Eur J Appl Physiol DOI 10.1007/s00421-014-2850-9

Original Article

Prediction of peak oxygen uptake from differentiated ratings of perceived exertion during wheelchair propulsion in trained wheelchair sportspersons Victoria L. Goosey‑Tolfrey · Thomas A. W. Paulson · Keith Tolfrey · Roger G. Eston 

Received: 5 December 2013 / Accepted: 11 February 2014 © Springer-Verlag Berlin Heidelberg 2014

Abstract  Purpose To assess the validity of predicting peak oxygen ˙ 2peak) from differentiated ratings of perceived uptake (VO exertion (RPE) obtained during submaximal wheelchair propulsion. Methods Three subgroups of elite male wheelchair athletes [nine tetraplegics (TETRA), nine paraplegics (PARA), eight athletes without spinal cord injury (NON-SCI)] performed an incremental speed exercise test followed ˙ 2peak test). Oxyby graded exercise to exhaustion (VO ˙ 2), heart rate (HR) and differentiated RPE gen uptake (VO (Central RPEC, Peripheral RPEP and Overall RPEO) were obtained for each stage. The regression lines for the perceptual ranges 9–15 on the Borg 6–20 scale ratings were ˙ 2peak. performed to predict VO Results There were no significant within-group ˙ 2peak (mean mean differences between measured VO 1.50  ± 0.39, 2.74 ± 0.48, 3.75 ± 0.33 L min−1 for TETRA, PARA and NON-SCI, respectively) and pre˙ 2peak determined using HR or differentiated dicted VO

Communicated by Jean-René Lacour. V. L. Goosey‑Tolfrey (*) · T. A. W. Paulson · K. Tolfrey  The Peter Harrison Centre for Disability Sport, School of Sport, Exercise and Health Sciences, Loughborough University, Leicestershire, Loughborough LE11 3TU, UK e-mail: [email protected] R. G. Eston  Sansom Institute for Health Research, School of Health Sciences, University of South Australia, Adelaide, SA, Australia R. G. Eston  Sport and Health Sciences, University of Exeter, Exeter, UK

RPEs for any group (P > 0.05). However, the coefficients of variation (CV %) between measured and pre˙ 2peak using HR showed high variability for all dicted VO groups (14.3, 15.9 and 9.7 %, respectively). The typical error ranged from 0.14 to 0.68 L min−1 and the CV % ˙ 2peak using differenbetween measured and predicted VO tiated RPE was ≤11.1 % for TETRA, ≤7.5 % for PARA and ≤20.2 % for NON-SCI. Conclusions Results suggest that differentiated RPE may be used cautiously for TETRA and PARA athletes when ˙ 2peak across the perceptual range of 9–15. predicting VO ˙ 2peak is not recommended for the However, predicting VO NON-SCI athletes due to the large CV %s (16.8, 20.2 and 18.0 %; RPEC, RPEP and RPEO, respectively). Keywords  Exercise prescription · RPE · Paraplegic · Tetraplegic · Aerobic capacity Abbreviations AB Able-bodied BLa− Blood lactate concentration CV Coefficient of variation GXT Graded exercise test to exhaustion HR Heart rate NON-SCI Athletes without spinal cord injury PARA Paraplegic RPE Ratings of perceived exertion RPEC Central RPE RPEP Peripheral RPE RPEO Overall RPE SCI Spinal cord injury TE Typical error TETRA Tetraplegic ˙ 2 Oxygen uptake VO ˙ 2peak Peak oxygen uptake VO

