Introduction

Although a growing number of studies suggest that exercise can improve physical fitness in people with spinal cord injury (SCI), currently there are no rigorously developed, clinical practice guidelines for prescribing exercise to this population. Without guidelines, exercise prescription and promotion are challenging. High-quality guidelines are developed using systematic methods that combine rigorous methodology with the meaningful engagement of a multidisciplinary team of stakeholders.1 Methodologically, guidelines begin with a clinical practice question that informs a systematic review of relevant research evidence. In a systematic review, procedures used to identify studies for inclusion are clearly stated and reproducible. An appraisal of the quality of evidence accompanies the evidence synthesis. This evidence base provides the foundation for the guideline team to make recommendations.1

Two systematic reviews have examined the effects of exercise on physical fitness, with the intention of providing recommendations for prescribing physical activity to people with SCI. The first2 aimed at identifying the characteristics of training regimens associated with physical capacity changes. Virtually all of the included studies utilized upper-body training. Although the authors did not formally evaluate study quality, they did report tremendous variability in the training regimens utilized across 25 included studies. Based on their review, they recommended 30 min of rhythmic exercises, 3 days per week at 70% of the maximum heart rate (HR) for 8 weeks, to improve cardiovascular endurance.

The second systematic review3 focused on determining the effects of different modes of upper body exercise on physical capacity, reflected by peak oxygen uptake and power output. Methodological quality was evaluated for each of the 25 studies included. The authors concluded that upper-body exercise may increase physical capacity among people with SCI. However, given the overall low quality of evidence, it was impossible to generate evidence-based conclusions regarding the most effective training modes or the relative effects for those with paraplegia versus tetraplegia.

Although these systematic reviews contribute to the evidence base for developing exercise guidelines, they are restricted to one aspect of fitness—physical capacity—and, for the most part, upper-body exercise training programs. No systematic review has examined the effects of all types of exercise training on all four components of physical fitness: physical capacity, muscular strength and endurance, body composition and functional performance. Given the importance of all four fitness components for health, independence and quality of life (QoL),4 it is important to determine whether exercise can improve each of these in people with SCI.

The present systematic review was undertaken as part of a larger project to develop exercise guidelines for people with SCI. In 2009, a guideline development team of stakeholders was assembled in Canada. A subgroup undertook a systematic review to address the following questions: ‘Can exercise improve each fitness component?’ and, if so, ‘What types of exercise improve each component?’ A second objective was to catalog data regarding the effects of different exercise prescriptions on fitness outcomes. Together, this information would provide the evidence base for the team's subsequent deliberations and recommendations in the guideline development process.5

Methods

Scope of the review/study inclusion criteria

The review focused on published, English-language studies of the fitness benefits of a physical activity or exercise training intervention in persons diagnosed with an SCI (paraplegia or tetraplegia). Studies had to include at least one of the following fitness measures in the analyses: strength, oxygen uptake/consumption, power output, peak work capacity, body composition, exercise performance or functional performance. Case studies, experimental and quasi-experimental designs were included.

Literature search strategy

Two authors (BF, DLW) conducted searches in the following electronic bibliographical databases:

  • MEDLINE (1950–March 2010, OVID Interface);

  • PsycINFO (1840– March 2010, Scholars Portal Interface);

  • EMBASE (1980–March 2010, OVID Interface);

  • CINAHL (1982–March 2010, OVID Interface);

  • SPORTDiscus (–March 2010).

Medical Subject Headings included the following search terms (in combination with ‘spinal cord injury’ OR ‘tetraplegia’ OR ‘paraplegia’): exercise, physical activity, training, exercise conditioning, physical fitness, exercise prescription, adaptation, effect, benefit, strength, aerobic capacity, endurance and body composition.

