Purpose To investigate the effect of functional electrical stimulation-assisted cycle ergometry (FES-cycling) on muscle strength, cognitive impairment and related outcomes.
Methods Mechanically ventilated patients aged ≥18 years with sepsis or systemic inflammatory response syndrome were randomised to either 60 min of FES-cycling >5 days/week while in the intensive care unit (ICU) plus usual care rehabilitation versus usual care rehabilitation alone, with evaluation of two primary outcomes: (1) muscle strength at hospital discharge and (2) cognitive impairment at 6-month follow-up.
Results We enrolled 162 participants, across four study sites experienced in ICU rehabilitation in Australia and the USA, to FES-cycling (n=80; mean age±SD 59±15) versus control (n=82; 56±14). Intervention participants received a median (IQR) of 5 (3–9) FES-cycling sessions with duration of 56 (34–63) min/day plus 15 (10–23) min/day of usual care rehabilitation. The control group received 15 (8–15) min/day of usual care rehabilitation. In the intervention versus control group, there was no significant differences for muscle strength at hospital discharge (mean difference (95% CI) 3.3 (−5.0 to 12.1) Nm), prevalence of cognitive impairment at 6 months (OR 1.1 (95% CI 0.30 to 3.8)) or secondary outcomes measured in-hospital and at 6 and 12 months follow-up.
Conclusion In this randomised controlled trial, undertaken at four centres with established rehabilitation programmes, the addition of FES-cycling to usual care rehabilitation did not substantially increase muscle strength at hospital discharge. At 6 months, the incidence of cognitive impairment was almost identical between groups, but potential benefit or harm of the intervention on cognition cannot be excluded due to imprecision of the estimated effect.
Trial registration number ACTRN 12612000528853, NCT02214823.
- critical care
Data availability statement
Deidentified participant data are available on reasonable request to Associate Professor Sue Berney (https://orcid.org/0000-0003-1633-805X).
Statistics from Altmetric.com
What is the key question?
Does the addition of functional electrical stimulation-assisted cycling (FES-cycling) to usual care rehabilitation improve strength at hospital discharge; and reduce the incidence of cognitive impairment at 6-month follow-up in mechanically ventilated patients with sepsis or systemic inflammatory response syndrome?
What is the bottom line?
The addition of FES-cycling to usual care rehabilitation in these study site intensive care units (ICUs), which are experienced in providing physical rehabilitation, early use of FES-cycling did not substantially improve muscle strength at hospital discharge, or cognitive impairment, physical function and patient-reported outcomes at 6 and 12 months follow-up.
Why read on?
Usual care rehabilitation practised in ICUs with a strong rehabilitation culture and completed by experienced health professionals achieved outcomes that were favourable when compared with previous interventional trials in similar populations.
Survivorship after critical illness is characterised by significant morbidity, including muscle wasting and weakness and related physical impairment.1 The genesis of muscle wasting starts early with impaired muscle protein synthesis2 associated with systemic inflammatory conditions such as sepsis.3 Substantial muscle wasting occurs,2 resulting in an estimated 40% incidence of intensive care unit (ICU)-acquired weakness (ICU-AW).4 Delirium is also common5 and an independent risk factor for cognitive impairment at 6–12 months follow-up.6 In two randomised controlled trials (RCTs), rehabilitation started within 72 hours of mechanical ventilation improved physical performance and delirium.7 8 However, subsequent RCTs, with substantial heterogeneity in the timing and type of rehabilitation interventions, have demonstrated variability in results, with some evidence indicating that earlier start of intervention may improve patient outcomes.9 Generally, non-volitional rehabilitation interventions, such as in-bed cycle ergometry (cycling) and neuromuscular electrical stimulation (NMES), are most feasible to start within 72 hours of the onset of mechanical ventilation.9
In critically ill patients, cycling can improve quadriceps strength and physical functioning at hospital discharge,10 preserve muscle mass11 and better regulate oxidative stress12 and inflammatory pathways.11 An RCT of cycling, NMES and early rehabilitation, delivered as separate interventions, improved function at ICU discharge,13 but another RCT of cycling and NMES, delivered separately, demonstrated no improvement in muscle strength at hospital discharge.14 However, in these RCTs, cycling and NMES were not combined as a synchronised functional activity (ie, functional electrical stimulation (FES)-assisted cycling). In mechanically ventilated patients, FES-cycling can produce a greater physiological exercise response compared with either cycling or NMES delivered alone,15 enabling non-volitional resistance exercise that may stimulate muscle protein synthesis. Our preliminary data16 suggested a potential benefit of FES cycling on quadriceps muscle strength and delirium in critically ill patients. Hence, our objective was to conduct a multicentred RCT to investigate the effect of early implementation of FES-cycling plus usual care rehabilitation versus usual care rehabilitation alone on physical and cognitive outcomes in critically ill patients.