13



Introduction The use of ratings of perceived exertion (RPE) during dynamic exercise has been shown to be a valid indicator of the degree of physical effort (Borg 1970). In able-bodied (AB) populations, the Borg 15-point category scale (6–20) (Borg 1970) has demonstrated a linear relationship with physical measures (power output and speed) and meta˙ 2)] bolic measures [heart rate (HR) and oxygen uptake (VO of exercise intensity during running (Eston et al. 1987; Lamb et al. 1999), cycling (Faulkner and Eston 2007; Skinner et al. 1973) and arm cranking (Borg et al. 1987; Eston and Brodie 1986). On this basis, RPE has been suc˙ 2max) cessfully used to predict maximal oxygen uptake (VO during the aforementioned exercise modalities for both AB and paraplegic (PARA) persons (Eston et al. 2005; Lambrick et al. 2009; Al-Rahamneh and Eston 2011). Using arm-crank ergometry obviously violates a key principle of ‘specificity’ when considering the application to high performance training in wheelchair sports, such as rugby and basketball. However, despite the recent interest on the topic of RPE and exercise in persons with a spinal cord injury (SCI) (Al-Rahamneh et al. 2010; Al-Rahamneh and Eston 2011; Goosey-Tolfrey et al. 2010; Leicht et al. 2011; Lewis et al. 2007; Müller et al. 2004; Paulson et al. 2013a), no research has examined the efficacy of RPE for predicting ˙ 2peak) during wheelchair propulpeak oxygen uptake (VO sion exercise. Injuries are common place during wheelchair sports (Stöhr and Zimmer 1997), and muscular imbalances can occur due to the over use of the upper body extensor muscle groups (Wilson and Washington 1993). Moreover, potentially harmful technical changes may occur during fatiguing wheelchair propulsion that makes the shoulder increasingly susceptible to musculotendinous-type overuse injuries (Rodgers et al. 1994). Consequently, during periods of high intensity training, it may not be possible for all wheelchair sportspersons to complete a test to volitional exhaustion due to their concerns with the aggravation of a ˙ 2peak from previous injury. For that reason, predicting VO submaximal bouts of exercise may be welcomed if a test has to be terminated early. ˙ 2max Despite the usefulness of RPE for predicting VO for AB persons, the aforementioned studies use protocols that typically use ‘overall’ RPE which is the integration of central and peripheral sensations of effort. It is important to note that during wheelchair propulsion a smaller degree of muscle mass is activated when compared to whole body exercise. Consequently, several review articles have emphasised the fact that exercise capacity is limited by peripheral factors rather than central circulatory factors (Hoffman 1986; Figoni 1993; Sawka et al. 1981). These include

13

Eur J Appl Physiol

restriction to muscle perfusion (Davis 1993), higher blood lactate concentrations (Borg et al. 1987) and muscular fatigue (van der Woude et al. 2001). Additionally, in lesstrained individuals such as those with recent SCI or novice/ development athletes, peripheral fatigue will play a more prominent role in limiting wheelchair exercise performance (van der Woude et al. 1999). Accordingly it may be more appropriate to use a differentiated RPE scale where ‘peripheral’ RPE (attributed to sensations of muscular strain) is also recalled (Pandolf 1978), which has been supported by recent work examining wheelchair exercise (Lenton et al. 2008b; Paulson et al. 2013b). Of the aforementioned studies, Lenton et al. (2008b) found that during wheelchair propulsion, ‘peripheral’ RPE dominated fatigue sensations and Paulson et al. (2013b) suggested that ‘peripheral’ RPE enabled more precise self-regulation during moderate-intensity wheelchair propulsion. Therefore, the aim of the present study was to assess ˙ 2peak from differentiated the accuracy of predicting VO RPE obtained during submaximal wheelchair propulsion in persons eligible to participate in wheelchair rugby and basketball. Specifically, the aim was to identify whether central RPE (RPEC), peripheral RPE (RPEP), overall RPE ˙ 2peak in three subgroups (RPEO) or HR best predicted VO of wheelchair athletes (tetraplegic, TETRA), PARA or athletes without spinal cord injury (NON-SCI). Novel aspects of the study include: (1) the use of differentiated RPE for prediction purposes, and (2) the use of wheelchair exercise with experienced participants.

Materials and methods Participants Twenty-six elite male wheelchair athletes (9 C6-C7 TETRA, 9 T6-L1 PARA and 8 wheelchair NON-SCI) volunteered to participate in the study. All participants were actively competing in international level wheelchair basketball or wheelchair rugby (team sports dominated by short bursts of high intensity activity, superimposed on aerobic activity). A summary of the participants’ characteristics is presented in Table 1. All procedures were approved by the Universities’ ethical advisory committee, and performed in accordance with the declaration of Helsinki. All participants provided written informed consent prior to the exercise tests. Participants reported to the laboratory on one occasion between 09:30 and 11:30, to perform the submaximal and maximal wheelchair propulsion tests. All exercise tests were performed in the participants’ competition court sports wheelchair on a motorised treadmill (HP Cosmos, Traunstein, Germany).

