The differential diagnosis between vegetative state (VS) and minimally conscious state (MCS) is a challenging task that requires assessment by experienced clinicians and the use of specific scales. A recent systematic review (Seel et al., 2010) concluded that only the Coma Recovery Scale-revised (CRS-R) (Giacino et al., 2004) could be used for the reliable assessment of patients with disorder of consciousness (DOC) in line with the Aspen Criteria (Giacino et al., 1997).
The CRS-R includes six items addressing auditory, visual, motor, oromotor, communication, and arousal processes, each with different response categories, and its total score is calculated considering the presence or absence of specific behavioral responses to sensory stimuli, with a range between 0 (worst) and 23 (best). Clinicians have to administer first stimuli related to the highest-level response categories for each item, which reflect cognitively mediated behaviors, and then proceed to lower-level categories, if a patient fails to respond. According to the Aspen Workgroup’s consensus-based diagnostic classifications (Giacino et al.,1997), the definition of VS (The Multi-Society Task Force on PVS, 1994), MCS, and emersion from MCS could be operationalized on the basis of the CRS-R items (Giacino, 2005). In detail, a diagnosis of VS is established if all the following criteria are fulfilled: auditory ≤2; visual ≤1; motor ≤2; oromotor/verbal ≤2; communication=0; and arousal ≤2. A patient is considered to be in MCS if auditory=3–4 and/or visual=2–5 and/or motor=3–5 and/or oromotor/verbal=3 and/or communication=1. Emersion from MCS (SD) is considered if a patient achieved a motor score of 6 or a communication score of 2.
This categorization is fundamental for clinical purposes and research, but presents two main limitations. First, the relationship between diagnosis and CRS-R score involves overlapping scores between VS and MCS. In particular, CRS-R scores for VS range from 0 to 9, whereas scores for MCS could ‘theoretically’ range from below 9 up to 21 (e.g. 1 point in communication subscale plus related scores collected in items arousal and auditory and/or, visual, motor, oromotor). Overlap generally occurs around CRS-R scores between 7 and 9 points and, for example, one patient with a CRS score of 8 could be either in VS or in MCS and this implies that two patients with the same score could not only have different overall functioning levels but could also show or not cognitively mediated behaviors, and this issue has a contentious implication for statistical analysis.
Another issue is related to the underlying theoretical assumption that if a patient is able to show higher-level behaviors, he/she is also able to show lower-level responses on each item that represents reflexive activity, or has progressed to a level of consciousness where such behaviors have extinguished (e.g. automatic motor behavior in motor item) (Giacino et al., 2004).
Considering that neurological structures that underpin reflex responses could be different from/or part of brain areas that allow expression of higher cognitive responses, this introduces bias in data analysis involving the CRS-R total score. An explicatory example emerges from analysis of motor functions item: hierarchical organization of response categories is ‘no response/flaccid’, ‘abnormal posturing’, ‘flexion withdrawal’, ‘localization to noxious stimulation’, ‘object manipulation’, ‘automatic motor response’, ‘functional object use’, from lower automatic motor response to higher cognitive-mediated behaviors, respectively. However, in this item, several different neurological functions have been merged without a clear hierarchical neurological/behavioral organization. An example could be given considering response categories related to nociceptive perception (such as ‘flexion withdrawal’) in relation to higher cognitive-mediated behaviors such as ‘objects manipulation’. harmful stimuli to the skin, in fact, activate several classes of nociceptor terminals, the peripheral endings of primary sensory neurons whose cell bodies are located in the dorsal root and trigeminal ganglia. Moreover, a central activation is shown in the brainstem, thalamus, and in the so-called ‘pain matrix’, comprising primary and secondary somatosensory, insula, anterior cingulate, and the parietal and prefrontal cortex (Melzack and Casey, 1968; Peyron et al., 2007; Mouraux et al., 2011; Oertel et al., 2012; Fomberstein et al., 2013). The nociception withdrawal reflex is a polysynaptic and multisegmental spinal reflex that induces a complex flexion synergy of the stimulated limb. The mechanical response is a rapid withdrawal movement that constitutes a protective mechanism against possible limb damage (Sandrini et al., 2005). It is well-known that supraspinal structures, such as the cerebral cortex, cerebellum, basal ganglia, and brainstem, are involved in the modulation of the nociception withdrawal reflex in animals (Schomburg, 1990) and that withdrawal reflex is inhibited by the dorsal reticulospinal tract and facilitated by corticospinal and rubrospinal pathways (Lundberg and Voorhoeve, 1962; Schwindt, 1981; Fleshman et al., 1988). Indeed, neural circuits involved in the ‘object manipulation’ response category, a task that requires the linking of several actions, such as reaching toward, grasping, lifting and transporting the object, object recognition, etc. seem to involve the frontoparietal, intraparietal sulcus, and S2 cortex (Binkofski et al., 1999; Lewis, 2006; Gallivan et al., 2015). Of course, ‘object manipulation’ must be considered a cognitive-mediated behavior that requires activation of several neural circuits, but these brain areas are not necessarily the same as those involved in the nociception withdrawal reflex, which is considered a behavioral response that needs ‘lower’ cognitive functions.
