Impact of Passive Leg Cycling in Persons With Spinal Cord Injury: A Systematic Review

We conducted a systematic review of the literature using PubMed, PEDro, Cochrane Library, Google Scholar, and EMBASE databases including all studies from the inception of the database until May 28, 2018. In the PubMed database, the following search strategy was used: “spinal cord injuries” [MeSH] AND “motion therapy, continuous passive” [MeSH] OR “passive cycling” [All Fields] OR “passive movement” [All Fields] OR “cycle ergometer” [All Fields]. A filter to only include studies with subjects over 18 years of age was applied: “Adult” [terms]. The search strategy was adapted to the other databases following the same principles to identify all relevant articles. We did not register the protocol for this systematic review. Corresponding authors were contacted to provide full texts when not accessible via electronic databases.

In the first phase, two reviewers (L.V. and C.P.) independently looked for the titles and abstracts of the articles retrieved. Articles that assessed passive leg cycling prospectively in adult participants (age ≥18 years) with SCI were included in this review. Studies with a retrospective design, single case studies, studies with healthy subjects or athletes, or studies using arm cycling were excluded. Next, both reviewers screened all remaining abstracts independently. Results of screening were compared, and any discrepancies between reviewers were discussed and a consensus-based decision was arrived at in consultation with a third reviewer (S.M.). Finally, full text screening was performed in a similar way (see Figure 1).

The study information was first independently extracted by two reviewers (L.V. and C.P.) using a standardized data extraction form. Then, individual cycling protocols among the included studies were summarized. The following data were extracted: study information year, author, study design, publication type, sample size, patient demographics (age, sex, and body mass index), completeness of injury, times since injury, functional status, parameters of intervention, and details of outcomes. The study outcomes were categorized into groups based on the body systems they targeted (cardiovascular, neurological, and musculoskeletal). When available, change scores and percent changes from baseline were calculated from the included studies so that they could be statistically analyzed. If pre/post outcome scores were only available, percent change from baseline was calculated using the formula: [(post - pre) / pre) × 100]. The extracted data are presented in Tables 1–3.

Effect size was calculated in all studies for the primary outcome measure pre and post passive cycling with Microsoft Excel using the following formulae7:

An effect size of 0.2 was considered small, 0.5 medium, and 0.8 and above large.8 

The Oxford levels of evidence (LOE) was used to indicate the methodological quality of the included studies.9 We assessed the risk of bias using the Study Quality Assessment Tools developed by the National Heart, Lung and Blood Institute10 to examine biases in the included studies. Sequence generation, allocation concealment, blinding, incomplete and selective outcome data, and other biases10 were assessed for each study by two reviewers (C.P. and S.M.) and discrepancies were resolved by consensus.

A summary of the studies is presented on Table 1. Two of the 11 studies were randomized controlled trials,11,12 one was a crossover trial,13 and eight studies were pre-post type including case series designs (see level of evidence categories in Table 1). The time since injury in persons with SCI tested in the included studies was at least 1 year post SCI and chronicity ranged from 1 to 25 years. Total number of subjects in the included studies was 179 (81 in the two RCTs and 98 in the non-RCT studies; see Table 1). The majority of patients (69%, n = 123/169) in the studies had a motor complete SCI (ASIA Impairment Scale levels A or B).

Summary of the cycling training parameters is presented in Table 2. Of the studies that examined effects of a single intervention, four studies tested the effects of <10 minutes of cycling duration: 5 minutes14 and 7 minutes6; and two studies tested the effect at 10 minutes15,16 of passive cycling. Four studies tested the effects of 20 to 100 minutes of a single cycling intervention (see Table 2). Multiple session studies ranged from 18 to 36 sessions of cycling intervention over a 6- to 12-week period. The pedaling frequency appeared to be lower for multiple session studies (range, 25–40 rpm) compared to single session studies (range, 35–60 rpm).

