Intensive care unit acquired weakness and physical rehabilitation in the ICU
BMJ 2025; 388 doi: https://doi.org/10.1136/bmj-2023-077292 (Published 27 January 2025) Cite this as: BMJ 2025;388:e077292
- Stephanie L Hiser
, assistant professor1, - Kelly Casey, rehabilitation therapy manager2,
- Peter Nydahl, lecturer3,
- Carol L Hodgson, professor4,
- Dale M Needham, professor5
- 1Department of Health, Human Function, and Rehabilitation Sciences, George Washington University, Washington, DC, USA
- 2Department of Physical Medicine and Rehabilitation, Johns Hopkins Hospital, Baltimore, MD, USA
- 3Department for Nursing Research and Development, University Hospital of Schleswig-Holstein, Kiel, Germany
- 4Australian and New Zealand Intensive Care Research Centre, School of Public Health and Preventive Medicine, Monash University, Melbourne, Australia
- 5Division of Pulmonary and Critical Care Medicine, Department of Medicine; and Department of Physical Medicine and Rehabilitation. Johns Hopkins University School of Medicine, Baltimore, MD, USA
- Correspondence to: S L Hiser stephanie.hiser{at}gwu.edu
Abstract
Approximately half of critically ill adults experience intensive care unit acquired weakness (ICUAW). Patients who develop ICUAW may have negative outcomes, including longer duration of mechanical ventilation, greater length of stay, and worse mobility, physical functioning, quality of life, and mortality. Early physical rehabilitation interventions have potential for improving ICUAW; however, randomized trials show inconsistent findings on the efficacy of these interventions. This review summarizes the latest evidence on the definition, diagnosis, epidemiology, pathophysiology, risks factors, implications, and management of ICUAW. It specifically highlights research gaps and challenges, with considerations for future research for physical rehabilitation interventions.
Introduction
Intensive care unit acquired weakness (ICUAW) is a clinical syndrome, defined as “clinically detected weakness in critically ill patients in whom there is no plausible etiology other than critical illness.”1A recently proposed term is critical illness weakness, which highlights that the weakness is not specifically associated with the healthcare setting (that is, intensive care).2 Most commonly, a patient’s strength is measured using physical examination (manual muscle testing), scored using the Medical Research Council (MRC) sum score (range 0-60), with <48 being considered to be ICUAW.3 Patients with ICUAW may have underlying polyneuropathy, myopathy, or a combination of both (that is, critical illness neuromyopathy), which is determined by electromyography and nerve conduction study (NCS).1 Isolated polyneuropathy is much less common than myopathy or combined neuromyopathy.345 Patients in the intensive care unit (ICU) experience prolonged immobilization, which may be an important modifiable risk factor contributing to ICUAW.67
Initial studies evaluating early rehabilitation in the ICU were published in the early 2000s.89 Early studies suggested improved outcomes for patients with early rehabilitation compared with usual care control groups that generally received bed rest/immobility while mechanically ventilated.101112 However, subsequent randomized trials have yielded inconsistent findings, which reflect substantial heterogeneity, including important differences in patient populations, types and doses of rehabilitation interventions, usual care control groups, and timing and measurement of primary and secondary outcomes.1314151617 This review synthesizes recent evidence on ICUAW and its epidemiology, evaluates randomized trials of physical rehabilitation interventions in the ICU (primarily focused on active, non-respiratory interventions), and discusses gaps/challenges and suggestions for future research.
Sources and selection criteria
We searched PubMed and Embase for peer reviewed articles from January 2019 to May 2024, to identify the most recent literature, using search terms “intensive care unit acquired weakness”, “critical illness myopathy”, “critical illness neuropathy”, “critical illness polyneuropathy”, “intensive care unit acquired paresis”, “early rehabilitation”, and “early mobilization”. These combined terms identified 4274 articles, of which we included 60 as the most relevant and important publications. We prioritized systematic reviews, meta-analyses, randomized controlled trials, and large studies from high impact general medicine and critical care journals. Additionally, we supplemented the articles with older publications to provide historical context or recognize landmark/unique publications. We hand searched the reference lists of included articles. We cite other study designs (for example, case series) if the literature on a specific topic was limited. Finally, we included relevant narrative review articles primarily when a relevant systematic review was not available.
Epidemiology
The estimated prevalence of ICUAW in critically ill patients is approximately 50%, with variability depending on the method of assessment and the time point of evaluation.1819
Definition and diagnosis
Ruling out other medical or neurological causes for generalized weakness is an important initial step in determining whether a patient has ICUAW. Diffuse symmetrical muscle weakness with sparing of the facial muscles in a critically ill patient should raise a suspicion of ICUAW.20 In the setting of generalized weakness, increased deep tendon reflexes or proximal sensory deficits are not consistent with ICUAW, and investigation should be directed to other causes.21 In evaluating weakness, the MRC sum score, based on physical examination via manual muscle testing, is often used. An MRC sum score <48 is a commonly used threshold for ICUAW,1 but <55 is also associated with long term morbidity and mortality.22 Handgrip strength (<11 kg in men; <7 kg in women23) can also identify ICUAW when the device and a trained assessor are available. Both manual muscle testing and handgrip strength are volitional measures, requiring patients to follow commands.
