Chest wall oscillation may cut bronchoscopy in respiratory failure, but the primary result was borderline

Design

Single-centre randomised controlled trial in a respiratory intensive care unit

Population

88 patients with acute respiratory failure on noninvasive support, retained secretions and weak cough

Primary outcome

Therapeutic bronchoscopy

Effect

13.6% vs 31.8% (unadjusted p = 0.042; adjusted OR 0.336, 95% CI 0.111 to 1.021)

This single-centre randomised trial in a respiratory intensive care unit included 88 patients with acute respiratory failure needing noninvasive support (high-flow oxygen or noninvasive ventilation), retained bronchial secretions and an ineffective cough (peak cough flow under 270 L/min). They were assigned to high-frequency chest wall oscillation plus standard care or standard care alone. The primary end point was therapeutic bronchoscopy.

Bronchoscopy was needed in 6 patients (13.6%) with oscillation and 14 (31.8%) without (unadjusted p = 0.042), but after adjustment the difference was not significant (OR 0.336, 95% CI 0.111 to 1.021; p = 0.055). In exploratory analyses without correction for multiple testing, hospital stay was shorter (12 vs 16.5 days; adjusted reduction 4.5 days), as was intensive care stay (8 vs 12 days; 3.6 days) and high-flow oxygen duration. Nobody was intubated and mortality did not differ.

The authors call the trial exploratory. It is a reason for a larger trial, not for buying the device.

• Bronchoscopy was needed in 13.6% with chest wall oscillation vs 31.8% without (unadjusted p = 0.042).
• After adjustment the result was not significant (OR 0.336, 95% CI 0.111 to 1.021).
• Hospital stay was 12 vs 16.5 days, an exploratory finding without correction for multiple testing.
• No patient needed intubation and mortality was unchanged.
• Airway clearance techniques remain a reasonable adjunct, but this trial is too small to change practice.

Why it matters

Retained secretions drive bronchoscopy and prolonged support, and there has been little trial evidence for non-invasive airway clearance in this setting.

Don't overread it

A single-centre, 88-patient trial with a non-significant adjusted primary result is hypothesis-generating only.

The statistics, in plain English

The crude result crossed the p = 0.05 line and the adjusted one just missed it. The confidence interval for the odds ratio (0.111 to 1.021) includes 1.0, so a benefit is likely but not shown. Secondary outcomes tested without correction for multiple comparisons produce chance findings more easily.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.