Automated oxygen titration kept saturation on target about a fifth more of the time

Design

Systematic review and meta-analysis of 7 randomised trials

Population

635 hospitalised adults on non-invasive oxygen therapy

Primary outcome

Proportion of time within target SpO₂ range

Effect

Mean difference +21.2 percentage points (95% CI 18.5–23.8) favouring closed-loop control

This meta-analysis pooled seven randomised trials with 635 hospitalised adults receiving non-invasive oxygen, comparing closed-loop systems that adjust oxygen flow automatically from continuous oximetry with manual adjustment by staff. Ventilated patients were excluded.

Patients on closed-loop control spent 81–95% of the time within their target saturation range, against 51–84% with manual control: a mean difference of 21.2 percentage points (95% CI 18.5–23.8). The gain was larger with low-flow oxygen (31 points) than with high-flow nasal oxygen (13 points).

Both too little and too much oxygen cause harm, and staff cannot watch every saturation continuously. The outcome here is time in range, a process measure; the trials did not show effects on mortality or length of stay. For busy emergency and ward areas, the case is that automation does what nursing staffing often cannot. Availability and cost in Indian hospitals are likely to limit uptake for now.

• Prescribe a target saturation range for every patient on oxygen, not just a flow rate.
• Use 88–92% for patients at risk of hypercapnia and 92–96% for most others, as in current practice.
• Consider closed-loop oxygen devices where available for patients on prolonged low-flow oxygen.
• Check that oxygen is weaned as saturation rises; over-oxygenation is the commoner manual error.

Why it matters

It shows how often manually titrated patients drift outside their target, which automation largely corrects.

Don't overread it

Time in target range is a process measure; the trials did not show better clinical outcomes.

The statistics, in plain English

I² of 88% means the trials differed a lot, but the difference was explained by oxygen type: low-flow studies showed about 31 points of improvement and high-flow studies about 13. All intervals were well clear of zero.

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.