THE ASSOCIATION BETWEEN THE HIGH-FLOW NASAL CANNULA TECHNIQUE DURING BRONCHOSCOPY AND THE MAINTENANCE OF PERIPHERAL OXYGEN SATURATION (SPO₂): A SYSTEMATIC REVIEW OF RANDOMIZED CONTROLLED TRIALS AND PRIMARY STUDIES

Abstract

Introduction: Peri-procedural hypoxaemia is the most frequent physiological complication of flexible bronchoscopy and endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA), particularly under procedural sedation, bronchoalveolar lavage (BAL) and in patients with limited respiratory reserve. High-flow nasal cannula (HFNC) delivers heated, humidified gas at up to 60–80 L/min with a stable fraction of inspired oxygen (FiO₂), nasopharyngeal dead-space washout and a flow-dependent positive airway pressure. We aimed to quantify the association between HFNC use during bronchoscopy and the maintenance of peripheral oxygen saturation (SpO₂), with emphasis on evidence published in the last 15 years.^1–3

Methods: The study strictly adhered to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) 2020 guidelines. Randomised controlled trials (RCTs), cohort, case–control, cross-sectional and other prospective or retrospective observational studies comparing HFNC with conventional oxygen therapy (COT), non-invasive ventilation (NIV) or different HFNC flows during bronchoscopy were eligible.^4–6

Results: Thirty-two primary studies (23 RCTs and 9 observational/non-randomised studies; 4,572 participants; 2012–2026) were included. In 12 adult RCTs (2,318 participants), HFNC reduced intraprocedural desaturation (SpO₂ < 90%) from 39.9% to 15.0% (risk ratio [RR] 0.34, 95% CI 0.25–0.47; I² = 72%; prediction interval 0.12–0.95), corresponding to 262 fewer events per 1,000 patients and a number needed to treat of 3.8. Including the PROSA-2 trial and two observational studies (odds ratio 0.20, 95% CI 0.14–0.28) did not change the inference. HFNC increased nadir SpO₂ by 4.95 percentage points (95% CI 2.89–7.00; five RCTs), reduced procedure interruption (RR 0.39, 0.28–0.54; I² = 13%), airway manoeuvres or support escalation (RR 0.40, 0.24–0.67) and severe hypoxaemia/advanced rescue (RR 0.31, 0.16–0.63). A monotonic flow–response relationship was observed, with a reported ED₉₅ of approximately 43 L/min. In moderate-to-severe hypoxaemic respiratory failure, NIV provided better oxygenation than HFNC. The effect was robust across all sensitivity analyses; certainty was moderate for desaturation and procedure interruption.^7–13

Discussion: The benefit of HFNC was consistent in direction across sedation depth, procedure type (EBUS-TBNA vs. non-EBUS bronchoscopy), baseline risk and comparator device, although heterogeneity in magnitude was substantial and was driven chiefly by trials with matched nominal FiO₂ or very high control-arm risk. Neutral trials under conscious sedation and matched FiO₂ suggest that part of the benefit derives from superior and more stable oxygen delivery, while physiological studies support additional contributions from preserved end-expiratory lung volume and dead-space washout.^14–16

Conclusion: HFNC during bronchoscopy is associated with clinically meaningful preservation of SpO₂, fewer hypoxaemic events and fewer procedure interruptions, with moderate certainty. HFNC at ≥ 45–60 L/min should be preferred for sedated, prolonged or BAL-containing procedures and for patients at increased risk, whereas NIV/CPAP remains preferable in established moderate-to-severe hypoxaemic respiratory failure. Adequately powered, blinded-outcome multicentre trials with standardised desaturation definitions and patient-important endpoints are required.^2,17