Participants

We studied a convenience sample including 63 persons with SCI; 37 (27 men and 10 women) with motor-complete paraplegia (MCP) AIS A-B (T7–T12), and 26 persons (19 men and 7 women) with motor-complete tetraplegia (MCT) AIS A-B (C5–C8). All were recruited through SCI-specific websites or by word of mouth. Inclusion criteria were SCI injury level C5–C8 and T7–T12; AIS A and B motor-complete, ≥1 year post-injury, age ≥18 years, with minimal spasticity (Baclofen treated), as reported on the spasm frequency scale (Penn)[17]. All participants used manual wheelchairs except for one who used power-assisted wheels indoors the same participant did none of the outdoor activities due to the rain. Exclusion criteria were coronary artery disease, angina pectoris and chronic congestive heart failure, chronic obstructive pulmonary disease, or shoulder pain. Persons were excluded if they were on pharmaceutical treatment such as beta blockers or hormone replacement therapy (thyroid hormone). However, those on Baclofen were included. All data were collected in a rehabilitation setting. The participants were asked to avoid heavy exercise 12 hours before the testing, to refrain from coffee and nicotine and to empty the bladder before testing. Each provided written informed consent, and ethical approval was given by the Stockholm region ethics committee, reference number 2011/1989-31/1. We choose to not include persons with an injury level between T1-T6, to have a clear difference between MCP with sufficient ANS functioning versus MCT with reduced or lack of response from ANS. This is based on that the group T1-T6 produces fuzzy results on HR due to shady differences in ANS function, whereas persons with injury level below T7 have an almost normal functioning cardiovascular response from ANS[18, 19].

Assessment of VO₂ during the standardized activities

The VO₂ assessment procedure has been described previously[20, 21]. In brief, VO₂ was assessed using the same mobile system as for the VO_2peak test. The system was calibrated using built-in automated procedures. All data were collected the same day (except for seven participants who did the VO_2peak test on a different day). The sedentary activities included television-watching and desk-based computer work; light intensity included setting a table following a standardized procedure, wheeling a manual wheelchair indoors (in a training hall, wooden floor, 25-meter track with two turns) at their own individual pace perceived as 10–11 (light exertion) on the Borg RPE scale, and individually-paced wheelchair wheeling outside on asphalt also perceived as Borg 10–11. The exercise activities included wheeling the wheelchair outside on asphalt at exercise pace Borg 13–14. Arm ergometer work at 60 rpm was performed at low level (10W or 15W for MCT and 18W or 24W for MCP) and high level (20W or 25W for MCT and 36W or 42W for MCP). For weight training, the instruction was to select a weight that the participant was able to lift ten times, comfortably at an even and controlled pace. This was tested/practiced before data collection began. The participants rested 15–45 seconds between each machine (including transfer and set-up times to the next gym machine). Circuit-resistance training was performed by one second for the concentric phase and two seconds for the eccentric phase, (rowing machine, pulldowns, Pec Dec for MCP and, rowing machine, pulldowns, external, internal rotation for MCT)[22]. The VO_2peak tests were performed as the last activity of the day.

Information about each activity was given as a standardized verbal instruction together with (Borg RPE 10–11 for light and 13–14 for moderate) and weight selection in weight training. Each test lasted for 6–7 minutes, and the time between each activity varied between a few minutes (sedentary) and >30 minutes (exercise activities). Before testing started the tire pressure was checked to be between 7–10 BAR depending on manufacturer.

Assessment of VO_2peak

A Monark arm ergometer (Ergomedic 891E Monark, Sweden) was used for the VO_2peak test. It was attached to a height-adjustable and the participants were seated in their own wheelchair. Prior to the test, placement and positioning of table and arm ergometer were individually adjusted. Individuals with poor hand function brought their own gloves to be able to hold on to the handlebars. All were asked if they wanted to be strapped to the wheelchair to retain upper-body balance. The test began with approximately three minutes of warm-up, followed by a short break, and was then incremented until exhaustion[23, 24]. The protocol used to achieve VO_2peak was individually designed according to international laboratory procedures and previous studies[24] with self-paced cadence starting between 70 and 90 revolutions per minute (rpm) and finishing around 100 to 120 rpm[25]. The starting resistance was chosen depending on level of injury and exercise status based on the resistance during the “low” and “high” arm ergometer activity”, which were performed at the beginning of the test-day. The resistance was subsequently increased every minute by 0.25kg for MCT subjects with low (10W or 15W) resistance during arm ergometer work and 0.5kg for those with high (20W or 25W) resistance. For MCP participants the resistance was increased by 0.5kg for those with 36W resistance during arm ergometer exercise and by 0.75kg for those with 42W. The increase during the last 2–3 minutes was individually managed according to the participant’s state of exhaustion as evaluated from visual/auditory contact. This gave an anticipated time to exhaustion of between 6 and 12 min. VO₂ and HR were measured continuously during the test using a mobile open-circuit system (Jaeger Oxycon Mobile system (Hoechberg, Germany). VO₂ was analyzed as the average of 10-second averages. VO_2peak was determined as the mean of the highest 30 seconds. Criteria for acceptance of the VO_2peak measurement were: “levelling off” of VO₂ despite increased resistance, RPE above 16, test time more than 6 minutes, supported by a respiratory quotient (RQ/RER) greater than 1.1. All participants reached the criteria of acceptance. None of the participants wore leg wraps and/or abdominal binders.

