Continuous Positive Airways Pressure (CPAP) is considered to be the cornerstone of therapy for obstructive sleep apnoea (OSA). However, compliance with this treatment is frequently poor, which may lead to ongoing symptoms of sleep disruption, daytime sleepiness and poor waking cognitive function. Mechanical interventions which involve changing the way that positive pressure is delivered, and the addition of humidification, might improve compliance. To determine the efficacy of pressure level modifications and additional humidification in increasing CPAP machine usage. We searched the Cochrane Airways Group Specialised Register (September 2008). Randomised controlled trials (RCTs) assessing interventions to improve compliance with CPAP usage. Control groups received fixed pressure CPAP. Two authors assessed articles for inclusion in the review and extracted data. We made attempts to obtain additional unpublished data from the trialists. Forty-five studies met the inclusion criteria (1874 participants). Auto-CPAP (30 studies, 1136 participants): a statistically significant difference in machine usage of 0.21 hours/night (0.08 to 0.35) was observed in favour of auto-CPAP from cross-over studies. This difference is of questionable clinical significance. Pooled effect estimates from parallel group trials detected a similar sized difference for average nightly machine usage, but this was not statistically significant. Evidence from parallel group studies did not identify a statistically significant difference between pressure modes in Epworth Sleepiness Scores, but there was an overall reduction of 0.64 units with cross-over studies (-0.12 to -1.16) in favour of auto-CPAP. Parallel group studies did not identify a significant difference. More participants preferred auto-CPAP to fixed CPAP where this was measured. Bi-level PAP (six studies, 285 participants): no significant differences were observed in machine usage. One small study found no difference in preference. C-Flex (six studies, 318 participants): no significant difference was observed in machine usage. Humidification (three studies, 135 participants): there were conflicting findings between the studies. Two parallel group trials found no significant difference in machine usage, whereas a cross-over study found a significant difference. Improvement in average machine use of auto-CPAP was superior in studies with a cross-over design; the point estimate in parallel group trials was similar, but did not reach statistical significance. It is uncertain how use of machines in study settings relates to 'real world' use. Where preference was measured participants preferred auto-CPAP to fixed pressure CPAP. Further studies are required to assess the evidence for Bi-PAP, C-Flex(TM) and humidification. The studies assembled were characterised by high machine usage in the control groups, and low withdrawal rates. Future studies need to consider the effects of treatment in participants with more mild disease, and those who struggle to accept therapy despite persistent symptoms.