pmc2985158R833Arch Phys Med RehabilArch Phys Med RehabilArchives of physical medicine and rehabilitation0003-99931532-821X401396291235375410.1016/j.apmr.2025.03.038HHSPA2068694ArticleBehavioral fluctuation in disorders of consciousness: a retrospective analysisBarraAlicePhD12BodienYelena GPhD345TanCan OzanPhD6MartensGeraldinePhD23MaloneChristopherPsyD3http://orcid.org/0000-0002-7916-9698GiacinoJoseph TPhD34IRENEA – Instituto de Rehabilitación Neurológica, Fundación Hospitales Vithas, València, SpainComa Science Group, GIGA Consciousness - GIGA Research, University of Liège, Liège, BelgiumSpaulding Rehabilitation Hospital, Charlestown, Massachusetts, USAHarvard Medical School, Department of Physical Medicine and Rehabilitation, Boston, MassachusettsCenter for Neurotechnology and Neurorecovery, Department of Neurology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA.RAM Group, Electrical Engineering, Mathematics, and Computer Science, University of Twente, the NetherlandsCorresponding author : Joseph T. Giacino, PhD, Spaulding Rehabilitation Hospital, 300 1st Avenue, Charlestown, MA 02129, jgiacino@mgh.harvard.edu, Reprints can be obtained from the Corresponding Author1242025102025243202524320261061015051513Objectives:

To assess the frequency of behavioral fluctuations in patients with prolonged disorders of consciousness (DoC), characterize the stability of consciousness ratings, and characterize the stability of behavioral signs of consciousness.

Design:

Prospective observational analysis.

Setting:

Specialized DoC Program in an inpatient rehabilitation facility and a long-term acute care hospital.

Participants:

Patients in vegetative state/unresponsive wakefulness state (VS/UWS), minimally conscious state and emerged from MCS followed weekly on the Coma Recovery Scale- Revised (CRS-R) between 28 and 90 days post-injury (N=241).

Main Outcome Measure:

Change in CRS-R subscale scores and consciousness ratings.

Results:

Behavioral fluctuation was observed in more than 80% of patients, and was most common on the CRS-R Motor subscale and least common on the Communication subscale (83% and 54% of patients experienced ≥1 fluctuation over the 3-week study period, respectively, with a 1-point change observed most frequently). Among patients who were conscious on baseline assessment, 25% were subsequently rated as unconscious at least once. Localization to pain and object manipulation were the most stable signs of consciousness, recurring at least 3 times after the first occurrence in ≥ 97% of the sample. Reproducible command-following and intelligible verbalization were the least stable, recurring at least 3 times after the first occurrence in ≤ 27% of the sample.

Conclusions:

Patients with prolonged DoC who undergo serial assessment demonstrate a high rate of fluctuation in behavioral signs of consciousness. These findings highlight that repeated assessments are essential in this population, both to capture the highest level of consciousness and to help distinguish spontaneous fluctuation from response to treatment in interventional studies.

Disorders of consciousnessassessment scaletraumatic brain injuryrehabilitationminimally conscious statevegetative stateunresponsive wakefulness syndrome

