© 2004 American Thoracic Society
ConSleep Apnea Does Not Cause Cardiovascular DiseaseOxford Centre for Respiratory Medicine Oxford, United Kingdom Obstructive sleep apnea has been linked to cardiovascular disease since the first case series published by Guilleminault in 1976 reported the high prevalence of hypertension (1). Indeed, it was the Stanford group that provided the first long-term data suggesting that treating obstructive sleep apnea reduced cardiovascular deaths (2). Partinen and Guilleminault compared patients with obstructive sleep apnea who accepted advice to have a tracheotomy with those who rejected this surgical approach (and opted for conservative treatment). Over 10 years there was a significant 10% difference in mortality favoring tracheotomy. This was not a randomized trial, however, and patients who accept medical advice tend to do better than those who do not. For example, in a study comparing clofibrate with placebo, the 5-year mortality was 15% in good compliers, but 24.6% in poor compliers, regardless of the treatment they received (3).
Over the 1980s and 1990s many studies were published looking at the relationship between hypertension (occasionally stroke) and obstructive sleep apnea (4), usually finding correlations, but rarely controlling adequately for potential confounding variables. Upper-body obesity, smoking, alcohol, exercise levels, and caffeine consumption (patients with obstructive sleep apnea drink nearly three times more coffee) (5) could all be postulated to co-correlate with obstructive sleep apnea severity (6) and influence blood pressure. Using the "Framingham" predictors of cardiovascular risk on their own shows that patients with obstructive sleep apnea, on average, are rather unhealthy (10-year risk of coronary heart disease and stroke being The most quoted epidemiological study exploring the relationship between obstructive sleep apnea and hypertension comes from the Wisconsin cohort of Young and coworkers (8). The degree of obstructive sleep apnea on an initial sleep study predicted the development of new hypertension during the subsequent four years, with an odds ratio of between two and three, even after controlling for most confounders, such as age, sex, upper-body obesity, and alcohol and cigarette usage. The measured blood pressures, however, were actually lower four years later in the subjects with more severe obstructive sleep apnea, and the incident hypertension had been defined on the basis of new drug treatment for hypertension, not the actual blood pressures. It is conceivable that a positive result on the initial study influenced subsequent screening for hypertension and thus its more frequent diagnosis and treatment in the subgroup with the most severe sleep apnea. As with all such studies there is also the possibility of an unrecognized confounding variable. In the oft quoted Sleep Heart Health Study (9), the presence of obstructive sleep apnea did appear to be an independent risk factor for hypertension in 6,132 subjects, even at very low levels of sleep apnea activity (apnea–hypopnea index of 10 per hour). At moderate levels of obstructive sleep apnea or sleep hypoxia, the dose–response relationship was flat or even negative, with a further rise in hypertension likelihood only at the highest levels of apnea frequency (apnea–hypopnea index of greater than 70 per hour). Overall the odds ratio for having hypertension was only 1.37 (apnea–hypopnea index greater than 30 versus less than 1.5), with 95% confidence intervals of 1.03–1.83. In epidemiological circles this would be uncompelling evidence as it would only require a very small effect from an unrecognized confounder, such as caffeine intake, to render this study negative. Medical science is littered with very plausible hypotheses generated from cross-sectional studies which proved to be incorrect when subjected to randomized controlled trials. And there is no shortage of potential hypotheses as to why obstructive sleep apnea might lead to increased cardiovascular morbidity and mortality (10, 11), including excess catecholamines, oxidative stress, alterations in clotting, and even snoring vibrating the carotid arteries enough to loosen atheromatous plaques! Recent dramatic examples of cross-sectional studies that have deceived us have been the use of hormone replacement therapy (in women) and antioxidants to prevent cardiovascular disease. A meta-analysis of cross-sectional studies showed that higher vitamin E consumption lowered cardiovascular risk by over 30% (12). A meta-analysis of over 80,000 subjects in seven interventional studies, however, showed no benefit whatsoever (13). Hormone replacement therapy, which had always been associated with reductions in cardiovascular risk in women, produced no such reduction in a large interventional study (14, 15). More recently, a 1-year interventional trial of nonsteroidal anti-inflammatory drugs for Alzheimer's disease (16) failed to support their apparent benefit in cross-sectional studies (17). The taking of medications, or different dietary habits, presumably act as markers of other unrecognized protective factors (for example, a healthier lifestyle). The lifestyle of an average patient with obstructive sleep apnea is likely to carry many potential confounders for cardiovascular risk. Thus cross-sectional studies are hypothesis generators, not proof of cause and effect. To demonstrate causal links requires controlled interventional studies. So far, such studies have only assessed short-term changes in blood pressure (18–20): and the treatment of obstructive sleep apnea with nasal continuous positive airway pressure does indeed produce a small fall in 24-hour blood pressure, which is largest in patients with severe obstructive sleep apnea, and probably absent in mild to moderate obstructive sleep apnea (18, 19). It is not established yet whether this fall is sustained beyond a few weeks or will translate into reduced long-term cardiovascular morbidity and mortality. It would be fair to point out, however, that in every other situation in which blood pressure has been lowered, a fairly predictable decline in cardiovascular mortality has followed. But this might not be so for obstructive sleep apnea. Thus, it is a highly attractive and plausible hypothesis (with many potential and exciting mechanistic explanations available) that obstructive sleep apnea is a risk factor for adverse cardiovascular outcomes. The data, however, are not yet robust: we have been fooled before with similarly plausible hypotheses. Hence treating patients to reduce cardiovascular risk is not yet evidence based. This field desperately needs the kind of long-term interventional trials that have finally proven the place of drugs such as the statins and ACE inhibitors in cardiovascular risk reduction. FOOTNOTES Conflict of Interest Statement: J.S. has no declared conflict of interest. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
30%, rising to 36% in those with the most severe obstructive sleep apnea) (
