Abstract Rationale In patients receiving invasive mechanical ventilation, determining the optimal time to transition from controlled to assisted mode is key for weaning. Inappropriate transition may be associated with intensive care unit-acquired weakness, ventilator-induced diaphragmatic dysfunction and patient self-inflicted lung injury. A recent study found that two-thirds of attempts to transition to assisted ventilation fail. However, published models relying on time-based criteria to define failure are constrained by important technical limitations. We performed an analysis using physiological criteria to discriminate between transition success and failure. Methods We retrospectively analysed granular intensive care (ICU) data from three London centres (2014-2025), including demographics, ventilatory, blood gas, laboratory, vital signs, vasopressor use, and fluid balance. Derived variables including PaO2/FiO2 (PF) ratio, driving pressure, rapid shallow breathing index and dynamic compliance, as well as Sequential Organ Failure Assessment scores, were calculated. Patients with hypoxaemic respiratory failure (one PF ratio 300mmHg, within the first 24 hours post-intubation) ventilated for at least 24 hours, were included. Ventilation modes were dichotomised into controlled or assisted. Transition failure was determined using physiological criteria (measured three hours after transition) outlined in European Respiratory Society guidelines. Comparisons between groups were performed using the Mann-Whitney U (continuous variables) and Chi-square tests (categorical variables). Transition outcome was predicted using LASSO regression. Results The cohort’s median age was 61 years and 66% were male. 1634/ 4199 (39%) patients failed their first transition from controlled to assisted mode. The most common criterion was blood pressure outside range (46%). The transition failure group had a longer time to first transition (78 39-140 vs 60 27-114 hours); ICU length of stay (median 16 IQR 9-27 vs 13 7-24 days), duration of mechanical ventilation (11 6-21 vs 9 4-18 days); and ICU mortality (22.1% vs 16.3%), all p 0.01. Before the first transition, patients in the failure group had a lower PF ratio (250mmHg 178-327 vs 308mmHg 244-381) across all PEEP strata; and more cardiovascular stress, indicated by higher heart rate, systolic and mean blood pressure (all p 0.01). Logistic regression showed transition success was significantly associated with PEEP settings (Figure 1, all p 0.01). LASSO regression modelling predicting transition failure showed moderate performance (AUC = 0.72). Conclusions Predicting transition outcome may optimise patient weaning by identifying potential candidates early. Around two-fifths of patients fail their first transition to assisted ventilation, reflecting worse pre-switch physiology and poorer outcomes. These findings need external validation and refinement of physiological criteria. This abstract is funded by: Asthma and Lung UK/ HCA Healthcare
Murali et al. (Fri,) studied this question.