Screening

Figure 1 shows the flow of articles through the search and screening process. After removing duplicate citations, two authors (BF, KAMG) independently scanned the title and abstract of each citation (n=3226) to determine its suitability. Eighty-one articles were retained. Additional articles were identified from references in other review articles (n=69) and through hand searches of the authors’ personal databases (n=16). These 166 articles were read by BF and KAMG (unblinded) to determine whether they met the inclusion criteria. Discussion with a third author (ALH) ensued in cases of disagreement. Full (100%) consensus was achieved for all citations. Eighty-four articles were excluded after review (reasons available from the first author). Eighty-two studies were retained for review, divided into studies involving people with acute SCI (<1 year post injury; n=13) and studies involving people with chronic SCI (>1 year post injury; n=69).

Figure 1
figure 1

Literature search strategy.

Data extraction

Two authors (BF and CP) completed data extraction forms for the 82 studies. Extracted data were verified by two additional authors (ALH and KAMG), and included study design, participant characteristics, methodology, outcomes related to our defined scope and conclusions. Reviewers were not blinded to the journal or authors.

Assessment of evidence quality

The quality of each study was determined using the procedures outlined in SCIRE (Spinal Cord Injury Rehabilitation Evidence).6 Specifically, randomized controlled trials (RCTs) were evaluated using the 10 internal validity items of the Physiotherapy Evidence Database (PEDro) tool. Consistent with the quality-assessment approach used in Valent et al.'s review,3 two items—‘blinding of all therapists’ and ‘blinding of all subjects’—were considered irrelevant when comparing a training group with a no-exercise control group; such studies were credited the two points for these items. The maximum PEDro score is 10. Higher scores reflect better-quality trials. All other studies were evaluated using a modified Downs and Black scale,7 identical to that used by the SCIRE process, with a maximum score of 28. Higher scores indicate better methodological quality. Articles were independently evaluated by three raters (DLW, BF and CP). Any scoring discrepancies were resolved through discussion. The level of evidence associated with each study was then coded using the SCIRE system, which is a 5-level system that distinguishes between studies of differing quality and incorporates the types of research designs common in SCI rehabilitation research (Table 1).6

Table 1 The levels of evidence-scaling criteria applied to the articles

Results

Extracted data and quality assessment for each study are presented in Tables 2 and 3. Evidence summaries are presented in Tables 4 and 5. Studies of acute (12 months post injury) and chronic patients (>12 months post injury) are summarized separately. These groups have been divided based on accepted knowledge that the greatest degree of functional recovery occurs in the first 12 months following SCI. Owing to the limited evidence base, studies and results have not been divided based on level of injury or completeness of injury.

Table 2 Evidence from acute studies
Table 3 Evidence from chronic studies
Table 4 Summary of studies and evidence for each fitness component (acute spinal cord injury)
Table 5 Summary of studies and evidence for each fitness component (chronic spinal cord injury)

Studies of acute SCI patients

Physical capacity

Decreased physical capacity is common post SCI and can be attributed to decreased sympathetic drive, muscle atrophy, loss of motor control and relative inactivity.8, 9 Six training studies examined changes in exercise capacity using measures of aerobic capacity and power output.

Aerobic capacity: Increases in peak oxygen consumption (VO2peak) following training is a consistent finding across the level 1 and level 4 studies. One study10 noted significant increases in VO2peak following 6 weeks of wheelchair interval training 3 × per week. Significant increases were also observed following a 6-week arm ergometry training program.11 Participants training at a high intensity (70–80% HR reserve) showed a significantly greater improvement in VO2peak compared with those training at a low intensity (40–50% HR reserve) over 8 weeks.12 Injury level seems important in predicting change in VO2max. Hjeltnes et al.13 found participants with paraplegia had a 28% increase in peak oxygen uptake throughout the 4-month training period, but there was no change in those with tetraplegia.13