Disorders of consciousness (DoC) are common after severe acquired brain injury. The three major DoCs are coma, vegetative state (VS), also referred to as unresponsive wakefulness syndrome (UWS), and minimally conscious state (MCS). The defining feature of coma is the complete loss of wakefulness1. The eyes are continuously closed and the patient cannot be aroused. There are also no observable signs of cognition or purposeful motor activity. In VS/UWS, sleep-wake cycles return along with spontaneous eye-opening in the absence of behavioral signs of self or environmental awareness2. In MCS, there is minimal but definite behavioral evidence of self or environmental awareness demonstrated by simple command-following, intelligible verbalization, recognizable but unreliable verbal or gestural “yes/no” responses or behaviors that selectively occur in the presence of specific environmental stimuli and cannot be attributed to reflexive activity (e.g., visual tracking of a moving object)3. MCS has been subcategorized into “MCS-plus (MCS+)” and “MCS-minus (MCS−),” based on the presence or absence, respectively, of receptive or expressive language function4. Emergence from MCS (EMCS) is signaled by the recovery of reliable communication or functional object use3. Patients who emerge from traumatic MCS typically transition into the post-traumatic confusional state (PTCS)5.Diagnostic assessment of patients with DoC is complex. Prior studies suggest that approximately 40% of patients who are believed to be unconscious based on routine bedside examination actually retain conscious awareness6,7. Fluctuation in behavioral responsiveness, a hallmark feature of DoC, is among a myriad of confounding factors that can contribute to misdiagnosis. This phenomenon has been described in both acute and prolonged (i.e., > 28 days post-injury) DoC8,9. Patients with DoC may demonstrate clear and convincing evidence of command-following on one examination, and then fail to show this sign of consciousness when the examination is repeated shortly afterward. Although behavioral fluctuation in DoC is widely recognized by investigators and clinicians10, the frequency of these fluctuations and their impact on diagnostic assessment have not received adequate attention. Behavioral fluctuation can lead to significant consequences, including misdiagnosis and misattributing normal behavioral variability to treatment effects. Misdiagnosis in DoC is common6 even when using a standardized behavioral assessment instrument such as the Coma Recovery Scale-Revised (CRS-R),11 which is considered the most reliable behavioral tool to evaluate patients with DoC11,12. Wannez and colleagues13 found that at least five CRS-R exams were necessary to reduce the risk of misdiagnosing a conscious patient as unconscious to 5% or less. When a single exam was performed, consciousness was missed in 36% of the cases assessed. Cortese and colleagues found that signs of consciousness also fluctuate diurnally and are more likely to be observed in the morning than in the afternoon14. Daily fluctuations in visual pursuit, which often marks the transition from an unconscious to conscious state15, were noted in 38% of cases followed.16 Given the ubiquity of fluctuation in DoC17 and the central role that diagnosis plays in clinical decision-making, a deeper understanding of the base rate of spontaneous behavioral fluctuation in DoC is needed.

In view of the knowledge gaps described above, the aims of this study were to: 1) determine the frequency of spontaneous fluctuation in behaviors indexed by the CRS-R among adult patients with prolonged DoC, 2) characterize the stability of ratings of consciousness among patients diagnosed with minimally conscious state (MCS) and emerged from MCS (eMCS); and 3) characterize the stability of specific behavioral signs of consciousness over repeated CRS-R assessments.

MethodsParticipants

We enrolled a convenience sample comprised of patients admitted to a DoC neurorehabilitation program co-located in an inpatient rehabilitation facility and a long-term acute care hospital. We identified patients who met the following inclusion criteria:

At least 18 years old;

Admitting diagnosis of vegetative state (also known as unresponsive wakefulness syndrome (VS/UWS), MCS or eMCS from either traumatic or non-traumatic (i.e., stroke, hemorrhage, hypoxic, ischemic, metabolic) mechanism of injury;

Assessed at least 6 times with the CRS-R over a maximum of 21 days;

Between 28 and 90 days post-injury at the time of initial assessment.

Procedures

All participants underwent serial standardized behavioral assessment of level of consciousness and received multidisciplinary rehabilitation provided by speech, physical, and occupational therapists in accord with standard clinical procedures. Trained clinicians administered the CRS-R twice per week during regularly scheduled therapy sessions. Examinations were always performed between 8:00 am and 3:00 pm, either in the patient’s room or in a therapy room within the hospital. CRS-R scores were recorded in the electronic medical record by the clinician and then extracted and transferred by research staff into the Research Electronic Data Capture (REDcap)18,19. The research protocol was approved by the local Institutional Review Board and written informed consent was waived because all data were collected as part of routine clinical care for the inpatient DoC Program.