Six studies2,12,14–17 examined outcomes in the cardiovascular category, two studies11,18 used musculoskeletal outcome measures, and four studies6,11,13,19 used neurological outcome measures (see Table 2). Rayegani et al11 used outcomes in two categories – neurological and musculoskeletal. Cardiovascular outcomes measured during cycling were (1) blood volume pulse signals, (2) blood flow velocity, (3) oxygen uptake (VO2), (4) stroke volume, (5) cardiac output, (6) pulmonary ventilation, (7) heart rate, (8) mechanical efficiency, and (9) peripheral vascular resistance. Musculoskeletal outcome measures were (1) thigh girth, (2) joint range of motion, and (3) messenger ribonucleic acid (mRNA) to examine myosin heavy chain expression of the vastus lateralis muscle. Neurological outcome measures were (1) modified Ashworth Scale, (2) stretch reflex peak torque, (3) pendulum test, (4) spasticity perception, (5) H-reflex amplitude, and (6) short-interval intracortical inhibition. None of the included studies measured activity or participation as described in the ICF domains.

The impact of cycling on various outcome measures is summarized in Table 3. Although not all studies reported a statistical significance, seven out of nine studies reported large effect sizes and the other two reported moderate effect sizes (Tables 3A and B).

Single session studies: In general, studies showed a trend toward improvement in acute responses to passive cycling. Ballaz et al16 showed a medium effect size (Cohen's d = −0.75) increase in blood flow velocity and decrease in peripheral vascular resistance after cycling. Muraki et al15 (n = 4, Cohen's d = −1.5), Figoni et al14 (n = 30, Cohen's d = −1), and Ter Woerds et al2 (n = 8; Cohen's d = not available) also showed large effect sizes with the small increase in blood flow, stroke volume, cardiac output, and pulmonary ventilation. The fourth study by Tran et al17 did not show any significant difference in the markers of vascular resistance between healthy controls and persons with SCI.

Multiple session studies: Only one study by Ballaz et al12 showed a large effect size (Cohen's d = 1.01) in blood flow velocity and decrease in peripheral vascular resistance after cycling.12 

Single session studies: Overall, most studies found some form of improvement in the acute responses to passive cycling. Two studies found a large effect size with an improvement in outcomes such as short-interval intracortical inhibition (Cohen's d = −1.84) and spasticity (Cohen's d = 1.53) measured using the relaxation index and the modified Ashworth Scale.6,13 Kakebeeke et al19 found no change in spasticity measured by the peak knee torque evoked as a result of knee stretching; however, in this study, 6 of 10 participants reported subjective improvements in spasticity. None of the studies except one6 tested whether these acute responses to passive cycling persisted. Nardone et al6 showed that short-interval intracortical inhibition seen immediately after passive cycling had disappeared at 30 minutes post cycling.

Multiple session studies: One study found an improvement in spasticity measured using the pendulum test and the modified Ashworth Scale as well as H-reflex amplitude (Cohen's d for H-reflex = 1.22).11 None of the studies that tested cardiovascular, musculoskeletal, or neurological outcomes after multisession cycling training tested whether the benefits persisted for any duration after the end of training.

Single session studies: None of the studies that examined musculoskeletal outcomes were single session designs.

Multiple session studies: In general, there was a trend for improvement after passive cycling among the studies reviewed. Willoughby et al18 reported a medium effect size with an increase (Cohen's d = −0.79) in genetic expression of myosin heavy chain filament proteins (MHC types IIa and IIx), which are associated with fast-twitch muscle fiber type, after 24 sessions of passive cycling intervention. In addition, there was also a concomitant decrease in genetic expression of markers of protein degradation (ubiquitin). Thigh girth was reported to have not changed statistically after training, however a small positive change was noted.18 Rayegani et al11 reported an improvement in joint range of motion with passive cycling intervention with a large effect size (Cohen's d = −1.71) that occurred over a 2-month period of regular training.

The LOE in the studies included in this review ranged from levels 2 to 4. Two papers were randomized controlled trials (LOE 2),11,12 and the remaining studies were case series or case control type trials (LOE 4). Ratings of the risk of bias for individual studies are reported in Tables 4 and 5. Quality and the risk of bias of the studies ranged from poor to good, but were in general in the fair risk of bias category. Missing information that typically led to fair or poor ratings was small sample size, lack of prespecified inclusion criteria, lack of assessor blinding, lack of statistical analysis or p value reporting, lack of clarity in the eligibility criteria in the pre-post trials, lack of treatment allocation (so that assignments could not be predicted), lack of double-blinding, inadequate sample size, and lack of baseline between-group comparisons in the randomized controlled trials.