This example renders unsatisfactory the idea of CRS-R items as a subscale with completely hierarchical scores, and our data suggest a more blurred situation in which the presence of lower/upper categories does not directly correspond to lower/upper motor/cognitive levels. In fact, some studies reported that the capacity of ‘object manipulation’ may be present in some pathologies without manifestation/behavioral of nociception withdrawal reflex, supporting the idea that some higher cognitive-mediated behaviors could be present without the lower ones (Cox et al., 2006; Sandroni et al., 2006; Umeda et al., 2013; Peddareddygari et al., 2014).
The demonstration of no linearity/causality between higher and lower behaviors was not cited in several studies that evaluated the validity of the CRS-R (Schnakers et al., 2008; Løvstad et al., 2010; La Porta et al., 2013; Estraneo et al., 2014), although a recent article (Gerrard et al., 2014), using Item Response Theory analysis, reported the strength of the relationship between specific behaviors and the underlying unidimensionality construct that they represent. However, this conclusion is suitable for statistical analysis that considered the CRS-R total score, but for analysis of a single item or for clinical and rehabilitation purposes, could lead to overestimation of the functioning level of patients with DOC. Two patients in MCS, for example, both with ability in ‘object manipulation’, a higher cognitive behavior, could equally show or not some signs of response to noxious stimulation and this is important for clinical management or rehabilitation of a single patient.
Considering these two issue, the aim of this study is to avoid the reported problems presenting a new standardized score for CRS-R.
Sixty adults with DOC were consecutively enrolled at the Coma Research Centre (CRC) of the Neurological Institute C. Besta IRCCS Foundation of Milan between January and December 2011. The CRC provides a 1-week program of clinical, neurological, neurophysiological, and neuroradiological assessments for the diagnosis and prognosis of patients with DOC. The CRC multidisciplinary protocol aims to assess levels of behavioral responses to external stimuli and to evaluate brain and neurological systems using advanced technologies to improve knowledge of disorders of consciousness following brain injuries.
All patients were assessed by two experienced raters, who evaluated every patient using the CRS-R scale four times a week. The assessment procedure was the same as that described by Giacino et al. (2004) for one rater (rater 1), but the other rater also collected information on lower response categories for that items where patients showed higher cognitive-mediated behaviors (rater 2). The best performance CRS-R total scores were calculated considering the best total scores recorded during each assessment.
The study was approved by the Ethics Committee of C. Besta Neurological Institute in Milan, Italy, and was carried out in accordance with the Declaration of Helsinki.
The Coma Recovery Scale-revised
The CRS-R items are organized on the basis of the neurologic complexity of the behaviors of interest and the entire examination takes 15 to 30 min to be completed. The most complex behavior on each item is administered first and is assigned the highest score, whereas the most primitive behavior is administered last and assigned the lowest score. At the beginning of the examination, a brief baseline observation period should be made by raters to collect data on the nature and frequency of behaviors that occur at rest to help differentiate volitional from random behaviors.
