Musculoskeletal complications following critical illness: A scoping review

Abstract

Purpose

To explore the extent to which musculoskeletal (MSK) complications have been reported following critical illness, identifying evidence gaps and providing recommendations for future research.

Materials and methods

We searched five databases from January 1st 2000 to March 31st 2021. We included published original research reporting MSK complications in patients discharged from hospital following an admission to an intensive care unit (ICU). Two reviewers independently screened English language articles for eligibility. Data extracted included the MSK area of investigation and MSK outcome measures. The overall quality of study was evaluated against standardised reporting guidelines.

Results

4512 titles were screened, and 32 met the inclusion criteria. Only one study included was interventional, with the majority being prospective cohort studies (n = 22). MSK complications identified included: muscle weakness or atrophy, chronic pain, neuromuscular dysfunction, peripheral joint impairment and fracture risk. The quality of the overall reporting in the studies was deemed adequate.

Conclusions

We identified a heterogenous body of literature reporting a high prevalence of a variety of MSK complications extending beyond muscle weakness, therefore future investigation should include evaluations of more than one MSK area. Further investigation of MSK complications could inform the development of future post critical illness rehabilitation programs.

Highlights

• •Musculoskeletal complications are a potential source of physical impairment following critical illness.
• •The shoulder is the most commonly reported location of post-ICU pain.
• •Association of musculoskeletal complications with function is unclear due to poor reporting.
• •Future research should ensure assessments of upper limb and lower limb are included.

1. Introduction

Survivors of critical illness frequently experience long-term physical impairment, persistent exercise limitation and decreased health related quality of life (QoL) [1,2]. The subsequent socioeconomic burden of critical illness is also high [3], with patients reporting significant healthcare utilisation and high rates of hospital re-admission within the first year after discharge [4]. Rates of return to employment following admission to an intensive care unit (ICU) are extremely low, with up to 31% of patients not returning to work within five years of admission to ICU [5]. Long term poor physical function following critical illness may be partially due to muscle weakness which occurs rapidly in ICU [6] and is associated with worse health related QoL [7] and five year survival [8]. However, there has been limited investigation into other potential physical complications that may be underlying poor patient reported physical function following critical illness.

Musculoskeletal (MSK) conditions are wide ranging, include problems affecting bone, muscle and joint, and are the leading cause of pain and disability in the UK [9]. ICU survivors may be at increased risk of developing long-term MSK complications given the rates of muscle mass loss seen in ICU [6]. Recent studies evaluating post-ICU rehabilitation interventions aimed at increasing general strength and exercise tolerance have provided mixed results [[10], [11], [12]]. This may be, in part, due to a lack of identification of specific MSK complications patients are presenting with, and therefore a lack of targeted intervention. Identification of the MSK complications present following critical illness, may help to develop future interventions aimed at improving physical function in ICU survivors. However, the types and severity of MSK complications encountered by ICU survivors are unclear, as are the outcomes used to assess them. Therefore the purpose of this review is to examine the MSK outcomes reported following critical illness, and the characteristics of the studies that include them. Our research question was: What is the extent of the original research reporting MSK outcomes following discharge from hospital after critical illness, and how are these outcomes reported and measured?

2. Material and methods

This scoping review was conducted according to the methodology described by Arksey and O'Malley [13] and Levac et al. [14], and reported using the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) [15]. The review protocol and PRISMA-ScR checklist are available as Supplemental Digital Content (S1,2). We consulted a health research librarian, identified relevant databases and subsequently developed a search strategy (Supplemental Digital Content S3). To ensure studies were a representation of current ICU patient population, we searched the following databases from January 1st 2000 until March 31st 2021: OVID Medline, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Allied and Complementary Medicine Database (AMED), Exerpta Medica database (EMBASE), Physiotherapy Evidence Database (PEDro). Authors' personal files were also searched, and reference lists of studies deemed eligible for inclusion were scanned for other studies of relevance.