Power output: Maximal power output achieved during a maximal exercise test is a component of physical capacity that provides an indirect indicator of muscle strength and aerobic capacity. Arm ergometry training programs of 6 and 8 weeks duration have produced significant improvements in maximal power output, regardless of training intensity.11, 12 The addition of arm ergometry training to inpatient rehabilitation has been shown to significantly increase peak power in patients with paraplegia and tetraplegia.13 Mixed exercise programs, including strength, aerobic and mobility training, significantly increased peak power output following 16 weeks of training.14 Wheelchair ergometry has also been shown to have a similar effect, resulting in significant increases in peak power output following 6 weeks of training.10

Muscle strength

Decrease in muscle strength—the result of pronounced muscle atrophy and decreased neural drive—is one of the most significant and rapidly occurring consequences post SCI. Four studies included a measure of muscle strength. Significant increases in weight lifted and number of repetitions completed during various upper-body exercises were observed following a 16-week mixed exercise training program that incorporated mobility, strength and aerobic training.14 The addition of 30 min of arm ergometry, 3 × per week, to inpatient rehabilitation programs also resulted in significant increases in muscle strength scores by more than 6 points in five upper-extremity muscles.13

Body composition

In people with acute SCI, lean mass is only 60–65% that of able-bodied controls and body fat can increase to levels corresponding to 100–113% of that of controls.15 Reduced muscle mass and increased fat mass are both considered risk factors for secondary health complications and chronic disease.

Body weight: Only three level 4 studies examined body weight as an outcome. One study16 reported no significant changes in body weight or body mass index following a 6-week arm ergometry program. Likewise, an arm ergometry program during 4 months of inpatient rehabilitation had no significant effect on body weight.13 Similar results were noted following a 16-week mixed exercise program.14 When interpreting these non-significant changes, it is important to note that the maintenance of body weight for people who are wheelchair dependent is an important clinical finding given the fact that obesity is a significant secondary complication in this group.17

Lean (muscle) and fat mass: Two level 4 studies examined body composition changes. Giangregorio et al.15 examined the effects of a 2 × per week body-weight-supported treadmill training (BWSTT) program and reported increases in thigh and calf muscle cross-sectional area (4–57%) and increased leg fat. Another study involving arm ergometry showed no significant changes in lean or fat mass.16

Functional performance

Post SCI, people often lack sufficient fitness to perform basic activities of daily living.4 Poor functioning can compromise independence and QoL.4

Wheelchair skills: In a level 4 study of acute and chronic SCI patients, time to complete various wheelchair skills significantly decreased following an exercise program with mobility-, strength- and aerobic-training components.14 Similar results were reported in a level 5 study that used a primarily aerobic-based training program.18

Walking measures: Level 1, 4 and 5 studies have evaluated the effects of BWSTT in people with acute SCI, with inconsistent results for walking outcomes. A single level 1 study demonstrated improvements on a variety of clinically relevant walking outcomes,17, 18 supported by level 4 studies showing increased gait speed.13, 16 However, improvements on these outcomes were not significantly different as compared with conventional intensive overground locomotor training. Non-significant improvements in walking cadence,19 stride length19 and percentage of body-weight support required15 have also been reported.

Studies of people with chronic SCI

Physical capacity

Physical capacity—reflected in measures of power output and aerobic capacity—is relevant to the independence, health and QoL of individuals with SCI. Power output, for instance, can impact the ability to effectively perform transfers and propel one's wheelchair on inclines and other challenging surfaces. Aerobic capacity is pertinent for cardiovascular health, functional independence and fatigue resistance. Forty-six studies have investigated the effects of exercise on power output and aerobic capacity. Although the magnitude of improvements varies across studies, exercise appears to be effective in improving these fitness outcomes.

