Twelve trials (all RCTs; n=745) were included in the review; 5 were crossover trials.
Effect on subjective measures (11 trials).
The pooled estimate of effect indicated a statistically significant beneficial effect of CPAP on sleepiness, as assessed by the reduction in ESS scores (weighted mean difference 2.87, 95% confidence interval: 1.48, 4.25). However, this analysis was subject to significant statistical heterogeneity and the pooled result cannot be considered reliable. Subgroup analyses found that heterogeneity could not be explained by differences in age, gender ratio, body mass index, study location, the mean number of hours of CPAP per study, or the nature of the control. When only the moderate- or good-quality trials were pooled the result was not statistically significant, but the level of statistical heterogeneity was not reported.
When trials of participants with severe sleep apnoea and significant sleepiness (6 trials) were analysed separately from those less severely affected (5 trials), the pooled reduction in mean ESS score was 4.75 (95% CI: 2.97, 6.53; statistically significant). This was different from the results from trials of participants less severely affected (5 trials; 1.10, 95% CI: 0.13, 2.32; not significant). The authors did not state if these two subanalyses were subject to statistical heterogeneity or not.
Effect on objective measures (8 trials).
The pooled estimate of effect indicated a statistically significant beneficial effect of CPAP on sleepiness, as assessed by the MSLT or MWT scores (weighted mean difference 0.93, 95% CI: 0.10, 1.76). This analysis was subject to some statistical heterogeneity, although this was not significant. Subgroup analyses of studies evaluating MSLT or MWT studies or by study quality did not alter the result.
Funnel plots did not indicate the presence of publication bias.