Underlying causes of weakness can include disuse atrophy, and/or critical illness polyneuropathy, critical illness myopathy, or critical illness neuromyopathy.12124 These conditions require electromyography/NCS for diagnosis, which can be a challenge to conduct in sedated or delirious patients in the ICU environment.21 Critical illness polyneuropathy primarily consists of a reduction in amplitudes of compound muscle action potentials and sensory nerve action potentials. For critical illness myopathy, decreased compound muscle action potential amplitudes with preserved sensory nerve action potentials are seen, along with other potential changes in motor unit action potentials and abnormal spontaneous activity on electromyography. The combined findings of critical illness neuromyopathy seem to be the most common presentation.345
The use of bedside ultrasonography to evaluate muscle thickness, cross sectional area, pennation angle, and echointensity has been increasingly used in the ICU.2526 One systematic review of 52 studies (n=3251) reported skeletal muscle loss of about 2% a day during the first week in the ICU, with ultrasonography being the most common method of assessment.18 Notably, ultrasonography requires a machine with an appropriate probe and rigorous training for reproducible results. Results of ultrasonography can be affected by adipose tissue or edema.27
Finally, muscle biopsy may be performed but is more invasive. A nerve biopsy may be considered in mechanistic research and/or for determining a clinical diagnosis in complex cases.28 Histologically, muscle biopsy may reveal loss of myosin and muscle necrosis with critical illness myopathy.2129 Skin biopsy, which is minimally invasive, may help to assess for histological evidence of degeneration of the intraepidermal fibers, which may be associated with acute axonopathy of large nerve fibers and myopathy.30
Pathophysiology
The pathophysiology of ICUAW is complex and not fully understood. Mechanistic studies have suggested structural changes (for example, axonal nerve degeneration, muscle myosin loss), functional changes (for example, electrical inexcitabiltiy of nerves), microvascular changes (for example, cytopathic hypoxia), and sodium channelopathy.29 Globally, the structural and functional changes in nerves and muscles indicate sequelae of systemic inflammation.3132 Electrophysiological studies have highlighted a rapid onset of peripheral nerve and muscle changes and the reversibility of these changes3334; however, these findings are not fully understood.29
More recently, bioenergetic failure (that is, mitochondrial dysfunction) is a hypothesized mechanism for ICUAW.3335 Mitochondrial function may be impaired owing to ischemic and cytopathic hypoxia.3637 Furthermore, in critical illness polyneuropathy, impairment in the blood-nerve barrier may be a mechanism for axonal depolarization with alteration in the microvasculature, resulting in lack of oxygen or increased toxins to the peripheral nerve.323839 Another hypothesis is endoneurial membrane depolarization due to endoneurial hyperkalemia.31 Muscle membrane inexcitability as a result of impaired sodium channels (that is, sodium channelopathy) may result in impaired muscle contraction.40
In critical illness myopathy, muscle atrophy from an imbalance in muscle proteolysis and synthesis and muscle contractile dysfunction are proposed mechanisms.4142 Furthermore, studies suggest anabolic resistance to dietary protein and impaired skeletal muscle incorporating amino acids.43 Another possible mechanism of critical illness myopathy is impaired insulin signal transduction, specifically failure of glucose transporter-4 translocation to sarcolemma membrane.44 Lastly, muscle atrophy often persists in patients even after discharge from hospital and may be due to different pathophysiological mechanisms, including reduced satellite cell content, which may affect the regenerative capacity of skeletal muscle.45
Risk factors
Studies suggest many risk factors for developing ICUAW, both modifiable and non-modifiable. Determining which factors have the greatest effect on ICUAW is challenging, with some inconsistency across studies.464748 Notably, some therapeutic interventions commonly used in managing critically ill patients may contribute to ICUAW.2949
In survivors of acute respiratory distress syndrome (ARDS), patient’s age and duration of ICU bed rest were positively associated with weakness at hospital discharge and at three, six, 12, and 24 months of follow-up.7 After adjustment for other risk factors, for every additional day of bed rest, a 3-11% relative decrease in muscle strength occurred over longitudinal two year follow-up.7 A five year follow-up analysis of this prospective cohort reported that pre-ARDS comorbidity and organ failure were associated with subsequent declines in strength.50
Recent systematic reviews have specifically evaluated risk factors for ICUAW. One systematic review reported risk factors for ICUAW in the categories of personal, therapeutic, disease, and laboratory indicators. The following showed significant associations with ICUAW: female sex (odds ratio 1.34, 95% confidence interval (CI) 1.06 to 1.71), duration of mechanical ventilation (3.04, 1.82 to 4.26), age (6.33, 5.05 to 7.61), ICU length of stay (LOS) (3.78, 2.06 to 5.51), infection/sepsis (1.67, 1.20 to 2.33), renal replacement therapy (1.59, 1.11 to 2.28), aminoglycoside antibiotics (2.51, 1.54 to 4.08), Sequential Organ Failure Assessment score (1.07, 0.24 to 1.90), and hyperglycemia (2.95, 1.70 to 5.11).51
One systematic review reported that neuromuscular blocking agents (NMBAs) were not a significant risk factor51; however, this was evaluated in only five studies (n=512). A second systematic review and meta-analysis of 30 studies (n=3839) reported that NMBA was associated with ICUAW (odds ratio 2.77, 95% CI 1.98 to 3.88). However, 26 of the studies were observational with high heterogeneity in measurement of the exposure and outcome (for example, type/dose of NMBA and timing of strength assessment).52 Furthermore, only one multicenter randomized trial (n=340) evaluated the effect of NMBAs versus placebo on ICUAW (MRC sum score <48), as a secondary outcome, reporting no difference at ICU discharge (among 201 participants: 64% v 69% respectively; P=0.51).53 The dose and duration of NMBA, along with patient and critical illness related co-factors (for example, deep sedation, systemic corticosteroids) may be important considerations in understanding any possible effect of neuromuscular electrical stimulation (NMES) on ICUAW.49 In general, our recommendation for NMBA would be to avoid routine use and use the lowest dose for the shortest duration possible.49
Finally, corticosteroids are another potential risk factor for ICUAW. However, the literature has been inconclusive in evaluating associations, particularly confounded by changes in clinical practice over time (for example, reduced use of corticosteroids combined with prolonged duration of infusions of NMBA medications).3254
Implications of ICUAW
In-hospital outcomes
Patients with ICUAW often experience a range of adverse short term and long term outcomes. Importantly, some potential risk factors for ICUAW may also be negative outcomes associated with ICUAW; for example, ICU LOS is reported as both a risk factor for and a consequence of ICUAW.23555657 This finding may be due to the relation between ICUAW and a longer duration of mechanical ventilation (for example, owing to respiratory muscle weakness slowing ventilator liberation) or survival bias (that is, patients who survived and developed ICUAW may have been sicker and needed longer ICU level care).