Measures

The CRS-R served as the primary outcome measure. The CRS-R is a traumatic brain injury (TBI) common data element recommended for monitoring recovery of consciousness in patients with DoC. The six subscales of the CRS-R assess auditory, visual, motor, oromotor/verbal, communication and arousal functions and all administration and scoring procedures are manualized. Subscale items are hierarchically-arranged and correspond to brain stem, subcortical and cortically-mediated functions20. The highest score on each subscale represents a cortically-based, cognitively-mediated behavior while the lowest score represents reflexive brain stem-mediated responses. The subscale score profile obtained on each assessment is used to establish the DoC diagnosis (VS/UWS, MCS, eMCS). The auditory subscale score ranges from 0 to 4, visual score from 0 to 5, motor score from 0 to 6, oromotor score from 0 to 3, communication score from 0 to 2, and arousal score from 0 to 2. Test Completion Codes (TCCs) were assigned to standardize validity ratings for each subscale administered. Performance was coded as valid, if the measure was completed in full with no threats to validity, attempted but not completed, if the measure was not completed or judged invalid or, not attempted, if no attempt was made to administer the subscale. The TCCs also indicate why performance was judged invalid (eg, sensory impairment, language barrier).

We also collected data on demographic (i.e., age, sex) and injury characteristics (i.e., mechanism of injury, time since injury [TSI],).

Statistical Analyses

Data analyses only included CRS-R assessments that were assigned a valid Test Completion Code (1.0) to eliminate the influence of confounding factors on subscale scores. For analytic purposes, we collapsed “consistent command-following” and “reproducible command-following” on the Auditory subscale and “visual fixation” and “visual pursuit” on the Visual subscale.

To evaluate the frequency of spontaneous fluctuations in CRS-R subscale scores for each participant (Objective 1), we calculated the number of times a CRS-R subscale score changed over six consecutive assessments and the percentage of patients by the number of fluctuations in subscale score captured (min=0, max =5) across the six assessments. Fluctuations in the arousal subscale score were excluded from this analysis because scores on this subscale, unlike the other five CRS-R subscales, are not used to establish DoC diagnosis. We defined fluctuation as a change (i.e., increase or decrease) in a CRS-R subscale score between consecutive assessments. See Table 1 for a simulated illustration of the number of fluctuations in a single patient. We calculated the magnitude of behavioral fluctuations by identifying the absolute difference in the subscale score (increase or decrease) between consecutive examinations. See Table 2 for an illustration of how the magnitude of fluctuation was calculated. We also explored whether the level of arousal, indexed by the Arousal subscale score, was associated with fluctuations in other subscale scores, and compared the directionality of the fluctuations and changes in arousal in a 2×2 matrix showing the frequency with which a fluctuation was associated with an increase or decrease in the CRS-R Arousal subscale score. We tested the distribution of the matrix with a Chi-squared test.

To characterize “stability of consciousness” (i.e., unconscious vs. conscious) (Objective 2), we included only patients who were conscious (i.e., diagnosed with MCS or eMCS) on the first (i.e., baseline) CRS-R assessment. We evaluated the number of times each of these participants was rated as conscious over the five CRS-R examinations that followed the baseline assessment. We also explored the relationship between stability of consciousness and arousal ratings. To determine whether level of arousal was associated with stability of consciousness, we used a 2×2 matrix to compare the frequency with which changes in consciousness were accompanied by changes in the CRS-R Arousal subscale score. We tested the distribution of the matrix with Fischer’s exact test.

Lastly, to determine the stability of individual CRS-R behavioral signs of consciousness (Objective 3), we calculated the frequency with which each of these behaviors recurred over the 5 evaluations that followed the first occurrence of the behavior. We operationally-defined signs of consciousness as consistent or reproducible response to command; visual fixation or pursuit; object reaching; object recognition; localization to noxious stimulation; intelligible verbalization; intentional communication and functional communication. We considered a behavioral sign of consciousness to have recurred any time the same behavior or a higher-scoring behavior was scored as present after the first occurrence. This approach aligns with the evidence-based hierarchical structure of the CRS-R in which higher-scoring behaviors have been shown to subsume lower-level behaviors21,22. We also report the percentage of the sample that showed each behavior one to five times after the first occurrence. This analysis included only participants showing at least one behavioral sign of consciousness. Most patients demonstrated more than one conscious behavior and were included in more than one analysis. For example, if a patient showed reproducible response to command, visual pursuit and object recognition on the first CRS-R assessment, their data were included in the analyses for all three of these behaviors. All data were analyzed using RStudio Version 4.3.1.

