Abstract
This cohort study examines the association of time to command-following with death or dependency at 1 year among individuals with moderate-severe traumatic brain injury (TBI).
Introduction
After a traumatic brain injury (TBI), the persistent absence of command-following (ie, behavioral responses to verbal instructions) is often considered to be an indicator of poor prognosis.1,2,3 However, the accuracy of prognoses based on the absence of command-following at specific postinjury time points is unclear.
We measured the association between time to command-following and 1-year outcomes in more than 9000 participants enrolled in 2 large prospective TBI studies, the TBI Model Systems (TBIMS) National Database4 and the University of Pittsburgh Brain Trauma Research Center (BTRC) Database.5
Methods
Study Cohorts
This cohort study was approved by institutional review boards at all sites, and participants’ surrogates provided informed consent. This report follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies. TBIMS6 includes participants aged at least 16 years with moderate-severe TBI admitted to 1 of 16 US inpatient rehabilitation hospitals. The BTRC database5 includes participants aged 16 to 80 years with severe TBI (admission Glasgow Coma Scale [GCS] score ≤8 with motor GCS ≤5) admitted to 1 US level-I trauma center, excluding those with GCS of 3 and bilaterally fixed and dilated pupils. TBIMS collected self-reported race and ethnicity data, categorized as Asian or Pacific Islander, Black, Hispanic, White, or other (ie, people who did not self-report into any of these categories). BTRC collected self-reported race data as Asian or Pacific Islander, Black, White, or other and used a separate category for Hispanic or Latino or not Hispanic or Latino ethnicity. Race and ethnicity were included to describe the cohorts.
This retrospective analysis included participants who did not follow commands on the day of acute hospital admission and survived to discharge (eMethods in Supplement 1). The primary outcome was a 1-year Glasgow Outcome Scale Extended (GOSE) score less than 4, indicating death or dependency.
Trained study staff followed standard operating procedures to review electronic health records and identify the date of command-following (eMethods in Supplement 1). Among participants who followed commands during acute care, we fit a logistic regression model in each cohort to quantify the association between time to command-following and outcomes. A third model tested for an interaction between cohort and time to command-following. In each cohort, we used linear models to estimate the increase in the proportion of participants with 1-year death or dependency for each additional day without command-following. P values were 2-sided, and statistical significance was set at P < .05. Data were assessed using R version 4.3.1 (R Project for Statistical Computing) from January 2023 to August 2024.
Results
This analysis included 9052 participants (mean [SD] age, 38 [18] years; 6841 [76%] male) from the TBIMS cohort and 228 participants (mean [SD] age, 37 [16] years; 174 [76%] male) from the BTRC cohort. Cohorts differed by TBI severity, the proportion receiving inpatient rehabilitation, and 1-year GOSE (Table). Among participants who followed commands during acute care (Figure, A and B), each additional week without command-following was associated with increased odds of death or dependency at 1 year, and the finding was consistent across cohorts (TBIMS: odds ratio [OR], 1.3 [95% CI, 1.3-1.4]; P < .001; area under the receiver operating curve [AUC], 0.6; BTRC: OR, 1.5 [95% CI, 1.2-2.0]; P = .003; AUC, 0.6; combined model: cohort × command-following interaction, 0.9 [95% CI, 0.7-1.2]; P = .41).
Table. Cohort Characteristics.
| Characteristic | Participants, No. (%) | P value | |
|---|---|---|---|
| TBIMS (n = 9052) | BTRC (n = 228) | ||
| Age, mean (SD), y | 38 (18) | 37 (16) | .30 |
| Sex | |||
| Male | 6841 (76) | 174 (76) | .86 |
| Female | 2208 (24) | 54 (24) | |
| Missing | 3 (<1) | 0 | |
| Racea | |||
| Asian or Pacific islander | 224 (2) | 3 (1) | <.001 |
| Black | 1554 (17) | 16 (7) | |
| Hispanic | 1075 (12) | NA | |
| White | 6040 (67) | 207 (91) | |
| Other | 151 (2) | 2 (1) | |
| Missing | 8 (<1) | 0 | |
| Ethnicityb | |||
| Hispanic or Latino | NA | 2 (1) | <.001 |
| Not Hispanic or Latino | NA | 192 (84) | |
| Missing | NA | 34 (15) | |
| Marital status | |||
| Single | 4636 (51) | 130 (57) | <.001 |
| Married | 2820 (31) | 55 (24) | |
| Other | 1592 (18) | 10 (4) | |
| Missing | 4 (<1) | 33 (14) | |
| Injury mechanism | |||
| High-velocity | 5169 (57) | 166 (73) | <.001 |
| Fall-related | 1987 (22) | 47 (21) | |
| Low-velocity or other | 1895 (21) | 14 (6) | |
| Missing | 1 (<1) | 1 (<1) | |
| Injury year, median (IQR) | 2010 (2006-2015) | 2010 (2008 2013) | .85 |
| ED GCSTotal | |||
| Median (IQR)c | 10 (4-13) | 6 (5-7) | <.001 |
| Missing | 5920 (65) | 0 | |
| Craniectomyd | |||
| Yes | 1986 (22) | 106 (46) | <.001 |
| No | 6924 (76) | 122 (54) | |
| Missing | 142 (2) | 0 | |
| SDH or SAH | |||
| Present | 7217 (80) | 186 (82) | .01 |
| Absent | 1680 (18) | 25 (11) | |
| Missing | 155 (2) | 17 (7) | |
| EDH | |||
| Present | 1127 (12) | 38 (17) | <.001 |
| Absent | 7768 (86) | 85 (37) | |
| Missing | 157 (2) | 105 (46) | |
| IVH | |||
| Present | 2737(30) | 41 (18) | .87 |
| Absent | 6161 (68) | 88 (39) | |
| Missing | 154 (2) | 99 (43) | |
| Contusions | |||
| Present | 6249 (69) | 91 (40) | .31 |
| Absent | 2645 (29) | 47 (21) | |
| Missing | 158 (2) | 90 (39) | |
| Followed commands | 8141 (90) | 144 (63) | <.001 |
| Time to command-following, Median (IQR), d | 5 (2-12) | 10 (3-18) | <.001 |
| Acute care LOS, mean (SD), d | 25 (19) | 27 (15) | .01 |
| Attended acute rehabilitation | |||
| Yes | 9052 (100) | 159 (70) | <.001 |
| Missing | 0 | 0 | |
| GOSE <4 | 2112 (23) | 96 (42) | <.001 |
| Died | 275 (3) | 29 (13) | <.001 |
Abbreviations: BTRC, Brain Trauma Research Center; ED, emergency department; EDH, epidural hematoma; GCS, Glasgow Coma Scale; GOSE, Glasgow Outcome Scale Extended; IVH, intraventricular hemorrhage; LOS, length of stay; NA, not applicable; SAH, subarachnoid hemorrhage; SDH, subdural hematoma; TBIMS, TBI Model Systems.
