In search of a better CPAP interface: A network meta-analysis comparing nasal masks, nasal pillows and oronasal masks

Until now, no study has directly network meta-analysed the impact of nasal masks, nasal pillows and oronasal masks on continuous positive airway pressure therapy in patients with obstructive sleep apnea. This study aimed to meta-analyse the impact of three kinds of nasal interfaces with both network meta-analysis and pairwise comparison. PubMed, EMBASE, CENTRAL and ClinicalTrials.gov were systematically searched from inception to December 2020 for studies that compared the three types of nasal interfaces for treating obstructive sleep apnea with continuous positive airway pressure. The outcomes were residual apnea-hypopnea index, continuous positive airway pressure, and nightly average usage. The network meta-analysis was conducted using multivariate random-effects in a frequentist framework where three interfaces were ranked with the surface under the cumulative ranking probabilities. The pairwise comparison was conducted using random-effects meta-analysis. Twenty-nine articles comprising 6378 participants were included. The pairwise comparison showed both nasal masks and nasal pillows were associated with lower residual apnea-hypopnea index, lower continuous positive airway pressure, and higher continuous positive airway pressure adherence compared with oronasal masks. The surface under the cumulative ranking confirmed that nasal masks were associated with the lowest residual apnea-hypopnea index and highest adherence, while pillows were associated with the lowest continuous positive airway pressure. The meta-regression identified that lower pretreatment apnea-hypopnea index and continuous positive airway pressure determined during continuous positive airway pressure titration (versus determined during continuous positive airway pressure therapy) was associated with lower continuous positive airway pressure with nasal masks and nasal pillows. In conclusion, compared with oronasal masks, nasal masks and nasal pillows are better interfaces, especially in patients with lower pretreatment apnea-hypopnea index and those with the therapeutic pressure determined during continuous positive airway pressure titration.