Results

Of 361 patients included in the REDCap database, 241 met all eligibility criteria, 121 had TBI, 96 were female, mean age was 65.3 (SD= 19.0) years and mean time since injury (TSI) was 35.4 (SD= 13.0) days (see Table 3 for demographic and injury characteristics and Figure 1 for the Study Flowchart).

Frequency and magnitude of fluctuation

Behavioral fluctuation occurred most frequently on the Motor subscale and least often on the Communication subscale (83% and 54% patients with ≥1 fluctuation, respectively). On the Auditory, Visual, and Oromotor/Verbal subscale, ≥1 fluctuation was observed in 80%, 80%, and 74% of participants, respectively. The frequency of fluctuation by subscale is illustrated in Figure 2 and the proportion of participants by number of fluctuations (0 to 5) is shown in supplementary table S1. A 1-point fluctuation in CRS-R subscale score was observed most frequently, however, fluctuations as high as 5 points (i.e., Motor and Visual subscales) were observed. See Figure 3 for details on the magnitude of fluctuation. The direction of fluctuation (i.e., increase versus decrease in CRS-R subscale score) was not related to the direction of change in arousal (increased versus decreased arousal, see table 4 for details). Changes in subscale scores were evenly distributed across increases and decreases in arousal and the Chi-squared test was not significant (p= .27).

Stability of ratings of consciousness

Among the 192 participants who were diagnosed as conscious on the first assessment, 25% had at least one subsequent assessment on which consciousness was not detected. Figure 4 shows the proportion of patients who were diagnosed as conscious 1 to 5 times after being diagnosed as conscious on the baseline exam.

The results of the Fisher’s exact test (p< .001) indicate a significant association between ratings of level of consciousness and level of arousal. Ratings of level of consciousness were associated with increases but not decreases in arousal level (see Table 5). In 100% of the examinations in which arousal level was observed to increase, the consciousness rating also improved. However, in 86% of the examinations, when the level of arousal decreased, the consciousness rating remained unchanged.

Stability of behavioral signs of consciousness

Figures 5A, 5B, and 5C show the frequency with which specific CRS-R behavioral signs of consciousness recurred across the five examinations that followed the baseline exam. Motor behaviors (i.e, localization to pain, object manipulation) were the most stable signs of consciousness, recurring at least 3 times after the first occurrence in ≥ 97% of the sample. Language-related behaviors (i.e., reproducible command-following, intelligible verbalization) were the least stable, recurring at least 3 times after the first occurrence in ≤ 27% of the sample. Approximately one half of the sample demonstrated intentional (55%) or functional communication (51%) at least 3 times after the first occurrence. Supplementary table S2 shows the most stable behavioral signs of consciousness in rank order.

Discussion

Behavioral fluctuation is commonly observed in patients with prolonged DoC10. Our objective was to systematically characterize the frequency and magnitude of behavioral fluctuation among patients with DoC, and explore the impact of these fluctuations on diagnostic assessment. We found that behaviors indexed by CRS-R subscales fluctuated at least once across six consecutive examinations. The magnitude of fluctuation was typically 1 point, although up to 20% of participants showed a change of 2 or more points, depending on CRS-R subscale. Among participants who were found to be conscious on the first examination, one in four failed to demonstrate consciousness on a subsequent examination. In line with this finding, most behavioral signs of consciousness fluctuated at least once over the 5 examinations performed after the baseline assessment. Given the frequency with which behavioral signs of consciousness appear to fluctuate, our results support existing practice guidelines recommending serial assessment before establishing a DoC diagnosis13,23 Our findings also suggest that one- to two-point increases or decreases in CRS-R subscale scores commonly occur in the absence of treatment interventions, suggesting that changes of this magnitude may reflect the base rate of fluctuation rather than true clinical improvement or decline. Serial monitoring provides a means to better understand the trajectory and clinical significance of CRS-R subscale changes as many participants showed increases or decreases in scores that subsequently reverted back to the original score over the course of 21 days or less.