2.1. Inclusion/exclusion criteria

We included: Population – adult patients discharged from hospital following an admission to ICU; Concept – the assessment and reporting of MSK outcomes (e.g. muscle weakness, peripheral joint range of movement (ROM), chronic MSK pain, neuromuscular function); Context – Prospective and retrospective original research.

In line with our research question, we excluded studies which only reported on MSK outcomes prior to hospital discharge. We also excluded studies that evaluated handgrip strength alone without additional evaluation of peripheral muscle strength, as it would not indicate the presence of an MSK complication. Case studies, case series and abstracts from conference proceedings were excluded, For pragmatic reasons, we did not include non-English language studies.

One reviewer (OG) assessed all titles and abstracts, removing obvious exclusions. The full text citations selected for potential inclusion were assessed independently by two reviewers (OG, MR) with a third reviewer (MW) available where there was disagreement.

2.2. Data collection and analysis

One reviewer (OG) reviewed all publications and extracted data, prior to a second independent review (MR) and data extraction to ensure accuracy. We extracted data on study characteristics (study design, sample size and duration of follow-up) and MSK assessment (area of MSK evaluation, outcome measure, main findings). We undertook a narrative synthesis [16] to describe our findings of the characteristics and results of the studies. The overall quality of study reporting was independently assessed by two reviewers (OG, MR) using the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) or Consolidated Standards of Reporting Trials (CONSORT) guidelines [17,18]. The completeness of reporting for each study was calculated as the proportion of the reported items from the relevant guideline, divided by the total items included in the guideline minus items not applicable to the study. We classified the reporting as: ≥70% adequate, 51–69% moderate, and ≤ 50% poor [19].

3. Results

We identified 4512 potentially eligible citations with 32 quantitative and no qualitative studies meeting the inclusion criteria, as described in our PRISMA study flow diagram [20] (Fig. 1). The most common reasons for exclusion at full text were not including a MSK assessment and an inpatient assessment only. Table 1 summarises the characteristics of the studies which included 31 observational and 1 interventional. There were 11 studies which were part of larger longitudinal or interventional studies [[21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31]], and 13 multicentre studies [7,21,23,[26], [27], [28], [29], [30], [31], [32], [33], [34], [35]]. The majority of studies included general ICU patients, however some studies investigated specific ICU patient populations including: Acute Respiratory Distress Syndrome (ARDS) [7,23,28,34,35], sepsis [21,36], blunt chest trauma [37] and severe acute respiratory syndrome (SARS) [25]. Almost all studies were published after 2010 (n = 29), and there was wide variation in both the duration of follow-up (Median 12 months, IQR 6–30) and the sample size (Median 87, IQR 50–167). Most of the studies were conducted in Europe (n = 17) or North America (n = 10).

Fig. 1.

PRISMA flow diagram.

Table 1.

Study characteristics.