The CRS-R items include auditory, visual, motor, oromotor/verbal, communication, and arousal.
The four-response category auditory item ranges from basic sensory detection of auditory stimuli (lower limit) to language comprehension (upper limit). The five-response category visual item examines simple to complex visuo perceptual functions. The six-category motor item includes tests assessing controlled movements (functional object use) as well as reflexive responses to noxious stimulation (flexion withdrawal, abnormal posturing). The three-response category oromotor/verbal item not only investigates the capacity for intelligible speech but also monitors unintelligible vocalizations and oral reflexive movements. The two-category communication item attempts to prompt yes–no responses using personal and situational orientation questions. Finally, the three-response category arousal item assesses the level of alertness: from basic wakefulness to sustained attention. Brain stem reflexes are also assessed for prognostic purposes and to aid the interpretation of specific item response categories (McDonnell et al., 2015).
An Arousal Facilitation Protocol, reported by Giacino in the CRS-R instructions, is administered by applying deep pressure to specific muscle groups with the aim of increasing alertness when patients appear sleepy or obtund.
A new scoring for the Coma Recovery Scale-revised
To develop a new scoring for the CRS-R, a procedure has been developed in three steps.
Step 1: extract two new indices for CRS-R scoring
For each CRS-R item, we subdivided responses into two categories: ‘noncognitively mediated reflex/behavior’ raw score (RB_rs) and ‘cognitively mediated behavior’ raw score (CMB_rs) as reported in Table 1. This distinction was made following the CRS-R classification of response categories for each item for the operationalization of diagnosis of VS and MCS (Giacino et al., 2004).
To determine the total RB_rs and total CMB_rs scores, the sums of each response category completed in both groups were considered.
The response categories ‘functional object use’ in the Motor Function Scale and ‘functional: accurate’ in the Communication Scale were not included in RB_rs and CMB_rs scores because these behaviors clearly require a recovered consciousness. The maximum score depended on the number of response categories as reported in Table 1. The total maximum scores for RB_rs and CMB_rs are 7 and 11, respectively. In detail, for the Auditory Function Scale – RB_rs range 0–2, CMB_rs range 0–2; for the Visual Function Scale – RB_rs=0–1, CMB_rs=0–4; for the Motor Function Scale – RB_rs=0–2, CMB_rs=0–3; for the Oromotor Function Scale – RB_rs=0–2; CMB_rs=0–1; and for the Communication Scale – CMB_rs=0–1.
According to CRS-R conceptualization, the arousal item score was not divided into RB and CMB as it is an item independent of the diagnosis of VS or MCS. However, reflecting the important role of arousal level during clinical evaluation, an arousal score (AS) ranging from 0 to 1, was developed (independently from RB and CMB) considering 0.333 points for each response category.
Step 2: calculation of standardized values of the two new indices
The total RB_rs and CMB_rs were transformed considering the following formulae:
For example, an RB_rs total value of 6 and a CMB_rs total value of 2 determine a RS=0.86 and CMB=0.18.
The authors developed a matrix in which all possible RB and CMB values are reported along the vertical and the horizontal axis, respectively. This matrix, called transposition matrix, comprises 96 cells. The authors decided to transform the CRS-R original total score from a range 0–23 to a range 0–100 (transformed score); thus, we considered an equal value of 1.0421 (derived from the formulae: 100 points−1/96 cells−1) for each cell in progressive order from RB=0; CMB=0 to RB=1; CMB=1 as reported in Table 2.
Step 3:calculation of the CRS-R Modified Score
To calculate the final score for each patient [called CRS-R Modified Score (MS)], we considered the value reported in the transposition matrix crossing RB and CMB values and then summing to this value the AS value of each patient (e.g. RB=0.86, CMB=0.18 and a supposed AS value of 0.33 determines a CRS-R MS=22.93+0.33=23.26). The CRS-R MS score ranges from 0 to 100 and scores above the cutoff value of 8.34 indicate that the patient shows at least one behavioral responses in line with a diagnosis of MCS according to Giacino rules (Giacino et al., 2004) and independent of what kind of behavioral response it is. In other words, if one response category marked with ‘*’ is reported in the CRS-R scale, the CRS-R MS will be greater than 8.34.
