Recovery and Long Term Functional Outcome in Subjects With Critical Illness Polyneuropathy and Myopathy: a Scoping Review

Background: Intensive care unit acquired weakness (ICUAW), embraces an array of disorders labeled “critical illness polyneuropathy” (CIP), “critical illness myopathy” (CIM) or “critical illness polyneuromyopathy” (CIPNM). Several studies have addressed the various characteristics of CIPNM, but the recovery is still unclear. Objective: The present review investigated the recovery and the long-term functional outcome of subjects with CIPNM, whether the types of CIPNM have different outcomes and whether there is any supporting evidence. Methods: Literature search was performed from MEDLINE/PubMed, CINAHL, EMBASE, PeDro, Web of Science and Scopus. Inclusion criteria were: i) sample size including ve or more subjects; ii) subjects who suffered from CIPNM and/or CIP, CIM and CIP/CIM; iii) CIPNM ascertained by EMG. Follow-ups longer than one year were dened as long-term. Results: Twenty-nine studies met the inclusion criteria. In total, 788 subjects with CIPNM were enrolled: 159 (20.1%) died and 588 (74.6%) were followed. Of all the included patients, 613 (77.7%) had CIP, 82 (10.4%) CIM and 56 (7.1%) CIP/CIM. Overall, 70.3% of the subjects with CIPNM fully recovered. Seven (24.1%) studies had a follow-up longer than 1 year (range 2-8) with 173 (21.9%) subjects enrolled globally and 108 followed. Of these subjects, 88.8% gained full recovery. Most of the studies did not use proper functional scales and only 4 and 3 studies employed the Barthel scale and the FIM scale. Differentiation between the types of CIPNM was performed in 7 studies, but only 3 studies reported that subjects with CIM had a better prognosis and earlier recovery than subjects with CIP/CIM. Conclusions: Subjects with CIPNM could achieve good recovery and could further improve at follow-up. The quality of the published studies due to short follow-ups and the paucity of dened outcome measures limit the evidence. patients having been discharged with the ICD-9/ICD-10 code for CIP and CIM (53-55) and did not use functional scale scores [52–53]. The included studies varied in aim, methodology design, sample size, case mix, inclusion/exclusion criteria, timing of the examination, follow-up and denition of recovery. Seventeen (58.6%) studies had a prospective design. Twenty-four (82.7%) studies concerned case series or small cohorts that had a mean sample size of 19.1±7.7 patients and did not exceed the total number of 30 subjects. The other 5 studies [35, 41, 44- 45, 48] had samples greater than 30 patients (range 36-119). A total of 788 subjects with CIPNM were enrolled; of these 159 (20.1%) died and 588 (74.6%) were followed. All studies except 7 [6, 22–23, 35, 41, 44, 46] did not perform the differentiation between the types of CIPNM and considered the disorder as a unique entity, labelling it as CIP or, more generically, as polyneuropathy or neuromuscular disorder. In particular, 613 (77.7%) patients had CIP, 82 (10.4%) CIM and 56 (7.1%) CIP/CIM. Diagnoses requiring ICU admission were widely and due to variable medical and surgical disorders. Two studies investigated the functional outcome in patients with CIPNM and coexistent brain lesions [45, 48]. Twenty (68.9%) studies were performed on subjects during their ICU or post-ICU stay, 1 in neurology and 8 (27.5%) in rehabilitation or in neuro-rehabilitative settings.