We conducted a randomised, parallel-group, allocation-concealed, assessor-blinded, controlled trial in ICUs at four hospitals in Australia and the USA.
Trial reporting follows Consolidated Standards of Reporting Trials (CONSORT) guidelines and the intervention is reported according to the Template for Intervention Description and Replication (TIDieR).17
Mechanically ventilated patients in an ICU were eligible if they were >18 years of age, had sepsis or severe sepsis18 (later expanded to include systemic inflammatory response syndrome (SIRS), as described in online supplementary 1), and were expected to require mechanical ventilation ≥48 hours and an ICU stay >4 days post randomisation. Participants were not eligible if they did not meet safety criteria to exercise (online supplementary table 1) within 72 hours of meeting inclusion criteria; had a primary neurological diagnosis and were not expected to survive to ICU discharge. Enrolled participants were not eligible for cognitive outcome assessment at 6-month follow-up if surrogates reported a score of >3.3 on the Informant Questionnaire on Cognitive Decline in the Elderly Short Form or a score of >10 on the Alcohol Use Disorders Identification Test prior to randomisation. For further details of eligibility criteria, see online supplementary 1.
Intervention and control group
The intervention commenced as early as possible, subject to safety guidelines (online supplementary table 1). FES-cycling involved synchronised stimulation of four muscle groups. Among intervention participants, one leg was randomised to FES-cycling and the other leg to cycling without FES. The intervention was performed for up to 60 min/day, for ≥5 days/week (figure 1). This intervention was provided in addition to usual care rehabilitation (online supplementary 1).
The control group received ‘usual care’ rehabilitation in accordance with existing local practices, delivered by experienced critical care physiotherapists (PT).
To aid in standardised reporting, usual care rehabilitation in both the intervention and control group was mapped to the ICU mobility scale.19 After ICU discharge, both groups continued to receive usual care rehabilitation on the hospital ward. PTs documented time spent directly engaging in FES-cycling and usual care rehabilitation, excluding time required for set-up, patient rest or recovery or passive activities.
Primary outcomes—quadriceps muscle strength and cognitive impairment
Two primary outcomes were designated, on an a priori basis, as part of the original national grant application: quadriceps muscle strength at hospital discharge and the prevalence of cognitive impairment at 6 months (online supplementary 1). Quadriceps muscle strength was reported as maximal isometric torque, measured in Newton metres (Nm), calculated as the product of: (1) muscle force (in Newtons (N)) from a hand held dynamometer and (2) the participant’s leg length from the tibiale mediale to sphyrion (m). For analysis purposes, in both the intervention and control groups, the highest of three strength measurements was used for each leg.
Cognitive function was evaluated via a battery of tests validated and extensively used in prior studies of ICU survivors20–23 (online supplementary 1). Cognitive impairment was defined as one cognitive test score >2 SD below population norms or two tests with a score >1.5 SD below norms, as done in previous studies.20 22 23
Secondary outcomes included: all-cause mortality, incidence and duration of delirium, manual muscle testing, hand grip strength, Physical Function in ICU Test scored; Functional Status Score for the ICU; the Short Physical Performance Battery; 6 min walk test; Katz Index of independence in activities of daily living, Lawton’s instrumental activities of daily living; Hospital Anxiety and Depression Scale; Impact of Events Scale-Revised; Short Form Health Survey SF-36 v2 and 5-level EQ-5D version (online supplementary 1; online supplementary table 2).
At baseline (prior to randomisation) and hospital discharge, ultrasound measurement of the cross-sectional area (CSA) of the rectus femoris was performed. For analysis, the mean value of three CSA measures (mm2) for the FES, cycle and control legs was used24 (online supplementary 2).
A sample size of 40 participants per group was calculated based on published data10 from a prior RCT of cycling in the ICU demonstrating a mean difference in quadriceps strength of 20 N, a within-group SD=31 (N) assuming 80% power and α=0.05. For the cognitive impairment outcome, a sample size of 46 participants per group was calculated based on a between-group difference in proportions of 0.26 (0.36 in control and 0.10 in intervention)20 evaluated using the same test battery. Attrition from mortality and loss to follow-up, occurring, from randomisation to the 6-month follow-up, was estimated at 40%25 yielding a final sample size of 154 (72 participants per group).