Continuous variables are presented as mean and SD, and Pearson correlation was calculated between the CRS-R MS and CRS-R scores.
Sixty adult patients with DOC were enrolled in the study and ten patients were excluded from the final analysis because after CRC Besta professionals evaluation, their diagnosis was severe disability. In the final sample, 30 patients were in VS and 20 in MCS. Sociodemographical and clinical information is reported in Table 3.
In terms of the distribution of diagnosis in relation to CRS-R scores, 16 patients (32%) had equal CRS-R total scores, but different diagnoses (eight VS and three MCS with a CRS-R score of 7; three VS and two MCS with a CRS-R score of 8), whereas when using the CRS-R/MS, no overlap was found. As reported in Fig. 1, CRS-R MS between VS and MCS patients are clearly distributed in two separated scatter plot areas (separated from the theoretical value of 8.34, indicated with a continuous vertical gray line), whereas CRS-R scores for VS and MCS patients were overlapping, between scores 7 and 8, as reported in the area comprised of horizontal dotted lines.
A correlation coefficient of 0.91 (P<0.001) was found between CRS-R scores and CRS-R MS.
In three patients, differences between responses in a single item reported by rater 1 (who followed the original CRS-R methodology) and rater 2 (who considered all response categories for each item using the MS) were found. In detail, in two patients, no ‘visual startle’ reflex was found, although they showed more complex cognitive behaviors for the same CRS-R item, and one patient did not show ‘localization to noxious stimulation’, although he was able to perform ‘object manipulation’.
The aim of this study was to develop a procedure to modify CRS-R scores first to avoid CRS-R scores overlapping between different diagnoses (VS and MCS) and, second, to define reflexes and cognitive-mediated behaviors as a continuum from no response to more complex reproducible behavior as conceptualized in CRS-R.
Pilot results showed that the CRS-R MS completely avoids the overlapping of scores in respect to the 32% of patients reported using the standard CRS-R total score. Moreover, for 6% of patients, for whom the original assumption of CRS-R implies scores for lowest behaviors if a patient shows a higher response category in the same item, the CRS-R MS highlighted that patients did not show the lowest response categories for visual or motor function items, although they presented cognitively mediated behaviors.
An accurate diagnosis is indeed crucial both for guiding clinical decision making and for clarifying the prognostic signs and clinical evolution of patients with DOC. The importance of CRS-R for the diagnosis of patients with DOC is shown, and also that a multidisciplinary team needs to integrate repeated administrations of CRS-R with data from structural and functional integrity of the brain and clinical information. The possibility of completely describing behavioral functioning of a patients with DOC, without any theoretical assumption, could be useful for clinicians to achieve the best possible clinical management for each patient. Furthermore, as reported in the literature (Schnakers et al., 2008; Sacco et al., 2011; Estraneo et al., 2014; Tamashiro et al., 2014), CRS-R is a valid scale for clinical and research uses from the sub acute to chronic stages of patients with DOC. Considering that patients in VS and MCS could show long-term recovery of responsiveness, the monitoring of any behavioral responses to stimuli, also those defined ‘noncognitively mediated’ during the long-term care, could be useful for detecting all possible signs of possible functional recovery.
Some limitations should be taken into account in this study. Patients in VS and MCS often show fluctuations in arousal level that make the behavioral assessment difficult and decrease the probability of observing signs of consciousness. The research of lower behavioral responses could prolong the time dedicated to the evaluation of patients by raters and, hence, the probability that patients are tired or show arousal fluctuations could be higher. However, the complete functioning profile of one patient obtained using CRS-R MS is in line with the principle that clinical assessment is useful to describe patients’ features, and not only for making a diagnosis, and to differentiate patients who have the same diagnosis but different characteristics.
Another limitation could be the fact that CRS-R is considered the gold standard for making a diagnosis, but a diagnostic error is always possible (Childs et al., 1993; Andrews et al., 1996; Schnakers et al., 2009) when considering only behavioral responses. Future studies combining data from multiple techniques (such as neuroimaging, electrophysiological, etc.) and CRS-R MS are needed to compare its usefulness to the score of CRS-R or of other scales.