Other measurements

Body weight was measured to the nearest 0.1 kg, and height was self-reported, with subsequent calculation of . Self-reported PA was assessed using a validated questionnaire[26], where time spent in moderate and vigorous PA was reported and subsequently calculated by multiplying the minutes (reported as 15 min bouts) each person spent in moderate, vigorous and leisure-time activity (LTA) by the number of days per week. The questionnaire was further dichotomized according to current SCI guidelines for cardiorespiratory fitness (0–44 or 45–450 minutes per week)[27]. Heart rate was measured with chest-strap (Polar) connected to the Oxycon Mobile.

Statistical analysis

Statistical analysis used SPSS (SPSS for Windows Version 23.0; Inc. Chicago, IL, USA). All data was tested for normality using Shapiro-Wilk tests and Q–Q plot analyses. For descriptive statistics, mean, SD or median and range were used. An independent sample two-tailed t-test for group comparisons and statistical significance was set at α = 0.05. Non-normally-distributed data is presented as median interquartile range, Q1–Q3 (IRQ). Categorical data was analyzed with the Mann-Whitney U test. Spearman’s rho correlation coefficient was assessed to describe the association between VO_2peak and anthropometrics (body weight and height), demographics (age, level of injury and gender) and PA questionnaire (level, duration and intensity), within the cohorts. Further, stepwise (both forward and backward) multiple linear regressions were used to identify factors that could explain variance in VO_2peak within each cohort, since VO_2peak on group level was normally distributed. The probability for entry into the model was set to F value = p<0.05 and the probability for removal was F value = >0.10. The collinearity statistics variance inflation factor (VIF) was 1.0–1.53 and the variables’ correlation to each other was below 0.6, with a correlation to VO_2peak above 0.48. All data for percentage of VO_2peak were stratified into standardized levels of activity according to ACSM[9, 28]. The levels for MCP were light (37–45%VO_2peak), moderate (46–63% VO_2peak) and vigorous (64–91%VO_2peak) [9, 28, 29]. The levels for MCT were light (44–51% VO_2peak), moderate (52–67%VO_2peak) and vigorous (68–94%VO_2peak). For an activity to be categorized into an intensity all variables, % of VO_2peak, Borg RPE, % of HR_peak and speed (kmh), were analyzed and median and interquartile range was used to classify each activity. The MET-value for VO_2peak, calculated with help of the individual REE from the same cohort, previously published by our research group, shows a MET ≈8 for MCP and ≈5 for MCT[20]. Categorization of the activities was made by examining how many persons (in each activity) that were categorized by stratifying for % of VO_2peak Borg RPE and HR. At least 66% of the participants in each activity needed to be categorized correct.

VO_2peak

The median for absolute VO_2peak for the persons with MCP was 1.36 L·min^-1 (IQR 1.14–1.65) and 0.74 L·min^-1 (IQR 0.60–0.89) for those with MCT (Table 1 and Fig 1A). Both men with MCP and MCT had significant higher VO_2peak than women, p≤0.001. Men with MCP had a VO_2peak of 1.57 L·min^-1 vs. women 1.02 L·min^-1 (54% higher) and men with MCT had a VO_2peak of 0.84 L·min^-1 vs. woman 0.53 L·min^-1 (53% higher). Relative VO_2peak was 18.5 ml·min^-1·kg^-1 (IQR 17.0–20.4) for MCP and 11.1 ml·min^-1·kg^-1 (IQR 9.6–13.5) for persons with MCT (Table 1 and Fig 1B). Men with MCP in general had a 7% higher relative VO_2peak than women p = 0.02, while there was no significant gender difference within the MCT group p = 0.43. Gross mechanical efficiency during submaximal arm-ergometer was ≈8% for 10-15W and ≈13% for 20-25W (MCT) and for MCP ≈10% (18-24W) and ≈14% 24-36W).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1.

a. Individual absolute VO_2peak L·min-1 (IQR) values and group median for tetraplegic (left) and paraplegic (right) women and men. b. Individual relative VO_2peak (ml·min-1·kg-1) values and group median for tetraplegic (left) and paraplegic (right) women and men. * Significant difference between men/women, p&<0.002 ** Significant difference between men/women, p&<0.03.

https://doi.org/10.1371/journal.pone.0222542.g001

Relationship between VO_2peak and anthropometric, demographic and PA variables

Persons with MCP that were physically active at vigorous intensity level at least 45 min per week had a significantly higher VO_2peak 1.7 L·min^-1 compared to 1.3 L·min^-1 p = 0.007. For persons with MCT, the difference was not significant: 0.87 L·min^-1 (more than 45 mins/week) vs. 0.68 L·min^-1 (p = 0.02).

