The efficacy of high-flow nasal cannula (HFNC) versus non-invasive ventilation (NIV) in patients at high risk of extubation failure: a systematic review and meta-analysis

BACKGROUND: Studies suggest that high-flow nasal cannula (HFNC) and non-invasive ventilation (NIV) can prevent reintubation in critically ill patients with a low risk of extubation failure. However, the safety and effectiveness in patients at high risk of extubation failure are still debated. Therefore, we conducted a systematic review and meta-analysis to compare the efficacies of HFNC and NIV in high-risk patients. METHODS: We searched eight databases (MEDLINE, Cochrane Library, EMBASE, CINAHL Complete, Web of Science, China National Knowledge Infrastructure, Wan-Fang Database, and Chinese Biological Medical Database) with reintubation as a primary outcome measure. The secondary outcomes included mortality, intensive care unit (ICU) length of stay (LOS), incidence of adverse events, and respiratory function indices. Statistical data analysis was performed using RevMan software. RESULTS: Thirteen randomized clinical trials (RCTs) with 1457 patients were included. The HFNC and NIV groups showed no differences in reintubation (RR 1.10, 95% CI 0.87–1.40, I(2) = 0%, P = 0.42), mortality (RR 1.09, 95% CI 0.82–1.46, I(2) = 0%, P = 0.54), and respiratory function indices (partial pressure of carbon dioxide [PaCO(2)]: MD − 1.31, 95% CI − 2.76–0.13, I(2) = 81%, P = 0.07; oxygenation index [P/F]: MD − 2.18, 95% CI − 8.49–4.13, I(2) = 57%, P = 0.50; respiratory rate [Rr]: MD − 0.50, 95% CI − 1.88–0.88, I(2) = 80%, P = 0.47). However, HFNC reduced adverse events (abdominal distension: RR 0.09, 95% CI 0.04–0.24, I(2) = 0%, P < 0.01; aspiration: RR 0.30, 95% CI 0.09–1.07, I(2) = 0%, P = 0.06; facial injury: RR 0.27, 95% CI 0.09–0.88, I(2) = 0%, P = 0.03; delirium: RR 0.30, 95%CI 0.07–1.39, I(2) = 0%, P = 0.12; pulmonary complications: RR 0.67, 95% CI 0.46–0.99, I(2) = 0%, P = 0.05; intolerance: RR 0.22, 95% CI 0.08–0.57, I(2) = 0%, P < 0.01) and may have shortened LOS (MD − 1.03, 95% CI − 1.86–− 0.20, I(2) = 93%, P = 0.02). Subgroup analysis by language, extubation method, NIV parameter settings, and HFNC flow rate revealed higher heterogeneity in LOS, PaCO(2), and Rr. CONCLUSIONS: In adult patients at a high risk of extubation failure, HFNC reduced the incidence of adverse events but did not affect reintubation and mortality. Consequently, whether or not HFNC can reduce LOS and improve respiratory function remains inconclusive. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s40001-023-01076-9.