Eur J Appl Physiol Table 1  Participants’ characteristics TETRA (n = 9)

PARA (n = 9)

NON-SCI (n = 8)

Age (year) Body mass (kg) Lesion level/disability ASIA impairment scale TSI (year) WC sport Training (h week−1)

30 ± 5 70.6 ± 10.1† C6/7 A 11 ± 5

29 ± 9 70.3 ± 12.6† T6-L1, spina bifida A 19 ± 8

27 ± 8 84.8 ± 10.7 Amputeea, clubfootb and hip problemc – 11 ± 5

Rugby (n = 9) 13 ± 3

Basketball (n = 9) 14 ± 3

Basketball (n = 8) 16 ± 2

Time in sport (year)

10 ± 4

13 ± 6

10 ± 6

TETRA cervical SCI, PARA thoracic SCI below T6, NON-SCI non-spinal injured, ASIA American spinal injury association, TSI time since injury, WC wheelchair †

  Significant difference, TETRA and PARA vs. NON-SCI, P < 0.05

a

  Six participants ranging from double leg transtibial amputee to single leg transfemoral

b

  One participant with talipes equinovarus and

c

  One participant with avascular necrosis of the hip

Measurements A health, training and disability questionnaire was completed by all participants, and body mass was obtained to the nearest 0.1 kg using double-beam seated scales (Marsden MPWS-300, Oxfordshire, UK). Before the exercise testing, participants were introduced to the Borg 6–20 scale and given standardised instructions in detail with a clear explanation of how to rate the differentiated RPE (Borg 1998). To determine RPEC, participants were asked to rate their perceived exertion for the heart, lungs and breathing. To determine peripheral RPE (RPEP), participants were asked to rate exertion only from the exercising muscle groups and joints. Overall RPE (RPEO) was then determined as the combination of RPEP and RPEC. The RPE scale was visible to participants throughout the duration of the test and ‘differentiated RPE’ was prompted to the participants by the investigator during the last 15 s of each 4-min bout while the participant was still exercising. These RPE values across the range RPE 9–15 were obtained during the submaximal exercise test and used for the prediction purposes. Expired air was collected during the last minute of each exercise stage and analysed using the Douglas bag technique. The concentration of oxygen and carbon dioxide in the expired air samples was determined using a paramagnetic oxygen analyser (Series 1400; Servomex Ltd, Sussex, UK) and an infrared carbon dioxide analyser (Series 1400; Servomex Ltd). Expired air volumes were measured using a dry gas meter (Harvard Apparatus, Kent, UK) and ˙ 2, corrected to standard temperature and pressure (dry). VO carbon dioxide output, minute ventilation, ventilatory equivalent and respiratory exchange ratio were calculated. HR was monitored continuously using radio telemetry (Polar PE 4000, Kempele, Finland). A small capillary blood

sample was obtained from the earlobe prior to exercise, following each submaximal exercise stage, immediately after the graded exercise test to exhaustion (GXT) and immediately after the verification test to determine blood lactate concentration (BLa−) using a YSI 1500 SPORT Lactate Analyser (YSI Inc, Yellow Springs, OH). At the same time points, participants were asked to indicate the rating of perceived exertion (RPEC, RPEP, RPEO) using the 15-point Borg scale according to previous instructions (Borg 1998). Submaximal exercise test Following a 5-min warm up at 1.2 m s−1, participants performed ~6 submaximal constant load 4-min exercise stages at ascending speeds at a fixed gradient of 1.0 %. The protocol was performed according to Goosey-Tolfrey (2007) to elicit submaximal physiological responses covering a range ˙ 2peak. Based on previous experiences from 40 to 80 % VO of using RPE in TETRA groups (Paulson et al. 2013a), tests were terminated when an RPEO above 15 was noted to avoid any effects of anaerobic energy production on HR and mechanical efficiency (Ekblom-Bak et al. 2012). Peak exercise test Following 15-min passive recovery, a GXT was then performed at a constant speed according to the protocol described by Leicht et al. (2011). Briefly, the gradient at the start of the GXT was 1.0 % for all subgroups, with subsequent increases of 0.3 % every minute for PARA and NON-SCI and 0.1 % every 40 s for TETRA to account for the functional differences between groups and ensure a minimum GXT duration of ~8 min (Leicht et al. 2013a, b). After the GXT, participants recovered actively at a low intensity (1.2 m s−1) at a 1.0 %

13



Eur J Appl Physiol

gradient for 5 min. Participants then performed a verification test, designed as a test to exhaustion at the same constant speed but 0.3 and 0.1 % higher than the maximal gradient achieved during the GXT for NON-SCI/PARA and TETRA, respectively (Leicht et al. 2013a, b). The GXT and the verification test were terminated when participants were unable to maintain the speed of the treadmill. Verbal encouragement was given throughout the test. Expired air was collected for at least the final three consecutive minutes of the GXT and for 2 min during the verification test. The higher of the two − ˙ 2peak, HRpeak and BLapeak VO values obtained in the GXT and the verification test was taken as the peak value.

in the sport. Other than the notable differences in impairment level as characterised by the ASIA scale and lesion level (Table 1), the groups differed by body mass, where the NON-SCI group were considerably heavier than both the TETRA and NON-SCI individuals (P < 0.05). Percent˙ 2peak at the corresponding RPEs 9, 11, 13 and 15 age of VO across the three differentiated RPE scores for the TETRA, PARA and NON-SCI groups is shown in Fig. 1. There was a main effect for group at RPEs 9, 11 and 13, showing that the TETRA athletes were working at a significantly higher exercise intensity than PARA and NON-SCI athletes (P < 0.05).