A prospective, multicenter cohort study (n=95) reported that presence versus absence of ICUAW (MRC sum score <48) was associated with longer mechanical ventilation (median 6 (interquartile range (IQR) 1-22) versus 3 (1-7) days; P=0.01).58 Furthermore, another study (n=377) reported that limb muscle weakness was associated with extubation failure (median MRC sum scores with versus without extubation 14 (IQR 12-16) versus 16 (12-18); P=0.024).59 These findings may be due to patients with ICUAW also having weaker respiratory muscles, which are predictive of longer duration of mechanical ventilation.60 However, findings are inconsistent regarding associations of ICUAW and diaphragm dysfunction.61
Studies have reported an association of increased short term mortality among critically ill patients with ICUAW. One prospective multicenter cohort study across 12 ICUs in Australia and New Zealand reported that in 94 (52% of study population) patients with ICUAW (MRC sum score <48) a decreased survival from ICU discharge to day 90 was seen.62 This finding is consistent with a previous prospective observational study (n=115) in France which reported that ICUAW (MRC sum score <48) was associated with greater ICU and hospital mortality.63 Finally, another, prospective multicenter cohort study of 136 patients needing mechanical ventilation for at least five days reported that ICUAW (assessed by grip strength) was independently associated with in-hospital mortality.23
Studies also report an association between ICUAW and short term functional impairment. One secondary analysis (n=83 with MRC sum score at ICU discharge) of a previous randomized trial reported median Functional Independence Measures scores (measure of disability; 18 items including motor and cognitive domains; total score ranging from 18 to 126, with lower scores indicating more disability) at hospital discharge of 24 (IQR 21-34), 31 (27-46), and 42 (35-58) for those with severe ICUAW (MRC sum score <36), moderate ICUAW (MRC sum score 36-47), and no ICUAW (MRC sum score >48), respectively (P<0.001).6465 This impact on physical functioning means that patients with ICUAW are often discharged from hospitals to inpatient rehabilitation facilities.6667 For instance, in one study, patients with versus without ICUAW were more likely to be discharged to rehabilitation units (18% v 10%; P=0.017).68
Post-hospital outcomes
In addition to its association with in-hospital mortality, ICUAW is associated with longer term mortality.62 A propensity matched analysis of 227 patients from a multicenter randomized trial reported higher one year mortality in those with ICUAW (MRC sum score <48) versus without ICUAW (31% v 17%; P=0.015).68 Additionally, one large (n=883) sub-analysis of a prospective study, involving patients from a clinical trial on early versus late parenteral nutrition in the ICU,69 reported that lower MRC sum scores were independently associated with increased five year mortality (hazard ratio 0.96, 95% CI 0.93 to 0.98; P=0.001).22
ICUAW is associated with prolonged impairment in physical functioning after discharge from hospital.48 A single center prospective cohort study (n=156) reported that ICUAW was independently associated with worse physical functioning (SF-36 physical function score) at six months’ follow-up.70 Similarly, a prospective cohort study across 13 ICUs at four hospitals evaluating 222 ARDS survivors at three, six, 12, and 24 months reported significantly worse physical functioning (SF-36; P≤0.001) and six minute walk distance (P≤0.01) at six, 12, and 24 months) among those with ICUAW (MRC sum score <48).7
Health related quality of life for survivors of critical illness is lower than population norms,71 with both physical and mental quality of life domains even lower for survivors of ARDS.72 A post hoc analysis evaluating 128 patients reported that ICUAW was associated with impairments in quality of life (Nottingham Health Profile and SF-36 questionnaire) at six months after discharge.73 The quality of life domains primarily affected were related to physical function rather than psychological status.
Prevention and treatment of ICUAW and impaired physical function
A variety of physical rehabilitation interventions have been evaluated in critically ill adults. Here, we focus primarily on non-respiratory interventions that generally target active rehabilitation (rather than passive or non-volitional interventions) in the ICU setting.16
Functional mobility/multicomponent physiotherapy interventions
Functional mobility (that is, people’s ability to move around their environment74) typically includes a stepwise rehabilitation approach from in-bed exercises through sitting on the edge of the bed, transfer and standing, and marching in place to ambulation.75 Multicomponent interventions often have at least three components, which typically include a combination of neuromuscular electrical stimulation, passive exercises (that is, no voluntary effort or no muscle contraction) or active exercises (that is, voluntary effort or active muscle contraction), in-bed cycle ergometry, functional mobility, muscle strengthening (for example, use of weights or resistance bands), education, cognitive training, and proprioceptive neuromuscular facilitation techniques (that is, recruiting muscles via stimulation of muscle proprioceptors76).77
Clinical trials of functional mobility and multicomponent rehabilitation interventions in the ICU began to emerge after a landmark 2008 two center randomized trial that reported improved functional outcomes with early combined physical and occupational therapy intervention that incorporated a progressive mobility approach.11 A recent scoping review reported that 73 (62%) of rehabilitation interventions included in 117 studies were either functional mobility or multicomponent interventions.77 These interventions show the potential to prevent ICUAW on the basis of a meta-analysis of seven randomized trials reporting (risk ratio 0.49, 95% CI 0.26 to 0.91; P=0.025).78 Another meta-analysis of nine studies also reported that early rehabilitation decreased ICUAW (odds ratio 0.63, 95% CI 0.43 to 0.92).79 On the basis of this evidence, we believe that functional mobility and/or multicomponent interventions (with avoidance of early high intensity functional mobility80) should be implemented in patients who are critically ill to decrease ICUAW.