Introduction
A number of studies have been published about the muscle weakness that may affect intensive care unit (ICU) survivors. This disorder, which in the intensive care literature is better known as ICU acquired weakness (ICUAW), embraces a spectrum of clinical conditions. All of these conditions show variable levels of muscle strength impairment, from weakness to paralysis, which involve, bilaterally, the upper and lower limbs of ICU subjects. Although the clinical pictures are generally indistinguishable, ICUAW encompasses different pathological forms that damage the muscular and the peripheral nervous system. Different types of this pathology have been described and labeled according to the histological aspects and the electrophysiological ndings and depending on the predominant structure involved. In particular, the de nitions include: i) critical illness polyneuropathy (CIP), if the peripheral nervous system is affected; ii) critical illness myopathy (CIM), if the muscles are predominantly involved, and iii) critical illness polyneuropathy and myopathy (CIP/CIM), critical illness neuromyopathy [1], and polyneuromyopathy (CIPNM) [2][3], if the pathological process affects both muscles and nerves. Although ICU specialists prefer to use the term ICUAW, the de nition "critical illness polyneuromyopathy" (CIPNM) is widely diffuse, and better suits the purposes of the present study. After the rst description by Bolton et al. at the beginning of the '80s [4], a number of studies have been published that have contributed to making remarkable advances in the understanding of the complex aspects of CIPNM, such as the electrophysiological [5] and histological features [6] as well as the pathogenic mechanisms. The occurrence of this disorder has been variously detected with a range from 45 to 80% [7][8][9]. A systematic review reported a median prevalence of 43% (10). Furthermore, many risk factors have been suggested to favor the development of CIPNM, including sex (female), sepsis, ICU length of stay and multiple organ failure [11][12]. Several therapeutic approaches and strategies have been proposed and evaluated for the management of CIPNM subjects, but pharmacological treatments have failed to prevent the occurrence and were ineffective in treating the disorder [13][14][15]. However, recent reviews and meta-analyses have demonstrated that early mobilization is associated with a lower likelihood of developing this clinical condition [16][17]. Some reviews have addressed the various aspects of CIPNM, but they have not highlighted recovery, functional outcome and quality of life [18][19][20]. Despite the lack of treatments and the limitations of rehabilitative strategies, it was reported that 55-70% of subjects reached a full recovery after ICU discharge [21][22], and that recovery depended on the type of CIPNM, as con rmed by the fact that CIM had an earlier and better functional outcome than CIP [22][23]. In 2005, an extensive literature review of the neuromuscular sequelae of ICU subjects with critical illness reported that 68.8% of them made a complete recovery and regained the ability to walk independently. However, such review was limited by an insu cient number of patients with a long follow-up, leaving unsettled the question of whether de cits following CIPNM were persistent [21]. Since then, no extensive study investigating the functional outcome in these subjects have been carried out, and the issue remains unsolved. The aim of the present review was to investigate recovery, in particular the longterm functional outcome of subjects with CIPNM, whether the types of CIPNM have different outcomes and whether there is any supporting evidence.

Materials And Methods
A search of the studies having tested the functional outcome in subjects with CIPNM was conducted using MEDLINE/PubMed, the Cochrane Central Register of Controlled Trials, CINAHL, EMBASE, PeDro, Web of Science and Scopus databases. The search was restricted to English language reports published between January 1984 and April 2021. The search terms varied slightly from database to database but included "intensive care unit acquired weakness", "ICUAW", "critical illness polyneuropathy", "CIP", "critical illness myopathy", "CIM", "critical polyneuropathy and myopathy", "CIPNM", "CIP and CIM"; "CIP/CIM", "acute tetraplegia", "rehabilitation", "functional outcome", "recovery", "physical therapy" and "mobilization". Search limits included only adults. Conference abstracts/posters or articles that were not peer-reviewed were excluded. The literature search was conducted by three independent authors (MC, MB, FDR). Inclusion criteria were: i) sample size including ve or more subjects; ii) subjects who suffered from CIPNM and/or the following types: CIP, CIM and CIP/CIM; iii) CIPNM ascertained by EMG; iv) studies with mixed samples that used the de nition of ICUAW, but subjects with CIP, CIM or CIP/CIM were also considered; v) follow-up and outcome.
In order to avoid confounding results, studies were excluded if: i) they contained only the de nition of the ICUAW, without any reference to the types of CIPNM or to CIP, CIM or CIP/CIM; ii) ICUAW was not ascertained by EMG; iii) reviews concerned ICUAW/CIPNM but the main aim of the study was not the outcome. Studies concerning subjects with CIPNM and COVID-19 were also excluded.
We de ned as long-term follow-ups those follow-ups longer than 1 year. In this review, the pathological condition was counted as CIP if this acronym or de nition was not speci ed in the studies analyzed. Due to the variability of the study designs, the functional measurements, the follow-ups, and the lack of data on the score of measurements, quantitative analysis was not possible. The research was conducted according to the Preferred Reporting Items for Systematic Reviews and the Meta-Analyses (PRISMA) diagram, depicting the selection of the articles searched for the study.