Data and statistical analysis

Peak exercise responses

The mean and standard deviations were computed for all variables. All data were analysed using either SPSS 21 for statistical package (SPSS Inc., Chicago, IL, USA) or the analysis of reliability spread sheet (beta version) developed by Hopkins (2010). Normality and homogeneity were checked with the Shapiro–Wilk and Levene’s statistics, respectively. A one-way analysis of variance (ANOVA) was applied to explore the between-group differences in participants’ char˙ 2peak data at RPE 9, 11, acteristics. For the percentage of VO 13 and 15, three-way (group by differentiated RPE) mixed measures ANOVA were applied to identify differences between the three groups (TETRA, PARA and NON-SCI) across the three differentiated RPE. Where the relevant statistical significance for all analysis was accepted at P < 0.05.

˙ 2 HR, BLa−, VE/VO ˙ 2 The peak physiological (i.e. VO RER) values observed at the termination of the GXT are presented in Table 2. Significant differences between TETRA individuals and the subgroups PARA and NON˙ 2peak and peak HR. FurtherSCI were found in absolute VO more, significant differences between PARA and NON-SCI ˙ 2peak. Peak BLa were individuals were found in absolute VO significantly lower in TETRA when compared with PARA and NON-SCI individuals (Table 2). Participants reported similar RPEs at the termination of the GXT (19–20).

˙ 2peak Measured vs. predicted VO ˙ 2 was Using individual participant linear regression, VO regressed against the corresponding differentiated RPE(C, P, O) values of 9–15 and extrapolated to the theoretical maxi˙ 2peak mal RPE (RPE 20) on the Borg scale to predict VO The same procedures were applied to when HR was used for the prediction. ˙ 2peak Analysis of the validity of the predictions in VO The coefficient of variation (CVs) was computed for each ˙ 2 measurement across the RPE range (9–15), predicted VO and unbiased typical error (TE) was also measured to quantify the relationship between predicted and measured ˙ 2peak values using the differentiated RPE(C, P, O) and HR VO values, for future details please refer to Hopkins (2010).

Results There was no significant difference between groups with respect to age, time since injury, training hours and time

13

Submaximal exercise responses The group mean (SD) coefficient of determination R2 of 0.86 (0.09), 0.97 (0.02) and 0.96 (0.05) suggests that there ˙ 2 and HR durwas a strong linear relationship between VO ing the incremental test for all groups (TETRA, PARA and NON-SCI, respectively). The various differentiated RPE parameters also demonstrated a narrow range in the R2 val˙ 2 and HR, ues (0.88–0.95) when correlated with both VO with lowest values observed for the TETRA group.

˙ 2peak Measured vs. predicted VO As subgroups, there was no significant difference between ˙ 2peak and either of the method the mean measured VO ˙ 2peak. However, the CV %s employed for predicting VO ˙ 2peak across the predicted methods of comparing the VO differentiated RPE(C, P, O) were two- to threefold higher in the NON-SCI subgroup (≤20.2 %) with the TE ranging from 0.54 to 0.68 L min−1, when compared with the other subgroups (Table 3). For the TETRA and PARA’s, the CV %s spanned 6.5–11.1 %, with the TE ranging 0.14– 0.22 L min−1 when the differentiated RPE(C, P, O) was used. ˙ 2peak was noted when For all groups, a larger CV % of VO HR was used, where the TE of 0.18–0.35 L min−1 translated to CV %s of 14.3, 15.9 and 9.7 %, respectively, for TETRA, PARA and NON-SCI groups.

Eur J Appl Physiol

˙ 2peak at each RPE (Borg 9–15) across the three differentiated RPE scores for the TETRA, PARA and NON-SCI groups. Fig. 1  Percentage of VO Note Asterisk denotes significant main effect for group (P < 0.05)

Table 2  Physiological and exertional responses TETRA ˙ 2peak (L min−1) VO −1

1.50 ± 0.39† †

PARA

NON-SCI

2.74 ± 0.48‡

3.75 ± 0.33

˙ 2peak values from the graded exercise test for Table 3  Measured VO ˙ 2peak values from the differenticomparison against the predicted VO ated RPE (9–15) and HR including the typical error (ratio CV)