Neuromuscular electrical stimulation
NMES may be used for patients who are unable to be mobilized out of bed owing to sedation, unstable clinical status, or other medical or surgical reasons. NMES may be used in patients who are unable to elicit a muscle contraction themselves. It provides an electrical current, applied via electrodes placed on the skin, to elicit a contraction of targeted muscle groups, typically the lower extremities (fig 1).8182
Neuromuscular electrical stimulation with electrodes placed to elicit contraction of targeted muscles
A recent randomized trial from Egypt reported that among 124 patients randomized to one of four groups (range of motion plus NMES, NMES only, range of motion only, and control group), ICUAW (MRC sum score <48) occurred in 0%, 13%, 60%, and 100%, respectively (P<0.001).83 Additionally, a recent systematic review, which did not include the aforementioned trial, reported that among six trials (n=274) NMES reduced the risk of ICUAW (risk ratio 0.48, 95% CI 0.32 to 0.72).84 However, a previous meta-analysis of six randomized trials (n=718) evaluating NMES versus usual care reported no difference in overall muscle strength (mean difference in MRC sum score 0.45, 95% CI −2.89 to 3.80; P=0.79).85 A recent secondary analysis of muscle biopsies (n=42) from two randomized trials evaluating functional electrical stimulation in mechanically ventilated patients reported that inflammation and altered substrate utilization may contribute to the observed lack of benefit of this intervention on physical functioning.86 Notably, a network meta-analysis of 23 randomized trials of 1312 mechanically ventilated adults reported that NMES alone or NMES combined with physical therapy interventions improved success of extubation (odds ratio 1.85, 95% CI 1.11 to 3.08); however, it reported no significant improvement in ICU LOS, duration of ventilation, or mortality.87 As our mechanistic understanding of ICUAW improves, re-evaluation of the effects of NMES in certain subpopulations or the timing of implementation may be important. However, evidence to support routine use of NMES as a primary rehabilitation intervention for critically patients is not adequate.
In-bed cycle ergometry
Similarly to NMES, in-bed cycling may be safe and feasible if patients are not suitable for out of bed activity as it can be performed in a semi-recumbent position.8889 In-bed cycling can be passive or active depending on the patient’s ability, with or without incremental resistance (fig 2).
In-bed cycle ergometry facilitating passive or active movement of lower extremities
The first randomized trial evaluating in-bed cycle ergometry versus a usual care control group, in medical/surgical patients (n=90) in ICU at a single center, reported increased median six minute walk duration (196 (IQR 126-329) m versus 143 (37-226) m; P<0.05) and higher SF-36 physical function scores (21 (18-23) versus 15 (14-23) points; P<0.01) at hospital discharge.12 A recent publication of a randomized trial of 56 patients with ICUAW reported that in-bed cycle ergometry decreased ICUAW at discharge from ICU (87% v 61%; P=0.039).90
Recently, the largest randomized trial of cycle ergometry in the ICU has been reported.91 This 16 center, international trial of 360 mechanically ventilated patients evaluated in-bed cycling plus usual care physiotherapy versus usual care physiotherapy alone. The cycling intervention, which was started at a median 2 (IQR 2-3) days after admission to ICU with 3 (1-4) sessions per patient, occurred with a fixed resistance (0.6 Newton-meters) and a fixed rate of five revolutions per minute during passive cycling time periods. The control group started rehabilitation at 2 (IQR 2-4) days after ICU admission, with 4 (2-7) sessions per patient and a mean duration of 29 (standard deviation (SD) 13) minutes. The primary outcome was physical functioning, measured via the Physical Function ICU Test scored at three days after discharge from ICU, with no difference between the intervention and control groups (mean 7.7 (SD 1.7) v 7.5 (1.8); absolute difference 0.23, 95% CI −0.19 to 0.65; P=0.29). Additionally, no significant difference was seen in ICUAW at hospital discharge (9.6% v 12.1%; odds ratio 0.87, 95% CI 0.39 to 1.90).91 Notably, another multicenter randomized trial is in progress to evaluate the combination of in-bed cycling plus protein supplementation versus usual care.92 Although evidence does not show that predominantly passive cycle ergometry improves outcomes, further investigation of its roles within multifaceted rehabilitation interventions, especially in patients who are otherwise unable to engage in active rehabilitation, and/or in combination with other therapies (for example, nutrition; see Emerging therapy section) is warranted.
Functional electrical stimulation
NMES and in-bed cycle ergometry can be delivered in a combined manner, known as functional electrical stimulation cycling (fig 3).93 Functional electrical stimulation is synchronized to elicit muscle contractions in a particular coordinated pattern, such as allowing cycling of an in-bed ergometer.94 Two randomized controlled trials of mechanically ventilated patients (one conducted in four ICUs in Australia and the US (n=162) and another in two ICUs in the Czech Republic (n=150)) evaluated functional electrical stimulation assisted cycle ergometry. Both randomized trials reported no benefit on muscle strength or measures of physician functioning.9596
Functional electrical stimulation assisted cycling with synchronized electrical stimulation to contract appropriate muscles for cycling
Adjuvant devices
Limited evaluations have been conducted of tilt table, dynamic tilt table, and multifunctional patient positioner devices (that is, a device that allows sitting, sit to stand, and tilting) for physical rehabilitation in the ICU.9798 Some hospital beds have these integrated functions. Typically, however, a patient must be transferred from their hospital bed onto the device and secured with straps to perform the interventions. Tilt table devices allow for the gradual progression from supine to standing position (fig 4). A dynamic tilt table can incorporate patient movement, and a multifunctional patient positioner device can position a patient in several supported sitting positions. One single site randomized study of 145 patients reported no difference in MRC sum score at discharge from ICU with versus without a tilt table intervention (50 (IQR 45-56) v 48 (45-54); P=0.555).97 However, this study reported muscular strength recovery (change in MRC sum score from baseline to ICU discharge) to be significantly greater in the tilt table group than the control group, with a median change in MRC sum score of 14 (IQR 10-24) v 10 (5-15) (P=0.004).
