1. Introduction
In recent decades, important advances have been made in emergency medicine and neurosurgical procedures, leading to improved survival of victims of severe acquired brain injuries. After a comatose state, patients may progress to a clinical condition of disorders of consciousness (DoC), which includes unresponsive wakefulness syndrome (UWS), minimally conscious state (MCS), and emergent from the MCS (E-MCS) [
1]. These diagnostic categories are defined by the presence and nature (reflex in UWS vs. intentional in MCS) of behavioral responses to multisensorial stimuli. MCS patients can in turn be subcategorized into two distinct entities: on one hand, MCS minus (MCS−), patients showing low-level purposeful behaviors (e.g., visual pursuit, object localization (reaching), and automatic motor response), and MCS plus (MCS+), patients showing a movement to command.
An accurate diagnosis of the consciousness state is crucial since it can influence several aspects of the care pathway, such as treatment and end-of-life decisions, the planning of the rehabilitation path, and the prognosis to be communicated to the patient’s relatives [
2]. The Coma Recovery Scale-Revised (CRS-R) [
3,
4] is the gold standard of consciousness assessment in patients with DoCs [
5]. It consists of 23 hierarchically organized items parcellated into six sub-scales assessing different functions with various numbers of hierarchically arranged items. These sub-scales include auditory, visual, motor, oromotor/verbal functions, communication, and arousal scales. The clinical diagnosis of consciousness is based on the presence of the highest level of consciousness assigned in the evaluated sub-scales [
3]. Therefore, for each sub-scale, a specific threshold value allows the consciousness stratification into UWS, MCS, or E-MCS [
3]. Psychometric studies have widely demonstrated a strong inter-rater and test-retest reliability [
3,
4] of the CRS-R. However, the scale is subject to inaccuracy, attributable to examiner error and other confounding factors, including comorbidities [
6], severe spasticity [
7], diffuse pain [
8], neuropsychological deficits [
9], psychomotor agitation [
10], inertia/akinetic mutism [
11], iatrogenic effects, presence of tracheostomy [
12], and caregiver exclusion during the consciousness evaluation [
13], which can lead to misinterpretation of results. In addition, the accuracy of the CRS-R assessment may be affected by extreme motor deficits as reported in previous studies [
7].
Critical illness polyneuropathy and myopathy (CIPNM) is the most frequent intensive care unit acquired weakness defined as “clinically detected weakness in critically ill patients in whom there is no plausible etiology other than critical illness” [
14]. It often occurs during critical disease [
15] and is associated with an increase in intensive care unit morbidity and up to 1 year mortality [
16,
17]. In the case of cooperative patients, a Medical Research Council scale score < 48 is a clinical criterion for diagnosis of CIPNM [
18]. However, the development of clinical diagnostic criteria for CIPNM is hampered by the low reliability of the clinical evaluation of patients with a lack of cooperation, such as patients with DoCs. Therefore, in these conditions, the diagnosis is principally based on the neurophysiological evaluation and electro-diagnostic studies of the peroneal and sural nerves. Among those, the amplitude reduction of the compound muscle action potential of the peroneal nerve has been recognized as the most sensitive and specific neurophysiological parameter [
19]. Other electrophysiological tools have been proposed, such as direct muscular stimulation, but a diagnosis of CIPNM subtype can be performed only by biopsy.
In our previous work [
20], CIPNM was found in about 50% of patients with severe acquired brain injuries at admission in the intensive rehabilitation unit (IRU). All enrolled patients with severe acquired brain injuries significantly improved during their IRU stay in terms of consciousness state and functional and swallowing abilities, but those with CIPNM achieved lower outcomes. In this study, we aimed to investigate the prevalence of CIPNM only in patients with severe acquired brain injuries admitted with DoC in the IRU and the possible relationship of this peripheral disease with the consciousness state and its evolution between admission and discharge, assessed by the CRS-R.
4. Results
During the study period, 181 patients with DoCs entered the IRU of the IRCCS Don Gnocchi Foundation (Florence, Italy); 4 (2%) were excluded due to polyneuropathy other than CIPNM (
Figure 1).
