Wheelchair Propulsion Training

The propulsion, or “pushing”, of a manual wheelchair can be used as a form of exercise. There are multiple approaches to performing such exercise, including simple over-ground pushing (as is used for mobility/locomotion) or the use of devices that allow for stationary pushing. The most common devices are roller systems that the wheelchair’s rear wheels are placed over that allow for power output to be modulated by controlling the amount of rotational resistance on the rollers. Less common are wheelchair-compatible treadmills. These devices are usually large, require different belts than running treadmills, and are often outfitted with catch systems so that users do not roll off the back of the device during use. It should be noted that if a person’s primary means of mobility/locomotion is wheelchair propulsion, the additional repetitions of this approach during exercise are likely to increase the risk for upper extremity overuse injury. Conversely, if the person’s primary means of mobility is an electric wheelchair then there can be a considerable learning effect with wheelchair pushing on a treadmill and/or rollers.

Author Year

Research Design

Total Sample Size



Keyser et al. (2003)




Population: Mean age: 41±10yr; Gender: males=20, females=6.

Intervention: Individuals completed wheelchair ergometer tests using a 1min JUMP protocol that resulted in volitional exhaustion in 6-12min. They underwent 12wk of simulated wheelchair rolling exercise using elastic straps positioned to mimic the motion of propulsion. JUMP and constant work rate tests were performed before training and after 6 and 12wk of exercise.

Outcome Measures: HR monitor, EKG electrodes, Microprocessor, Pneumotachometer.

·         No significant differences in VO2peak, anaerobic threshold, or peak HR were observed at 6-12 weeks of the training program (p>0.05).

·         Substantial improvement in peak constant work rate tests time was noted at 6 and 12wk (p<0.001), with no significant difference between 6 and 12wk and no significant intergroup difference (p>0.05).

Le Foll-de Moro et al. (2005)




Population: Mean age: 29±14yr; Mean time since injury: 94±23 days; Gender: males=5, females=1; Level of injury: T6-12.

Intervention: Spirometric values at rest and dynamic ventilatory responses were studied before and after the wheelchair propulsion training program. The training program consisted of three 30 min exercise sessions/wk for 6 weeks including 6 successive work bouts 5 min each. During each work bout, a 4 min period of moderate work (base) was followed by a 1 min period of intense work (peak). Initially, the base was set at 50% of the maximal tolerated power obtained in the initial maximal exercise test, and the peak was set at 80%. The intensity of each training session was designed to lead up to almost 80% of maximal heart rate at the end of the sixth peak.

Outcome Measures: Spirometric test, Breath-by-breath gas analyzer system, ECG, Telemetry.

·         Vital capacity (VC) (Δ3%), forced expiratory volume in 1 sec (FEV1) (Δ7.1%), peak expiratory flow rate (PEFR) (Δ7.6%), and ratio FEV1/VC (4.1%) showed small positive change, which was not statistically significant (p>0.05).

·         Maximal tolerated power (MTP) and VO2peak significantly ↑ after training (p<0.05).

·         At maximal exercise, there was a non-significant improvement in peak ventilation (VEpeak), peak tidal volume (VTpeak), peak breathing frequency (Fpeak), and ventilatory reserve after training (p>0.05).

·         Oxygen cost of VE ↓ significantly after training (p<0.05).

At the same workload after training, VE and f significantly ↓ and VT significantly ↑ (p<0.05).

Bougenot et al. (2003)




Population: Mean age: 35±13yr; Gender: males=7; Level of injury: T6-L5; Level of severity: AIS A.

Intervention: Individuals performed 45min of wheelchair ergometry 3 times/wk for 6wk. Nine successive exercise bouts consisted of a 4min period of moderate exercise (base level) and a 1min period of intense exercise (peak level). The peak and base loads were ↑ by 10 W/2 min until the individuals were no longer able to maintain the required speed.

Outcome Measures: Breath-by-breath gas analyser system, Telemetry.

·         Maximal tolerated power (MTP), VO2peak, and VCO2peak production, and oxygen pulse (O2p), ↑ significantly after wheelchair training (p<0.05).

·         HR peak and respiratory rate were unchanged after training (p>0.05).

·         Changes in ventilation (VE) peak and tidal volume (Vt) were not significantly different (p>0.05).

·         At the ventilatory threshold, there was a significant improvement in power output (W), VO2, VCO2, VE, and O2 after training (p<0.01).