Randomisation and blinding
Random allocation (1:1 ratio) was performed via a random number generator, stratified by study site, with a random block size. Randomisation was concealed via use of opaque sealed envelopes (online supplementary 1). It was not possible to mask the participants or ICU clinical team, but all study outcomes measured at hospital discharge and 6 and 12 months follow-up were performed by blinded assessors.
Statistical analyses were conducted according to an a priori statistical analysis plan (online supplementary 2). Outcomes were assessed among survivors, without imputation of outcomes for participants who died during their ICU stay or prior to the 6-month follow-up, as a modified intention-to-treat analysis. For missing data, multiple imputation was performed with 200 data sets using the multivariate imputation using chained equations package in R statistical software for both primary and secondary analyses.
The primary outcome of muscle strength was evaluated using linear mixed models with the participant as a random effect and the randomised treatment allocation as a fixed effect (FES-cycling leg, non-FES-cycling leg vs control). For the second primary outcome, cognitive impairment, differences across the groups were based on the Fisher’s exact test, with the OR comparing the treatment with control group computed via logistic regression. As an a priori decision, analyses of differences between the randomised groups in strength and other physical secondary outcomes were adjusted for the following three baseline covariates: Charlson Comorbidity Index, Functional Comorbidity Index and baseline ultrasound measurement of CSA of rectus femoris. No adjustment for baseline covariates was conducted for any other outcomes. Secondary analyses of the primary outcomes included a per-protocol analysis as well as sensitivity analyses using intention-to-treat principle, that is, included all randomised participants, where the outcome for deaths was recorded as worse than any living patient, and distribution-free methods were used (online supplementary 2).
Secondary analyses were performed in accordance with the statistical plan and were considered exploratory (online supplementary 2). All-cause mortality was analysed using the log-rank test and Cox’s proportional hazards model.
No adjustment for multiple comparisons was made given that the interpretation of the impact of the intervention would be applied independently to each primary outcome.26
A total of 162 participants were randomised (80 intervention; 82 control; figure 2A). The number of randomised participants exceeded the original sample size (N=154), with ethics approval, due to greater-than-anticipated participant exclusion and attrition in reaching the sample size goal (N=92) for the co-primary outcome of cognitive impairment at 6-month follow-up. With 162 participants randomised, we exceeded the sample size goal (N=80) for the co-primary outcome of muscle strength; however, given funding limitations, the sample size for the 6-month cognitive outcome could not be reached (figure 2B). Demographics and ICU-related variables for the two groups are presented in table 1, and online supplementary tables 3 and 4.
The FES-cycling intervention was delivered on 511 of 894 (57%) scheduled treatment days, for a median (IQR) of 5 (3–9) sessions and 56 (34–63) min/session (online supplementary table 5). A visible or palpable contraction in the stimulated muscles occurred, on average (SD), in 91% (5%) of intervention sessions. FES-cycling was initiated a median (IQR) of 3 (2–4) days from time of intubation (online supplementary table 5). The most common reason for no intervention on the 383 scheduled treatment days was not meeting safety criteria (online supplementary table 6).
For the intervention and control groups in the ICU, respectively, usual care rehabilitation was provided on 309 of 909 (34%) and 336 of 901 (37%) days; initiated at a median (IQR) 5 (3–8) days and 4 (2–8) days from intubation; and had a median (IQR) duration of 15 (10–23) and 15 (8–15) min/day. Excluding those who died, 12 (15%) and 9 (11%) participants, respectively, did not receive any usual care rehabilitation in the ICU, with non-responsiveness (58%) as the most common reason (online supplementary table 7). On the ward post-ICU, usual care rehabilitation was provided on 423 of 703 (60%) and 453 of 835 (54%) days, respectively (online supplementary tables 8 and 9; online supplementary figures 1 and 2).
No serious adverse events were attributable to the intervention. Sixteen (1.7%) and nine (3.0%) adverse events occurred in the intervention and control group, respectively. All events resolved with interruption/cessation for the FES-cycling intervention or with return to bed for usual care rehabilitation, without clinical consequences. In the intervention and control group, respectively, death in the ICU occurred for 20 of 80 (25%) and 14 of 82 (17%) participants, respectively (p=0.456 for between-group comparison), with 17 of the 20 (85%) decedents in the intervention group never having receiving FES-cycling prior to death due to never meeting safety criteria (online supplementary table 1) to commence the intervention.
