Home and outpatient versus in-hospital initiation of non-invasive ventilation in people with neuromuscular and restrictive thoracic disorders: a network meta-analysis

Introduction There is a growing trend towards outpatient or home initiation of long-term non-invasive ventilation (NIV) instead of in-hospital initiation. However, evidence supporting the non-inferiority of this strategy is limited in people with neuromuscular diseases (NMDs) or restrictive thoracic disorders (RTDs). Methods Systematic review and network meta-analysis (NMA) to compare the effect of three NIV initiation settings (home, outpatient and in-hospital) on long-term NIV efficacy in people with NMD or RTD. We used MEDLINE, Web of Science and the CENTRAL (Cochrane Central Register of Controlled Trials) to identify randomised controlled trials (RCTs), controlled and uncontrolled prospective studies comparing NIV efficacy between home and outpatient, and in-hospital initiation. Studies published between January 2000 and February 2023 with endpoints ≥1 month were included. The main outcomes were diurnal arterial carbon dioxide pressure (PaCO2) (primary), oxygen partial blood pressure (PaO2), bicarbonate (HCO3−), quality-of-life and NIV adherence. Results Eight studies (five RCTs), comprising 350 individuals with PaCO2 measurements (144, 119 and 87 individuals for in-hospital, home and outpatient NIV initiation, respectively) were included. Home and outpatient initiation were not inferior to in-hospital initiation in terms of mean PaCO2 change (Δ) based on the adjusted Bayesian NMA model. No significant difference was found for any secondary outcomes: PaO2, HCO3−, adherence and quality of life. Conclusion Despite the limited number of studies included in this NMA, we found no evidence suggesting that home or outpatient NIV initiation is less effective than in-hospital initiation for individuals with NMD or RTD who are deemed suitable for home or outpatient care. Home NIV initiation may therefore be a particularly valuable strategy, especially when hospital resources are limited or when patient autonomy is compromised. PROSPERO registration number CRD42023403339. Data are available in a public, open access repository. Data are available upon reasonable request. We deposed all data set along with an R Markdown HTML sheet allowing to reproduce all statistical analyses for the main outcome in a dedicated repository in Open Science Framework: .