The endpoint examined in this analysis was rate of incident type 2 diabetes (prespecified secondary endpoint in the main SOS study).

Rate of diabetes incidence, the endpoint examined in this analysis, was 6.8 cases/1,000 person-years in the bariatric surgery group vs 28.4 cases 1,000 person-years in the control group based on 15-year incidence data (P<0.0001).

There was a 78% reduction in the long-term incidence of type 2 diabetes with bariatric surgery vs usual care (HR with surgery: 0.22 [95% CI, 0.18-0.27]; P<0.001).

Development of type 2 diabetes occurred among 110 subjects in bariatric surgery group and 392 subjects in the control group.The strongest predictors of diabetes outcome in addition to treatment were baseline blood glucose level and presence/absence of impaired fasting glucose.

All bariatric surgery was associated with reduced type 2 diabetes incidence.

  • Banding (n=311): 20 events; HR 0.20 (95% CI, 0.13-0.32); P<0.001
  • Vertical banded gastroplasty (n=1,140): 84 events; HR 0.25 (95% CI, 0.19-0.31); P<0.001
  • Gastric bypass (n=207): 6 events; HR 0.12 (95% CI, 0.05-0.27); P<0.001

No difference in type 2 diabetes incidence was seen in the control group for weight loss with or without professional guidance.

  • No professional guidance (reference group; n=871): 195 events; HR (95% CI, 1.00)
  • Professional guidance (n=900): 197 events; HR 0.89 (95% CI, 0.74-1.06); P=0.20

Type 2 diabetes risk and the preventive effect of bariatric surgery increased with increasing levels of baseline glucose (P value for interaction=0.007) and serum insulin (P value for interaction=0.007); no significant interaction was seen with body mass index (BMI; P value for interaction=0.55).

The effect of surgery on diabetes incidence was highly significant in all subgroups. Significant interactions between baseline risk factors and treatment group were only seen only in subgroups defined by parameters listed below:

  • Presence/absence of impaired fasting glucose; P value for interaction=0.002
  • Fasting blood glucose; P value for interaction=0.007
  • Fasting serum insulin; P value for interaction=0.007
  • HOMA-IR; P value for interaction=0.001

When data were examined at 2 and 10 years of follow-up, sensitivity analyses showed that the effects of surgery on diabetes incidence were at least as strong as at 15 years of follow-up.

Over the first 90 days after surgery, 245 subjects (14.8%) reported at least one surgery-related complication; 46 subjects (2.8%) reported complications requiring reoperation. The mortality rate was 0.2% within 90 days after surgery.


October 2012  

This overview was created by KnowledgePoint360 Group, LLC, and was not associated with funding via an educational grant or a promotional/commercial interest.  

Related content: 

SOS Analysis: Design  

SOS Analysis: Mean Baseline Characteristics  

SOS Analysis: Rate of Diabetes Incidence  

SOS Analysis: Rate of Diabetes Incidence by Subgroup  

SOS Analysis: Interaction Between Risk Factors and Treatment  

SOS Analysis: Sensitivity Analyses; Adverse Events  


Last Modified: 11/15/2013