Parameter

TETRA

PARA

NON-SCI

Measured ˙ 2peak (L min−1) VO

1.50 ± 0.39

2.74 ± 0.48

3.75 ± 0.33

RPEC

1.47 ± 0.43 0.15 (10.8 %) 1.52 ± 0.45 0.16 (11.1 %) 1.50 ± 0.44 0.14 (10.1 %) 1.54 ± 0.38

2.71 ± 0.63 0.22 (7.5 %) 2.73 ± 0.57 0.19 (6.5 %) 2.68 ± 0.57 0.19 (6.5 %) 2.97 ± 0.70

3.68 ± 0.81 0.54 (16.8 %) 3.92 ± 1.01 0.68 (20.2 %) 3.80 ± 0.88 0.59 (18.0 %) 3.82 ± 0.40

HRpeak (b min ) BLa–peak (mmol L−1) Overall RPEpeak ˙ 2 VE/VO

127 ± 11 4.99 ± 0.63† 20 (19, 20) 36.9 ± 4.4

181 ± 10 7.69 ± 1.87 19 (19, 20) 36.4 ± 3.9

183 ± 8 8.29 ± 1.64 20 (19, 20) 34.1 ± 3.5

RER GXT time (min)

1.12 ± 0.07 8.8 ± 0.9

1.19 ± 0.05 8.8 ± 0.9

1.15 ± 0.06 9.5 ± 0.7

RPEP

VER time (min)

2.0 ± 0.0

1.9 ± 0.3

1.9 ± 0.4

RPEO

TETRA cervical SCI, PARA thoracic SCI below T6, NON-SCI nonspinal injured, V˙ O2peak peak oxygen uptake, HRpeak peak heart rate, BLa− peak peak blood lactate concentration, RPEpeak peak rating of overall perceived exertion, VE pulmonary ventilation, V˙ O2 oxygen uptake, VE/V˙ O2 ventilatory equivalent for oxygen, RER respiratory exchange ratio, GXT graded exercise test, VER verification test

HR

0.18 (14.3 %) 0.47 (15.9 %) 0.35 (9.7 %) All values in L min−1  with unbiased typical error and CV (%)

†

  Significant difference, TETRA vs. PARA and NON-SCI, P < 0.05

‡

  Significant difference, PARA vs. NON-SCI, P < 0.05

Discussion The main findings of this study confirm that, during wheelchair propulsion for trained wheelchair athletes at low

through to high relative exercise intensities (39 ± 8 to ˙ 2peak), the use of HR obtained from a sub79  ± 8 % VO maximal test is the less-preferred method for predicting ˙ 2peak Instead, the use of differentiated RPE may be used VO cautiously for TETRA and PARA athletes when predicting ˙ 2peak if a GXT is terminated early due to an injury or VO

13



if submaximal testing (RPE 9–15) is conducted leading up to a major competition. This is possible, since the CV %s of these two subgroups compare well with the day-to-day variability of 9.3 and 4.5 % for TETRA and PARA, respec˙ 2peak tively (Leicht et al. 2013a, b). However, predicting VO is not recommended for the NON-SCI athletes due to the ˙ 2peak found in this study (16.8, 20.2 and large CV %s of VO 18.0 %; RPEC, RPEP and RPEO, respectively) which when ˙ 2peak compared to the day-to-day variability data of VO (3.3 %) reported for similar athletes (Leicht et al. 2013a, b) is considerably higher. ˙ 2peak for the participants in this study The mean VO compare well to the previous literature of highly trained wheelchair athletes (Goosey-Tolfrey 2005; Leicht et al. 2011; Paulson et al. 2013a). Therefore, we can assume that the participants recruited for the present study are representative of wheelchair athletes with sports experience and high levels of training volume. There is a great ˙ 2max in AB populations, interest in the prediction of VO and one of the underlying features within past experimental designs has been with the manipulation of the perceptual ranges investigated (e.g., RPE 9–17, 9–15 and 11–17) (Eston et al. 2008, 2012). Subsequently the accuracy of ˙ 2max is highly influenced by the perceptual predicting VO range (Eston et al. 2008). Despite the recommendation that removing RPE 17 from the predictive analysis may ˙ 2max, from previous experience TETRA underestimate VO athletes have reported low RPE even at moderate-high ˙ 2peak, Paulson exercise intensities (RPE 12 at 70 % VO et al. 2013a). The authors therefore deemed it appropriate to remove RPE 17 from the prediction equation, since many of the TETRA athletes may have been approaching maximal exercise between RPEs 15–17. Furthermore, since the use of RPE for the self-regulation of exercise has been validated in trained PARA and TETRA groups (RPEs 12 ± 1 and 16 ± 1) (Goosey-Tolfrey et al. 2010; Paulson et al. 2013a), and that all the participants were previously familiar to the protocol, the authors were confident that the experienced participants could accurately recall RPE at the lower exercise intensities. As said, it is evident that the efficacy of implementing such procedures in wheelchair athletes without a SCI (NON-SCI) is lacking, and further investigation of RPE and training prescription is warranted for those athletes who do not rely on a wheelchair for daily ambulation. One of the objectives of this study was to explore whether using the differentiated RPE (RPEC, RPEP, RPEO) ˙ 2peak in either of provided a more accurate prediction of VO the three subgroups (TETRA, PARA and NON-SCI). It is noteworthy that a recent study has examined the reliability of peak physiological variables in wheelchair athletes during wheelchair propulsion on a motorised treadmill (Leicht et al. 2013a, b). This study provides an important criterion