Results
The Prisma diagram of the studies' selection is shown in Figure 1. After studies were searched for and collected, 36 of them were considered eligible; of these, 29 [2,6,22, were included according to inclusion and exclusion criteria (Table 1). Seven studies [49][50][51][52][53][54][55] were excluded even if they included subjects with CIPNM (Appendix 1). In particular, 4 of these 7 studies were excluded because they contained a duplicate of the data already included in the 29 studies, where they were analyzed in a greater sample [48-51]; 1 study with a large sample, due to the diagnosis of CIPNM being based predominantly on the clinical examination [52]; 2 studies including ICU patients having been discharged with the ICD-9/ICD-10 code for CIP and CIM (53)(54)(55) and did not use functional scale scores [52][53]. The included studies varied in aim, methodology design, sample size, case mix, inclusion/exclusion criteria, timing of the examination, follow-up and de nition of recovery. Seventeen (58.6%) studies had a prospective design. Twenty-four (82.7%) studies concerned case series or small cohorts that had a mean sample size of 19.1±7.7 patients and did not exceed the total number of 30 subjects. The other 5 studies [35,41,[44][45]48] had samples greater than 30 patients (range 36-119). A total of 788 subjects with CIPNM were enrolled; of these 159 (20.1%) died and 588 (74.6%) were followed. All studies except 7 [6, 22-23, 35, 41, 44, 46] did not perform the differentiation between the types of CIPNM and considered the disorder as a unique entity, labelling it as CIP or, more generically, as polyneuropathy or neuromuscular disorder. In particular, 613 (77.7%) patients had CIP, 82 (10.4%) CIM and 56 (7.1%) CIP/CIM. Diagnoses requiring ICU admission were widely and due to variable medical and surgical disorders. Two studies investigated the functional outcome in patients with CIPNM and coexistent brain lesions [45,48]. Twenty (68.9%) studies were performed on subjects during their ICU or post-ICU stay, 1 in neurology and 8 (27.5%) in rehabilitation or in neuro-rehabilitative settings. Table 1 Studies about functional outcome in CIPNM subjects. The pathological condition was counted as CIP if differentiation between types of CIPNM was not performed and if this acronym or de nition was not speci ed in the studies analyzed.       Legend         number of patients who had polyneuropathy to EMG; £ the sample included children and CIP was not de ned; *including 12 patients described in a precedent paper; ** other forms of myopathy or motor axonopathy could not be excluded; $ description is reported in appendix 1.
Fourteen (48.2%) studies did not have patient's global ability as primary end-point and did not use functional scales. Recovery was evaluated from a clinical point of view by neurological examination or on the basis of the improvement in muscle strength by the MCR scale. This measure was used in 11 studies [22-23, 28, 30, 36, 38, 40, 42-44, 47]. The severity of muscle weakness was not correlated with the clinical and electrophysiological diagnosis, and there was no correlation between the degree of the nerve conduction and the clinical ndings [22,37].

Long-term outcomes
Seven studies (24.1%) had a follow-up longer than 1 year, ranging from 2 to 8 years (Table 2) for a total of 173 (21.9%) subjects, of whom 37 (21.3%) died and 108 were followed. Overall 124, 43 and 6 subjects had CIP, CIM and CIP/CIM, respectively. Of followed subjects with CIPNM 96 (88.8%) achieved a full recovery. Almost all studies included small samples characterized by case series including 7 to 22 patients, and only 2 studies had greater samples consisting of 49 [35] and 42 subjects [41], respectively. Furthermore, the investigation of long-term functional outcomes in subjects with CIPNM as main purpose was addressed only in 1 of these 7 studies [41]. This study had a mixed sample and included 42 patients with different etiology of ICU admission and showed that 73.8% of the entire sample of patients made a good recovery, as shown by the functional measures at follow-up: mean Barthel and median mRS score of 86.7±15.9 and 1 (IQR: 0-5), respectively. The differentiation between the types of CIPNM was performed only in 2 studies [35,41]. Both studies showed that subjects with CIP/CIM had the worst outcome. Functional measures were employed only in 2 studies that used the Barthel scale [39,41] and the mRS [41]. Both studies had a long follow-up of 3.5 and 5 years, respectively, and a good recovery was detected in a high percentage of subjects. Five (71.4%) studies did not use proper scales, and the recovery was evaluated by motor improvement, such as the ability to walk without support or aid.

CIPNM type and outcome
The differentiation between the types of CIPNM was performed in 7 (25%) studies [6, 22-23, 35, 41, 44, 46]. Among these, 3 investigations concerning the long-term functional outcome reported that the CIM type had a better prognosis and an earlier recovery than CIP/CIM [22][23]41].  (Table 1). Five subjects with CIM recovered within 6 months, whereas the presence of CIP alone or in addition to CIM was associated with a more delayed recovery between 6 and 12 months, and more than 50% of those individuals had persistent de cits at 1 year of followup [22]. Likewise, the study by Intiso et al. reported that CIM patients had a better recovery than subjects with CIP or CIP/CIM and did not show differences in their health status compared to the Italian normative data. Of the remaining 4 studies, 3 did not report data on the recovery of CIPNM types, since CIPNM subjects were evaluated in comparison to controls independently of the type of CIPNM [6, 44,46], and the last one had severe limitations despite the differentiation between the types of CIPNM, since other forms of myopathy or motor axonopathy could not be excluded [35].

