STAMPEDE: surgery vs medicines for diabetes

Overview

Schauer PR et al.. N Engl J Med 2017 · PMID 28199805 · DOI 10.1056/NEJMoa1600869

After 5 years, 23-29% of operated patients reached near-normal HbA1c vs 5% with medicines alone.

Does the study support the specific claim? No · Methodological quality: moderate

Original publication

Plain-language summary

Scientific in-depth review

1. Full reference

Authors

Schauer PR, Bhatt DL, Kirwan JP, Wolski K, Aminian A, Brethauer SA, Navaneethan SD, Singh RP, Pothier CE, Nissen SE, Kashyap SR, STAMPEDE Investigators

Journal

N Engl J Med

Year

2017

Identifier

PMID 28199805 DOI 10.1056/NEJMoa1600869 · PMC5451258

Verification

verified – PubMed E-utilities efetch, 2026-10-25

Retracted

no

Corrections

none

Extent checked

full text

Conflicts of interest and funding

Investigator-initiated; supported by Ethicon Endo-Surgery (grant EES IIS 19900), LifeScan, Cleveland Clinic and NIH (R01 DK089547). Sponsors had no role in data accrual, analysis or manuscript. Individual disclosure forms at NEJM.org (not in text).

2. Research question

Surgery plus intensive medical therapy vs intensive medical therapy alone for T2D

3. Study design and population

Study design

Randomised controlled trial – Randomized controlled trial (STAMPEDE; NCT00432809), three groups 1:1:1, stratified by insulin use, non-blinded, single centre (Cleveland Clinic); final 5-year report

Population

Adults 20-60 years with uncontrolled T2D (HbA1c >7.0%) and BMI 27-43; analysed patients: mean age 49, 66% women, HbA1c 9.2%, BMI 37 (37% BMI <35), diabetes duration 8.4 years, 44% on insulin

Country

USA

Period

Randomization March 2007 - January 2011; 5-year follow-up

Sample size

150 randomized (50 per group); 134 analysed at 5 years (38 medical, 49 gastric bypass, 47 sleeve); 9 never started treatment (8 medical, 1 sleeve), 6 lost to follow-up, 1 death

Inclusion and exclusion criteria

Age 20-60, HbA1c >7.0%, BMI 27-43

4. Methods and endpoints

Exposure / intervention

Intensive medical therapy plus RYGB or sleeve gastrectomy (with daily multivitamin, B12, vitamin D, calcium, iron)

Comparator

Intensive medical therapy alone (adjusted every 3 months for 2 years, then every 6 months; target HbA1c <=6.0%)

Endpoints

Primary: HbA1c <=6.0% with or without medications at 5 years. Secondary: glycaemic measures, weight, BP, lipids, renal, ophthalmologic outcomes, medication use, adverse events, RAND-36 quality of life

Statistical methods

Chi-square/Fisher's exact test in completers (primary analysis); Bonferroni step-down adjustment; imputed ITT (multiple, minimum and mean imputation); ANOVA for secondary endpoints; mixed models; stepwise logistic regression for predictors

5. Actual results

Results

Primary endpoint (completers): 2/38 (5.3%) medical vs 14/49 (28.6%) RYGB (unadjusted P=0.01; adjusted P=0.03) vs 11/47 (23.4%) sleeve (P=0.03; adjusted 0.07). Imputed ITT (multiple imputation): 7.3% vs 26.4% vs 20.4% (P=0.08 and 0.17); minimum/mean imputation P=0.002-0.02. HbA1c <=6.0% without medications: 0 vs 22.4% vs 14.9%; <=6.5% without medications 0 vs 30.6% vs 23.4%. HbA1c change -0.3 vs -2.1 vs -2.1 percentage points. Weight change -5.3 kg vs -23.2 kg vs -18.6 kg (RYGB > sleeve, P=0.01). About 89% of surgical patients off insulin (mean HbA1c 7.0%) vs 61% of medical (8.5%). Diabetes duration <8 years and assignment to RYGB predicted the primary endpoint; 1-year weight loss OR 1.10 (95% CI 1.04-1.16). No differences in BP, LDL, retinopathy. Adverse events (through 5 y): GERD 21% medical, 10% RYGB, 27% sleeve; anaemia 16%, 28%, 49% (sleeve P<0.05); excessive weight gain 19% vs 0 vs 0; one late reoperation (sleeve converted to bypass for fistula); 1 fatal MI (medical), 1 stroke (sleeve).