·         During the same training session performed at the same loads before and after training, HR was significantly lower at the last base (p<0.01). There was a significant difference at the last peak in in O2p and in VE (p<0.05).

·         Base intensity and peak intensity were significantly lower after training in VE (p<0.05).

·         Base intensity was significantly lower after training in VO2 (p<0.05).

Total physical work (TPW) was improved between the first and last training session (p<0.01).

Yim et al. (1993)




Population: Mean age: 30.9yr; Gender: males=11, females=0; level of lesion: T8-T12=11; Mean time post injury: 20.6mo.


Participants completed wheelchair ergometer training three times a week for five weeks. Each exercise session was 30min long, consisting of three, 10-minute exercise sets with 5 minutes rest in between. Participants were encouraged to maintain a velocity of more than 3 km/h at each resistance. Participants were assessed before and after the full training period.

Outcome Measures:

Mean heart rate, mean peak systolic blood pressure, time for 100m wheelchair propelling.


·         The mean HRpeak, the mean peak systolic blood pressure and the mean time required for 100m wheelchair propelling at resistance level 1 were significantly ↓ at the end of 5 weeks of training as compared with those at the pre-training (p<0.05).

·         There was no statistically significant difference in the mean resting heart rate, the mean resting systolic and diastolic blood pressure when running 100m at resistance level 1 of wheelchair ergometer at pre- and post-training.

·         There was no statistically significant difference in pulmonary function pre- and post-training.

The mean peak torque of shoulder flexor and the mean total work of the elbow flexor was shown with significant ↑ from pre to post exercise (p<0.05). There were no statistically significant differences in the mean peak torque and the mean total work for flexors and extensors of shoulder and elbow at the end of the training period.

Hooker & Wells (1989)


Prospective Controlled Trial


Population: Low-intensity group: n = 6, 3 male, 3 female, C5-T10, age 26–36yr, 3 months to 19yr post-injury; moderate-intensity group: n = 5, 3 male, 2 female, C5-T9, age 23–30yr, 2–19yr post-injury.

Intervention: Wheelchair ergometry 20 min/d, 3 d/wk, 8wk: low-intensity (50%–60% peak HRR) and moderate intensity (70%–80% peak HRR).

Outcome Measures: total cholesterol (TC), triglycerides, HDL, LDL.

·         No change in VO2peak or POpeak.

lipid levels in low-intensity group. Significant ↑ in HDL and ↓ in triglycerides, LDL, and the TC/HDL ratio in the moderate intensity group.

Tordi et al. (2001)


Pre Post


Population: Mean age: 27yr; Gender: males=5, females=0; Level of injury: Th6-L4.

Intervention: Paraplegic men performed 30-min wheelchair ergometry 3x/wk for 4wk. The training program based on the the Square Wave Endurance Exercise Test (SWEET) consisted of six successive workouts of 5 min each. During each workout, a 4-min period of moderate exercise, named `base’ level, was followed by a 1-min period of intense exercise, named ‘peak’ level. Maximal dynamic performance and endurance capacity were studied before and after the training program with an incremental test (10 W/2 min) until volitional fatigue and a constant work rate test, respectively. Cardiorespiratory responses were continuously studied during each of these tests.

Outcome measures: Ventilation, respiratory rate, tidal volume (VT), oxygen uptake (V02), VCO2 and CO2 production, Heart rate (HR).




·         Wheelchair training produced a significant ↑ in peak tolerated power, in VO2peak VCO2, O2p (all p<0.05). Peak HR was significantly lower after the training.

Pre to post training participants were able to maintain the load applied during the constant test for a significantly longer period of time.

Gauthier et al. (2018)




NInitial=11, NFinal=9


Population: HIIT Group (n=4): Mean age=33.9yr; Gender: males=3, females=1; Level of injury range: C7-T10; Mean time since injury: 6.0yr. MICT Group (n=5): Mean age=43.2yr; Gender: males=3, females=1; Level of injury range: C6-T11; Mean time since injury: 15.5yr.

Intervention: Participants were randomized to a home-based self-managed manual wheelchair program. Weelchair ergo! The high-intensity interval training (HIIT) group alternated 30s high-intensity intervals (6 – 8, Borg CR10 scale) and 60s low-intensity intervals (1 – 2, Borg CR10 scale). The moderate-intensity continuous training (MICT) maintained a constant moderate intensity (4 – 5, Borg CR10 scale). The programs were six wks, consisting of three 40-min wheelchair propulsion training session/wk.