Author, year, country	Musculoskeletal area of evaluation	Study design	Sample size at (max) follow-up	STROBE/CONSORT reporting score	Musculoskeletal outcome measure
Chan et al. [34] 2018 USA	Muscle mass and strength	Multicentre Prospective cohort	n = 91 12 months	77.4%	DXA; MRC SS; handgrip strength,
dos Santos et al. [29] 2016 Canada	Muscle strength, mass and function	Multicentre Nested prospective cohort	n = 11 6 months	54.8%	MRC SS; isometric peak torque (quadriceps); CSA quadriceps; muscle specific force; NCS; EMG.
Nedegaard et al. [22] 2021 Scandinavia	Muscle strength, mass and function	Single centre Nested randomised clinical trial	n = 116 3 months	93.8%	Handgrip MVC; rectus femoris CSA; knee-extension MVC, RFD and endurance at 25% of MVC.
Fan et al. [7] 2014 USA	Muscle strength	Multicentre Longitudinal prospective cohort	n = 127 24 months	83.3%	MRC SS; hand grip dynamometry
Borges et al. [36] 2015 Brazil	Muscle strength	Single centre Prospective cohort	n = 51 3 months	63.6%	Maximal isometric voluntary contraction (quadriceps); hand grip dynamometer;
Pfoh et al. [35] 2016 USA	Muscle strength	Multicentre Longitudinal prospective cohort	n = 115 5 years	90.3%	MRC SS
Duarte et al. [38] 2017 Brazil	Muscle strength	Single centre Retrospective cohort	n = 591 3 months	56.6%	MRCSS
Dinglas et al. [32] 2018 USA	Muscle strength	Multicentre Longitudinal prospective cohort	n = 103 4 years	81.8%	MRCSS
Van Aerde et al. [27] 2020 Belgium	Muscle strength	Multicentre Sub-analysis of prospective cohort	n = 205 5 years	84.8%	Hand grip dynamometry; Hand held dynamometry for peripheral muscle groups; MRC SS;
Poulsen et al. [39] 2013 Denmark	Muscle function	Single centre Case-control	n = 16 12 months	83.8%	Biomechanical testing during isometric knee extension; EMG.
Kiriella et al. [31] 2018 Canada	Muscle function, gait and postural control	Multicentre Nested prospective longitudinal	n = 16 6 months	70.9%	Isokinetic dynamometry (KEs, PFs, DFs); Gait assessment; quiet standing using force plates.
Fletcher et al. [30] 2003 UK	Neuromuscular function	Multicentre Prospective cohort study	n = 22 5 years	60.0%	NCS; EMG.
Angel et al. [23] 2007 Canada	Neuromuscular function	Multicentre Nested prospective cohort	n = 16 6–24 months	56.6%	NCS; EMG; muscle biopsy; biochemical tests.
Semmler et al. [21] 2012 Switzerland	Neuromuscular function	Multicentre Prospective cohort	n = 51 6–24 months	53.3%	MRC SS; EMG; NCS.
Koch et al. [50] 2014 Germany	Neuromuscular function	Single centre Prospective cohort	n = 26 12 months	46.6%	NCS.
Meyer-Freisen et al. [40] 2020 Germany	Pain and weakness	Single centre Retrospective cohort	n = 149 6 months − 10 years	93.5%	NRS; bespoke questionnaire.
Boyle et al. [48] 2004 Australia	Chronic pain	Single centre Prospective cohort	n = 52 6 months	50.0%	Pain scale; PSEQ.
Battle et al. [41] 2013 UK	Chronic pain	Single centre Retrospective cohort	n = 196 6–12 months	79.3%	Bespoke pain questionnaire
Baumbach et al. [45] 2016 Germany	Chronic pain	Single centre Case-control	n = 202 6 months	64.5%	German pain questionnaire;
Baumbach et al. [44] 2017 Germany	Chronic pain	Single centre Cross-sectional	n = 84 6 months	70.9%	Quantitative sensory testing
Baumbach et al. [46] 2018 Germany	Chronic pain	Single centre Prospective cohort	n = 159 12 months	77.4%	German pain questionnaire; quantitative sensory testing
Hayhurst et al. [26] 2018 USA	Chronic pain	Multicentre Nested prospective cohort	n = 253 12 months	84.3%	BPI
Langerud et al. [47] 2018 Norway	Chronic pain	Single centre Prospective cohort	n = 89 12 months	64.5%	BPI – short form
Carrie et al. [37] 2019 France	Chronic pain	Single centre Prospective cohort	n = 64 12 months	80.0%	BPI – short form DN4
Devine et al. [49] 2019 UK	Chronic pain	Single centre Prospective observational	n = 47 up to 12 months	56.6%	BPI
Demoro et al. [43] 2020 Italy	Chronic pain	Single centre Retrospective cohort study (mixed methods)	n = 116 6 months	77.4%	BPI
Koster-Brouwer et al. [42] 2020 The Netherlands	Chronic pain	Single centre Retrospective cohort study	n = 958 12 months	87.9%	NRS; bespoke questionnaire; body map; DN4
Probert et al. [28] 2021 USA	Chronic pain	Multicentre Longitudinal Prospective cohort ARDS	n = 737 12 months	83.9%	Bodily Pain subscale of SF-36
Orford et al. [24] 2011 Australia	Fracture risk	Single centre Retrospective longitudinal case-cohort	n = 739 4 years	70.9%	Incident fracture rate; age adjusted fracture risk
Zhang et al. [25] 2020 China	Femoral head necrosis	Single centre Prospective cohort	n = 57 15 years	66.6%	MRI; Harris hip score
Herridge et al. [33] 2011 Canada	Heterotopic ossification and frozen shoulder	Multicentre Longitudinal prospective cohort	n = 64 5 years	87.1%	MMT; physical examination
Gustafson et al. [51] 2018 UK	Shoulder impairment	Single centre Prospective cohort study	n = 61 6 months	75.0%	Pain (VAS); ROM; CMS; QD.