The root cause of fluctuation in behavioral responsiveness in patients with DoC is unknown and likely multi-determined. Schiff and Purpura have proposed a pathophysiological model,8 hypothesizing that when a lesioned brain area is unable to provide excitatory input to distant brain regions- a mechanism referred to as “disfacilitation”, the result may be diffuse neural network inhibition. This inhibitory state produces hyperpolarization of neuronal membranes that, together with generally lower global cerebral metabolism post-injury, may result in functional impairments that appear disproportionate to the extent of the structural injury. The combined effects of diffuse inhibition, hyperpolarization and reduced cerebral metabolic rate are dynamic and may drive or contribute to behavioral fluctuation.

We found an association between fluctuations in arousal level and ratings of consciousness. While changes from an unconscious to conscious state were always associated with increases in level of arousal, consciousness ratings remained stable in nearly 90% of the examinations associated with a decline in arousal ratings. This finding suggests that declines in level of consciousness are not necessarily driven by decreases in arousal level and may instead reflect naturally-occurring oscillations in resting brain state as have previously been reported in cortical deafferentation, sleep and anesthesia studies8,24. The absence of a consistent bi-directional relationship between arousal level and changes in level of consciousness may also be due to the influence of potentially-sedating medications as their impact on behavioral responsiveness varies based on drug class, dosage and time of administration. However, these effects are difficult to assess given the variability in dosing and metabolism across patients. Additionally, this was a retrospective study and the timing of CRS-R administration in relation to these data elements was not consistently available in the medical record. We are also unable to rule out that the absence of a relationship between decreased arousal and level of consciousness is attributable to the lack of sensitivity of the Arousal subscale, although the unidirectional nature of this finding makes this possibility less likely. 24

One of the most important findings from this study is that 25% of participants failed to show any sign of consciousness on at least one examination performed after initial detection of consciousness. This points to the importance of conducting regular systematic assessment over the course of two to three weeks, especially during the acute to subacute phase of recovery as originally recommended by Wannez and colleagues.13 We are unable to draw firm conclusions about the stability of specific behavioral signs of consciousness as small sample sizes for some behaviors prevented comparative analysis. Nevertheless, our study provides descriptive evidence that localization to pain and object manipulation on the CRS-R Motor subscale, and object reaching and object recognition on the Visual subscale, may be more stable than other behavioral signs of consciousness.

Study Limitations

This study has several important limitations that should be considered. First, we were unable to control for extraneous factors that may have influenced changes in CRS-R subscale scores. Many of these factors have been well-documented and include time of day, exposure to potentially-sedating medications, seizure activity, occult medical complications, metabolic instability and sleep disturbance14,17. Nonetheless, these circumstances are characteristic of the real-world conditions that clinicians must contend with. Future studies should incorporate study design features that control or adjust for these factors. Second, we included patients who were 28–90 days post-injury. The pattern of fluctuation preceding and following this time period may differ from what we observed here. Finally, it is important to recognize that our definition of fluctuation is deliberately agnostic as to whether the observed changes in behavior (and CRS-R subscale scores) reflect true clinical improvement or decline. For example, a participant whose Motor subscale score continuously improved over the course of the six examinations was coded as having experienced five fluctuations. While our data do not allow us to identify the underlying cause of these changes, this frequency was only observed in 2%–5% of the sample and, as such, may represent a non-random change in clinical status. Measures of CRS-R responsiveness have recently been developed25 and are expected to help clinicians discern when changes in behavioral responsiveness exceed measurement error and reflect true change.

Conclusions

Serial standardized assessment of level of consciousness conducted over 21 days among patients with prolonged DoC suggests that behavioral fluctuation is frequently observed. When the CRS-R is used to monitor behavioral recovery, a 1-point change in subscale score in either direction appears to be the norm, although larger changes (i.e., 3 points) are not uncommon. Behavioral fluctuation does not appear to be closely tied to arousal level and may reflect injury-related pathophysiologic perturbations in corticothalamic connectivity. These findings suggest that it is essential to conduct repeated assessments in this population to capture the highest level of consciousness retained and to help distinguish spontaneous fluctuation from response to treatment in interventional studies.

Supplementary MaterialPresentation acknowledgment

The findings from this study were summarized in a poster presented at the annual meeting of the American Congress of Rehabilitation Medicine (ACRM) in November 2019 in Chicago, IL, USA.