Finally, future studies are needed to analyze the psychometric properties of CRS-R MS to avoid statistical differences with the original version (Estraneo et al., 2014; Gerrard et al., 2014) and to test some possible limitations related to the time needed for clinicians for testing all responses categories in each patient. CRS-R MS is a new procedure to score the CRS-R scale that could be optimal for future studies on patients with DOC.
The authors are grateful to all patients and healthy participants who participated in the study. The Besta Coma Research Centre (CRC) team, on behalf of which the present publication was submitted, would like to acknowledge its other members: A. Andronache, P. Fazio, R. Benti, G. Marotta, D. Caldiroli, F. Molteni, F. Panzica, B. Reggiori, G. Varotto, J. Vela Gomez, and E. Visani for valuable suggestions, and F.Ciaraffa, for support with the sample recruitment. The study was carried out in collaboration with the European Biomedical Research Federation (FERB).
The Coma Research Centre (CRC) project was funded by grant no. IX/000407–05/08/2010 awarded by the Lombardy regional government.
The study was coordinated by the Carlo Besta Neurological Institute IRCCS Foundation (Milan, Italy) between January and December 2011 and approved by the Institutional Review Board/ethical committee (ref. CRC). Written informed consent was obtained by the legally appointed patient guardians. The study was carried out in accordance with the Declaration of Helsinki.
Conflicts of interest
There are no conflicts of interest.
Andrews K, Murphy L, Munday R, Littlewood C (1996). Misdiagnosis of thevegetative state: retrospective study in a rehabilitation unit. BMJ 313:13–16.
Binkofski F, Buccino G, Posse S, Seitz RJ, Rizzolatti G, Freund H (1999). A fronto-parietal circuit for object manipulation in man: evidence from an fMRI-study. Eur J Neurosci 11:3276–3286.
Childs NL, Mercer WN, Childs HW (1993). Accuracy of diagnosis of persistent vegetative state
. Neurology 43:1465–1467.
Cox JJ, Reimann F, Nicholas AK, Thornton G, Roberts E, Springell K, et al. (2006). An SCN9A channelopathy causes congenital inability to experience pain. Nature 444:894–898.
Estraneo A, Moretta P, De Tanti A, Gatta G, Giacino J, Trojano L (2014). An Italian multicentre validation study of the Coma Recovery Scale-Revised. Eur J Phys Rehabil Med [Epub ahead of print].
Fleshman JW, Rudomin P, Burke RE (1988). Supraspinal control of shortlatency cutaneous pathway to hindlimbmotoneurons. Exp Brain Res 669:449–459.
Fomberstein K, Qadri S, Ramani R (2013). Functional MRI and pain. Curr Opin Anaesthesiol 26:588–593.
Gallivan JP, Johnsrude IS, Randall Flanagan J (2015). Planning ahead: object-directed sequential actions decoded from human frontoparietal and occipitotemporal networks. Cereb Cortex pii:bhu302.
Gerrard P, Zafonte R, Giacino JT (2014). Coma Recovery Scale-Revised: evidentiary support for hierarchical grading of level of consciousness. Arch Phys Med Rehabil 95:2335–2341.
Giacino JT (2005). The minimally conscious state: defining the borders of consciousness. Prog Brain Res 150:381–395.
Giacino JT, Zasler ND, Katz DI, Kelly JP, Rosenberg JH, Filley CM (1997). Development of practice guidelines for assessment and management of the vegetative and minimally conscious states. J Head Trauma Rehabil 12:79–89.
Giacino JT, Kalmar K, Whyte J (2004). The JFK Coma Recovery Scale-Revised: measurement characteristics and diagnostic utility. Arch Phys Med Rehabil 85:2020–2029.
La Porta F, Caselli S, Ianes AB, Cameli O, Lino M, Piperno R, et al. (2013). 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.
Lewis JW (2006). Cortical networks related to human use of tools. Neuroscientist 12:211–231.
