OSA and Cancer Link

I've been poking into the literature a bit recently (using PubMed, recent reviews on OSA) and have found many papers showing that untreated or poorly treated OSA results in inflammation of the airway and other associated structures (hardly surprising) and perhaps also some systemic inflammatory effects. There are a number of blood biomarkers associated with inflammation (e.g. cytokines, free radical-associated molecules and some biomarkers I'd never heard of) that correlate very strongly with active OSA vs treated OSA or normal subjects. I wish that I'd discovered cancer biomarkers that were as strongly associated with disease or response to therapy as some of these - that was my field for a few years until I retired.

But I have never, not in over 20 years of cancer research, come across a paper with good evidence for an association between OSA and cancer. I'm not saying that such research doesn't exist, I just haven't seen it. If someone else knows of a paper, please post the reference here and we can look into the science.

But severe, untreated OSA can cause sudden nighttime death. I'll never forget one phone conference I was participating in with a group of cancer experts (outside investigators, not pharma employees) in which we were discussing an ongoing Phase 2 oncology clinical trial. The docs on the call were from various well known cancer institutions and had patients on the trial; I was the biomarker specialist for the pharma company on the call. The issue of cause-of-death for patients on study came up in the conversation - of course all such events are thoroughly evaluated and reported to FDA, and prolongation of overall survival time ("OS", any cause) is often the key measure of efficacy for a test compound and the basis for FDA approval or rejection. Anyway, one subject died in the middle of the night on drug treatment for a reason unrelated to his cancer (in the opinion of the responsible physican/investigator) and immediately the other physicians on the phone call discussed the possibility of sudden death due to an apnea-induced crisis. This was soon after my own OSA diagnosis and I was most impressed that this possibility was given such a priority. These were eminent docs in their specialty (Harvard Med Sch etc), very well published and extremely knowledgeable.

Untreated severe OSA is undoubtedly a very serious condition but there's no reason to fear cancer as far as I know - you probably won't live long enough!

moonslice wrote:she's a little nutz! <snip> I think it did something to my Mom's brain!

Maybe she has untreated sleep apnea? It does cause problems with the brain.

Pneumophile wrote:one subject died in the middle of the night

I've heard that, if you divide the day into three-hour periods, you will find significantly more deaths from heart attacks and unexplained causes during the 3:00 - 6:00 a.m. period. The explanation for this is that sleep apnea is most severe during this time when the most REM sleep is reached.

Sorry, no references on that claim.

{{{Direct link between UNTREATED sleep apnea and cardiac problems (heart attacks)}}}
I guess you can pick your poison.
Or just ignore every dumb thing your mother says, give her a hug, and know she means well.

Chevie wrote:

Pneumophile wrote:one subject died in the middle of the night

I've heard that, if you divide the day into three-hour periods, you will find significantly more deaths from heart attacks and unexplained causes during the 3:00 - 6:00 a.m. period. The explanation for this is that sleep apnea is most severe during this time when the most REM sleep is reached.

Sorry, no references on that claim.

Very interesting and highly plausible. My own poly study showed much more frequent apnea events during REM than non-REM.

Pneumophile wrote:much more frequent apnea events during REM than non-REM.

I've been told that the muscles in the body are at maximum relaxation during REM. This would include the muscles that hold the airway open.

Arlene1963 wrote:
Moonslice, here's a large Canadian study on OSA and the development of Cancer, published in 2014. The full text version indicates that the authors took into account reported CPAP versus non CPAP use, and this didn't alter the conclusion. The full text version is available to read at this site.

http://www.cmaj.ca/content/186/13/985.s ... 186/13/985

Thank you for this article Arlene. I've read it carefully and here is my take:

- this is a very large (>10,000 patient), statistically well powered study looking for a possible relationship between OSA and cancer-of-any-kind. This is important because it seems that earlier studies have involved fewer patients and are therefore less statistically sound.
- the paper is well written. That helps a lot.
- follow-up was long (8 years) and complete - also important if you're interested in a delayed onset of cancer resulting from OSA that was untreated (until time of diagnosis of course) or indeed in cancer induced by OSA that was inadequately treated.
- the authors clearly consulted a statistician or perhaps one of them is a statistician. They used all the right analyses imho (log rank test, cumulative incidence test, linear regression, multivariate ....) and conducted a variety of checks and follow-up analyses to answer certain interesting questions. I've published similar work myself - working with statistician co-authors, I'm not a statistician - and the authors in this study get full marks for methodology.
- this study concludes that there was NO EVIDENCE for a link between cancer prevalence at diagnosis of OSA and severity of OSA, or for a link between cancer incidence after onset of treatment for OSA and severity of OSA. Table 2 shows that on multivariate analysis of the data Odds Ratios (cancer prevalence) and Hazard Ratios (cancer incidence) were all close to 1.0 and that the 95% confidence interval ranges all included 1.0. Nothing doing.
- as I've found in my own research on cancer, initial analysis of the data just looking at single variables ('univariate analysis') did show some statistically significant - and rather large - associations but when all the variables known to be associated with cancer (age, BMI, etc) were corrected for by multivariate analysis there was no significance anywhere except for a relationship between oxygen desaturation (oximetry data) and smoking related cancers.