Absolute risks

HbA1c <=6.0% at 5 years: 5% medical, 29% RYGB, 23% sleeve; without diabetes medications 0%, 22%, 15%. Anaemia 16%/28%/49%; GERD 21%/10%/27%.

Authors' conclusion

Surgery plus medical therapy more effective than medical therapy alone in decreasing or resolving hyperglycemia.

6. Strengths

• Randomized, stratified, three-arm design
• Pre-specified primary endpoint and published protocol
• Objective primary outcome (HbA1c)
• 89-90% of randomized patients assessed at 5 years
• Includes patients with BMI 27-34

7. Weaknesses

• Single centre, small; underpowered for hard clinical endpoints and for RYGB vs sleeve differences (authors)
• Non-blinded
• Differential attrition: 12 of 50 medical patients missing at 5 years vs 1 and 3 in surgical arms; ITT analysis with multiple imputation not significant (P=0.08 and 0.17)
• Adjusted and ITT analyses not prespecified
• Medication use in the medical arm declined after year 3 (reduced adherence), possibly widening the gap (authors)
• Industry funding (Ethicon)

8. Risk of bias

Orienting tool

RoB 2

Type of assessment

structured assessment based on the full text by a single reviewer (no formal double review)

Full formal assessment

no

Overall risk of bias

some concerns

Rationale

Structured single-reviewer judgement from full text; a formal assessment would require two independent reviewers and the protocol.

Full-text review (25 October 2026): Added patient flow and differential attrition, ITT imputed rates (7.3/26.4/20.4%; P=0.08/0.17), remission without medications (0/22.4/14.9%), predictors, full adverse event table (GERD 21/10/27%, anaemia 16/28/49%), funding detail; RoB 2 domains added, overall unchanged (some_concerns).

9. Assessment of the statistical interpretation

Direction and size of effect favour surgery consistently (HbA1c -2.1 vs -0.3 points, P=0.003), but the binary primary endpoint is fragile in ITT with multiple imputation. Only 22% (RYGB) and 15% (sleeve) achieved HbA1c <=6.0% without medication, so 'full cure' was uncommon; most surgical patients still improved substantially.

10. Relevance to the specific claim

Limits C-BAR-08.

11. Independent contrary findings

12. Transferability to Germany and Europe

Applicable to similar patients; single US centre, BMI 27-43 including patients below usual surgical thresholds.

13. Overall methodological judgement

Randomized evidence that surgery plus medical therapy improves glycaemic control more than intensive medical therapy at 5 years; small, single-centre, with differential attrition. Shows bypass with less reflux and anaemia than sleeve in this trial, but complete remission in a minority.

14. Open questions

Hard outcomes (MI, stroke, renal failure, death); durability beyond 5 years; RYGB vs sleeve differences with adequate power.

15. Linked fact checks

• The diabetes remission achieved by bariatric surgery erodes over time (SOS: 72% at 2 years to 30.4% at 15 years; Adams: 75% to 51%). – Studies contradicting or limiting the claim
• Sleeve gastrectomy is a physiologically wrong operation and a 'transgression against God's creation', because the stomach is not a bag but an endocrine, muscular and immune organ whose functions are lost. – Studies contradicting or limiting the claim
• Gastric bypass is preferable to sleeve gastrectomy (less reflux) and can fully treat type 2 diabetes; for patients who fail dieting, bypass rather than sleeve should be chosen (2021 statement). – Studies with mixed or unclear results

Use by the Tayebat archive

Note: according to the archive, the studies were compiled by the archive operators. Whether Dr. Elawady himself cited this study is not established.

What the archive says about the study

Listed in the archive bibliography only; the archive gives no specific interpretation.

A. What was actually studied?

Surgery vs medical therapy for T2D

B. What did the authors find?

Surgery superior at 5 years

C. What conclusions do the authors themselves draw?

Surgery more effective

D. What does the archive derive from it?

Not used in narrative.

E. Is this derivation scientifically justified?

Not used; the trial contradicts the narrative's implication that diabetes benefit is illusory.

Adequacy of the interpretation (by the archive)

not applicable