Pardon me for my skepticism, but large statistical studies of this type tend to confuse correlation with causation and are somewhat flawed by patients with significant co-morbidities. One might suspect that the controls for making such a sweeping assessment of risk may have been inadequate to the task.
"Participants who then consented to continue into the study's main phase, were randomly allocated (like the toss of a coin) to either receive treatment with CPAP, or to continue with usual medical care. In this main phase, the effects of CPAP on cardiovascular disease will be evaluated." One might assume that continuing with the "usual medical care" involved interventional drugs that might have skewed the results.
This is the study where the conclusion should be prefaced with: "It generally appears that within this specific cohort, these findings seem statistically relevant, but may not be generalized to apply to individual patients or extended to a more global population.
From the New England Journal of Medicine ...
"Obstructive sleep apnea causes episodic hypoxemia and nocturnal sympathetic nervous system activation and elevates blood pressure and markers of oxidative stress, inflammation, and hypercoagulation. Large negative intrathoracic pressure swings also impose mechanical stress on the heart and great vessels. Population-based and sleep-clinic–based cohort studies have shown an association between obstructive sleep apnea and cardiovascular events, particularly stroke. Randomized, controlled trials have shown that treatment with continuous positive airway pressure (CPAP) lowers systolic blood pressure by 2 to 3 mm Hg in patients with normotensive obstructive sleep apnea18 and by 6 to 7 mm Hg in patients with resistant hypertension, improves endothelial function, and increases insulin sensitivity. Observational clinical studies have shown that the use of CPAP is associated with lower rates of cardiovascular complications and of death from cardiovascular causes, especially among patients who are adherent to treatment."
"Obstructive sleep apnea is a common condition among patients with cardiovascular disease, affecting 40 to 60% of such patients. Because the risks of recurrent cardiovascular events among these patients remain high despite contemporary therapies, CPAP could be a useful additional treatment for the prevention of these events. We describe the main results of the Sleep Apnea Cardiovascular Endpoints (SAVE) study, a secondary prevention trial that was designed to evaluate the effectiveness of CPAP in reducing the rate of cardiovascular events among patients with obstructive sleep apnea."
"A primary end-point event was confirmed in 436 participants — 229 (17.0%) in the CPAP group and 207 (15.4%) in the usual-care group (hazard ratio with CPAP, 1.10; 95% confidence interval [CI], 0.91 to 1.32; P=0.34)"
"Participants in the SAVE study who were assigned to CPAP adhered to the treatment for a mean of 3.3 hours per night over several years, which is similar to the mean adherence in other reports of CPAP use in patients who had no or minimal daytime sleepiness and which is consistent with CPAP use in clinical practice.
However, although this overall level of adherence to CPAP therapy exceeded the estimates in our power calculations, it may still have been insufficient to provide the level of effect on cardiovascular outcomes that had been hypothesized.
For practical reasons and to ensure efficient recruitment and consistency of data across multiple sites, we used a simple screening device (ApneaLink) that was based on oximetry and nasal pressure recordings and used automated algorithms to analyze signals, rather than the conventional standard test for obstructive sleep apnea in which polysomnographic data from an overnight stay in a hospital or clinic are scored manually."
Sorry to say, I find the difference between the two groups 229 vs. 207 to be statistically irrelevant to me although you can compute statistical values from any sample. Just reading through the methodology made me question the assertiveness of the statement that was released to the media. But, then again, nobody ever said the headline had to be supported by the body text.
I suggest you read the study for yourself and draw your own conclusions. Mine were "ho hum" let's get funding for the next study.
"Participants will then be followed up by hospital staff for 2-7 years until the end of 2015, and their health and the impact of the OSA treatment will be evaluated." Were these clinical evaluations against baselines, were they simply self-administered, or guided questionnaires that concentrated on events?"