Summary of findings 2. Bi‐PAP compared to fixed CPAP for sleep apnoea in adults.

Bi‐PAP compared to fixed CPAP for improving usage of continuous positive airway pressure machines in adults with obstructive sleep apnoea
Patient or population: adults with a diagnosis of sleep apnoea Setting: Europe and USA Intervention: Bi‐PAP Comparison: fixed CPAP
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with fixed CPAP	Risk with Bi‐PAP
Machine usage (hours/night) Follow‐up: 4 to 52 weeks	Average nightly machine usage 5.5 hours per night across CPAP	0.14 hours/night higher (0.17 lower to 0.45 higher)	‐	268 (4 RCTs)	⊕⊕⊝⊝ Low 1 2
Machine usage assessed by number of participants using machine for 4 or more hours per night ‐ not measured	‐	‐	‐	‐	‐
Symptoms assessed with ESS from 0 to 24 Follow‐up: 4 to 12 weeks	The mean symptoms ranged from 8 to 11 ESS units	0.49 ESS units lower (1.46 lower to 0.48 higher)	‐	226 (4 RCTs)	⊕⊕⊝⊝ Low 1 3	MCID of between 2 to 3 has been proposed by Patel 2018.
Withdrawals (parallel group trials/first arm cross‐over trials) Follow‐up: 4 to 52 weeks	Study population		OR 0.55 (0.26 to 1.17)	261 (3 RCTs)	⊕⊕⊝⊝ Low 3 4
	188 per 1000	113 per 1000 (57 to 213)
Quality of life assessed with FOSQ: scale from 5 to 20 Follow‐up: 8 weeks	Mean change from baseline 5.1 units	0.8 FOSQ units lower (6.08 lower to 4.48 units higher)	‐	151 (1 RCT)	⊕⊕⊝⊝ Low 3 5	MCID for FOSQ has not been confirmed; a change of 1 unit has been proposed as representing a possible meaningful change (Billings 2014).
AHI measured with number of events/hr Follow‐up: 4 to 8 weeks	The mean AHI was 6.6 events/hour	1.36 events/hour lower (6.92 lower to 9.63 higher)	‐	179 (2 RCTs)	⊕⊕⊝⊝ Low 6
Blood pressure ‐ not measured	‐	‐	‐	‐	‐
Adverse events (machine tolerability outcomes) Follow‐up: 4 to 52 weeks	Five studies provided information on tolerability outcomes but data could not be combined because they were measured and analysed inconsistently across the studies. One study (N = 62) reported 5 withdrawals in CPAP group due to mask discomfort or device intolerance. 20 participants across both arms complained of nasal dryness. 1 trial (N = 151) reported similar rates of dry mouth (4.2% and 7.5%) and mask intolerance (11% and 10%) in Bi‐PAP and CPAP groups, respectively. Two small studies (N = 46) reported non‐specific adverse events. One reported that telephone contact did not identify the need for further interventions, and the other that there were similar rates of non‐specific adverse events. One study (N = 28) used a global treatment comfort score on a 0 to 00 VAS. There was insufficient evidence to determine whether Bi‐PAP improved comfort scores (69 versus fixed CPAP 60, P = 0.16).		‐	239 (5 RCTs)	⊕⊝⊝⊝ Very low 1 7
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). AHI: Apnoea Hypopnoea Index; Bi‐PAP: bilevel positive airway pressure; CI: confidence interval; CPAP: continuous positive airway pressure; ESS: Epworth Sleepiness Scale; FOSQ: Functional Outcomes of Sleep Questionnaire; MCID: minimal clinically important difference; OR: odds ratio; RCT: randomised controlled trial; VAS: visual analogue scale
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

¹ Downgraded one level due to serious risk of bias. Open‐label design from one study and insufficient detail available to assess blinding and allocation process in other studies.

² Downgraded one level due to serious imprecision. Wide confidence interval including higher average usage with either device.

³ Downgraded one level due to serious imprecision. Wide confidence interval.

⁴ Downgraded one level due to serious risk of bias. Lack of detailed description of randomisation process and lack of blinding in two studies contributing data to the analysis.

⁵ Downgraded one level due to risk of bias. Single study at high risk of bias from lack of blinding.

⁶ Downgraded two levels due to very serious imprecision. Wide confidence interval and small sample size.

⁷ Downgraded due to very serious inconsistency. Variation in the measurement of tolerability across the studies meant that no data could be combined,