Bariatric surgery treats obesity, not diabetes. In a randomised trial followed beyond 10 years, mean total weight loss was about 25.5% after sleeve and 27.7% after bypass. The choice of operation then depends on reflux, associated conditions, eating pattern and operative risk.
If you are considering metabolic care or surgery for obesity, Noria Health Hub can arrange a full check-up or a second medical opinion in Brussels, with a report within 24 h, coordinate your care where needed across more than 40 medical pathways in 48 h, then, if your medical condition allows, arrange surgery within 7 days.
Obesity, how much weight stays off at 10 years?
In the Swiss randomised trial SM-BOSS, followed beyond 10 years, mean total weight loss was 25.5% after sleeve and 27.7% after bypass. The 2 operations can therefore produce durable weight loss. Bypass, however, gave less new reflux and fewer conversions to another anatomy than sleeve. It is these differences, and not diabetes alone, that should guide the choice of operation.
Diabetes, what randomised trials show
| The trial | What it measures | The result |
|---|---|---|
| 150 adults with diabetes, 5 years | Share reaching a glycated haemoglobin of 6.0% or less | 5% on medical therapy alone, 29% after bypass, 23% after sleeve |
| 262 participants followed, pooled analysis of 4 trials, 7 years | Diabetes remission | 18.2% against 6.2% on medical therapy |
| The same, at 12 years | Diabetes remission | 12.7% against 0.0% |
| The same | Glycated haemoglobin | A fall of 1.6 points against 0.2. Gap maintained at 12 years |
These are randomised trials, the highest level of evidence. But the numbers are modest, 150 to 262 patients, which limits what they can demonstrate.
Remission fades with time
In a prospective cohort, remission after bypass was seen in 71.2% of patients at 1 year, then 69.4% at 3 years, 64.6% at 5 years and 60.2% at 7 years. Among the patients initially in remission, a share therefore relapse as time passes.
And in the randomised trial at 12 years, remission concerns only about 1 operated patient in 8. Surgery durably shifts the glycaemic trajectory. It does not permanently cure diabetes. Saying one without the other would be an overpromise.
Mortality, encouraging results, but not from randomised trials
| The study | The result | The limit |
|---|---|---|
| 2,007 operated against 2,040 controls, median follow-up 24 years | All-cause death 22.8% against 26.4%. Median life expectancy higher by 3.0 years | But still 5.5 years less than the general population. And the study is not randomised |
| 21,837 matched pairs, mean follow-up 13.2 years | All-cause mortality reduced by 16% | A matched cohort, not a randomised trial |
| Pooled analysis of 4 randomised trials | No difference in major cardiovascular events | The trial was not powered to settle it |
No randomised trial has the size or the duration to measure an effect on mortality. The figures in circulation all come from matched cohorts.
The bias must be named, because it matters. Deciding to have surgery, being judged operable and returning for follow-up are three markers of health, motivation and access to care that statistical matching does not neutralise. This is healthy-candidate bias, and it mechanically inflates the apparent benefit.
Sleeve or bypass, what the head-to-head trials show
| What is compared | At 5 years | At 10 years |
|---|---|---|
| Weight loss | No significant difference between the 2 techniques | An 8.4 point gap favouring bypass, equivalence not demonstrated |
| Diabetes remission | Not measured at this point | 26% after sleeve against 33% after bypass, not significant |
| Hypertension remission | Not measured at this point | 8% against 24%, favouring bypass |
| Reinterventions | 15.8% against 22.1% | 15.7% against 18.5%, not significant |
The follow-up horizon changes the conclusion. At 5 years the 2 techniques do not separate on weight, at 10 years a gap appears. And about 1 patient in 6 had a reintervention by 10 years, whichever technique was used.
Sleeve, reflux is a significant long-term risk
Reflux is one of the best documented differences between the two techniques. At 10 years, with systematic endoscopy, oesophagitis was seen in 31% of patients after sleeve against 7% after bypass, and proton pump inhibitor use was 64% against 36%. Another trial at 5 years likewise finds more worsening of reflux after sleeve.
On Barrett’s oesophagus a range must be given, not a single figure. The randomised trial finds 4% at 10 years, an observational study with systematic endoscopy finds 18.8% beyond 5 years. The gap is a factor of 4, it stems from recruitment and from the definition used, and it is unresolved.