13

Eur J Appl Physiol

reference to which we can compare our data and found ˙ 2peak was 13.2, 6.4 that the day-to-day variability of VO and 4.7 % for similarly trained TETRA, PARA and NONSCI athletes, respectively. They reported that TETRA athletes demonstrated greater variability, which was explained by their lower aerobic capacity. Incidentally, the general ˙ 2peak conclusion was that the day-to-day variability of VO obtained during upper body exercise (wheelchair propulsion) was similar to the CV % of 8.9 % that has been noted during arm-crank exercise in AB persons (Price and Campbell 1997). Hand–rim propulsion is a guided movement that is regulated predominantly by the rim curvature and its speed and direction of movement (Lenton et al. 2008a). As the exercise intensity increases, the participants usually adapt by changing their arm frequency, force application and/or propulsion technique in such a way that suits the given wheelchair propulsion velocity (Lenton et al. 2008a). Research to date would suggest that at higher arm frequencies the local factors (peripheral) are more important than central cues of exertion when using RPEO as an indicator of perceived effort (Lenton et al. 2008b). Our findings suggest that it was only the PARA group that favoured this perceptual cue (RPEP) as the CV % was at its lowest (6.5 %) and within the expected day-to-day variability. By creating subgroups, it was possible to consider the fact that the NON-SCI group was able to walk and did not rely on wheelchair propulsion for daily ambulation. Thus, we were able to explore further whether RPE from the exercising muscle mass and joints were the dominant perceptual signal during exercise. We know that manual hand–rim propulsion has been associated with neurologic and muscular pain in the wrist and shoulder joints because of high mechanical loads (Boninger et al. 2002). It was anticipated that there may have been a better association of the submaximal exercise using an RPE specific to the peripheral exertional signals (RPEP) in this group, since Paulson et al. (2013b) noted that RPEP enabled a more precise method for self-regulation during moderate-intensity wheelchair exercise in novice users. It can be concluded that for the NON-SCI group, the use of ˙ 2peak did not have an differentiated RPE for predicting VO effect on the signal dominance at relative exercise inten˙ 2peak as the variability sities ranging from 39 to 73 % VO ˙ 2peak was 16.8, 20.2, 18.0 %; RPEC, with the measured VO RPEP and RPEO, respectively. It is important to note that this group comprised of athletes with double amputation through to clubfoot, which created a greater degree of heterogeneity compared to the two groups with a spinal cord injury. Futures studies may wish to control for the nature of the physical impairment within this NON-SCI group. The peak exercise responses elicited during the GXT clearly demonstrated that the TETRA athletes showed a disrupted autonomic innervation of the heart with the

Eur J Appl Physiol

significantly reduced HRpeak values of 127 ± 11 vs. 181 ± 10 b min−1 and 183 ± 11 b min−1 of the PARA and ˙ 2 NON-SCI groups (P < 0.05). That said, linearity of the VO -HR relationship in this group was still observed as noted previously (McLean et al. 1995; Valent et al. 2007; Leicht ˙ 2peak from this physiologiet al. 2011). Yet, predicted VO cal parameter across the perceptual range of RPE 9–15, which corresponded to 89 ± 15 to 134 ± 29 b min−1 (for all groups) varied considerably with a TE of up to 0.47 L min−1 and CV % of 15.9 %. At this stage, it is unclear why the CV % was high for both the TETRA and PARA groups, although on close inspection it would appear that the higher HRpeak found with the NON-SCI athletes may have led to ˙ 2–HR relationship and consequently the lower a better VO CV % for NON-SCI athletes as 9.7 %. It is possible that a greater HR range is needed to improve the accuracy of this prediction method and that a better CV % would have been found if we had extrapolated using data with RPE 17.

