Incidence of nosocomial pneumonia and clinical outcomes of patients requiring non-invasive ventilation: A systematic review and meta-analysis.

Background: Non-invasive ventilation (NIV) is used for an increasing range of clinical conditions at various levels of care, including intensive care units (ICUs), and has been shown to carry a lower risk of nosocomial pneumonia (NP) compared with invasive mechanical ventilation (IMV). Objectives: To assess the incidence of NP and clinical outcomes (intubation rates and mortality) in patients receiving NIV as initial support. Methods: A systematic search of PubMed, Embase and Scopus for relevant research articles published in English was conducted up to 6 February 2025. Eligible studies included adult patients who received NIV for any respiratory condition and reported the incidence of NP. NP was defined as any new-onset pneumonia occurring at any point during the clinical course and ≥48 hours after initiating NIV. Furthermore, a subset of patients in whom NP was attributed solely to NIV support was defined as NIV-associated pneumonia (NIVAP). Two reviewers independently conducted database searches, data extraction and risk-of-bias assessment. A subgroup analysis was performed based on the indication for NIV, country and study design to identify heterogeneity. Results: We incorporated 30 studies, including 36 049 patients receiving NIV. Of these, 29 studies reported the incidence of NP, while only 22 reported the incidence of NIVAP. Overall, the incidence of NP was 6% (95% confidence interval (CI) 4 - 8; I2=89.4%), and the pooled incidence of NIVAP was 3% (95% CI 2 - 4; I²=32.9%). The rate of intubation was 28% (95% CI 22 - 35; I²=89.3%), and overall mortality was 18% (95% CI 15 - 23; I²=99.0%) among patients receiving NIV. Conclusion: NP, including NIVAP, remains a significant complication in patients receiving NIV. Our findings underscore the need for standardised diagnostic criteria for NP in patients receiving NIV. Contribution of the study: This systematic review and meta-analysis provide robust pooled estimates of nosocomial pneumonia (NP) (6%; 95% CI: 4-8) and NIV-associated pneumonia (NIVAP) (3%; 95% CI: 2-4) among patients receiving non-invasive ventilation (NIV), addressing an important gap in the literature. It shows that the risk of pneumonia persists despite avoidance of intubation, challenging the prevailing assumption that NIV substantially mitigates infection risk. It also reveals significant variability in NP incidence across study designs, geographic regions, and patient subgroups-particularly a higher risk in patients with acute respiratory distress syndrome (ARDS)-highlighting the importance of context-specific risk assessment. Importantly, it underscores the lack of standardised definitions and diagnostic criteria for NP and NIVAP, which likely contributes to under-recognition and heterogeneity in reported outcomes. Collectively, these findings emphasise the need for improved surveillance, uniform diagnostic frameworks, and targeted preventive strategies for pneumonia in patients receiving NIV, extending beyond the traditional focus on invasive mechanical ventilation.