The decision to withdraw or withhold life-sustaining therapies (WLST) is a significant factor in many ICU deaths, influenced by patient factors such as age, comorbidities, illness severity, and patient or surrogate preferences. In patients with acute brain injuries (ABIs), WLST decisions may be particularly influenced by expectations of poor neurological outcomes. However, concerns exist that these decisions are sometimes made too early, before long-term neurological prognosis can be accurately determined.
The timing and frequency of WLST in ABI patients versus those without ABI is not well understood, and exploring these differences could improve the quality of WLST decisions, as some patients with perceived poor prognoses may experience long-term neurological recovery.
Findings from a study presented at ISICEM 2025 of traumatic brain injury (TBI) patients in Canada showed 64% had life-sustaining therapies withdrawn by day 3, with notable variability across hospitals. This study compared WLST decisions in patients with and without ABI and analysed the variability across ICUs.
A secondary analysis was conducted using data from two prospective, international studies that recruited patients ventilated either invasively or non-invasively between 2004 and 2016 across 40 countries. ABI was defined to include brain trauma, ischaemic stroke, intracranial haemorrhage, seizures, or meningitis-encephalitis. The comparator group consisted of patients with non-ABI conditions.
21,970 patients were recruited, with 16,791 included in the WLST analysis. Of these, 13,526 (61.6%) were male, 8,444 (38.4%) were female, and 2,896 (13.2%) had ABI. WLST occurred in 2,056 (12.2%) of 16,791 patients, with a higher incidence in those with ABI (17.0%) compared to those without ABI (11.5%), showing a risk difference of 5.5%. WLST decisions were made earlier for ABI patients (median 4 days) compared to non-ABI patients (median 6 days), with an absolute difference of 2 days. This trend was consistent across various ABI subgroups, world regions, and cohort years. There was high variability in WLST decisions across ICUs for both ABI and non-ABI patients, with median odds ratios indicating notable differences in decision-making.
This analysis found that patients with ABI had nearly 2.5 times higher odds of WLST compared to those without ABI, with substantial variability across ICUs. WLST decisions occurred earlier in ABI patients, with most followed by in-hospital death. The analysis showed that all major ABI subgroups had higher odds of WLST compared to non-ABI patients, while patients with conditions like sepsis and ARDS had similar or lower odds of WLST. These findings suggest that thresholds for WLST might differ for ABI patients.
WLST is common before death in patients with severe brain injuries, often occurring despite prognostic uncertainty. Early neuroprognostication may sometimes overestimate long-term disability, with some ABI patients possibly recovering to partial independence. The study highlighted the significant impact of ICU culture on WLST decisions, which appeared to be as influential as patient-specific factors. Variation in WLST was high across ICUs, and practices in neurological-neurosurgical ICUs may differ from general ICUs. While the study has strengths, including its large, diverse cohort, it also has limitations such as unmeasured confounding and lack of patient or surrogate preferences data.
In conclusion, ABI patients had higher odds of WLST, with decisions made earlier despite the difficulty of accurate long-term neuroprognostication. Further research is needed to understand the rationale behind early WLST decisions for ABI patients and to explore ways to improve decision-making in these cases.
Source: The Lancet Respiratory Medicine; ISICEM 2025 Presentation
Image Credit: ISICEM Congress 2025