The between-groups adjusted mean difference (95% CI) in muscle strength was 3.3 (−5.0 to 12.1; p=0.460) Nm in the intervention versus control group, based on the multiple imputation analysis. In the per-protocol analysis, results were similar: 2.7 (−8.7 to 14.0; p=0.645) Nm.
At 6-month follow-up, 22 (52%) and 15 (38%) participants in the intervention and control groups, respectively, were alive, eligible for assessment (ie, not excluded due to potential pre-existing cognitive impairment), and contributed cognitive impairment data. In the intervention versus control group, cognitive impairment at 6 months occurred in 9 (41%) and 6 (40%) (complete cases analysis), with an OR (95% CI) of 1.1 (0.30 to 3.8; p=0.929) in multiple imputation analysis.
Sensitivity analyses demonstrated similar results for muscle strength and for cognitive impairment.
All-cause mortality was not significantly different in the intervention versus control group (p=0.483 using log-rank test in the intention-to-treat analysis; HR 1.2 (95% CI 0.7 to 2.0)). Median survival could not be estimated since survival was better than 50% in each group during the 12-month follow-up period.
For the intervention versus control group, the prevalence of ICU-AW at ICU discharge was 19% vs 26% (p=0.068). The unadjusted mean quadriceps muscle forces at hospital discharge for the FES-cycling leg (intervention group) and control group were 147±58 N and 136±58 N, respectively (results for the ‘cycle without FES’ leg are presented in online supplementary table 10). Delirium was present in 47 (59%) and 44 (55%) of participants for a median (IQR) of 2 (2–6) and 3 (2–6) days in intervention and control groups, respectively. Other outcomes at ICU and hospital discharge are reported in(table 2).
For secondary outcomes at 6 and 12 months (table 3), there were no statistically significant differences.
This RCT of mechanically ventilated patients with sepsis or SIRS demonstrated no benefit of FES-cycling plus usual care rehabilitation, compared with usual care rehabilitation alone, on quadriceps strength at hospital discharge, cognitive impairment at 6-month follow-up, or secondary physical and psychological outcomes evaluated in-hospital or at 6 and 12 months follow-up.
The FES intervention in this trial was developed based on existing literature as well as local pilot data according to Medical Research Council complex intervention frameworks, including clearly defined components, timing and delivery methods.27 We did not set out to investigate the mechanistic effects of the FES intervention such as time spent cycling actively versus passively, but to compare the complete intervention with a control group. We have reported on all aspects of our intervention according to current guidelines (TIDieR and CONSORT).
The application of FES in this study resulted in a higher percentage of muscle contraction (91%) than previously reported (25%–65%).28 Both studies titrated electrical current to achieve a muscle contraction; however, in the current study, Sequential Organ Failure Assessment scores were lower at <13.5 (previously identified as a threshold for muscle response to electrical current28) and the application of FES involves stimulating muscle in a functional task rather than a leg at rest with NMES. Differences in the presence of oedema may have also impacted the ability to visualise contraction in some participants.
Notably, the four study sites in Australia and the USA had established ICU rehabilitation clinical programmes, as needed to ensure that this complex FES-cycling intervention could be safely and effectively delivered. The usual care rehabilitation, delivered in both the intervention and control groups, occurred at a higher frequency than reported in international point prevalence studies by Sibilla et al,29 Jolley et al,30 Hodgson et al,31 Nydahl et al 32 and Berney et al.33 In our trial, reports of usual care rehabilitation in both groups included only active participation time, excluding all passive activities, set-up and patient rest/recovery time. Compared with control groups in prior trials, usual care rehabilitation in our control group was initiated a minimum of 2 days earlier34–36 and achieved higher ambulation rates (39% vs 25%).7 Indeed, in our trial, usual care rehabilitation, provided in both groups resulted in a lower prevalence of ICU-AW at ICU discharge compared with other trial control groups (26% vs 49%)37 and greater quadriceps force at hospital discharge (mean 138–147 N in our study vs 73 N38 and 105–106 N34). These results highlight both that usual care rehabilitation in the context of this trial was sufficient to produce improved outcomes compared with previous work and the importance of measuring and reporting usual care practices to appropriately interpret ICU rehabilitation trials.