A total of 177 patients (UWS: 81 (45.8%); MCS: 96 (54.2%); women: 78 (44.1%)) were included in the analysis and presented the following characteristics at admission: median age: 67 years (IQR: 55–75), etiology: traumatic (
n = 44) 24.9%, anoxic (
n = 25) 14.1%, ischemic (
n = 25) 14.1%, hemorrhagic (
n = 70) 39.6%, and others 7.3%: (
n = 13: metabolic
n = 2; tumoral
n = 8; infective
n = 3), CRS-R median score: 9 (IQR: 5–14); median time post-onset: 43 days (IQR: 32–62). The median length of stay (LOS) was 105 days (IQR: 64–159). One hundred and eight (61.0%) patients were diagnosed with CIPNM at admission. The results of the total sample are summarized in
Table 1.
Among the 177 included patients, 117 (66.1%) presented an IR at discharge. When comparing demographics and clinical features of patients who improved their IR at discharge versus those who did not improve, a significantly higher median CRS-R score at admission (11 vs. 5,
p < 0.001), a higher percentage of MCS diagnosis at admission (64.1% vs. 35.0%,
p < 0.001), and a lower percentage of CIPNM diagnosis (54.7% vs. 73.3%,
p = 0.016) were found in patients with IR (
Table 1).
We then performed a multivariate regression analysis on the whole sample, introducing all variables that were significantly different between IR and No-IR patients in the univariate analysis, except for the clinical status at admission, which was not included in the multivariate logistic regression because it was highly correlated with the CRS-R total score at admission (ρ = 0.874,
p < 0.001). The results of the multivariate analysis on the whole sample showed that only a higher CRS-R total score at admission remained associated with a higher probability of IR at discharge (OR 1.169,
p < 0.001) (
Table 2).
Patients were then divided into two groups (UWS and MCS) to account for the influence of consciousness state at admission. The UWS and MCS patients at admission presented significantly different CRS-R scores (
p < 0.001,
Table 3). The percentage of patients with CIPNM diagnosis was higher in the UWS group compared to that in the MCS group (68% vs. 55%), although the difference was not statistically significant (
p = 0.085). A higher CRS-R score at admission (
p = 0.004), a shorter time post-onset (
p = 0.035), and a lower frequency of CIPNM diagnosis (
p = 0.009) were observed in IR compared to No-IR patients, but only in the UWS group (
Table 3), while none of the included features were significantly different between the IR and No-IR patients in the MCS group.
In the UWS group, when variables significantly different between IR and No-IR patients were included in a logistic regression analysis along with age and sex, a higher CRS-R score at admission resulted and was significantly associated with IR at discharge (OR = 1.462,
p = 0.006;
Table 4), while the presence of CIPNM was found to reduce the likelihood of IR (OR = 0.286,
p = 0.039;
Table 4).
5. Discussion
In the present study, the prevalence of CIPNM was 61% in patients with DoCs at their admission to the rehabilitative setting. To the best of our knowledge, except for a case study performed on 20 patients [
26], this is the first study assessing the prevalence of CIPNM in subjects with DoCs.
This condition continues to be widely underestimated in rehabilitative settings, probably because of the rarity of ENG/EMG examination in both acute and post-acute phases. Consequently, its possible interference both with functional outcomes and the clinical consciousness diagnosis is understated. Our results strongly confirm the recommendation that a systematic neurophysiologic assessment be performed in all patients with DoC, allowing the diagnosis of CIPNM, to improve prognostic assessment and individual rehabilitation project personalization.
