(1256) Noninvasive Ventilation Versus High-Flow Nasal Cannula in Preventing Reintubation: RCT Meta-Analysis

Introduction: The optimal post-extubation support strategy for high-risk patients remains debated. This meta-analysis evaluates the effectiveness of noninvasive ventilation (NIV) versus high-flow nasal cannula (HFNC) in preventing reintubation in high-risk patients and assesses 28-day mortality.

Methods: We conducted a systematic review and meta-analysis by searching MEDLINE, Scopus, and the Cochrane Database of Systematic Reviews to compare NIV and HFNC for reintubation prevention in high-risk patients. Given the homogeneity among randomized controlled trials, fixed-effect models were utilized to pool effect sizes, with a 95% confidence interval (CI) determined by the Jackson method. Sensitivity analysis used to assess the accuracy of the results.

Results: Our analysis included seven studies with 1226 patients for all-case intubation and 1202 for 28-day mortality. The patient cohort was 53% male, with a mean age of 65 years (SD = 16; 621 in the NIV group and 595 in the HFNC group). In the NIV group, 95 extubation failures (15.3%) occurred, compared to 126 reintubations (21.2%) in the HFNC group. The RCTs included had similar patient profiles, including comorbidities, BMI, and Apache II score, resulting in low population heterogeneity (I² = 30%). The between-study variability was small (τ² = 0.11 with p = 0.2). The reintubation rate was significantly lower in the NIV group compared to the HFNC group (pooled odds ratio [OR] = 0.65, 95% CI [0.49–0.88], z score = -2.78, p = 0.005) (Figure). Mortality was 12.7% in the NIV group and 13.6% in the HFNC group (pooled OR = 0.9652 [0.7269; 1.2815], z = -0.25, p = 0.8063). Median reintubation timing was not significantly different between NIV (46.12 hours) and HFNC (49.12 hours). Sensitivity analysis indicated robust results, with no significant impact from omitted studies. The symmetrical funnel plot suggested minimal publication bias.

Conclusions: This meta-analysis indicates a trend toward benefit with NIV (pooled OR = 0.65), and the result was statistically significant (z = -2.78, p = 0.005). No significant difference in 28-day mortality was found between NIV and HFNC. Further prospective studies are needed to clarify these findings.