1553 - Factors Impacting Mortality and Withdrawal of Life Sustaining Therapy in Severe Traumatic Brain Injury

Objectives: The objective was to evaluate the correlation between withdrawal of life-sustaining therapy (WLST) and established prognostic predictors in TBI, as well as identify additional factors that may influence the timing of WLST.

Design: This case-control study analyzed adults (aged 18-100) admitted between 8/1/2018 and 7/31/2021 with severe TBI. The case group consisted of 164 individuals who underwent WLST after TBI, while the control group included 164 individuals with TBI who did not undergo WLST. Case and control patients were matched 1:1 based on injury severity score, mechanism of injury, age, and sex. Prognostic indicators such as bilateral grade 3 diffuse axonal injury, absent pupillary responses, absent somatosensory evoked potentials, midline shift size, and duration of coma ( >28 days) were recorded. Other factors assessed included cerebral hemorrhage, herniation, Glasgow Coma Scale (GCS) score on admission, length of hospital stay, anticoagulation, craniectomy, Do-Not-Resuscitate code status, and palliative care consultation. Functional recovery of control subjects was evaluated using Glasgow Outcome Scale Extended (GOSE) scores at discharge and during follow-up 10-14 months later.

Results: The study involved 328 individuals (66±19.9 years). Significant findings revealed lower GCS scores and shorter hospital stays in the case group compared to controls. Bilaterally absent pupillary responses were more common in cases (54.8%) and associated with shorter time to WLST and poorer functional outcomes in controls. Brain herniation was more prevalent in cases (25%) and midline shift predicted WLST only in geriatric population. Palliative care consultations, craniectomy procedures, and DNR code status were associated with time to WLST and functional outcomes. In the control group, 58% overall and 97% in non-geriatric individuals had good outcomes (GOSE 4-8) at follow-up.

Conclusions: This study highlights the challenges of making WLST decisions, emphasizing the importance of standardized guidelines, objective indicators, and avoiding nihilism in WLST discussions as many control subjects experienced a good outcome.

Scientific Paper Best Paper Award Abstract:

Objectives: Traumatic brain injury (TBI) poses a significant public health concern, often placing patient families and clinicians in challenging situations when considering the withdrawal of life-sustaining therapy (WLST). There is a lack of established guidelines to guide decision-making regarding WLST following severe TBI. Furthermore, determining the appropriate timing for WLST is a complex and high-stakes decision. The objective was to evaluate the correlation between WLST and established prognostic predictors in TBI, as well as identify additional factors that may influence the timing of WLST.

Design: This case-control study included adults aged 18-100 admitted between 8/1/2018 and 7/31/2021 with a diagnosis of severe TBI. The case group comprised all 164 individuals who experienced severe TBI and underwent WLST, while the control group consisted of 164 individuals with severe TBI who did not undergo WLST. The case and control subjects were matched in a 1:1 ratio based on injury severity score (within five points), mechanism of injury, age (within five years), and sex. Age greater than or equal to 65 years was used to define the geriatric population.

Patient details were collected from the trauma registry and, when necessary, from the medical records. These measures included established predictors of prognosis such as bilateral grade 3 diffuse axonal injury (DAI), bilaterally absent pupillary responses, bilaterally absent N20 somatosensory evoked potentials (SSEPs), midline shift size, and duration of coma ( >28 days). Other patient characteristics were evaluated, including the presence of a subarachnoid hemorrhage, subdural or epidural hematoma, cerebral herniation, total Glasgow Coma Scale (GCS) score on admission, length of hospital stay, anticoagulation therapy, craniectomy, Do-Not-Resuscitate (DNR) code status, and whether the patient received a palliative care consultation. To assess functional recovery in control subjects following TBI, Glasgow Outcome Scale Extended (GOSE) scores at the time of discharge and during a follow-up evaluation conducted 10-14 months later.

Predictors of prognosis were analyzed using logistic regression and Fisher's Exact tests for continuous and categorical variables, respectively, in the case and control populations. Spearman's correlations and Wilcoxon Rank-Sum tests were employed to examine the relationship between continuous prognostic indicators and time to WLST in case subjects, as well as between continuous indicators and GOSE scores at discharge and follow-up in control subjects. Categorical predictors were also assessed using Wilcoxon Rank-Sum tests.

Results: The study included a total of 328 individuals, with a mean age of 66±19.9 years, comprising 128 adults and 200 geriatric individuals. Logistic regression analyses revealed several significant findings. The case population was associated with a lower GCS score on admission and had a shorter length of stay compared to control subjects. Furthermore, the case population had a higher cumulative number of negative prognostic indicators in each patient.

Regarding intracranial pathologies, there was no significant difference in the proportion of individuals with bilateral grade 3 diffuse axonal injury (DAI), subarachnoid hemorrhage, subdural hematoma, or epidural hematoma between the case and control groups. However, the case population had a significantly higher proportion of individuals with bilaterally absent pupillary responses (54.8%) compared to control subjects (3%). Bilaterally absent pupillary responses were significantly associated with a shorter time to WLST among case subjects and lower GOSE scores at follow-up among control subjects, emphasizing the importance of pupillary responses as prognostic indicators for both survival and functional outcomes. The proportion of individuals with brain herniation was significantly different between case patients (25%) and control patients (2.4%). None of the geriatric individuals in the study who had brain herniation or bilaterally absent pupillary responses were discharged without withdrawal of life-sustaining measures. Midline shift, when compared by amount of shift or whether >5mm shift was present, was associated with the WLSM in the geriatric population only. In the control group, 13 participants had a midline shift of >5 mm and survived, with an average age of 46±22.2 years and GOSE scores at 10-14 months ranging between 2 and 8. This suggests that midline shift may not be a reliable indicator of long-term outcome.

In terms of palliative care consultations, consultation was a significant indicator of WLST in the adult population but not in the geriatric population. This difference may be attributed to biases in clinician perception, with palliative care consultations being more commonly provided to adult subjects compared to geriatric subjects. Notably, indicators such as palliative care consultation, craniectomy procedure, and Do-Not-Resuscitate (DNR) code status were all significantly associated with time to withdrawal of life-sustaining measures among case subjects and with GOSE scores among control subjects.

Analyzing the control group, when considering a poor outcome at follow-up, defined as a GOSE score of 1-3 and a good outcome as a GOSE score of 4-8, 58% of the participants had a good outcome overall. However, within the non-geriatric population, 97% achieved a good outcome. The oldest participant with a GOSE score of 7 was 62 years old, and with a score of 8, was 45 years old. In contrast, only 34% had a good outcome among the geriatric population, with no geriatric participants reaching a GOSE score of 7-8. These findings align with previous studies indicating that a good functional recovery is unlikely in the geriatric population following severe TBI.

Conclusions: This study addresses the challenges faced by clinicians and families when deciding to WLST. The absence of established guidelines and the complexity of prognostication make these decisions difficult. Objective measures like pupillary responses and brain herniation were associated with WLST. Palliative care consultations influenced WLST decisions in the non-geriatric population, potentially due to biases. The study also found that good functional recovery after severe TBI is less likely in the geriatric population compared to younger individuals. It is important to consider these factors when discussing prognosis with families and they should be informed that a good functional recovery on the GOSE is unlikely in the geriatric population, but moderate disability may still be possible, which could be considered a favorable outcome by families. These findings highlight the importance of standardized guidelines, objective indicators, and avoiding nihilism in WLST discussions as many control subjects experienced a good outcome.