The benefits measured beyond weight loss
| What was measured | The result |
|---|---|
| Hypertension, randomised trial of 100 patients | 83.7% cut their antihypertensive count by at least 30% with controlled blood pressure, against 12.8% on medical therapy alone |
| Steatohepatitis proven on biopsy, randomised trial of 288 patients | Histological resolution without worsening fibrosis in 56% after bypass and 57% after sleeve, against 16% with optimised lifestyle and medical therapy |
| Sleep apnoea, 3-year sub-analysis | Share free of apnoea rising from 4.2% to 70.8%, severe forms from 20.8% to 0% |
| Sleep apnoea, randomised trial of 60 patients with polysomnography | A weight difference of over 20 kg, and yet no significantly greater reduction in the apnoea index |
The last trial is small and used a technique little practised today. But it rules out writing that the link between weight loss and correction of apnoea is mechanical and guaranteed.
Weight regain, watch which denominator is used
Two figures circulate and they look contradictory. A prospective cohort measures a mean regain of 3.9% of initial weight between year 3 and year 7. A 10-year cohort measures a mean regain of 28% of maximum weight lost, with 56.7% of patients regaining more than 20% of what they had lost.
These are not two opposite results, they are two different denominators for the same phenomenon. No consensus definition of weight regain exists, and the authors themselves call for standardisation. That is useful information in itself.
Risks, what the studies allow us to quantify
| What was measured | The result |
|---|---|
| 30-day mortality, 180,544 procedures | 0.06% in patients without heart failure, 0.74% in those with it |
| Serious 30-day complications | 2.4% without heart failure, 7.2% with |
| Long-term effects, randomised trial | Anaemia, fractures and digestive events more frequent than with medical therapy |
| Self-harm, national registry of 69,492 operated matched to 694,920 controls | 3.54 per 1,000 person-years against 0.81 per 1,000 |
| Alcohol use disorder arising after surgery, prospective cohort at 5 years | 20.8% after bypass against 11.3% after gastric band |
These events remain rare in absolute terms. But the self-harm signal is consistent across a national registry, a matched study, a US cohort and a meta-analysis, that is 4 independent sources.
The correct reading is not that surgery makes people suicidal. The factors identified are psychiatric history, psychotropic use, sleep disorders, young age, socioeconomic disadvantage, and not attending follow-up appointments. The practical consequence is structured and lasting psychological follow-up, not abstention.
New drug treatments, important results, but not directly comparable with surgery
Two recent randomised trials measure weight losses of 20.9% and 14.9% at 68 to 72 weeks, against 3.1% and 2.4% on placebo. These figures are accurate and striking.
But what is not compared must be said. No randomised trial has directly compared bariatric surgery with these treatments on hard endpoints. The horizons differ, 72 weeks against 5 to 24 years. The drug trials measure the effect under continuous treatment, and the question of stopping is not settled there. And the populations, the inclusion criteria and the adverse effect profiles are not the same. Placing the percentages side by side would be factually invalid, even though each figure taken alone is accurate.
What the studies do not allow anyone to claim
We will not claim that surgery cures diabetes, because remission falls to about 1 operated patient in 8 at 12 years. Nothing entitles us to write that it extends life, because no randomised trial has demonstrated it and matched cohorts carry a known bias. You will not read here that it corrects sleep apnoea, because the only randomised trial with polysomnography on that endpoint is negative.
And we will repeat that each of these figures describes a population, not a person. In the same trial, excess weight loss at 10 years ranged from 2.1% to 109.2% across patients. No average predicts an individual result.
In practice, if this concerns you
You set out your problem, your symptoms or your question. Reports, test results and images already available can be studied before you arrive, so that the pathway is prepared. The aim is not to run every test as a matter of course, but to select those that can genuinely help confirm a hypothesis, rule out a risk or guide the decision. The results are then brought together and the next steps are organised. Follow-up is coordinated afterwards.
Need a medical opinion quickly? For a concern that does not call for the emergency services, you can ask to speak to a doctor at any time through Noria Health Hub. This on-demand medical consultation is billed separately from the work-up pathway.
Further reading
The companion articles.