Rehabilitation after SCI requires addressing deficits across multiple body systems. In this systematic review, we examined the acute responses (single session) and effects (multiple session) of passive cycling intervention on cardiovascular, musculoskeletal, and neurological outcomes and summarized passive cycling training parameters such as intensity, duration, and type of passive leg cycling interventions. Our study builds on the previous systematic review that focused on neurological outcomes and animal studies4 by additionally reviewing cardiovascular and musculoskeletal effects of passive cycling in human studies. Two RCTs reported significant benefits of multiple sessions of passive cycling on cardiovascular (improved leg blood flow velocity), musculoskeletal (improved joint range of motion and markers of muscle hypertrophy), and neurological outcomes (improved spasticity and reflex excitability). No clear picture emerged with single session studies, with about half the studies showing a statistically significant improvement in acute responses in cardiovascular (blood flow velocity) and neurological outcomes (short interval intracortical inhibition and spasticity),6,13,16 while the rest reported no change. The studies reviewed here showed a diversity of cycling protocols and outcomes across multiple body systems.

Number of factors such as comfort level, tolerance, joint range of motion, stiffness, and ergometer motor capacity will determine the pedaling frequency. In the multiple session studies, frequency of the exercise sessions ranged from one to three times per day, 20 to 90 minutes per session, 2 to 6 days per week, and 6 to 12 weeks total duration. This large variation in training parameters makes it difficult to determine the optimal patterns for maximizing improvement. The mix of acute responses to a single session of passive cycling may be a result of the variation in the single session duration in these studies, which ranged from 5 to 100 minutes per session; however, we did not observe any patterns that would indicate that there was a relationship between length of the cycling session and statistically significant results. Thus, it is difficult to make conclusive statements about the nature of acute responses to a single session of passive cycling. It is important that future studies include a follow-up period after end of training to assess how long these effects can last. The duration of these effects will help in determining optimal training protocols to maximize the beneficial changes post training.

We identified three broad categories of outcome measures in this study – cardiovascular, neurological, and musculoskeletal – to assess the impact of passive cycling exercise. Although only a few outcomes were statistically significant, the moderate to large effect sizes warrant future larger studies.

Two parameters indicative of adaptive changes after SCI are the decreased blood flow to the lower limb muscles and increased peripheral vascular resistance.2,3 Thus, an improvement in these two parameters reported by Ballaz et al12,16 and Figoni et al14 after passive cycling is a desirable impact on the vascular system. Most studies that found decrease in peripheral vascular resistance used a more direct assessment of peripheral resistance by calculating the velocity index (difference in maximum and minimum blood velocity divided by the maximum velocity). Tran et al17 showed a nonsignificant 4% increase in peripheral vascular resistance (after 30 minutes of cycling); however, they used a different method of assessing peripheral resistance than the rest of the studies, which may explain the contradictory results. Tran et al17 used assessment of the dicrotic notch of the blood volume pulse (BVP) waveform, which is known to be influenced by physical stress even in healthy controls. It is possible that their findings are related to the duration of exercise (an indicator of physical stress), which was six times longer (30 minutes) than other studies and may reflect physical stress more than the peripheral vascular resistance. In the studies that showed no statistical change, the results may have been affected either by a small sample size2,15 or a shorter duration of cycling.14 Overall, it appears that passive cycling can improve blood flow to the lower limbs and can potentially help prevent and address conditions related to impaired blood flow such as skin breakdown lesions.20 Decreased circulation can also result in vascular disorders such as thrombosis and muscle atrophy,21 and thus passive cycling has the potential to induce preventive effects.

In addition to cardiovascular changes, three studies reviewed here found significant neurological benefits of passive cycling, namely decrease in spasticity and H-reflex amplitude and an increase in intracortical inhibition. A fourth study also examined a measure of spasticity (stretch-induced muscle torque) but did not find any significant changes.19 The lack of change in torque reported by Kakebeeke et al19 can be attributed to the slow stretch speed chosen (120 degrees per second) to assess spasticity. Knee velocity during walking can be up to 169 degrees per second in healthy controls and up to 153 degrees per second in persons with cervical SCI.22 Animal SCI model spasticity assessments indicate that spasticity can only be assessed at stretch speeds in a range of 350 to 612 degrees per second.23 Thus, it is possible that 120 degrees per second, certainly not a functional speed,22 may not have been fast enough to elicit a reflex response.23 Although, Rayegani et al11 reported a significant change in spasticity (with a large effect size), they did not report actual change scores. They also reported greater inhibition of H-reflexes, which are known to be hyperexcitable in the SCI population.24 Previous reports have indicated that H-reflex amplitude is correlated with functional independence and spasticity in the SCI population.24 Thus, these changes amount to positive changes with potentially meaningful functional impact on persons with SCI.

