A Cochrane network meta-analysis pooled 18 studies in adults with obesity and type 2 diabetes, all with at least three years of follow-up: 13 randomised trials contributing 911 people and five cohort studies contributing 14,371. It compared Roux-en-Y gastric bypass, sleeve gastrectomy, one-anastomosis gastric bypass, adjustable gastric banding and gastric plication against each other and against lifestyle or medical treatment.
At five years or more, Roux-en-Y gastric bypass gave 16.95% more weight loss than lifestyle or medical treatment (95% CI -24.19 to -9.71), 12.34 cm more waist reduction (CI -17.88 to -6.81) and an HbA1c 1.44 percentage points lower (CI -2.26 to -0.63). One-anastomosis bypass and sleeve gastrectomy also beat medical treatment, but on very uncertain evidence. The part worth reading twice is what the review could not establish: for complete diabetes remission, serious adverse events, all-cause mortality and diabetes complications, the certainty was very low. Only 536 people across nine trials contributed to the remission analysis.
This is not an argument against surgery. It is an argument for precision when counselling. The confident claims are about weight, waist circumference and HbA1c. The claims most often made across the desk — that surgery puts diabetes into remission, that it is proven to cut complications and death — rest on far thinner data than most people assume. Promise what is supported, and describe remission and hard outcomes as likely but not yet demonstrated at this level of evidence.
- When counselling for surgery, separate what is well supported (weight, waist, HbA1c) from what is not (complete remission, mortality, complications).
- Use ethnic-specific BMI cut-offs when assessing Indian patients — the review itself allowed for them.
- Document pre-operative HbA1c, weight, waist circumference and diabetes duration; these are what you will judge the result against.
- Arrange micronutrient monitoring and follow-up before referral, not after the operation.
- For a patient choosing between procedures, Roux-en-Y gastric bypass has the strongest evidence base here, despite sleeve gastrectomy being done more often.
The statistics, in plain English
The GRADE certainty ratings are doing the real work in this review. 'Very uncertain' does not mean the effect is absent; it means the trials are too small and too prone to bias for the estimate to be trusted. The remission analysis rested on 536 people across nine trials, and when events are few an interval can be wide enough that benefit and harm both fit the data. The wide interval on gastric bypass weight loss, -24.19% to -9.71%, tells the same story: the direction is certain but anything from about a tenth to a quarter of body weight is compatible. The cohort studies added 14,371 people and agreed with the trial findings, but were judged at high or critical risk of bias and so were excluded from the certainty rating altogether.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for diabetes & endocrinology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free