Tilt table allowing graduated weight bearing to assist with standing. A former critically ill patient, who reviewed this article, highlighted her experience with the tilt table intervention
Functional activity
Functional activity and interventions targeting activities of daily living are supported by a growing body of evidence.99100 Clinicians providing these interventions vary but may include physiotherapists or occupational therapists. A recent randomized trial (n=200) in the US reported combined treatment by physical and occupational therapists and showed improved long term cognitive impairments (Montreal Cognitive Assessment score <26) at one year (24% for intervention versus 43% for usual care; absolute difference −19%, 95% CI −32 to 6; P=0.004).101 This study is the first to show improved long term cognitive outcomes from early physical rehabilitation, which may be attributed to integrating activities of daily living and mobilization. Additionally, at one year’s follow-up, the intervention group had lower ICU acquired weakness (0% v 14% of patients; P<0.001) and a higher median SF-36 physical component summary score (52 (IQR 45-57) v 41 (32-49); P<0.001). Notably, the median numbers of days of delirium was shorter in the intervention group (0 (IQR 0-2) v 1 (0-3; P=0.005).
Additional considerations
Non-traditional interventions to mitigate ICUAW have been evaluated, including robotics, interactive video games, and hydrotherapy. We will briefly discuss each of these here, with recognition that further evidence is needed to evaluate their efficacy for managing ICUAW.
Robotics for ICU based rehabilitation is a subject of recent interest, especially during the covid-19 pandemic. One robotic assisted mobilization system can be attached to a hospital bed to allow for passive or assisted step movements of the legs in progressive vertical positions.102 Studies have primarily focused on safety and feasibility.102103104
Interactive video games that use wireless controllers and/or pressure sensitive balance boards allow patients to receive multisensory feedback, which may facilitate improved movement patterns, balance, and/or activity tolerance. This approach may provide enjoyable interventions and facilitate increased patient engagement and seems to be safe and feasible.105
Hydrotherapy involves rehabilitation performed in a swimming pool, which is a novel intervention for patients receiving mechanical ventilation.106 Hydrotherapy exploits the buoyancy of water to assist in early initiation of functional mobility that may not otherwise be possible with significant ICUAW.107 One single center study (n=25) determined that hydrotherapy in mechanically ventilated patients in ICU seemed to be safe and feasible.108
Post-ICU rehabilitation
Studies evaluating the effects of rehabilitation on survivors of critical illness after discharge from ICU are limited.109 However, a growing body of evidence evaluating rehabilitation interventions after discharge from hospital is available. A recent meta-analysis of 14 studies reported that exercise interventions improved aerobic capacity (nine studies; n=880; standard mean difference 0.20, 95% CI 0.03 to 0.30) and SF-36 physical component summary score (six studies; n=669; 3.3, 1.0 to 5.6).110
Additionally, post-ICU clinics have emerged and are growing in number, especially since the covid-19 pandemic.111 These clinics often have a multidisciplinary team that may include ICU or non-ICU physicians, rehabilitation specialists, psychologists, pharmacists, nurses, and social workers. However, evidence supporting the efficacy of post-ICU clinics on long term muscle weakness and physical functioning is unclear.112 Notably, a recent large multicenter randomized trial (n=540) reported that hospital based, face-to-face, intensivist led multidisciplinary consultation at ICU discharge and three months and six months after discharge was associated with a poor clinical outcome (defined as death or severe-to-extreme impairment of at least one EuroQoL-5D-5 dimension at 12 months; adjusted odds ratio 1.49, 95% CI 1.04 to 2.13; P=0.03).113 The reasons for this finding are not clear, but the authors noted that the multidisciplinary team did not include a physiotherapist or occupational therapist, who may have been able to tackle physical and cognitive impairments.114
Prevalence of early physical rehabilitation in the ICU
An international survey of 1484 ICU leaders in France, Germany, the UK, and the US reported that early mobilization practices were present in 40%, 59%, 52%, and 45% of ICUs in each country, respectively.115 Another survey of 687 randomly selected ICUs in the US (73% response rate) reported a median of 6 (IQR 5-7) days per week and 2 (2-3) daily sessions of early mobilization; however, details on the type and level of mobilization were not reported.116 Despite international guidelines with some support for early rehabilitation, challenges/barriers to successful implementation exist.117
Effects of early rehabilitation
Despite the many different types of rehabilitation interventions available, these interventions are often collectively referred to as “early rehabilitation” or “early mobility.” Furthermore, systematic reviews and meta-analyses often evaluate these interventions collectively, despite heterogeneity in their expected effects, as described below.