These results suggest that neurological outcomes may be more readily influenced after a single session and multiple sessions of passive cycling. The improvement in these outcome measures indicates that passive cycling can induce beneficial changes in the central nervous system. One of the underlying mechanisms of spasticity is thought to be increased sensitivity of the muscle spindles25 as well an increase in spinal excitability (higher H-reflex amplitude).26 Improvement in outcome measures such as modified Ashworth Scale and H-reflex amplitude are indicative of the positive effects of passive cycling. Cortical excitability is known to be higher after SCI and is associated with a decrease in intracortical inhibition.27 Thus, an increase in intracortical inhibition is an indicator of improved neuromodulation within the cortical neurons as a result of passive cycling. It is important to note that all these improvements in outcome measures such as spasticity,28 H-reflex,26 and cortical inhibition6 are all indicators of neuroplasticity associated with functional improvement after SCI.

Musculoskeletal improvements in range of motion after multiple sessions of passive cyclic movements suggest that passive movements may not only maintain but also improve joint mobility post SCI.29 In addition to overt musculoskeletal changes, one study examined the impact of multiple sessions of passive cycling on genetic expression of markers of muscle building and degradation.18 The results of the study by Willoughby et al18 are encouraging because they show an increase in the expression of fast-twitch myosin heavy chain and a concomitant decrease in genetic expression of markers of protein degradation (ubiquitin). These results suggest that passive cycling has the potential to attenuate muscle atrophy18; however, further studies are needed to determine whether a passive cycling intervention can protect muscle size. As anticipated, none of the single session studies assessed musculoskeletal outcomes because these changes most likely require a longer period of training.30 

The common mechanism likely involved in the circulatory effects seen after passive cycling is the mechanical effect of rhythmic cycles of stretch and shortening. Large muscles like plantar flexors and quadriceps can exert mechanical pumping action as well as stretch reflex activation to trigger muscle pump and improve venous return.31,32 Peripheral vasculature resistance is prone to decrease with a decrease in movement typically seen after SCI. It appears that even a single session of cycling can elicit an acute response of enhanced blood flow and improved circulation through an increase in arterial diameter possibly because of preserved endothelial function in arteries below the lesion. Range of motion probably improved as a result of decrease in hyperreflexia5 as well as decrease in muscle and tendon stiffness.33 Increased genetic expression of myosin heavy chain filament proteins and decreased expression of ubiquitin suggests that passive cycling may have upregulated pathways for muscle hypertrophy, which could help in the prevention of muscle atrophy and stimulate hypertrophy in people with SCI.34 

Majority of the studies reviewed here were nonrandomized studies with a fair risk of bias, and the results must be cautiously interpreted. This limitation should be weighed against the large effect sizes seen in the studies reviewed here. In addition, it is difficult to ascertain the clinical meaningfulness of the benefits of passive cycling reported here. The protocols used for cycling were diverse and so were the outcome measures precluding meta-analyses. In addition, all outcome measures were limited to the body structure and function category of the International Classification of Functioning, Disability and Health35 with no outcomes in the activity or participation domains. The cutoff values for a clinically meaningful difference in the outcome measures after passive cycling in the studies reviewed here have not been established. Most outcome measures reported in this study were research tools but not routine clinical measurements; future studies need to examine the impact of passive cycling on assessments routinely used in the clinical environment. Finally, none of the studies barring one6 examined the duration of the effects of passive cycling lasted.

Evidence from this systematic review indicates that multiple sessions of passive leg cycling showed some form of beneficial changes across cardiovascular, musculoskeletal, and neurological systems in persons with SCI, based on two RCTs. More evidence is required to understand the best parameters for single and multiple sessions of passive cycling. More randomized controlled trials are needed to systematically understand the effects of passive cycling in persons with SCI. We recommend that future studies not only assess outcomes for body function and structure, but also include outcomes for activity and participation and include both objective as well as patient-reported outcomes.