We also found that patients with UWS had a lower probability of showing an IR when they suffered from CIPNM. This finding is somewhat puzzling since the relationship between a peripheral disease such as CIPNM and the recovery of consciousness is not immediate. It can be hypothesized that a CIPNM diagnosis is a marker of a higher clinical severity that hinders consciousness recovery independent from the CRS-R score at admission [
20]. t was shown that critical illness polyneuropathy, myopathy, or both most commonly develop after acute respiratory distress syndrome, sepsis, systemic inflammatory response syndrome, or multiple organ failure, and that prolonged bed rest, medication, and infections are major risk factors for CIPNM; other major risk factors include long duration of organ dysfunction, parenteral nutrition, vasopressor and catecholamine support, and central neurologic failure (e.g., septic encephalopathy) [
27], all of which may affect the probability of recovery of consciousness [
28]. However, due to the motor impairment that it causes, the presence of CIPNM might hinder intentional motor responses during the consciousness evaluation, especially in patients with UWS. A previous study by Jorh et al. suggested that the recognition of overt consciousness recovery might be underestimated when severe damage to the motor system that affects motor planning and efferent motor pathways is concomitant, preventing the patient from partially or totally displaying any voluntary responses [
29].
Detecting subtle signs of consciousness may have important prognostic, therapeutic, and ethical implications. Several studies have shown that the functional prognosis is affected by the consciousness diagnosis [
30,
31,
32]. Consequently, a correct diagnosis of consciousness might influence pharmacological treatment, decision-making, the design of rehabilitation programs, and family counseling [
33]. Knowledge of those confounding factors and clinical situations that may lead to an underestimation of consciousness should be improved, especially when the consciousness diagnosis is exclusively based on clinical observation.
Although the CRS-R has strong evidence of reliability and validity for the assessment of patients with DoC, several studies have highlighted its limitations in some clinical situations [
34], identifying some possible confounding factors. To improve CRS-R scoring interpretation, some red flags have been identified to recognize impossible or improbable combinations of specific CRS-R sub-scales that should trigger additional data-quality review to exclude a misdiagnosis of consciousness [
35]. Among the relevant confounding factors, quadriplegia was found to generate a warning in the CRS-R scoring due to the combination of zero on the motor sub-scale with the maximum score on the visual or the communication sub-scales [
34]. However, since the maximum scoring of the visual or the communication sub-scores is enough to formulate a diagnosis of E-MCS, the presence of quadriplegia does not generate a misdiagnosis of consciousness in patients with a higher level of consciousness. By contrast, it is much more complex to recognize clinical signs of transition from UWS to MCS− [
3]. It was shown that five items, including three motor items of the CRS-R assessment, consent to detect 99% of the patients passing from UWS to MCS− [
36]. In short, for the nature of the CRS-R, the presence of motor impairment has a higher probability of inducing a consciousness misdiagnosis in the lower states of consciousness (UWS). Therefore, it would be desirable to also include ENG/EMG among the screening examinations in the early rehabilitation phase.
Finally, the use of a multimodal evaluation by combining clinical evaluation and instrumental tools (electroencephalography (EEG), somatosensory-evoked potentials, or functional neuroimaging) is now recommended to improve clinical classification and prognostication of people with DoC [
37,
38]. Neurophysiological tests, EEGs in particular [
39], have proven to be valuable tools to be used alongside the clinical scale to reduce the risk of misdiagnosis in patients with DoC [
40]; the presence of CIPNM should be a further reason to recommend such assessments. Furthermore, with a diagnosis of CIPNM, the application of complementary diagnostic tools, such as the motor behavior tool, as proposed for the motor/cognitive dissociation [
29], but also the practice of focusing CRS-R motor assessment on facial muscles may help to reduce misdiagnosis.
This study has some limitations that warrant discussion. First, as with all retrospective analyses, we could not control the training background and level of experience of the examiners. However, the agreement between trained investigators who perform the CRS-R evaluation in our center was proven high in an earlier prospective study (kappa coefficient for total scores: 0.827) [
30]. Second, data were collected in a single center; therefore, generalization of the obtained results should be made with caution. Finally, the absence of additional information for consciousness diagnosis, either neurophysiological or clinical, prevented us from better elucidating the possible role of severe CIPNM in producing a cognitive motor dissociation and, by this, possibly inducing a misclassification of the level of consciousness assessed by the CRS-R, particularly in patients with UWS. However, the results obtained in this study highlight the need for a deeper and more prospective analysis of this crucial issue; prospective studies combining clinical and instrumental diagnosis of consciousness are required to investigate whether and how CIPNM might impact the consciousness diagnosis in patients with DoC.