Outcome Measures: Cardiorespiratory Fitness: VO2peak, Heart Rate (HR), POpeak, RPEmuscu, RPEcardio; Upper Limb Strength: Shoulder (flexors, extensors, abductors, adductors, internal rotators, external rotators), Elbow (flexors, extensors). Shoulder pain measured via the WUSPI

·         There were no statistically significant within group differences from pre- to post intervention. There were also no statistically significant between group differences for the relative change (expressed as a percentage) in cardiorespiratory fitness.  (p>0.05).

·         Similarly, upper-limb strength did not significantly improve between groups for all outcome measures (p>0.05).

Special consideration should be given to shoulder pain when initiating a HIIT intervention, especially in those with prior shoulder pain. Besides this, a home-based manual wheelchair HIIT program appears feasible and safe.

van der Scheer et al. (2016)





Population:  Exercise Group (n=14): Median age=55yr; Gender: males=12, females=2; Level of injury range: C4-L5; Median time since injury: 16.0yr. Control Group (n=15): Median age= 57yr; Gender: males=10, females=5; Level of injury range:  C4-L5; Median time since injury: 20.0yr.

Intervention: Inactive manual wheelchairs (MWC) users were randomized to an wheelchair propulstion exercise group or no exercise group. The low-intensity training program (30–40% HRR or 1-3 RPE on a Borg CR10 scale) was 16wks, consisting of wheelchair treadmill propulsion 2x/wk for 30min.

Outcome Measures: peak aerobic work capacity: VO2peak, POpeak; Submaximal fitness: MEsub1, MEsub2; Anaerobic work capacity:  5s peak power output over a 15-m overground sprint (P5-15m); Isometric Strength; Wheelchair Skills Performance (WSP): performance time, ability score, strain score; Physical activity levels: Physical Activity Scale for Individuals with Physical Disabilities (PASIPD), distance.

·         Participants were, on average, able to ↑ power output and velocity over the training period.

·         10/14 participants felt that the training improved their fitness.

·         Most participants reported that wheelchair skill performance and physical activity levels had not changed.

·         No significant training effects were found in peak aerobic work capacity, WSP or Physical activity levels.

P5-15m was the only outcome measure that was statically significant between the control and intervention group (p=0.02).


Eight studies were identified that examined the effect of wheelchair propulsion training on cardiorespiratory fitness and/or endurance. There is level 1b evidence from 1 RCT study showing that low-intensity wheelchair propulsion for 2 sessions per week for 16 weeks fails to elicit an increase in cardiorespiratory fitness. Level 2 evidence from two studies (one RCT and one prospective control) shows that increasing exercise intensity, albeit over a shorter 6 to 8 wk training period, does not result in greater improvements in cardiorespiratory fitness compared to lower intensity wheelchair propulsion. Of five pre/post studies, providing weaker evidence, one showed 12 wk of wheelchair propulsion did not ↑ VO2peak while three showed that 4 to 6 wk of wheelchair population can modestly ↑ VO2peak. Results at various evidence levels show similar incongruent findings for the effect of wheelchair propulsion on endurance performance: Level 1b evidence from an RCT shows a benefit only on sprint performance (not endurance), while level 4 evidence from four pre/post studies shows evidence of benefit on endurance. One pre/post study suggests the potential for wheelchair propulsion training to benefit pulmonary function, while two other pre/post studies show no effect of this training.


Van der Scheer et al. (2016) provides Level 1 evidence that 16 wk of 2 day/wk of wheelchair pushing benefits endurance performance but can do so without improving VO2peak.

Hooker and Wells. (1989) provided Level 2 evidence that 8 wk of 3 day/wk of wheelchair pushing can have no effect of VO2peak and POpeak while benefitting some blood lipid measures.

Fauthier et al. provides Level 2 evidence that 6 wk of 3 day/wk of wheelchair pushing can result in no effect on cardiorespiratory fitness or performance.

Keyser et al. (2003) provided Level 4 evidence that 6 – 12 wk of wheelchair pushing benefits POpeak.

Le Foll-de Moro et al. (2005) provided Level 4 evidence that 6 wk of 3 day/wk of wheelchair pushing benefits VO2peak and some measures of pulmonary function.

Bougenot et al. (2003) provided Level 4 evidence that 6 wk of 3 day/wk of wheelchair pushing benefits VO2peak, endurance performance, and some measures of pulmonary function.

Tordi et al. (2001) provided Level 4 evidence that 4 wk of 3day/wk of wheelchair pushing benefits VO2peak and endurance performance.