ICU rehabilitation aims to reduce ICUAW and improve physical functioning, but heterogeneity in physical outcome measures (for example, strength, mobility, physical functioning—both patient reported and performance based physical testing) exists, with strength not always measured in trials.118 A recent meta-analysis evaluating the effects of rehabilitation in critically ill patients reported that among the 43 randomized trials (n=3548), decreased duration of mechanical ventilation, ICU LOS, and hospital LOS were seen, with a mean difference of −1.7 (95% CI -3.6 to −0.3), −1.2 (–2.5 to 0), and −1.6 (–4.3 to 1.2) days, respectively. However, effects on muscle strength were not reported. Notably, the significantly decreased duration of mechanical ventilation was observed only for interventions using protocolized physical rehabilitation (rather than other interventions, such as in-bed cycling and NMES) and was more prominent in patients with a longer ICU LOS and lower severity of illness (Acute Physiology and Chronic Health Evaluation II score).119 Another meta-analysis, including 60 trials (n=5352), reported that ICU rehabilitation improved physical function and reduced ICU and hospital LOS.120 This meta-analysis included subgroup analyses stratifying the amount of rehabilitation in the control group as either high dose (>5 days/week) or low dose (<5 days/week). The review reported that rehabilitation was more effective when the intervention was compared with a low dose control group. Notably, delivery of rehabilitation four days per week (that is, the “low dose” control) is higher than usual care at many institutions. A recent meta-analysis of 15 trials of ICU early mobility (n=2703) that evaluated six month follow-up data reported a 95% probability of improved physical function (patient reported outcome measure) at the six month follow-up but no difference in strength.15
Effects of early rehabilitation on delirium/cognition
Early rehabilitation has shown positive effects on other outcomes aside from physical impairments, such as delirium and cognition. One of the earliest randomized trials (n=109) evaluating co-treatment by occupational and physical therapists reported that duration of delirium was reduced (median 2 (IQR 0-6) days versus 4 (2-8) days; P=0.02).11 A meta-analysis of 13 studies (n=2164) reported a reduction in the incidence (odds ratio 0.53, 95% CI 0.34 to 0.83; P=0.01) and duration (mean difference −1.8 days, 95% CI −2.7 to −0.8; P<0.001) of delirium with early mobilization.121 However, a recent randomized trial evaluating an early, high intensity functional mobility intervention versus a usual care control group that received early and frequent lower intensity mobility reported no difference in delirium-free days and six month cognition (measured via Montreal Cognitive Assessment (MoCA-BLIND)) among 741 mechanically ventilated patients.80
Studies that evaluate combined physical and occupational therapy interventions, such as the previously mentioned US study (n=200) that reported improved cognitive outcomes at one year,101 suggest the importance of integrating functional cognition activities with functional mobility.122 Additionally, a randomized controlled trial of 140 non-ventilated patients in the ICU evaluated the effect of intensive occupational therapy interventions and reported a lower incidence (3% v 20%; P<0.001) and duration (risk incidence ratio 0.15, 95% CI 0.12 to 0.19; P<0.01) of delirium.123 In summary, understanding the potential synergistic benefit of occupational therapy as part of a comprehensive early rehabilitation program is an important research priority.
Gaps/challenges in current evidence and considerations for future studies
The PICO (patient, intervention, comparator, outcomes) framework assists in evaluating the body of evidence for physical rehabilitation in critically ill adults.124Figure 5 summarizes research gaps/challenges, and box 1 summarizes related considerations for future research. We further discuss these gaps and considerations below.
Research gaps and challenges for clinical trials evaluating rehabilitation interventions in critically ill adults, using PICO (patient, intervention, comparator, outcomes) framework. ICU=intensive care unit; OT=occupational therapist; PT=physical therapist; RN=registered nurse
Summary of considerations for future clinical trials
Patients
Include patients with impaired physical function and stratify randomization by baseline function
Minimize loss to follow-up via evidence based strategies and existing resources (eg, www.improveLTO.com)
Interventions
Detailed reporting, in study protocol and trial publication, of type of intervention(s), timing of initiation, type of clinician delivering intervention, and dose (frequency, duration, and intensity) for planned intervention and intervention actually delivered
Determine valid methods of measuring intervention intensity in critically ill patients
Evaluate optimal intervention duration across the continuum of recovery (eg, in-ICU, in-hospital, or after hospital discharge)
Standardization of definition and reporting of adverse events and any clinical consequences both per participant and per intervention
Comparator
Specify intervention for control group, rather than usual care, including specification of timing, type, and dose (frequency, duration, and intensity)
Outcomes
Use core outcome sets
Understand impact of timing/location on feasibility of outcome assessments (eg, ICU versus hospital discharge versus post-discharge)
Consideration of mechanistic (eg, muscle ultrasonography), performance based (eg, strength testing), and patient reported (eg, self-reported physical functioning) outcome measures
Selection of primary and secondary outcomes that can be affected by interventions (eg, strength/function versus mortality)
Study design
Stepped wedge or cluster RCT to test bundled interventions, including sedation and culture change
Bayesian adaptive platform trials to evaluate multiple types of interventions, timing of intervention, and dose (frequency, intensity, duration)
Use of multiphase optimization strategy framework125126127 for efficient and simultaneous testing of aspects of interventions by using fractional factorial design
Embedded mechanistic studies including relevant longitudinal measures (eg, protein metabolism, inflammation, body composition, muscle biopsy, electromyography/nerve conduction)
Secondary use of data
Encourage secondary use of data to allow pooling for individual patient data meta-analyses and understanding of heterogeneity of treatment effect (eg, emulation trials)
ICU=intensive care unit; RCT=randomized controlled trial
Patient related gaps
The inability to prospectively measure a patient’s baseline status given the acute onset of critical illness is a fundamental challenge in critical care research.128 To understand the effect of an intervention on ICUAW, many studies exclude patients with abnormal baseline function, via a proxy based survey (for example, Barthel Index129). However, in many ICUs, most patients have impaired baseline status; thus, such exclusions limit the sample size and the generalizability of the results. Future studies should consider including patients with mild or moderate pre-existing physical impairments and stratify randomization on the basis of baseline function.46
Intervention related gaps
Understanding the optimal dose of rehabilitation interventions is challenging owing to studies not reporting all parameters of dose (that is, frequency, duration, and intensity) for both their planned and their delivered intervention.130 Future studies should prioritize thorough reporting using available guidelines, such as the Template for Intervention Description and Replication,131 Consensus on Exercise Reporting Template,132 or Rehabilitation Treatment Specification System.133 Moreover, future research is needed to aid understanding of valid, reliable, and feasible methods of measuring the “intensity” aspect of dosing given the challenges in adapting existing guidance to critically ill patients.134
Timing of initiation is another important factor in the ICU setting. No consensus exists on the definition of “early,”135136 but German guidelines recommend starting mobilization within 72 hours of admission to ICU.137 Additionally, a systematic review defined early mobilization as within 72 hours of admission to ICU,16 consistent with other publications.138139 Finally, with novel rehabilitation interventions in critical illness, careful attention to standardized reporting of safety related events is needed, with consideration of differentiating an expected physiological response to exercise (for example, changes in heart rate and blood pressure) from changes that may be unsafe.140141142
Comparator related gaps
Substantial heterogeneity exists among control groups in the ICU rehabilitation literature. One scoping review of 125 studies reported that 88 trials reported “usual care” as their control group. However, 60 different activities were performed in the usual care groups.143 Given the evolution of early rehabilitation, no clear definition exists for usual care or for best practice.144145 Future trials will benefit from an established consensus on relevant control groups for comparison. Such control groups should aim to avoid complete bed rest and late initiation of rehabilitation; however, a suitable dose of intervention for control groups is unclear. Contemporary point prevalence studies of ICU rehabilitation in clinical practice may help to inform this question.