DXA = Dual-energy X-ray absorptiometry, ARDS = Acute Respiratory Distress Syndrome, MRC SS = Medical Research Council Sum Score, CSA = Cross sectional area, NCS = Nerve conduction studies, NRS = Numerical rating scale, EMG = Electromyography, MVC = maximal voluntary contraction, RFD = rate of force development, ICUAW = Intensive Care Unit Acquired Weakness, 6MWD = Six minute walk distance, SF-36 = Medical Outcomes Short-Form 36 Health Survey, KE = Knee extension, PF = Plantar flexion, DF = Dorsi flexion, CIP = Critical illness polyneuropathy, CIM = Critical illness myopathy, CIPNM = Critical illness polyneuromyopathy, PSEQ = Pain Self Efficacy Questionnaire, BPI = Brief pain inventory, DN4 = Douleur Neuropathique 4, MRI = Magnetic resonance imaging, MMT = Manual muscle testing, VAS = Visual Analogue Scale, ROM = Range of movement, CMS = Constant-Murley Score, QD = Quick DASH, UL = Upper limb.

The majority of the studies (n = 25) evaluated an aspect of muscle impairment [7,22,27,29,31,32,[34], [35], [36],[38], [39], [40]] or chronic pain [26,28,37,[40], [41], [42], [43], [44], [45], [46], [47], [48], [49]]. There was also evaluation of neuromuscular function [21,23,30,50], peripheral joint impairment [33,51], fracture risk [24] and femoral head necrosis [25]. Only six studies evaluated more than one aspect of MSK health [22,29,31,33,40,51], however the majority of studies (n = 21) did include a physical examination [7,[21], [22], [23],25,27,[29], [30], [31], [32], [33], [34], [35], [36],39,40,42,44,46,50,51].

The outcome measures used varied greatly, with the Medical Research Council Sum Score (MRC SS) the most common (n = 8). The neuromuscular and pain outcome measures used in the studies are summarised in Table 2, Table 3 respectively. Additional outcomes used to evaluate peripheral joint impairment were the: QuickDASH, Constant-Murley and Harris hip scores [25,51]. Three studies included outcome measures that were not validated [33,40,41].

Table 2.

Neuromuscular outcome measures.

	MRC SS	Handgrip	Dynamometry	USS	DXA	Bespoke questionnaire	NCS	EMG	Muscle biopsy	Physical examinationa
Chan [34]	x	x			x
dos Santos [29]	x		x	x			x	x
Nedergaard [22]		x	x	x
Fan [7]	x	x
Borges [36]		x	x
Pfoh [35]	x
Duarte [38]	x
Dinglas [32]	x
Van Aerde [27]	x	x	x
Poulsen [39]			x					x
Kiriella [31]			x
Meyer-Freisen [40]						x
Herridge [33]										x
Fletcher [30]							x	x
Angel [23]							x	x	x
Semmler [21]	x						x	x
Koch [50]							x

MRC SS = Medical Research Council Sum Score, USS = ultrasound scan, DXA = Dual-energy X-ray absorptiometry, NCS = Nerve conduction studies, EMG = Electromyography.