Conclusions Overall, these results suggest that differentiated RPE may be used cautiously for TETRA and PARA athletes when ˙ 2peak across the RPE perceptual range of predicting VO 9–15. It appears that RPEC (heart, lungs and breathing) and RPEP (exercising muscle groups and joints) mediate RPEO similarly during wheelchair propulsion. However, predict˙ 2peak using these methods is not recommended for ing VO the NON-SCI athletes due to the large CV %s (16.8, 20.2 and 18.0 %; RPEC, RPEP and RPEO, respectively). Future work is warranted to examine the impact of trained status, exercise modality (arm ergometery vs. wheelchair propulsion) and constant load vs. continuous and intermittent exercise protocols in this cohort of wheelchair games players. Acknowledgments  We thank Dr. Christof Leicht and Dr. John Lenton for their help during laboratory testing. Moreover, we thank the Great Britain Wheelchair Rugby Ltd, British Wheelchair Basketball and the Peter Harrison Centre for their support. Appreciation is also extended to all athletes who volunteered to participate in this study.

References Al-Rahamneh HQ, Eston RG (2011) Prediction of peak oxygen consumption from the ratings of perceived exertion during a graded exercise test and ramp exercise test in able-bodied participants and paraplegic persons. Arch Phys Med Rehabil 92:277–283 Al-Rahamneh HQ, Faulkner JA, Byrne C, Eston RG (2010) Relationship between perceived exertion and physiologic markers during arm exercise with able-bodied participants and participants with poliomyelitis. Arch Phys Med Rehabil 91:273–277 Boninger ML, Souza AL, Cooper RA, Fitzgerald SG, Koontz AM, Fay BT (2002) Propulsion patterns and pushrim biomechanics

in manual wheelchair propulsion. Arch Phys Med Rehabil 83:718–723 Borg GA (1970) Perceived exertion as an indicator of somatic stress. Scand J Rehabil Med 2:92–98 Borg GA (1998) Borg’s perceived exertion and pain scales. Human Kinetics, Leeds Borg GA, Hassmén P, Lagerström M (1987) Perceived exertion related to heart rate and blood lactate during arm and leg exercise. Eur J Appl Physiol 56:679–685 Davis GM (1993) Exercise capacity of individuals with paraplegia. Med Sci Sports Exerc 25:423–432 Ekblom-Bak E, Björkman F, Hellenius ML, Ekblom B (2012) A new submaximal cycle ergometer test for prediction of VO(2max). Scand J Med Sci Sports. doi:10.1111/sms.12014 Eston RG, Brodie DA (1986) Responses to arm and leg ergometry. Br J Sports Med 20:4–6 Eston RG, Davies BL, Williams JG (1987) Use of perceived effort ratings to control exercise intensity in young healthy adults. Eur J Appl Physiol 56:222–224 Eston RG, Lamb KL, Parfitt G, King N (2005) The validity of predicting maximal oxygen uptake from a perceptually-regulated graded exercise test. Eur J Appl Physiol 94:221–227 Eston RG, Lambrick D, Sheppard K, Parfitt G (2008) Prediction of maximal oxygen uptake in sedentary males from a perceptually regulated, sub-maximal graded exercise test. J Sport Sci 26:131–139 Eston RG, Evans H, Faulkner J, Lambrick D, Al-Rahamneh H, Parfitt G (2012) A perceptually regulated, graded exercise test predicts peak oxygen uptake during treadmill exercise in active and sedentary participants. Eur J Appl Physiol 112:3459–3468 Faulkner J, Eston RG (2007) Overall and peripheral ratings of perceived exertion during a graded exercise test to volitional exhaustion in individuals of high and low fitness. Eur J Appl Physiol 101:613–620 Figoni SF (1993) Exercise responses and quadriplegia. Med Sci Sports Exerc 25:433–441 Goosey-Tolfrey VL (2005) Physiological profiles of elite wheelchair basketball players in preparation for the 2000 paralympic games. APAQ 22:57–66 Goosey-Tolfrey VL (2007) BASES physiological testing guidelines: the disabled athlete. In: Winter EM, Jones AM, Davison R, Mercer TH (eds) Sport and exercise physiology testing. Routledge, Oxon, pp 358–367 Goosey-Tolfrey VL, Lenton JP, Goddard J, Oldfield V, Tolfrey K, Eston R (2010) Regulating intensity using perceived exertion in spinal cord-injured participants. Med Sci Sports Exerc 42:608–613 Hoffman MD (1986) Cardiorespiratory fitness and training in quadriplegics and paraplegics. Sports Med 3:312–330 Hopkins W (2010) Spreadsheet for the analysis of reliability (beta version). http://www.sportsci.org/. Accessed 2013 Lamb KL, Eston RG, Corns D (1999) Reliability of ratings of perceived exertion during progressive treadmill exercise. Br J Sports Med 33:336–339 Lambrick DM, Faulkner JA, Rowlands AV, Eston RG (2009) Prediction of maximal oxygen uptake from submaximal ratings of perceived exertion and heart rate during a continuous exercise test: the efficacy of RPE 13. Eur J Appl Physiol 107:1–9 Leicht C, Bishop NC, Goosey-Tolfrey VL (2011) Submaximal exercise responses in tetraplegic, paraplegic and non spinal cord injured elite wheelchair athletes. Scand J Med Sci Sports 22:729–736 Leicht C, Tolfrey K, Lenton J, Bishop N, Goosey-Tolfrey VL (2013a) The verification phase and reliability of physiological parameters in peak testing of elite wheelchair athletes. Eur J Appl Physiol 113:337–345