What” rel=”noopener” target=”_blank”>https://noriahealth.com/blog/what-a-health-assessment-can-actually-find/”>What a health assessment can actually find
What” rel=”noopener” target=”_blank”>https://noriahealth.com/blog/what-one-night-of-recording-actually-measures/”>What one night of recording actually measures
Sources
Schauer PR et al., STAMPEDE, Bariatric surgery versus intensive medical therapy for diabetes, 5-year outcomes, New England Journal of Medicine, 2017. https://doi.org/10.1056/NEJMoa1600869
Courcoulas AP et al., ARMMS-T2D, Long-term outcomes of medical management versus metabolic surgery in type 2 diabetes, JAMA, 2024. https://doi.org/10.1001/jama.2024.0318
Courcoulas AP et al., Seven-year weight trajectories and health outcomes in the LABS-2 study, JAMA Surgery, 2018. https://doi.org/10.1001/jamasurg.2017.5025
Carlsson LMS et al., Life expectancy after bariatric surgery in the Swedish Obese Subjects study, New England Journal of Medicine, 2020. https://doi.org/10.1056/NEJMoa2002449
Mortality after bariatric surgery, a matched cohort of 21,837 pairs, Obesity, 2023. https://doi.org/10.1002/oby.23646
Salminen P et al., SLEEVEPASS, Effect of laparoscopic sleeve gastrectomy versus Roux-en-Y gastric bypass at 10 years, JAMA Surgery, 2022. https://doi.org/10.1001/jamasurg.2022.2229
Peterli R et al., SM-BOSS, Effect of laparoscopic sleeve gastrectomy versus Roux-en-Y gastric bypass at 5 years, JAMA, 2018. https://doi.org/10.1001/jama.2017.20897
Barrett’s oesophagus after sleeve gastrectomy, multicentre endoscopic study, Obesity Surgery, 2019. https://doi.org/10.1007/s11695-019-03704-y
Schiavon CA et al., GATEWAY, Effects of bariatric surgery in obese patients with hypertension, Circulation, 2018. https://doi.org/10.1161/CIRCULATIONAHA.117.032130
GATEWAY, obstructive sleep apnoea at 3 years, International Journal of Obesity, 2021. https://doi.org/10.1038/s41366-021-00752-2
Dixon JB et al., Surgical versus conventional therapy for weight loss treatment of obstructive sleep apnea, JAMA, 2012. https://doi.org/10.1001/2012.jama.11580
Verrastro O et al., BRAVES, Bariatric-metabolic surgery versus lifestyle intervention plus best medical care in non-alcoholic steatohepatitis, The Lancet, 2023. https://doi.org/10.1016/S0140-6736(23)00634-7
Thirty-day outcomes of primary metabolic and bariatric surgery, national accreditation database, Obesity Surgery, 2026. https://doi.org/10.1007/s11695-025-08481-5
Weight regain 10 years after metabolic surgery, Obesity Facts, 2026. https://doi.org/10.1159/000549498
Jastreboff AM et al., SURMOUNT-1, Tirzepatide once weekly for the treatment of obesity, New England Journal of Medicine, 2022. https://doi.org/10.1056/NEJMoa2206038
Wilding JPH et al., STEP-1, Once-weekly semaglutide in adults with overweight or obesity, New England Journal of Medicine, 2021. https://doi.org/10.1056/NEJMoa2032183
Comparative effectiveness of surgical, endoscopic and pharmacological treatments, network meta-analysis, Obesity, 2026. https://doi.org/10.1002/oby.70083
Self-harm after bariatric surgery, Swedish national registry, Clinical Obesity, 2023. https://doi.org/10.1111/cob.12576
Suicide and non-fatal self-harm after bariatric surgery, Annals of Surgery, 2021. https://doi.org/10.1097/SLA.0000000000003742
King WC et al., Alcohol and other substance use after bariatric surgery, prospective evidence from the LABS-2 cohort, Surgery for Obesity and Related Diseases, 2017. https://doi.org/10.1016/j.soard.2017.03.021
Kraljevic M et al., Long-Term Outcomes of Roux-en-Y Gastric Bypass vs Sleeve Gastrectomy for Obesity: SM-BOSS randomized clinical trial, JAMA Surgery, 2025. https://doi.org/10.1001/jamasurg.2024.7052