Does Sleep Apnea Cause Cancer? What a 20,000-Patient Study Found
The headline association is real and the causal claim mostly is not. In a cohort of over 20,000 sleep-clinic patients, low overnight oxygen was independently linked to existing cancer — but apnea severity did not independently predict developing it.
If you have searched this because something frightened you, the headline answer is reassuring. The largest clinical cohort to look at it concluded that sleep apnea is not an independent risk factor for developing cancer.
The more interesting finding is the part underneath, where two different measures of the same condition gave two different answers.
What the study was
Researchers reporting in Chest analysed a Western Australian sleep clinic cohort of 20,289 people, linking in-laboratory sleep studies to state cancer and death registries.
They did two things that make the result worth taking seriously.
First, they measured apnea two ways: by apnea-hypopnea index, the standard severity count, and by nocturnal hypoxemia — the duration and percentage of the night spent below 90% oxygen saturation. Those are not the same thing, and a central argument of modern sleep medicine is that the second matters more. Hypoxic burden.
Second, they confined analyses within periods using consistent AHI scoring criteria, because the rules for scoring a hypopnea changed partway through the era covered — which is a methodological detail most studies in this field skate over.
Median follow-up for cancer incidence was 11.2 years.
What it found
The two measures diverged, and the divergence is the finding.
On cancer people already had: nocturnal hypoxemia — but not AHI — was independently associated with prevalent cancer, after adjusting for age, sex, BMI, smoking status, socioeconomic status and blood pressure.
On cancer people went on to develop: in univariable models, higher hazard ratios appeared for mild, moderate and severe apnea compared with no apnea. That is the association that generates the headlines.
Then the adjustment. After controlling for confounders, there was no independent association between apnea severity and cancer incidence. The authors' own interpretation: "OSA severity is associated with incident cancer, although this association seems secondary to other risk factors for cancer development. OSA is not an independent risk factor for cancer incidence."
What was doing the work instead was age and, in some cases, sex, BMI and smoking status — which are the things that drive both cancer risk and apnea risk simultaneously.
Evidence tier 1 for the conclusion. Twenty thousand patients, registry-linked outcomes, over a decade of follow-up, and adjustment for the obvious confounders. Tier 2 for the hypoxemia-and-prevalent-cancer signal, which is cross-sectional and cannot establish direction.
Why confounding matters so much here
This is the clearest illustration on this site of why an unadjusted association means very little.
Obesity raises the risk of sleep apnea. Obesity also raises the risk of several cancers. Smoking raises the risk of snoring and apnea through airway inflammation. Smoking is the single largest modifiable cause of cancer. Age raises everything.
So a cohort of people with apnea will contain more older, heavier smokers than a cohort without — and they will develop more cancer whether or not the apnea contributed anything. Removing those overlaps is what the multivariable model does, and when it did, the apnea signal went with it.
That pattern is worth recognising generally, because it recurs constantly in sleep headlines. How to read sleep research in the news.
What about the hypoxemia finding?
It is the part that is not dismissed, and it should be read carefully.
Low overnight oxygen was independently associated with having cancer, not with getting it. A cross-sectional association cannot tell you which came first. Cancer, its treatment, and the deconditioning that accompanies both can all plausibly worsen overnight oxygenation — so the arrow could run either way, or both could share a cause.
The biological argument for the other direction is not empty. Intermittent hypoxia is studied as a driver of angiogenesis and tumour growth in laboratory models, and there is separate work on hypoxemia and outcomes in people who already have cancer. But a laboratory mechanism plus a cross-sectional association is not evidence of causation in humans, and this study specifically failed to find the incidence signal that causation would predict.
So what should you take from it?
Do not treat your apnea because of cancer. On this evidence that is not the reason, and better reasons already exist.
Do treat it. The cardiovascular, metabolic, cognitive and accident risks of untreated apnea are considerably better established than any cancer link, and they are what the case for treatment rests on. What untreated apnea actually does.
And note which number the study found informative. Nocturnal hypoxemia carried the signal that AHI did not — which is the same lesson appearing across this field, and a reason to ask what your oxygen data showed rather than only your AHI. Reading your sleep study.
The line
No consumer product treats obstructive sleep apnea. If you have witnessed breathing pauses, gasping arousals, morning headaches, or sleepiness that more sleep does not fix, the next step is assessment — and that holds regardless of what this particular study found. Snoring versus sleep apnea.