Funding acknowledgment

This work was supported by the HORIZON EUROPE Marie Sklodowska-Curie Actions (PF-01-10110814), NIH National Institute of Neurological Disorders and Stroke (R21NS109627, RF1NS115268, UH3NS095554, U01 NS1365885, U01- NS086090), NIH Director’s Office (DP2 HD101400), National Institute on Disability, Independent Living and Rehabilitation Research (NIDILRR), Administration for Community Living (90DPCP0008-01-00, 90DP0039), James S. McDonnell Foundation, and Tiny Blue Dot Foundation, U.S. Department of Defense (W81XWH14-2-0176, X81XWH-18-DMRDP-PTCRA), National Science Foundation (1014552). Further support was given by the University of Liege, the Belgian National Funds for Scientific Research (FRSFNRS), the European Union’s Horizon 2020 Framework Programme for Research and Innovation under the Specific Grant Agreement No. 945539 (Human Brain Project SGA3), and the European Commission HORIZON-MSCA-2022-PF-01-101108141, the fund Generet, the Mind Care International Foundation, the King Baudouin Foundation, USA, DOCMA project [EU-H2020-MSCA-RISE-778234].

Conflicts of interest

Drs. Giacino and Bodien receive licensing royalties for the Coma Recovery Scale- Revised. Dr. Giacino occasionally receives honoraria for conducting Coma Recovery Scale- Revised training seminars for academic institutions and medical centers. The other authors have nothing to disclose.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