By patient or caregiver report. Other race includes individuals who did not self-report as any of the provided groups. BTRC did not collect Hispanic or Latino ethnicity as a racial category but as a separate ethnicity category.
By patient or caregiver report; TBIMS collected Hispanic or Latino as a racial category; ethnicity was only collected as independent from race starting in 2012.
In TBIMS, GCS was considered unscorable if a patient was intubated, sedated, or paralyzed.
TBIMS began collecting this variable January 1, 2007.
Figure. Command-Following and 1-Year Outcomes.
One-year death or dependency was defined as Glasgow Outcome Scale Extended less than 4. The solid line represents the line of best fit of the individual data points (dots); shading 95% CIs of the daily proportions calculated using the formula for the standard error of a proportion. BTRC indicates Brain Trauma Research Center; TBIMS, TBI Model Systems.
Among all participants, each additional day without command-following was associated with a 1.2% (95% CI, 1.2-1.2) (P < .001) increase in TBIMS and 1.1% (95% CI, 1.0-1.1) (P < .001) increase in BTRC in the proportion with death or dependency (Figure, C). A 90% or greater likelihood of this outcome was only observed in participants who did not follow commands for at least 50 days in the TBIMS cohort or at least 40 days in the BTRC cohort (Figure, D and E).
Discussion
In 2 TBI cohorts, time to command-following discriminated weakly between 1-year outcomes. Each additional day without command-following yielded only an approximately 1% increase in probability of death or dependency, a rate remarkably consistent between cohorts. A 90% likelihood of this outcome was observed only in participants failing to follow commands for more than 40 days, a rare occurrence in either cohort, even among those with 1-year death or dependency. These findings suggest that clinicians should avoid confidently assigning a poor prognosis based on the failure to follow commands within the first 5 weeks after TBI.
Limitations of this study include retrospectively determined command-following dates, variability in the number of command-following assessments across participants, and unmeasured confounders associated with critical illness. Additionally, to avoid self-fulfilling prophecy bias, we removed participants who died during acute hospitalization, which may have systematically excluded those more likely to remain dependent. When establishing a prognosis after severe TBI, clinicians may consider applying the 1% rule, that is, the probability of an unfavorable outcome increases by approximately 1% for each additional day without command-following.
eMethods.
Data Sharing Statement
References
- 1.Giacino JT, Sherer M, Christoforou A, et al. Behavioral recovery and early decision making in patients with prolonged disturbance in consciousness after traumatic brain injury. J Neurotrauma. 2020;37(2):357-365. doi: 10.1089/neu.2019.6429 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Helbok R, Rass V, Beghi E, et al. ; Curing Coma Campaign and its Contributing Members . The Curing Coma Campaign International Survey on Coma Epidemiology, Evaluation, and Therapy (COME TOGETHER). Neurocrit Care. 2022;37(1):47-59. doi: 10.1007/s12028-021-01425-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Whyte J, Cifu D, Dikmen S, Temkin N. Prediction of functional outcomes after traumatic brain injury: a comparison of 2 measures of duration of unconsciousness. Arch Phys Med Rehabil. 2001;82(10):1355-1359. [DOI] [PubMed] [Google Scholar]
- 4.Traumatic Brain Injury Model Systems National Data and Statistical Center. Accessed September 4, 2021. https://www.tbindsc.org/
- 5.Deng H, Nwachuku EL, Wilkins TE, et al. Time to follow commands in severe traumatic brain injury survivors with favorable recovery at 2 years. Neurosurgery. 2022;91(4):633-640. doi: 10.1227/neu.0000000000002087 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kowalski RG, Hammond FM, Weintraub AH, et al. Recovery of consciousness and functional outcome in moderate and severe traumatic brain injury. JAMA Neurol. 2021;78(5):548-557. doi: 10.1001/jamaneurol.2021.0084 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods.
Data Sharing Statement
