Who Should Get Treated for Sleep Apnea?

• Sleep apnea syndromes
• Sleepiness
• Continuous positive airway pressure
• Quality of life
• Cardiovascular diseases

This issue includes a report of a provocative randomized, placebo-controlled, parallel-group study by Barbé and colleagues (1), who found that nasal continuous positive airway pressure (CPAP) provided no benefit in patients with obstructive sleep apnea and a respiratory disturbance index greater than 30 episodes/h (currently regarded as severe sleep apnea [2]) but no self-reported sleepiness. This lack of benefit was demonstrated for self-reported and objectively measured sleepiness, as well as on generic and disease-specific quality-of-life instruments, neurocognitive tests, and 24-hour ambulatory blood pressure monitoring.

The sample size in this study, derived according to data collected in a previous placebo-controlled parallel-group study (3), was based on a mean difference of 9 in the change in the Medical Outcomes Study Short Form 36-item questionnaire score after treatment between study groups and assumed a standard deviation in change scores of 9.5 points. This is equivalent to a standardized effect size of 0.95, a value typically considered large. But patients in the previous study had self-reported sleepiness; the two patient groups are therefore not comparable, raising the possibility that Barbé and colleagues' study is underpowered to detect more moderate but still clinically important effect sizes with respect to sleepiness and other subjective and objective clinical outcomes. The investigators acknowledge this possibility but argue that the consistent pattern of small absolute differences between groups and the observed overlap between the 95% confidence intervals make it unlikely that biologically relevant differences would have been revealed with a larger sample.

Barbé and colleagues' findings raise at least three important issues. First, why are …