Løvstad M, Frøslie KF, Giacino JT, Skandsen T, Anke A, Schanke AK (2010). Reliability and diagnostic characteristics of the JFK coma recovery scale-revised: exploring the influence of rater’s level of experience. J Head Trauma Rehabil 25:349–356.
Lundberg A, Voorhoeve P (1962). Effects from the pyramidal tract on spinal reflex arcs. Acta Physiol Scand 56:201–219.
McDonnell E, Giacino JT, Kolakowsky-Hayner SA (2015). A brief overview of the Coma Recovery Scale-revised: updates from the COMBI. J Head Trauma Rehabil 30:143–145.
Melzack R, Casey KLKenshalo DR (1968). Sensory, motivational and central control determinants of pain: a new conceptual model. The Skin Senses. Springfield, IL: Thomas. 423–443.
Mouraux A, Diukova A, Lee MC, Wise RG, Iannetti GD (2011). A multisensory investigation of the functional significance of the ‘pain matrix’. Neuroimage 54:2237–2249.
Oertel BG, Preibisch C, Martin T, Walter C, Gamer M, Deichmann R, Lötsch J (2012). Separating brain processing of pain from that of stimulus intensity. Hum Brain Mapp 33:883–894.
Peddareddygari LR, Oberoi K, Grewal RP (2014). Congenital insensitivity to pain: a case report and review of the literature. Case Rep Neurol Med 2014:141953.
Peyron R, Kupers R, Jehl JL, Garcia-Larrea L, Convers P, Barral FG, Laurent B (2007). Central representation of the RIII flexion reflex associated with overt motor reaction: an fMRI study. Neurophysiol Clin 37:249–259.
Sacco S, Altobelli E, Pistarini C, Cerone D, Cazzulani B, Carolei A (2011). Validation of the Italian version of the Coma Recovery Scale-Revised (CRS-R). Brain Inj 25:488–495.
Sandrini G, Serrao M, Rossi P, Romaniello A, Cruccu G, Willer JC (2005). The lower limb flexion reflex in humans. Prog Neurobiol 77:353–395.
Sandroni P, Martin DP, Bruce BK, Rome JD (2006). Congenital idiopathic inability to perceive pain: a new syndrome of insensitivity to pain and itch with preserved small fibers. Pain 122:210–215.
Schnakers C, Majerus S, Giacino J, Vanhaudenhuyse A, Bruno MA, Boly M, et al. (2008). A French validation study of the Coma Recovery Scale-Revised (CRS-R). Brain Inj 22:786–792.
Schnakers C, Vanhaudenhuyse A, Giacino J, Ventura M, Boly M, Majerus S, et al. (2009). Diagnostic accuracy of the vegetative and minimally conscious state: clinical consensus versus standardized neurobehavioural assessment. BMC Neurol 21:9–35.
Schomburg ED (1990). Spinal sensorimotor systems and their supraspinal control. Neurosci Res 7:265–340.
Schwindt PCTowe AL, Luschei ES (1981). Control of motoneuron output by pathways descending from the brain stem. Motor Coordination, Handbook of Behavioral Physiology. New York: Plenum Press.
Seel RT, Sherer M, Whyte J, Katz DI, Giacino JT, Rosenbaum AM, et al., American Congress of Rehabilitation Medicine, Brain Injury-Interdisciplinary Special Interest Group, Disorders of Consciousness Task Force (2010). Assessment scales for disorders of consciousness: evidence-based recommendations for clinical practice and research. Arch Phys Med Rehabil 91:1795–1813.
Tamashiro M, Rivas ME, Ron M, Salierno F, Dalera M, Olmos L (2014). A Spanish validation of the Coma Recovery Scale-Revised (CRS-R). Brain Inj 28:1744–1747.
The Multi-Society Task Force on PVS (1994). Medical aspects of the persistent vegetative state
(first of two parts). N Engl J Med 330:1499–1508.
Umeda M, Corbin LW, Maluf KS (2013). Preliminary investigation of absent nociceptive flexion reflex responses among more symptomatic women with fibromyalgia syndrome. Rheumatol Int 33:2365–2372.