Cochrane on bariatric surgery: large weight and HbA1c gains, but the remission evidence is weaker than the reputation

A Cochrane network meta-analysis brought together 18 studies with 15,282 participants - 13 randomised trials totalling only 911 people and five cohort studies contributing 14,371 - comparing metabolic and bariatric surgery procedures against each other and against lifestyle or medical treatment in adults with obesity and type 2 diabetes, with at least three years of follow-up. Roux-en-Y gastric bypass gave the largest long-term weight loss against lifestyle or medical treatment (mean difference -16.95%, 95% CI -24.19 to -9.71) and a 12.34 cm greater fall in waist circumference. One-anastomosis gastric bypass and sleeve gastrectomy also beat non-surgical care, but the review rates that evidence as very uncertain. Gastric bypass reduced HbA1c by 1.44 percentage points against non-surgical care.

What is striking is where the certainty runs out. On complete diabetes remission - the outcome patients actually ask about - the evidence was rated very uncertain. The same applies to serious adverse events, all-cause mortality and total diabetes complications. The randomised evidence base for a widely performed operation is fewer than a thousand patients, and the cohort studies were judged at high or critical risk of bias and so excluded from the certainty assessment.

This does not argue against surgery. It argues for consenting honestly. Weight loss and HbA1c improvement are well supported and large. Durable, complete remission of diabetes, and any claim about living longer, are not - and neither is a confident statement that one procedure is safer than another. For Indian practice, where surgery is increasingly offered at lower BMI thresholds in line with ethnic-specific cut-offs, and where much of the trial evidence comes from higher-BMI Western cohorts, this uncertainty is wider still.

• Quote the weight and HbA1c benefit confidently; describe complete diabetes remission as possible but not reliably predicted
• Do not tell a patient that one procedure carries a lower serious complication rate than another - the comparative evidence does not support it
• Record duration of diabetes, C-peptide and insulin use before referral; these predict remission better than BMI does
• Plan lifelong micronutrient monitoring - iron, vitamin B12, vitamin D, calcium - at the point of referral, not after surgery
• Set the expectation of continued glucose-lowering therapy after surgery for most people, and review it rather than stopping it automatically

The statistics, in plain English

The confidence interval for weight loss after gastric bypass runs from -24.19% to -9.71%: the benefit is certainly real, but its size could be anything from very large to moderate. Cochrane's GRADE rating of very uncertain for remission, harms and mortality does not mean surgery lacks these benefits - it means the studies are too small, too varied and too prone to bias for the answer to be trusted at all, which is a different statement from a null result. The imbalance in this network matters too: 911 randomised patients against 14,371 in cohorts, and because those cohorts carried high or critical risk of bias they were kept out of the certainty grading, so the confident-looking total of 15,282 is not the number the conclusions rest on.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.