Power output: Sixteen studies reported on changes in power output. Most were level 4 studies. However, there were two level 1 RCTs and a level 2 non-randomized trial. Taken together, the 16 studies provide strong evidence that combined resistance and aerobic exercise, and functional electrical stimulation (FES)-assisted exercise, produced significant improvements in power output. Most studies prescribed exercise 3 × per week for 6–12 weeks, although two studies showed particularly large improvements in lower limb power output following 1 year of FES cycling, 3–5 × per week.20, 21 Unfortunately, it is difficult to pinpoint the exercise intensity required for improvements because the majority of studies (11 of 16) employed FES exercise; stimulation parameters vary widely among these studies and progression typically occurs as individually tolerated. However, moderate-intensity arm ergometry exercise (70–80% max HR or 50% HR reserve) combined with resistance training (progressed to repetitions with 70–80% of the one repetition maximum) seems sufficient for long-term improvements. When max HR cannot be confidently estimated (for example, in individuals with cervical injuries and thus, blunted HR responses to exercise), exercise intensities of 3–4 on the Borg scale of perceived exertion may be used. Of note, one level 1 RCT showed resistance and aerobic exercise performed 2 × per week at a moderate intensity led to significant increases in power output over a 39-week study.22

Aerobic capacity: Thirty studies reported on changes in aerobic capacity; virtually all were level 4 or 5 studies. Across these studies, it was clear that FES exercise of various forms (cycling and ambulation), as well as arm ergometry and wheelchair exercise, produced significant improvements in aerobic capacity. One level 4 study showed that 3 × per week circuit resistance exercise combined with arm ergometry improved aerobic capacity following 12 weeks of training.22 Most of the studies prescribed exercise 3 × per week for 6–12 weeks, although studies employing as little as 4–5 weeks of exercise have also shown improvements.23, 24, 25 Again, because almost half of these studies (13 of 30) employed FES exercise, it is difficult to discern the exercise intensity required for improvements in aerobic capacity. It is evident that moderate arm ergometry or wheelchair exercise at an intensity between 60–80% of max HR, or 60–65% VO2peak, seems sufficient to improve aerobic capacity.

There are a few important considerations when prescribing exercise to improve power output and aerobic capacity in individuals with SCI. First, when deciding between voluntary versus stimulated exercise, the potential for arm ergometry to improve power output and aerobic capacity depends in large part on the motor function in the upper limbs. For those with little to no arm function, a stimulated form of exercise such as FES is probably more effective. Second, when employing FES exercise, it is still unclear as to which form (for example, cycling, ambulation) is the most effective in improving aerobic capacity, although all have shown benefit.

Muscle strength

Muscle strength is a highly relevant fitness outcome in the chronic SCI population, as improvements in strength will have a significant impact on the ability to perform activities of daily living (for example, transferring, wheeling). If increased strength is associated with increased muscle mass, such changes can also have metabolic benefits. Numerous studies have evaluated the efficacy of various exercise-training protocols for improving muscle strength. These protocols can be categorized as ‘voluntary’ strength training (that is, using non-paralyzed muscles), and those utilizing some form of electrically stimulated exercise of paralyzed muscles.

Voluntary strength training: Eleven studies examined changes in strength following training of non-paralyzed muscles. With the exception of a single level 1 RCT, all were level 4 studies. Across studies, it was clear that the muscles responded to training in a similar manner as would be expected in the able-bodied population. Specifically, circuit resistance training paradigms, BWSTT, arm ergometry and training with specialized equipment (for example, shoulder kayak ergometry) significantly increased muscle strength of trained limbs after as little as 5 weeks. While the majority of studies used training volumes incorporating 3 × per week frequencies, significant strength changes were demonstrated with 2 × per week training in a high-quality RCT22 and a low-quality pre-post study.26 For studies employing resistance training (for example, lifting weights), training intensities varied between 50–80% of the one repetition maximum. There was clear evidence from the level 1 RCT that twice-weekly strength training employing 2–3 sets at 70–80% of the one repetition maximum is effective in increasing voluntary muscle strength.22

Functional electrical stimulation: FES is traditionally used in the SCI population to activate muscles that can no longer be fully activated voluntarily. Of the 11 FES studies reviewed, one was level 2 and the remainder were level 4 studies. There was a wide variety in the type of FES training employed (cycling, walking and resistance training) across studies. Most studies reported significant increases in leg (knee extensors, hip flexors and/or knee flexors) or arm strength following training programs of 6–12 weeks. Training frequency in the majority of studies was 3 × per week. Despite the consistent finding that FES-assisted training of the paralyzed musculature enhances strength, the heterogeneity in FES training modes makes it impossible to comment on the intensity of training needed to elicit improvements. It should also be noted that FES is not tolerated equally by all individuals with SCI, so it may not be appropriate for everyone.