Clinical care practices in the ICU (for example, sedation, delirium, nutrition) may need careful consideration as potential co-interventions with physical rehabilitation. Sedation and delirium can affect what types of rehabilitation interventions are possible (for example, interventions requiring active patient participation are not possible with deep sedation) and affect fidelity to the intervention protocol, as well as the safety of rehabilitation. Additionally, rehabilitation can have an effect on the cognitive status of patients as discussed above.101121
Outcome related gaps
To reduce heterogeneity in study outcomes and related measurement instruments, consensus based core outcome sets are being developed.146 Core outcome sets for ICU trials evaluating interventions related to each of rehabilitation, delirium, and long term outcomes after hospital discharge have recently been published.147148149150151152 These core outcome sets may assist with this research gap. However, another important methodological challenge is retaining ICU patients in studies evaluating longitudinal post-discharge assessments.153 Guidance on reporting methods and outcomes for retention of participants may improve reporting practice, along with use of existing “best practices” for optimizing patient retention in follow-up assessments (for example, see free National Institutes of Health funded resources at www.improveLTO.com).154155156157
Study design considerations
Incorporating contemporary study designs, with embedded mechanistic studies, into ICU rehabilitation research may help to close research gaps.158 For instance, a stepped wedge cluster randomized trial design can evaluate implementation of a bundle of interventions that require “culture change” at the level of the ICU, rather than for an individual patient (for example, a sedation/delirium intervention combined with early rehabilitation or the “ABCDEF Bundle”159 that also includes interventions for ventilator liberation and family involvement).160 A recent example of this study design evaluated a structured approach (daily goal setting, inter-professional communication, performance feedback) to improve mobilization of patients across 12 ICUs in four US hospitals.161 Moreover, a stepped wedge cluster randomized trial in pediatric ICUs is evaluating a bundle of interventions (tiered physical activity/mobility plan, sleep hygiene promotion, delirium screening) to optimize outcomes.162
To overcome the need for multiple randomized evaluations of different types of rehabilitation interventions and dosing (for example, different frequency, duration, and intensity), bayesian adaptive platform trials and fractional factorial design (for example, via a multiphase optimization strategy (MOST) framework125126127) may facilitate efficient comparison of multiple study arm studies evaluating different interventions or doses. Such a strategy has been used recently in evaluating different approaches to improving psychological outcomes of adult survivors of critical illness.163 Via a bayesian adaptive approach, study arms that show no effect can be dropped and new interventions can be added over the course of the platform trial.164165166
Finally, registry based randomized controlled trials are a pragmatic design with growing interest. In this design, patients are identified and recruited through a pre-existing registry, which can offer efficiency in data collection on baseline status and patients’ outcomes given existing registry infrastructure.167 Additionally, this design may offer a more timely completion and overall lower cost in comparison with traditional randomized trials.168
Emerging treatments
Great interest exists in combined nutrition and rehabilitation interventions among critically ill patients, as shown by it being the top-rated research priority in the field of critical care nutrition and metabolism.169 A growing number of phase 2 randomized trials evaluate a rehabilitation intervention with protein supplementation versus usual care control groups, showing some signal of potential reduction in muscle weakness.170171172 However, in a three arm, phase 2 randomized trial (usual care versus early mobilization versus combined early mobilization and guideline based early nutrition), no difference was seen in muscle weakness between the two intervention groups,173 raising the question of whether nutritional interventions have any incremental benefit. However, additional phase 2 data are pending from a multicenter randomized trial evaluating cycle ergometry combined with intravenous protein supplementation versus a usual care control group.92
Additionally, a multicenter, four arm, randomized trial (n=112) evaluated usual care control versus resistance training versus β-hydroxy-β-methylbutyrate alone versus combined resistance training and β-hydroxy-β-methylbutyrate.174 β-hydroxy-β-methylbutyrate was evaluated with a goal of stimulating muscle protein synthesis and inhibiting protein degradation.175176177 This trial reported that resistance training (alone) and the combination of resistance training and β-hydroxy-β-methylbutyrate, compared with control, showed improved physical functioning; however, no significant differences were reported in the β-hydroxy-β-methylbutyrate only group.174 Moreover, targeted nutritional intervention, via daily use of indirect calorimetry directed feeding combined with in-bed cycle ergometry, has been evaluated in a small single center randomized trial (n=21), with no significant differences in muscle mass (assessed by ultrasonography) reported.178 Evaluation of combined nutrition and rehabilitation interventions is needed.