AbbreviationsCRS-R

Coma Recovery Scale – Revised

DoC

Disorders of Consciousness

eMCS

Emergence from minimally conscious state

MCS

minimally conscious state

n-TBI

Non-traumatic brain injury

SD

Standard deviation

TCC

Test completion code

TSI

Time since injury

ReferencesPosnerJB Plum and Posner’s Diagnosis of Stupor and Coma. vol. 71 (OUP USA, 2007).JennettB & PlumF Persistent vegetative state after brain damage. A syndrome in search of a name. Lancet 1, 734–737 (1972).4111204 GiacinoJT The minimally conscious state: Definition and diagnostic criteria. Neurology (2002) doi:10.1212/WNL.58.3.349.ThibautA, BodienYG, LaureysS & GiacinoJT Minimally conscious state “plus”: diagnostic criteria and relation to functional recovery. J Neurol 267, 1245–1254 (2020).31773246 ShererM Post-traumatic Confusional State: A Case Definition and Diagnostic Criteria. Arch Phys Med Rehabil 101, 2041–2050 (2020).32738198 SchnakersC Diagnostic accuracy of the vegetative and minimally conscious state: Clinical consensus versus standardized neurobehavioral assessment. BMC Neurol 9, (2009).SchnakersC Update on diagnosis in disorders of consciousness. Expert Rev Neurother 20, 997–1004 (2020).32700583 SchiffND & PurpuraKP Towards a neurophysiological foundation for cognitive neuromodulation through deep brain stimulation. Thalamus Relat Syst 2, 55–69 (2002).GiacinoJT The minimally conscious state. Neurology 58, 349 LP – 353 (2002).11839831 PapadimitriouC “Fluctuation is the norm”: Rehabilitation practitioner perspectives on ambiguity and uncertainty in their work with persons in disordered states of consciousness after traumatic brain injury. PLoS One 17, e0267194 (2022).35446897 GiacinoJT, KalmarK & WhyteJ The JFK Coma Recovery Scale-Revised: measurement characteristics and diagnostic utility. Arch Phys Med Rehabil 85, 2020–2029 (2004).15605342 BodienYG, CarlowiczCA, ChatelleC & GiacinoJT Sensitivity and specificity of the coma recovery scale-revised total score in detection of conscious awareness. Arch Phys Med Rehabil 97, (2016).WannezS, HeineL, ThonnardM, GosseriesO & LaureysS The repetition of behavioral assessments in diagnosis of disorders of consciousness. Ann Neurol 81, 883–889 (2017).28543735 CorteseMD Coma recovery scale-r: Variability in the disorder of consciousness. BMC Neurol 15, 1–7 (2015).25595849 MartensG, BodienY, SheauK, ChristoforouA & GiacinoJT Which behaviours are first to emerge during recovery of consciousness after severe brain injury? Ann Phys Rehabil Med 63, 263–269 (2020).31783144 CandelieriA, CorteseMD, DolceG, RiganelloF & SannitaWG Visual pursuit: Within-day variability in the severe disorder of consciousness. J Neurotrauma 28, 2013–2017 (2011).21770758 GiacinoJT Practice Guideline Update Recommendations Summary: Disorders of Consciousness: Report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology; the American Congress of Rehabilitation Medicine; and . Arch Phys Med Rehabil 99, 1699–1709 (2018).30098791 HarrisPA The REDCap consortium: Building an international community of software platform partners. Journal of Biomedical Informatics vol. 95 Preprint at 10.1016/j.jbi.2019.103208 (2019).HarrisPA Research electronic data capture (REDCap)-A metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 42, 377–381 (2009).18929686 La PortaF Can we scientifically and reliably measure the level of consciousness in vegetative and minimally conscious States? Rasch analysis of the coma recovery scale-revised. Arch Phys Med Rehabil 94, 527–535 (2013).23127303 La PortaF Can we scientifically and reliably measure the level of consciousness in vegetative and minimally conscious States? Rasch analysis of the coma recovery scale-revised. Arch Phys Med Rehabil 94, 527–535 (2013).23127303 WeaverJA, LiuJ, GuernonA, PapeTB & MallinsonT Psychometric properties of the coma near-coma scale for adults in disordered states of consciousness: A rasch analysis. Arch Phys Med Rehabil 102, 591–597 (2021).33161008 GiacinoJT Behavioral assessment in patients with disorders of consciousness: gold standard or fool’s gold? in Coma Science: Clinical and Ethical Implications (eds. LaureysS., SchiffND. & OwenAMBT-P. in B. R.) vol. 177 33–48 (Elsevier, 2009).SchiffND Central thalamic deep-brain stimulation in the severely injured brain: rationale and proposed mechanisms of action. Ann N Y Acad Sci 1157, 101–116 (2009).19351360 WeaverJA Interpreting Change in Disorders of Consciousness Using the Coma Recovery Scale-Revised. J Neurotrauma (2024).Flowchart of inclusion

Study flowchart.

Frequency of fluctuation. Bars represent the proportion of the sample (N= 241) demonstrating 0 to 5 fluctuations in CRS-R subscale scores over the course of the 6 evaluations, specified by color (i.e., 0; 1; 2; 3; 4; 5).

Magnitude of fluctuation. Bars represent the proportion of CRS-R subscale score fluctuations among the full sample (N= 241) that were between 1 and 5 points. Note that the range of possible score fluctuations varied by subscale (Auditory: 0 to 4; Visual: 0 to 5; Motor: 0 to 6; Oromotor/Verbal: 0 to 3; Communication: 0 to 2). While the modal fluctuation was 1-point, we observed fluctuations as high as 5 points, depending on subscale.

Stability of ratings of consciousness. Colors depict the number of examinations on which the patient was rated as conscious after the first examination demonstrating consciousness (i.e., 0; 1; 2; 3; 4; 5). Among 192 subjects who demonstrated consciousness on at least one examination, one in four (25%) fluctuated from conscious to unconscious at least once over the 5 exams. 13% of the sample was diagnosed as in VS/UWS only once across the 5 exams (4 conscious assessments), whereas the remaining 12% was diagnosed as in VS/UWS on 2 to 5 instances.

A-C. Stability of behavioral signs of consciousness. Bars represent the frequency with which specific CRS-R language (A), visuoperceptual (B) and motor (C) signs of consciousness recurred across the six examinations. Sample sizes are indicated at the bottom of each bar and differ for each behavior, reflecting differences in the number of participants demonstrating the behavior at baseline. Color shading shows the proportion of participants who demonstrated each behavioral sign of consciousness 0 to 5 times after the first demonstration of the behavior. Localization to noxious stimulation (n= 21) was captured most consistently across examinations (75% of the sample showed this behavior on all 5 assessments after the first occurrence). In contrast, in 44% of the sample, intelligible verbalization was never captured after the first occurrence.