Body composition

Body composition measures can indirectly provide information about fitness status (that is, muscle strength, oxygen uptake). A combination of paralysis and inactivity often leads to an increase in body weight and fat mass, and a decrease in lean tissue mass in people with chronic SCI. This increases the risk of developing secondary health complications and can decrease QoL.

Body weight: Nine studies assessed changes in body weight. All were level 4 or 5 studies. Training protocols included resistance and aerobic exercise,14, 27, 28, 29 FES-assisted exercise30, 31 and BWSTT,32, 33 all performed at least 3 × per week. None of the level 4 studies reported significant changes in body weight.

Lean (muscle) and fat body mass: In all, 19 training studies reported changes in lean mass and 7 reported changes in fat mass. All were classified as level 4 or 5, with the exception of one level 2 trial. Training programs involving BWSTT,32, 34, 35 FES cycling,30, 31, 36, 37, 38 FES ambulation,39, 40 neuromuscular electrical stimulation resistance training41 and vibration exercise42 produced significant increases in muscle mass, with training frequencies ranging from 2–7 × per week, for 8–52 weeks duration. Of note, the level 2 trial showed significant increases in quadriceps muscle mass after 26 weeks of FES-assisted treadmill training, 2 times per week for 20 min per session.40 The majority of studies found no significant decreases in fat mass post training. However, in two lower-quality studies (Downs & Black score 12), significant reductions in fat mass were observed after FES cycling performed 7 × per week for 8 weeks36 and 5 × per week for 32 weeks.31

Functional performance

Measures of functional performance have included tests of wheelchair skills and propulsion, walking and standing. Improvements in these outcomes can translate into increased independence and QoL.

Wheelchair skills and propulsion: These nine studies were all classified as level 4. All showed significant improvements in their respective measures, including time taken to perform selected wheelchair skills, peak power output on a wheelchair ergometer, propulsion speed, distance traveled in 12 min and propulsion time to exhaustion. Training regimens consisted of either arm, wheelchair or rowing ergometry. One study also incorporated stretching and strengthening exercises.43 Participants exercised 3 × per week in all but one study that prescribed exercise 5 × per week.44 Sessions ranged from 5 to 45 min. Only one study had a group perform low-intensity exercise.45 Moderate- to heavy-intensity exercise was prescribed in all other studies.

Walking: Eleven studies reported on treadmill walking parameters (for example, speed, percentage of body-weight support required) and five reported independent, overground walking performance. All studies were level 4, except for one level 1 RCT46 and a level 5 case study.32 Significant changes in treadmill walking parameters emerged in most but not all studies, and varied across the parameters measured. Likewise, most studies reported significant improvements in overground walking. In general, there was tremendous variability in training programs across the studies, rendering it impossible to discern training parameters associated with improvements. Training modalities included various types of treadmill training (BWSTT, robotic assisted, manually assisted, FES assisted), FES-assisted overground walking training and lower-body strength training. Training frequency ranged from 2 to 5 sessions per week. Training session duration was participant-determined in some studies and experimenter-determined in others, ranging from 15 to 90 min per session. Of note, the only RCT46 showed significant improvements in overground walking speed and treadmill training speed across four types of BWSTT prescribed 3 × per week at an individualized intensity.

Standing: A single level 5 case study32 reported improvements in the participant's ability to stand with a walker following 9 months of progressive BWSTT, performed 3 × per week for an average of 55 min per session.