Guidelines
No recent international clinical practice guidelines specifically for the management of ICUAW are available. However, guidelines exist for ICU based physical rehabilitation and early mobilization. A systematic review appraised 10 clinical practice guidelines published between 2008 and 2020. This review reported on seven topics that were agreed upon, as follows: “1) early mobilization is safe and may reduce healthcare costs, 2) safety criteria should be provided, 3) a protocolized or structured approach should be used, 4) collaborative teamwork is required, 5) staff require specific skills or experience, 6) patient and family engagement is important, and 7) program evaluation and outcome measurement are a key component of implementation.”179 Additionally, this review reported that the Society of Critical Care Medicine’s Clinical Practice Guidelines for the Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption (PADIS Guideline, published in 2018) should be the foundation for implementing early mobilization programs given its rigorous methods and completeness. The PADIS guideline reported a “conditional recommendation” (as per Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology) for performing rehabilitation or mobilization in critically ill adults.180181 Furthermore, as part of the 2021 “Choosing Wisely” campaign, an initiative to reduce waste and overuse by recommending evidence based practices, early mobilization was one of five recommendations: “Don’t delay mobilizing ICU patients.”182
More recent guidelines, published in 2022-24 by the German Association of the Scientific Medical Societies,137 the National Health and Medical Research Council (NHMRC) of Australia,183 the Japanese Society of Intensive Care Medicine,184 the Korean Society of Critical Care Medicine,185 and the Union of Rehabilitologists of Russia with the Federation of Anesthesiologists and Resuscitators of Russia,186 provide added guidance. The German and Russian guidelines recommend that early mobilization should begin within 72 hours.137186 With respect to frequency of intervention, the Russian and Japanese guidelines recommend once daily and multiple sessions per day, respectively.184186 Notably, the NHMRC considered safety with respect to both initiation and frequency in good practice statement 1: “All patients admitted to the ICU should be assessed and screened daily for suitability to deliver physical rehabilitation and/or mobilization. The initial screening should occur as early as possible after admission to ICU, within 24 hours if feasible.” With respect to duration of intervention, the Russian guideline recommended >30 minutes a day186; however, the German guideline favors an individual approach depending on the patient’s condition and recommends more research for the optimal rehabilitation dose.137 Finally, regarding types of interventions, the Japanese guideline recommends cycle ergometry (with very low certainty of evidence as per GRADE methodology),184 and the German guideline recommends cycling when functional training is not possible (weak recommendation).137 The German, Japanese, and Russian guidelines all recommend NMES with recognition of limitations in evidence.137184186 Lastly, the PADIS and NHMRC guidelines suggest comprehensive safety criteria for starting and stopping physical rehabilitation or mobilization,181183 with other guidelines providing similar guidance, including emphasizing that clinical judgment must be exercised, rather than strict adherence to specific safety related criteria.181184 The European Respiratory Society and American Thoracic Society are developing a new clinical practice guideline on physical rehabilitation and mobilization for critically ill adult patients, with expected completion in 2026.187
Conclusion
With an aging population and increasing multimorbidity, the number of patients admitted to critical care settings with increased risk of ICUAW is growing. Identification of ICUAW and its risk factors have improved, but further understanding of its mechanisms and pathophysiology is needed, particularly to inform the design of interventions to reduce ICUAW. With a rapidly growing number of studies focused on physical rehabilitation interventions in the ICU, we have greater clarity on existing gaps in knowledge and understanding of how interventions may mitigate ICUAW. Better understanding of such gaps, along with consideration of new study designs, can be important steps forward to improve the short term and long term outcomes of ICUAW.
Questions for future research
Who should be the target population for early rehabilitation interventions?
How do we validly and reliably measure “intensity” of interventions in patients who are critically ill?
How should we define the control group in early rehabilitation studies?
What outcomes should be prioritized, how should we measure them, and when should they be assessed?
How patients were involved in creation of this article
A patient, who was in the medical intensive care unit (ICU) at Johns Hopkins Hospital for 36 days and received early rehabilitation interventions, reviewed this article. She provided comments about how this content related to her experience in the ICU. The following are direct quotes from the patient’s perspective:
“I came into the hospital with weakness that was caused by sepsis and organ failure (kidney, spleen, and liver), so it was hard to differentiate which was weakness from that and what was additional weakness that was acquired during my intubation.”
“…the early intervention and the progressive mobility approach you discuss [in this paper] was very successful for me, especially combined with the occupational therapy. Specifically, the occupational therapy focused on the fine motor skills (opening bottles, caps, etc), which was as important as the physical therapy addressing the weakness of major muscle groups.”
“I used the tilt bed frequently during my ICU stay (fig 4)—very effective in gaining strength, feeling oriented again with my body.”
“…the frequency of my therapy (six days a week) was critical in improving ICU acquired weakness and helped me progress more quickly.”
“There are long term cognitive impacts of ICU acquired weakness (I still get very vivid dreams of my body feeling the weakness I felt both prior to the hospital, as well as the weakness I felt doing therapy).”
A video of patient and her family discussing ICUAW and rehabilitation can be found at https://www.youtube.com/watch?v=kxQhDJPj48U.
Acknowledgments
We acknowledge the peer reviewers, who provided very helpful feedback to strengthen the paper: Jennifer Krall, Nicola Latronico, Marc Moss, and Zudin Puthucheary.
Footnotes
Series explanation: State of the Art Reviews are commissioned on the basis of their relevance to academics and specialists in the US and internationally. For this reason they are written predominantly by US authors
Contributors: All authors participated in conception, selection of included articles, drafting the manuscript, and revising the manuscript for important intellectual content. SLH is the guarantor.
Competing interests: We have read and understood the BMJ policy on declaration of interests and declare the following interests: DMN is a principal investigator on an NIH funded, multicentered randomized trial (R01HL132887) evaluating nutrition and exercise in acute respiratory failure and, related to this trial, is in receipt of an unrestricted research grant and donated amino acid product from Baxter Healthcare Corporation and an equipment loan/donation from Reck Medical Devices and EnableMe; CLH is a principal investigator on a MRFF funded, multicenter randomized trial evaluating rehabilitation in patients on ECMO (NCT05003609); she led the TEAM trial which was an international trial of early rehabilitation in ICU (https://www.nejm.org/doi/full/10.1056/NEJMoa2209083) and is funded by an NHMRC investigator grant; PN has received travel funds for lectures about early rehabilitation by professional medical societies such as German Society of Intensive and Critical Care, and German Society for Critical Care Nursing.
Provenance and peer review: Commissioned; externally peer reviewed.