^aNon-descripted physical examination.

Table 3.

Pain outcome measures.

	NRS	VAS	BPI	BPI-SF	PSEQ	German PQ	Bespoke	QST	DN4	BP subscale
Boyle [48]	x				x
Battle [41]							x
Baumbach [45]						x
Baumbach [44]								x
Baumbach [46]						x		x
Hayhurst [26]			x
Langerud [47]				x
Carrie [37]				x					x
Devine [49]			x
Demoro [43]			x
Koster-Brouwer [42]	x								x
Probert [28]										x
Gustafson [51]		x
Meyer-Freisen [40]	x

NRS = Numerical rating scale, VAS = Visual Analogue Scale, BPI = Brief pain inventory, BPI-SF = Brief pain inventory short form, PSEQ = Pain Self Efficacy Questionnaire, German PQ = Pain Questionnaire, QST = Quantitative sensory testing, DN4 = Douleur Neuropathique 4, BP Subscale = Bodily pain subscale of SF36.

3.1. Musculoskeletal complications

Muscle impairment was reported as atrophy [29,34], weakness [7,27,29,35,36,38] or reduced muscle function [31,39], with reported prevalence varying greatly (9–73%). Neuromuscular impairment was described as motor sensory deficits [30], mononeuropathies [23] and critical illness polyneuropathy (CIP) [21]. The majority of physical assessments evaluated the lower limbs, with only three studies assessing the upper limb [27,33,51]. The prevalence of chronic pain was reported as between 16 and 74% at 6 months to 5 years following discharge from ICU [26,40,41,[44], [45], [46], [47], [48]]. The shoulder was the most commonly reported site for chronic pain [41,42,47], and was the only joint to undergo a detailed investigation [51]. Fig. 2 demonstrates the musculoskeletal complications reported using available pooled data.

Fig. 2.

Summary of the musculoskeletal complications based on the available pooled data from the studies included in the review.

The single interventional study [22] investigated the effect of sedation versus no sedation in invasively ventilated patients on a series of physical outcomes. They study demonstrated no difference in musculoskeletal outcomes between the groups.

Several studies investigated risk factors for the development of MSK complications [21,28,37,41,42,45,47,49,51]. Only pre-admission co-morbidity and severity of illness were reported as independently associated with weakness [35]. In contrast, age, sepsis, smoking history, in-ICU opioid use, duration of ventilation, hospital length of stay, female gender and days in ICU with hyperinflammation were all reported as associated with chronic pain [28,41,42]. However, there was disagreement between studies with several reporting no risk factors for the development of chronic pain [37,45,47,49].

A number of studies also investigated the association between MSK complications and physical function [7,28,29,31,32,34,46,48], although these were not reported in detail in several studies. However, both pain and weakness were reported as correlating with the physical function component score of the SF-36 [7,46,48].

3.2. Quality of reporting

The reporting within the majority of the studies was classified as adequate according to evaluation against STROBE and CONSORT (Median score 76%, IQR 63–84), with consistently well reported MSK assessment (Table 1). Sample size and association between MSK complications and function/quality of life were the areas commonly reported poorly.

4. Discussion

This scoping review has explored the extent to which MSK complications have been reported following critical illness, including the outcome measures used and the standard of the reporting. We evaluated 32 studies which reported a high prevalence of a variety of MSK complications through the use of a number of different outcome measures. The association between MSK complications and function is unclear due to poor reporting.

