Three phenotypes of weaning failure, and a first failed attempt that predicts the decision to stop

Design

secondary analysis of a prospective multinational observational cohort

Population

5,664 patients receiving invasive mechanical ventilation; 664 failed weaning by day 90

Primary outcome

ICU mortality, with weaning trajectory and withdrawal decisions characterised

Effect

ICU mortality 78% in failed-wean versus 2% in successful-wean; three phenotypes by separation attempt count and timing of withdrawal decisions

A secondary analysis of the WEAN SAFE cohort followed 5,664 ventilated patients. Of these, 1,270 (22.4%) never had a separation attempt at all; of the 4,394 who did, 3,730 (65.9%) weaned successfully and 664 (15.1%) had failed to wean by day 90. Failed-wean patients had the longest ventilation and the longest ICU stay, with more separation attempts but fewer attempted extubations, and more reintubations and tracheostomies. ICU mortality was 78%, against 2% in those who weaned.

Three phenotypes emerged, separated by how many separation attempts were made and when life-sustaining therapy was withdrawn or withheld: a single attempt followed by a withdrawal decision, a single attempt without one, and more than one attempt. A failed first separation attempt was strongly associated with a subsequent withdrawal decision, and that decision was associated with higher hazard of ICU death and a shorter time to it.

The authors are careful, and so should the reader be. That sequence can be read two ways: as clinicians correctly recognising futility early, or as a first failure setting an expectation that then shapes everything after it. The data cannot separate those, and the honest use of this paper is to notice how much weight a single failed attempt is carrying in your own unit's decisions.

• Notice how much a first failed separation attempt influences the subsequent plan on your unit
• Document the reason a separation attempt failed - load, weakness, delirium, secretions - rather than recording failure alone
• Reassess reversible contributors before the second attempt rather than after the third
• Distinguish never attempted from failed: 22.4% here never had an attempt at all
• Keep withdrawal discussions explicitly separate from the weaning trajectory, so one does not silently determine the other

Why it matters

It raises the possibility that a single failed separation attempt is doing more work in end-of-life decisions than anyone intends.

Don't overread it

These are observational phenotypes - they do not show that more attempts would have changed any outcome.

The statistics, in plain English

The 78% versus 2% mortality contrast is not a treatment effect - failing to wean is a marker of how ill the patient is, and the two groups are not comparable. The important limitation is on the association between a failed first attempt and withdrawal of therapy: in an observational cohort, treatment decisions and outcomes influence each other in both directions, so a self-fulfilling pattern is indistinguishable from accurate prognostication. Phenotypes derived from an existing dataset also need validating elsewhere before they mean anything at the bedside.

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.