13

Leicht C, Tolfrey K, Lenton J, Bishop N, Goosey-Tolfrey VL (2013b) Erraturn to: the verification phase and reliability of physiological parameters in peak testing of elite wheelchair athletes. Eur J Appl Physiol 113:347 Lenton JP, Fowler N, van der Woude L, Goosey-Tolfrey VL (2008a) Efficiency of wheelchair propulsion and effects of strategy. Int J Sport Med 29:384–389 Lenton JP, Fowler NE, van der Woude LHV, Goosey-Tolfrey VL (2008b) Wheelchair propulsion: effects of experience and push strategy on efficiency and perceived exertion. Appl Physiol Nutr Metab 33:870–879 Lewis JE, Nash MS, Hamm LF, Martins SC, Groah SL (2007) The relationship between perceived exertion and physiologic indicators of stress during graded arm exercise in persons with spinal cord injuries. Arch Phys Med Rehabil 88:1205–1211 McLean KP, Jones PP, Skinner JS (1995) Exercise prescription for sitting and supine exercise in subjects with quadriplegia. Med Sci Sports Exerc 27:15–21 Müller G, Odermatt P, Perret C (2004) A new test to improve the training quality of wheelchair racing athletes. Spinal Cord 42:585–590 Pandolf KB (1978) Influence of local and central factors in dominating rated perceived exertion during physical work. Percept Mot Skills 46:683–698 Paulson T, Leicht C, Bishop N, Goosey-Tolfrey VL (2013a) Perceived exertion as a tool to self-regulate exercise in individuals with tetraplegia. Eur J Appl Physiol 113:201–209 Paulson T, Bishop N, Eston R, Goosey-Tolfrey VL (2013b) Differentiated perceived exertion and wheelchair exercise in novice ablebodied users. Arch Phys Med Rehabil 94:2269–2276

13

Eur J Appl Physiol Price MJ, Campbell IG (1997) Determination of peak oxygen uptake during upper body exercise. Ergonomics 40:491–499 Rodgers MM, Gayle GW, Figoni SF, Kobayashi M, Lieh J, Glaser RM (1994) Biomechanics of wheelchair propulsion during fatigue. Arch Phys Med Rehabil 75:85–93 Sawka MN, Glaser RM, Laubach LL, Al-Samkari O, Suyraprasad AG (1981) Wheelchair exercise performance of the young, middleaged, and elderly. J Appl Physiol 50:824–828 Skinner JS, Hutsler E, Bergsteinova V, Buskirk ER (1973) The validity and reliability of a rating of perceived exertion. Med Sci Sports Exerc 5:110–115 Stöhr H, Zimmer M (1997) Wheelchair basketball from the orthopaedic viewpoint. Sportverletz Sportchaden 11:109–115 Valent LJ, Dallmeijer AJ, Houdijk H, Slootman J, Janssen TW, Hollander AP, van der Woude LH (2007) The individual relationship between heart rate and oxygen uptake in people with tetraplegia during exercise. Spinal Cord 45:104–111 van der Woude LHV, van Croonenborg JJ, Wolff I, Dallmeijer AJ, Hollander PA (1999) Physical work capacity after 7 wk of wheelchair training: effect of intensity in able-bodied subjects. Med Sci Sports Exerc 31:331–341 van der Woude LHV, Dallmeijer AJ, Janssen TWJ, Veeger D (2001) Alternative modes of manual wheelchair ambulation: an overview. Am J Phys Med Rehabil 80:765–777 Wilson PE, Washington RL (1993) Pediatric wheelchair athletics: sports injuries and prevention. Paraplegia 31:330–337