Discussion
The present review detected that 70.3% of a large number subjects with CIPNM could achieve a full recovery. This nding is similar to that obtained from a previous review by Latronico et al., who reported that 68.8% of patients made a complete recovery [21]. Furthermore, a higher percentage of 88.8% gained full recovery at long term follow-up. However, because of strong limitations of studies that had small samples and were widely variable in aim, methodology design, case mix, and outcome measures, the nding should be considered with caution. The investigation of the functional outcome was the main purpose in 31% of the studies analyzed; in addition, a variety of measures was employed, and only 7 (24.1%) studies used a proper functional scale such as the Barthel scale and the FIM.
Of great importance, immediately after the rehabilitation treatment, is the overall health status of the patients, culminating in a return to active daily living, socialization and participation. The ICF recommends a new approach to evaluate disabled people, which is based on a holistic model in which activities and participation represent essential aspects. In this respect, only 2 studies investigated the outcome according to the ICF, but because of the small size of the samples and the limited duration of the follow-ups, it was not possible to draw de nitive conclusions [40,42]. Most studies (71.4%) evaluated recovery on the basis of the motor improvement and the ability to walk without support, but no data was reported on the functionality in activity daily living. Furthermore, though the present review retrieved a larger number of 788 patients with CIPNM, those who had long term follow-up  [45]. The authors reported that patients with sABI and CIPNM achieved a good recovery, but the magnitude of these improvements was better in the subjects with sABI alone. However, it is still not clear if the residual disability in these subjects is due predominantly to CIPNM, to CNS damage or to both disorders. Likewise, the effect that each disease might have on the course of the disability still needs to be clari ed. People with both disorders might require different rehabilitation approaches and strategies.
Outcome of the types of CIPNM It has been suggested that subjects with CIM have a better prognosis, reaching early and full recovery, than subjects with CIP or CIP/CIM, but this nding was reported only in 3 studies including a total of 20 patients [22][23]41]. Therefore, it is not possible to achieve a de nitive conclusion about this matter due to the very small number of investigated subjects. Several di culties hamper and make it hard to make real and objective comments on this issue. Some of these di culties are the lack of unique and shared de nitions of muscle weakness that may affect ICU subjects as well as an insu cient differentiation between the types of CIPNM. Today, a wide range of de nitions are still being used including ICUAW, CIP, CIM or CIP/CIM, and this aspect can complicate the analysis, and the results of the studies may be consequently biased. Most investigations considered this disorder as a single entity, therefore it is not possible to exclude that different forms of CIPNM were present in the sample of the studies investigating only subjects with CIP. Furthermore, although EMG is able to differentiate between the subtypes of CIPNM, it does not allow the quanti cation of the muscle impairment and the related disabilities, and no electrophysiological exams or imaging have helped to solve this point.

Limitations
The present study has limitations that must be acknowledged. This is a review of the literature having the aim to discuss the recovery and the longterm functional outcome of CIPNM subjects. Cohort studies, case series as well as functional measurements and follow-up were highly heterogeneous regarding the functional outcome. A further limitation concerns the recovery of subjects with CIPNM; such limitation is due to the paucity of trials focusing on rehabilitative interventions [41][42]59]. Even if this issue is beyond the scope of the present review, the majority of the studies analyzed did not de ne whether the subjects followed any kind of rehabilitation treatment after hospital discharge. Currently, apart from the early neuromuscular electrical stimulation that might prevent ICUAW and improve the quality of life by enhancing muscle strength in ICU patients [59], no de nitive studies have evaluated the effects of rehabilitation programs in inpatient or outpatient settings in this population [60]. Therefore, several questions remain unanswered and further research should be carried out on this matter.

Suggestions and implications for the future
Given the protean aspects of CIPNM, a closer collaboration as well as a more active participation of multiple specialists and experts has been suggested [61]. In particular, ICU specialists, neurologists and physiatrists should collaborate more to properly evaluate and follow these subjects.
Specialists who manage CIPNM patients should adopt unique and shared terminology and de nitions, and future studies should be planned considering the following aspects: the aim should be focused on functional recovery; the methodology design should include a large sample of patients, proper functional measures and de ned long-term follow-up; differentiation between the types of CIPNM; rehabilitation interventions and their effect on functional outcome and quality of life, given that few studies have evaluated the effects of rehabilitation programs in this population.; Occurrence and recovery of CIPNM in subjects with CNS damage should be investigated through dedicated studies.

Conclusion
A percentage of 70.3 % of survivor subjects with CIPNM could achieve a full recovery and a higher percentage was detected at long term follow-up.
However, the quality of the published studies due to short follow-ups, and the absence of clearly de ned outcome measures did not allow de nitive