After controlling for ten different variables (Table 2) sex, body weight, body height and self-reported time spent in vigorous activity per week showed a significant weak- (r = 0.40 to 0.59)-to-moderate (r = 0.60 to 0.79) correlation[30] with VO_2peak for both cohorts. These four variables were further introduced into a stepwise regression analysis that revealed that sex (b = 0.26, p = 0.02), weight (b = 0.01, p<0.001) and time spent in vigorous intensity activity (b = 0.001, p = 0.03) explained 63% (R² = 0.63) of the VO_2peak variance in MCP. For MCT, sex (b = 0.20, p = 0.02) and time spent in vigorous intensity activity (b = 0.001, p = 0.001) explained 55% of the variance (R² = 0.551).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Association between different demographical, anthropometric and physical activity variables with VO_2peak.

https://doi.org/10.1371/journal.pone.0222542.t002

Standardized activities as percentages of VO_2peak

The categorization of the eleven activities into different intensity levels, expressed as % of VO_2peak, Borg RPE (seven activities) and % of HR_peak (Fig 2). Some activities in Fig 2 have a 95% CI within two intensity levels (wheeling indoors, setting table for MCP), which makes them harder to categorize. However, wheeling indoors could be stratified by speed (kmh), which is done in Table 3 and the heart rate for setting table was 51% HR_peak which was categorized as light intensity (Fig 2). Moreover, for MCT setting a table and strength training were within the limits for moderate intensity and wheeling indoors and the other non-exercise activities was categorized between moderate and vigorous (Fig 2). Vigorous intensity included all exercise activities for MCP and MCT.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Dual axis for VO_2peak (mean 95% CI) and Borg RPE or % of HR_peak during activities’, for MCT/MCP.

Wheeling indoors (Borg 10–11) Km h^-1, mean (range). Paraplegia 4.4 (2.9–5.8). Tetraplegia 4.2 (2.0–6.7). Wheeling outdoors (Borg 10–11) Km h^-1, mean (range). Paraplegia 6.6 (4.7–8.5). Tetraplegia 5.2 (1.9–7.4). Wheeling outdoors (Borg 13–14) Km h^-1, mean (range). Paraplegia 8.9 (6.3–12.0). Tetraplegia 7.6 (5.5–8.6).

https://doi.org/10.1371/journal.pone.0222542.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Eleven activities categorized into intensity level based on the result for percentage of VO_2peak, Borg RPE and percentage of HR_peak.

https://doi.org/10.1371/journal.pone.0222542.t003

Borg RPE correlated strongly with %VO_2peak (correlation coefficient = 0.59, p<0.001) in MCP and significantly but weakly in MCT (correlation coefficient = 0.32, p<0.001). The Borg RPE for light intensity had a median 9 (IQR 8–10) for persons with MCP and a median of 12 (IQR 10–13) for MCT. For moderate intensity the Borg RPE median was 10.5 (9.5–13 IRQ) for MCP, and 13 (12–14 IQR) for MCT. For vigorous intensity the Borg RPE median was 14 (13–15 IQR) for MCP, and 14 (13–15 IQR) for MCT (Fig 2). Further, % HR_peak correlated strongly to %VO_2peak (correlation coefficient = 0.79, p<0.001) for MCP and (correlation coefficient = 0.63, p<0.001) for MCT. There were no significant differences for HR_peak between men and women within the group of persons with MCT and group of persons with MCP. Hence, there was a significant difference between men and women in resting HR among individuals with MCP (Table 1). Percent of HR_peak for light intensity showed a median of 52% (46–59 IQR) for MCP versus 74% (66–78 IQR) for MCT. The moderate intensity median was 65% (58–75 IQR) for MCP and 79% (71–85 IQR) for MCT. Vigorous intensity showed a median for relative HR of 75% (68–84 IQR) for MCP and 84% (75–88 IQR) for MCT (Fig 2).