Discussion

Our systematic review included 13 exercise-training studies involving people with acute SCI and 69 studies involving those with chronic SCI. The review yielded only eight RCTs (five in acute population; three in chronic). Most studies utilized a pre–post study design, with scores <20/28 on the Downs & Black evaluation scale. As such, overall, evidence regarding the effects of exercise on fitness is characterized as low quality. Evidence quality is taken into consideration in the following sections addressing our primary research questions: ‘Can exercise improve each fitness component?’ and, if so, ‘What types of exercise improve each component?’

Conclusions from studies of acute SCI

There is insufficient evidence to draw meaningful conclusions regarding the effects of exercise, or specific types of exercise, on any of the four fitness components. It should be noted, however, that a single RCT showed the effectiveness of exercise in improving physical capacity with supporting evidence from four pre–post studies. These encouraging findings require replication in high-quality studies. It is also noteworthy that no study reported fitness decrements. Given the profound deconditioning that follows an SCI, the maintenance and/or absence of significant fitness losses could be interpreted as a positive outcome.

Conclusions from studies of chronic SCI

Based on level 1 and 2 evidence, with consistent and substantial supporting level 4 and 5 evidence, we conclude that exercise training increases physical capacity. This conclusion is consistent with the other two systematic reviews.2, 3 Regarding exercise types, level 1 and 4 evidence showed a combination of resistance and arm ergometry training, performed 2–3 × per week at a moderate intensity (60–80% of max HR, or 60–65% VO2peak), improves physical capacity. In addition, level 2 evidence with consistent and substantial supporting level 4 and 5 evidence showed the effectiveness of 3 × per week FES-assisted exercise. Given the variability across studies, it is impossible to draw conclusions about the relative effectiveness of different types of FES or to recommend a particular intensity. Finally, there is low-quality level 4 and 5 evidence of the effectiveness of wheelchair ergometry and arm ergometry for improving physical capacity. These studies suggest 3 × per week training is effective, but there is insufficient evidence to make conclusions about other aspects of the prescription.

Based on level 1 evidence with consistent, supporting level 4 evidence, we conclude that exercise training increases muscle strength. Level 1 and 4 studies show that a variety of resistance training paradigms are effective, with consistent evidence for the effectiveness of training performed 2–3 × per week, at 50–80% 1 RM. There is also level 1 and level 4 evidence for FES exercise to increase muscle strength. However, the FES protocols are too varied to draw conclusions regarding the specific type or intensity of FES training needed to elicit improvements. Although the data are generally supportive, there is insufficient quantity and quality evidence to draw conclusions regarding the effectiveness or dose of other types of exercise (for example, BWSTT, kayak ergometry) for increasing strength.

There is insufficient evidence to conclude that exercise can affect body composition in people with chronic SCI. Currently, there is no evidence that exercise can decrease body weight. The evidence is mixed regarding the effects of exercise on muscle and fat mass. Although FES-assisted exercise looks very promising for increasing muscle mass and possibly decreasing fat mass, further quality research is needed before conclusions can be drawn regarding its effectiveness.

There is insufficient quality evidence to conclude that exercise improves functional performance. Evidence is mixed regarding the effects of exercise on standing and walking. Some studies showed improvements in overground walking and treadmill training parameters after step training, but there was too much variability across training protocols to draw conclusions about the type, intensity, or frequency of training that elicits improvements. Additionally, there is consistent, albeit low-quality level 4 evidence that ergometry performed 3 × per week at a moderate to heavy intensity can improve wheelchair propulsion. Further quality research on this promising training modality is needed before meaningful conclusions can be drawn regarding its effectiveness.

Summary

Exercise is effective in increasing physical capacity and muscular strength among people with chronic SCI. Although there is insufficient evidence at this time to conclude that exercise has similar fitness benefits for people with acute SCI, there is no evidence to suggest that exercise is harmful to this population. These conclusions, and the exercise protocol information catalogued in Tables 2 and 3, will provide the evidence base for the development of much-needed physical activity guidelines for people with SCI.