Overall study reporting as evaluated by STROBE and CONSORT guidelines was adequate with a median score of 76%, however there was a wide range of scores from 47 to 94%. This demonstrates an opportunity for improvement in reporting future studies, particularly in sample size calculations and the association between MSK complications and function or quality of life. Improved reporting will allow for better evaluation of the wider longer term impact of MSK complications on ICU survivors, and their ability to participate in everyday activities.

More consistent use of MSK outcome measures will also allow for improved evaluation. The outcome measures used for each MSK condition varied greatly between the studies with 23 different measures used, including 10 to measure varying aspects of pain. The most commonly used outcome measure was the MRC SS. However, this is a measure predominantly used in ICU and, given the large ceiling effect, may not be a useful long-term evaluation of muscle strength [52]. Several studies used dynamometry to evaluate muscle function, which may be a more appropriate measure to use in future studies [22].

Several different long term MSK complications were evaluated and reported. Decreased muscle mass, weakness and impaired muscle function is well documented during ICU [6] and muscle impairment was the most frequent and detailed area evaluated by studies in our review. However, there is an increasing number of studies specifically investigating chronic pain following critical illness, which is important for the development of future post-ICU rehabilitation studies as pain is not an outcome that has been included in their evaluation to date [53]. The most commonly reported location for pain following critical illness was the shoulder, which was also the joint identified has having reduced range of movement, and the focus of the only study to evaluate multiple aspects of MSK health in detail [51]. Despite this, the upper limb had minimal evaluation compared to the lower limb, with only three studies including an upper limb measure in addition to grip strength [27,33,51].

The studies included in our review demonstrate that MSK complications are an emerging potential source of long-term physical impairment in ICU survivors. There are several opportunities for further research, which should clearly report sample size calculations and investigation into associations between MSK complications and function or quality of life. Firstly, any future investigation should include a core set of assessments of multiple aspects of MSK health, as opposed to focusing on a single problem e.g. pain or weakness. These outcomes should be aligned to those recommended in other MSK services, which include MSK specific patient reported measures and work metrics [54]. To date, there has been very little evaluation of the upper limb and, as the shoulder was the most common location for pain, warrants further investigation. Evaluating the relationship between MSK complications and cardiopulmonary fitness would also help to guide the selection of rehabilitation outcomes for patients with MSK complications. Finally, further understanding of these MSK complications, including their clinical course, could inform the development of future successful post-ICU rehabilitation interventions.

This review was limited by the exclusion of non-English language studies, and as such, may not represent MSK complications reported in other languages. Grey literature such as conference abstracts were also excluded, however their limited word count means it is unlikely that they would report the information necessary for study evaluation [55]. This review has several important strengths. We evaluated a breadth of literature, encompassing all aspects of MSK complications following hospital discharge. We developed a comprehensive search strategy and searched five electronic databases from January 2000 until March 2021. We also used a robust methodology and evaluation encompassing the STROBE and CONSORT reporting guidelines which facilitated reviewer agreement.

5. Conclusions

MSK complications are an increasingly investigated potential source of physical impairment, and our review sought to evaluate the extent of the original research in this area. We identified a heterogenous body of literature reporting a high prevalence of a variety of MSK complications, over a varying time period, using a variety of outcome measures. Physical examinations undertaken in future studies should adopt a holistic approach and include assessments of the upper and lower limb. Future research should ensure clear reporting of sample size calculations and association between MSK complications and function or QoL.

Funding

Owen Gustafson, Clinical Doctoral Research Fellow, NIHR301569, is funded by Health Education England (HEE)/National Institute for Health Research (NIHR). Professor Helen Dawes is supported by the Elizabeth Carson Trust and the NIHR Oxford Biomedical Research Centre. The views expressed are those of the authors and not necessarily those of the NIHR, NHS or the UK Department of Health and Social Care.

Declaration of Competing Interest

The authors report no conflicts of interest.

Appendix A. Supplementary data

Supplementary material 1 Protocol

Supplementary material 2 PRISMA-ScR checklist

Supplementary material 3 Search strategy