Simulated illustration of the approach used to calculate the frequency of fluctuation.

CRS-R Assessment
Patient 1123456Frequency of fluctuation
CRS-R Visual Subscale score335*4*3*3
Fluctuation present (+) / absent (−)−+++−3

Note. This example shows a patient with CRS-R Visual Subscale scores of 3, 3, 5, 4, 3, and 3 on assessments 1 through 6. Fluctuations in score are marked with an asterisk (*). A plus sign (+) is used to indicate the presence of a fluctuation in score between consecutive assessments, and a minus sign (−) indicates the absence of fluctuation between assessments. As shown, fluctuations occur between CRS-R assessments 2 and 3 (score increased from 3 to 5), 3 and 4 (score decreased from 5 to 4) and 4 and 5 (score decreased from 4 to 3) for a total of 3 fluctuations across the 6 assessments performed. CRS-R: Coma Recovery Scale-Revised.

Simulated illustration of the approach used to calculate the magnitude of fluctuations.

Patient 1CRS-R Assessment
123456
CRS-R Visual Subscale score335*4*3*3
Magnitude of fluctuationNA0+2−1−10

Note. This example shows a patient with changes in the CRS-R Visual Subscale score (marked with an asterisk) between CRS-R assessments 2 and 3 (+2 points), 3 and 4 (−1 point), and 4 and 5 (+1 point). As indicated, the magnitude of fluctuation ranges from −1 to +2 points across the 6 assessments performed. CRS-R: Coma Recovery Scale-Revised.

Demographic and clinical characteristics of the sample.

Sample size
 N screened= 361
 N included= 241
Age
 Mean = 65.3 (SD= 19)
TSI
 Mean = 35.8 (SD= 13)
Sex
 Females = 96 (39.8%)
 Males = 145 (60.6%)
Etiology
 TBI = 121 (50.2%)
 n-TBI = 120 (49.7%)

Note. Age is reported as mean and (SD) in years. TSI is reported as mean and (SD) in days. SD: Standard Deviation; TSI: Time Since Injury; TBI: Traumatic Brain Injury; n-TBI: non- Traumatic Brain Injury.

Relationship between arousal and behavioral fluctuation.

Arousal UpArousal Down# Observations
Fluctuation Up333434767
Fluctuation Down290336626
# Observations6237701,393

Note. Fluctuation Up indicates there was an increase in the CRS-R subscale score (i.e., behavioral improvement) and Fluctuation Down indicates there was a decrease in the CRS-R subscale score (i.e., behavioral decline). Arousal Up indicates an increase in the Arousal subscale score (i.e., improvement in arousal level) and Arousal Down indicates a decrease in the Arousal subscale score (i.e., decline in arousal level). Increases and decreases in arousal level did not co-occur with improvement and decline in CRS-R Subscale scores, respectively. (χ2=1.1806, p= .277224).

Relationship between arousal and ratings of consciousness.

Arousal DownArousal Up# Observations
Consciousness Down13013
Consciousness Up103545
# Observations233558

Note. Consciousness Down indicates a decline in level of consciousness (i.e., change in diagnosis from MCS or eMCS to VS/UWS) and Consciousness Up indicates an improvement in level of consciousness (i.e., change in diagnosis from VS/UWS to MCS or eMCS). Arousal Down indicates a decrease in the Arousal subscale score (i.e., decline in arousal level) and Arousal Up indicates an increase in the Arousal subscale score (i.e., improvement in arousal level). As indicated in the table, while increases in the Arousal subscale score were associated with improvement in level of consciousness (i.e., Arousal Up/Consciousness Up= 35/35; Arousal Up/Consciousness Down= 0/35), decreases in the Arousal subscale score did not differentially affect level of consciousness (ie, Arousal Down/Consciousness Down= 13/23; Arousal Down/Consciousness Up= 10/23) (Fisher’s exact test: p< .001). Fluctuations in level of consciousness that did not occur in association with changes in level of arousal were excluded.