The only deficiency in this study that I can see - one that the authors point out clearly themselves and one they could do nothing about - is that it was not possible to analyze individual cancer types for possible associations with OSA. All of the analyses were done on "cancer-of-any-type" in this Canadian OSA population, presumably because there weren't enough patients with cancer of any particular type to make analysis statistically meaningful. Only about 6% of the OSA patients in this study had had a diagnosis of cancer prior to OSA diagnosis so perhaps this isn't surprising; an even larger study is called for. This is extremely important because of course cancer is not just one disease, it is in fact dozens or more like hundreds of different diseases, each having distinct causes (etiology) and distinct molecular pathologies (pathophysiology). So it is possible that one or more cancer types DO have a significant association with OSA, and one might suppose that cancers that are promoted or caused to metastasize more rapidly by chronic but episodic hypoxia would be among those. I can think of a few candidates but that's pretty speculative.

If this excellent study had shown a strong and highly significant relationship between cancer prevalence or cancer incidence and OSA in this population (let's say hazard ratio > 2.0 for severe vs mild OSA, p < 0.001), the article could have been published in a journal such as Nature Medicine because of the potential significance to clinical practice of such a finding. I'm not knocking CMAJ, I'm sure it's a decent journal, but that's just the way things go in clinical science.

Hi Pneumophile,

Thanks for your analysis!

I hope this study will help allay the concern Moonslice expressed re OSA and the development of Cancer. (And her mom's worry re CPAP use!! )

Last year I went digging for info re OSA and cancer after reading badly written articles on the subject on the internet, and was very pleased to come across this study.

Pneumophile wrote:one might suppose that cancers that are promoted or caused to metastasize more rapidly by chronic but episodic hypoxia

I would think the damage to hormone secretion would be every bit as big or bigger a risk to cancer.

Pneumophile wrote:- this is a very large (>10,000 patient), statistically well powered study looking for a possible relationship between OSA and cancer-of-any-kind. This is important because it seems that earlier studies have involved fewer patients and are therefore less statistically sound.
- the paper is well written. That helps a lot.

How did they know the patients had OSA? Generally, this is not known until a sleep study is conducted. Then CPAP is prescribed. Could this be a population that had reasonable control of their sleep apnea with CPAP? Would seem to invalidate the study.

Arlene1963 wrote:I hope this study will help allay the concern Moonslice expressed re OSA and the development of Cancer. (And her mom's worry re CPAP use!! )

Doesn't the study support the mother's worries? Since OSA did not cause the brain tumor, CPAP must have.

I'm not sure if you're joking, Chevie!

Anyway, researchers compared results between those who used CPAP (treated) and those who didn't (untreated), and the same negative results were obtained.

So no, according to this study, CPAP doesn't cause cancer either!! Phew.

Did I miss something? The article said that there was a link between sleep apnea and cancer. NOT CPAP use and cancer, but sleep apnea.

Suffice it to say, an aging mind is a funny thing.
One thing goes in the ears, followed by the opposite idea out the mouth.
Bless their hearts.
The scary thing is when this happens to those we might consider rational.

Chevie wrote:

Pneumophile wrote:one might suppose that cancers that are promoted or caused to metastasize more rapidly by chronic but episodic hypoxia

I would think the damage to hormone secretion would be every bit as big or bigger a risk to cancer.

Pneumophile wrote:- this is a very large (>10,000 patient), statistically well powered study looking for a possible relationship between OSA and cancer-of-any-kind. This is important because it seems that earlier studies have involved fewer patients and are therefore less statistically sound.
- the paper is well written. That helps a lot.

How did they know the patients had OSA? Generally, this is not known until a sleep study is conducted. Then CPAP is prescribed. Could this be a population that had reasonable control of their sleep apnea with CPAP? Would seem to invalidate the study.

Agree about the effects on hormone levels (cortisol, adrenaline, etc). The sympathetic system is greatly stimulated. On your question: the full, cancer prevalence analysis was done on ~ 10,000 patients in the Canadian database who had been diagnosed with OSA after polysomnography. They looked at any-cancer prevalence in this study population (about 6%) and analysed (retrospectively, as it were) by severe OSA vs mild OSA (see paper for details). In the incidence analysis, those patients who had NOT been diagnosed with cancer at time of OSA diagnosis were followed for 8 years for a cancer diagnosis and analysed for a possible relationship with OSA.

The study design makes sense. I can't think how else they could pull out interesting results from that database.

SewTired wrote:Did I miss something?

Yes, you did.