Distinct weaning phenotypes identified in patients failing to wean from ventilation

For the millions of patients who require a breathing tube each year, the moment of liberation from the ventilator is one of the most consequential events in critical care. Most patients eventually breathe on their own again, but a stubborn minority does not, and for them the prognosis has long been treated as uniformly grim. Now, a large international analysis has revealed that “weaning failure” is not a single condition but a spectrum of three clinically distinct phenotypes, each with its own trajectory, treatment pattern and outlook — a finding that could reshape how intensivists counsel families, plan care and, most strikingly, how and when decisions are made to withdraw life-sustaining therapy.

The study, published in the journal Intensive Care Medicine, is a secondary analysis of the WEAN SAFE cohort — the WorldwidE AssessmeNt of Separation of pAtients From ventilatory assistancE — a multicentre, prospective observational study that captured mechanical ventilation practices across 50 countries. Led by Reginald Caldecott and John G. Laffey of the University of Galway, together with statistician Andrew J. Simpkin and collaborators spanning Europe, Canada and beyond, the team set out to answer a deceptively simple question: what actually happens to patients who enter the weaning process but never succeed in separating from invasive mechanical ventilation within 90 days?

The scale of the dataset lends the findings unusual weight. Of the 5,664 patients in the study population, 1,270 — 22.4 percent — never underwent a separation attempt at all, meaning clinicians never even initiated a deliberate effort to discontinue the ventilator. The remaining 4,394 patients, or 77.6 percent, had at least one such attempt. Within this latter group, 3,730 patients, representing 65.9 percent of the total, successfully weaned. But 664 patients — 15.1 percent of those who attempted separation — ultimately failed to wean from invasive mechanical ventilation by day 90. These so-called Failed-Wean patients experienced the longest duration of invasive ventilation and the longest intensive care unit stay of any group in the entire cohort, confirming that weaning failure marks a population that consumes a disproportionate share of critical care resources while facing devastating outcomes.

The mortality figures are stark. ICU mortality was 78 percent among Failed-Wean patients compared with just 2 percent among those who weaned successfully. Yet the new analysis shows that this headline number conceals important heterogeneity. Failed-Wean patients had undergone more separation attempts on average than Successful-Wean patients, but paradoxically fewer attempted extubations — the actual removal of the breathing tube — and more reintubations and tracheostomies. In other words, many of these patients were repeatedly tried on the ventilator, tested for readiness to breathe independently, and then returned to full support without the tube ever being removed, a pattern that suggests clinicians sensed poor prognosis but could not articulate it within existing classification frameworks.

Using patterns of separation attempts and decisions around withdrawal or withholding of life-sustaining therapy, the researchers identified three clinically distinct Failed-Wean phenotypes. Phenotype A comprised patients who had a single separation attempt followed by a WLST decision — the withdrawal or withholding of life-sustaining therapy. Phenotype B included patients with a single separation attempt but no WLST decision. Phenotype C was defined by more than one separation attempt. Each phenotype carried a different clinical profile, weaning trajectory and outcome, implying that a one-size-fits-all label of “weaning failure” obscures three fundamentally different patient journeys through the intensive care unit.

Perhaps the most provocative finding concerns the very first attempt. The analysis found that a failed-first separation attempt was strongly associated with subsequent WLST decisions. In this cohort, a WLST decision was, in turn, associated with a higher hazard of ICU mortality and a shorter median time to ICU death. The causal direction of these relationships is impossible to establish in an observational dataset — clinicians may withdraw therapy because a patient is clearly deteriorating, and the withdrawal itself obviously ends the struggle to breathe independently — but the strength and consistency of the association raise uncomfortable questions about whether early failure on a first attempt is being read, consciously or not, as a prognostic verdict that it may not deserve.

The weaning process itself is technically intricate. Patients are typically assessed for readiness using criteria such as the fraction of inspired oxygen, positive end-expiratory pressure, and the ratio of arterial oxygen partial pressure to inspired oxygen — the P/F ratio — before undergoing a spontaneous breathing trial, in which ventilatory support is reduced or removed for a defined period. A separation attempt, as defined in the WEAN SAFE framework, encompasses any deliberate attempt to discontinue invasive mechanical ventilation, whether or not it culminates in extubation. Extubation itself carries risk: reintubation, the re-establishment of invasive ventilation after a planned extubation, is associated with increased mortality, prolonged stays and ventilator-associated complications. Supplementary analyses from the new study show that among the 489 reintubated patients, first reintubations clustered within the first few days after extubation, with distinct timing distributions between those who ultimately failed weaning and those who did not.

The classification of weaning outcomes has evolved considerably over the past two decades. Earlier frameworks, such as the consensus guidelines published in the European Respiratory Journal in 2007, distinguished simple, difficult and prolonged weaning, and subsequent studies validated categorisations such as the WIND classification — Weaning according to a New Definition — which separates patients who never wean, those who die before weaning, and those who fail. The new phenotyping approach builds on this lineage but adds granularity by incorporating the number of separation attempts and the timing of end-of-life decisions, dimensions that prior schemes ignored. Previous work from the same cohort has already shown how frailty, older age, comorbidities and even patient sex shape weaning management and outcomes, and the present analysis extends that programme to the most severely affected subgroup.

The clinical implications could be substantial. If Phenotype A patients — those who fail a single separation attempt and then have therapy withdrawn — represent a large fraction of weaning deaths, then the question arises whether some of these patients might have weaned had more attempts been made, or whether the first failure is genuinely a signal of irreversible decline. Conversely, Phenotype C patients, who undergo repeated attempts, may represent prolonged ventilator dependence whose outcomes could be improved by earlier tracheostomy, transfer to specialised weaning units, or more aggressive treatment of the underlying drivers of ventilator dependence, such as respiratory muscle weakness, cardiac dysfunction or persisting inflammation. Distinguishing these phenotypes at the bedside could allow clinicians to tailor both the intensity of weaning efforts and the candour of prognostic conversations with families.

The authors are careful about interpretation. Because WLST decisions were associated with mortality in a dataset where causality cannot be untangled, they call for further investigation of the relationship between failed first separation attempts, WLST decisions and the risk of death. The analysis was funded jointly by the European Society of Intensive Care Medicine and the WEAN SAFE Clinical Research Collaboration, with additional endorsement from the European Respiratory Society, and the trial was registered as NCT03255109. The investigators declare no conflicts of interest, and the underlying WEAN SAFE dataset remains available from the corresponding author upon reasonable request, subject to approval from the study steering committee.

For a field in which roughly one in six patients who attempt to leave the ventilator never manage it, the message from this analysis is that those patients are not interchangeable. Some die quickly after a single aborted attempt and a decision to redirect care; others linger through repeated failures; still others persist without any end-of-life decision at all. Recognising these patterns, the researchers argue, is the first step toward interventions that could shift trajectories — and toward ensuring that the pivotal first separation attempt is used not as a verdict but as information. With prolonged mechanical ventilation already known to carry poor long-term survival and to strain intensive care capacity worldwide, understanding who fails, how, and why has never mattered more.