Bariatric Surgery and Sleep Apnea: The Trial Nobody Quotes
A randomised trial gave one group five times the weight loss of the other. The difference in apnea severity between them did not reach significance. That result is worth understanding before anyone promises surgery will fix your breathing.
Obesity is the dominant modifiable risk factor for obstructive sleep apnea, weight loss reduces apnea severity, and bariatric surgery produces more weight loss than anything else short of a GLP-1 drug. The conclusion looks obvious.
There is a randomised trial that tested it directly, and the result is more awkward than the logic predicts.
The trial
Published in JAMA in 2012, it randomised 60 obese adults — BMI between 35 and 55 — with recently diagnosed obstructive sleep apnea and an AHI of at least 20. All had been prescribed CPAP.
Thirty went to a conventional weight loss programme with dietitian and physician consultations and very low-calorie diets as needed. Thirty went to bariatric surgery, specifically laparoscopic adjustable gastric banding. The primary outcome was change in AHI on polysomnography at two years, scored by staff blinded to which group the patient was in.
The weight result, then the apnea result
The weight difference was emphatic:
| Weight change at 2 years | |
|---|---|
| Conventional programme | −5.1 kg (95% CI 0.8 to 9.3) |
| Bariatric surgery | −27.8 kg (95% CI 20.9 to 34.7) |
That is more than five times the loss, P<0.001. Surgery did exactly what surgery does.
Now the apnea:
| AHI change at 2 years | |
|---|---|
| Conventional programme | −14.0 events/hour (95% CI 3.3 to 24.6) |
| Bariatric surgery | −25.5 events/hour (95% CI 14.2 to 36.7) |
Both groups improved substantially. But the between-group difference was −11.5 events per hour, 95% CI −28.3 to 5.3, P=0.18 — not statistically significant.
The authors' own conclusion: bariatric surgery compared with conventional weight loss therapy did not result in a statistically greater reduction in AHI despite major differences in weight loss.
CPAP adherence did not differ between the groups either. Surgery did produce a greater improvement in the SF-36 physical component score (mean 9.3, 95% CI 0.5 to 18.0, P=0.04).
What this does and does not mean
It does not mean bariatric surgery fails to help sleep apnea. Read the confidence interval: −28.3 to 5.3. That interval contains a 28-event advantage for surgery. With 60 patients the trial simply could not distinguish a large benefit from none — this is an underpowered null, not a demonstration of equivalence. Anyone citing this as proof that surgery does not work for apnea is misreading it as badly as someone citing it as proof that it does.
What it does establish is that the relationship between kilograms lost and apnea severity is far looser than assumed. The conventional group lost 5 kg and dropped 14 events per hour. If AHI tracked weight in any straightforward way, that should not happen.
The likely reason is that what matters is not mass but where fat sits — tongue fat, parapharyngeal fat, neck circumference — and how collapsible your airway is to begin with. Two people losing 20 kg can get entirely different airway results. Why neck circumference beats the scale.
It also matters that both arms were still substantially apneic at the end. Starting at an AHI of 20 or more and improving by 25 events is excellent and frequently still leaves measurable disease.
How this sits next to the drugs
This is the useful comparison, because the drug evidence arrived later and is cleaner.
SURMOUNT-OSA randomised 469 people with moderate-to-severe apnea and obesity to tirzepatide or placebo and produced an AHI reduction of roughly 20 to 24 events per hour more than placebo, with 42% to 50% reaching remission. That is a placebo-controlled effect on apnea specifically, in a trial powered to detect it. The detail, and what the approval covers.
So the position in 2026 is genuinely odd: the pharmacological route has better apnea-specific randomised evidence than the surgical route, despite surgery producing more weight loss. That is a statement about which trials were run and how large, not about biology. And the evidence is drug-specific rather than class-wide.
Evidence tier 1 that weight loss reduces apnea severity. Tier 2 that surgical weight loss beats medical weight loss for apnea specifically — plausible, mechanistically sensible, and not demonstrated in an adequately powered trial.
The practical version
If you are considering bariatric surgery, apnea improvement is a reasonable expectation and a poor primary justification. The trial group that had surgery improved by 25 events per hour. They also had greater physical quality-of-life gains. Those are good reasons. "It will fix my sleep apnea" is a promise the evidence does not support with confidence.
Do not stop CPAP on the strength of weight loss. CPAP adherence did not differ between arms in this trial, and both arms retained measurable apnea. Whatever route you take, the machine question is settled by a repeat sleep study, not by the scale.
Get retested afterwards. This is the recurring point across every weight-based treatment for apnea: severity changes, the change is not predictable from weight, and the only way to know your current AHI is to measure it. What a sleep test involves.
And if you are reading this as a snorer rather than a diagnosed patient, the order of operations is a diagnosis first. Witnessed pauses, gasping arousals, waking unrefreshed, or sleepiness that more sleep does not fix all point at testing. Snoring versus apnea.