We hypothesised that if FES-cycling reduced delirium, as demonstrated in our pilot data,16 there could be a reduction in cognitive impairment at 6-month follow-up. However, the intervention did not affect delirium in the ICU or subsequent cognitive impairment. The prevalence of delirium and cognitive impairment were similar to prior reports in mechanically ventilated patients.5 14 20 22 30 Notably, our study was underpowered despite recruiting more than the original sample size due to exclusion related to potential baseline cognitive impairment and attrition preventing assessment of the cognitive impairment primary outcome.
This study had several limitations. Similar to prior RCTs,34 36 the intervention was only delivered on 57% of scheduled days. Despite the study being conducted at hospitals experienced in ICU rehabilitation, and having safety guidelines (online supplementary table 1) permitting intervention with moderate-to-high levels of ventilator and/or vasopressor support as per published expert consensus,39 the severity of illness of some participants meant intervention could not be safely delivered, as demonstrated by 17 (85%) of 20 ICU deaths in the intervention group occurring before any intervention started. It is possible that there could be a systemic effect of FES cycling that may introduce confounding of our outcomes, although, if present, this should not have influenced our primary outcome of strength. Inability to complete outcomes assessments in all patients, especially for the 6-month cognitive impairment outcome, is common in many ICU trials,25 34 35 but can be a source of potential bias. The use of multiple imputation for missing data aimed to address this limitation. Notably, the number of participants providing data for the primary muscle strength outcome exceeded that sample size target. However, participants excluded from cognitive assessment, due to potential baseline impairment from pre-existing cognitive impairment (9%) and alcohol abuse (12%), was higher than anticipated based on prior studies (6%).6 The association between alcohol abuse and the incidence of sepsis may have contributed to this finding.40 Such exclusions were considered necessary to help ensure that the cognitive impairments detected at follow-up could be plausibly linked with the ICU stay and therefore potentially be impacted by the in-ICU intervention.
In this multicentre RCT of mechanically ventilated patients with sepsis or SIRS, the addition of FES-cycling to usual care rehabilitation did not substantially increase muscle strength at hospital discharge. The intervention did not reduce the incidence of cognitive impairment at 6-month follow-up but the estimate of the effect was imprecise and cannot definitively exclude benefit or harm of the intervention on cognitive impairment. These findings must be interpreted in the context of the study sites being experienced in providing ICU rehabilitation as part of usual care and the study participants’ muscle strength being greater than reported in prior studies.
Data availability statement
Deidentified participant data are available on reasonable request to Associate Professor Sue Berney (https://orcid.org/0000-0003-1633-805X).
The trial had ethical approval at all participating sites. Written informed consent was obtained from the participant or legally authorised representative (if participant was incapable, with continuation of consent obtained once capable). Study conduct was consistent with Good Clinical Practice Guidelines and the Declaration of Helsinki for protection of human subjects. This study was overseen by an independent data monitoring committee.
The authors would like to thank Ally Macdonell for trial coordination and site liaison, Elizabeth Hibbert, Hannah Verpuy and Alan Moss for early trial set-up and coordination, Jennifer Jones for data cleaning, all physiotherapists who provided intervention and usual care, trial personnel who provided outcome assessments and the members of the Data Safety Monitoring Committee (Professor David Berlowitz, Associate Professor Peter Bragge and Associate Professor Adam Deane).
DMN and LD are joint senior authors.
Twitter @SueBerney, @PastvaAmy
DMN and LD contributed equally.
Contributors All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by SB, JWR, IG, EC, DMN and LD. The first draft of the manuscript was written by SB and all authors provided critical review and revision. All authors read and approved the final manuscript. SB and LD are guarantors of this paper.
Funding National Health and Medical Research Council of Australia, the Mrs. Sheila S. Pakula and Dr Lawrence C. Pakula Patient Recovery Fund in Pulmonary and Critical Care Medicine at Johns Hopkins University, The American Thoracic Society Foundation Unrestricted Research Grant (AMP); Intensive Care Foundation of Australia and Austin Health Medical Research Fund. Restorative Therapies provided the RT 300 supine cycle ergometer to the Austin Health site.
Competing interests ZP reports personal fees from Faraday Pharmaceuticals, Lyric Pharmaceuticals, Fresenius Kabi, Nestle, Orion, GlaxoSmithKline, outside the submitted work.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.