VO_2Peak

As in previous literature[31, 32], our results show differences in absolute (L·min^-1) and relative (ml·min^-1·kg^-1) VO_2peak between MCP and MCT. This is most likely related to the difference in cardiorespiratory response from the autonomic nervous system[13, 33] and the individual’s functional muscle volume. The results for absolute VO_2peak are comparable to previous results for persons with similar levels of injury and completeness[32, 34–39]. Hence our results are lower compared to most male athletes in Olympic sitting disciplines except for shooting[39]. We choose to use arm-ergometer for VO_2peak testing. Recent research suggests no difference in VO_2peak between arm ergometer peak testing and wheelchair tests for individuals with SCI (T3 –L1 and completeness AIS A-C)[36, 40–43]. Asynchronous arm ergometer was used during testing and has been reported to be produce equal or higher values compared to synchronous during VO_2peak testing[44]. Regardless of method, these two assessments differ from the traditional VO_2peak for the general population, which tests cardiac output. Individuals with MCP and MCT have limited muscle mass, so these tests assess aerobic capacity to a higher degree[36]. Moreover, the loss of descending sympathetic control is not always equal to the lesion level in MCT as shown when comparing high-performance athletes to non-athletes[11].

Intensity of standardized activities

This study presents data on the intensity of different standardized activities for persons with MCP and those with MCT, expressed as % of VO_2peak, % of HR_peak and self-reported perceived exertion. Further, in Table 3 we categorize activities based on % of VO_2peak, % HR_peak and Borg RPE into intensity levels (sedentary, light, moderate and vigorous). Our results are based on general physiological parameters [29, 45] and show that we slightly altered Borg RPE and % HR_peak for both groups. The results for % HR_peak for persons with MCP are comparable to those in previous studies, where 40% of VO_2peak corresponded to 61–66% HR_peak[13, 46] compared to the present ≈59% HR_peak. Likewise, 60% of VO_2peak corresponded to 73–77% HR_peak[13, 46] compared to ≈72% HR_peak in our study. For MCT, relative HR at moderate and vigorous intensity is comparable to that in previous studies[47, 48]. However, it’s difficult categorize relative HR for MCT due to attenuated heart rate responses due to impairment in descending sympathetic control. However, it’s still of interest to describe the HR for MCT on the basis of the recommended intensity levels from American college of sports medicine[9]. Consequently, it was more difficult to categorize the activities for persons with MCT and not as accurate as for MCP. So, we decided to merge the activity columns for moderate and vigorous intensity for MCP, which means that all daily activities and arm-crank 20W can be moderate or vigorous depending on % of HR, Borg RPE and speed km/h. The large variations and relative high intensity level for daily activities for MCT has also been reported earlier and that study used % heart rate reserve (HRR) to describe intensity level [49]. Moreover, we explored the use of % of HRR and it showed a lower correlation coefficient (0.6), and the already small span of HR become even smaller for MCT. The result regarding Borg RPE for MCP for different activity and intensity levels is also comparable [13, 38, 46]. For MCT there are fewer comparable studies[38, 48]. However, Goosey-Tolfrey et al[38] found similar levels for light intensity but also had difficulties to distinguish between moderate and vigorous intensity levels using Borg RPE for MCT. Previous research showing cardio metabolic benefits for SCI used a relative intensity between 50–70% of VO_2peak or 50–80 HR_peak to represent MVPA [50], this indicates that wheeling wheelchair outdoors at 4.8–5.8 km/h or arm-cranking 24-36W at Borg 12–13 could be a way of reaching that intensity level for person with MCP. Hence, for person with MCT wheeling outdoors at 2–6 km/h or arm-crank 10-15W at Borg 13–15, might be more applicable.

Strengths and limitations

The strengths of this study include the large homogenous groups, for level and severity of injury, and the extensive protocol for several different activities performed by almost all participants. Also, nearly 30% of the study cohorts were women, which enable important and clinically relevant gender comparisons to be made. There was large intra-individual variation for VO_2peak, affecting the activity-related relative VO₂ (% of VO_2peak). This reflects the heterogeneous sample of individuals included in these homogenous subsamples, which means that the included participants have different backgrounds of physical activity. It also shows the complexity of metabolic processes within persons with motor-complete SCI.

The study was limited, however, in that not all activities were represented for both Borg RPE and HR. Additionally, individual PA level may have affected how we categorized the activities. Moreover, variation in mechanical efficiency between the more standardized arm-cranking and the more variable wheeled mobility may have influenced the variation in relative intensity level of performance between these two different modes of activity. Unfortunately, we were not able to compare mechanical efficiency between the wheeled mobility and arm-cranking, as valid measurements of the power output. The results apply only to non-elite athletes within the same limits of BMI and levels of injury and injury severity. Another limitation that could influence the result is that nerve roots projecting from T7-T10 innervate the adrenal gland, and this might contribute to varying degrees of catecholamine spillover during the task. This could contaminate the observed heart-rate responses even when they were standardized to a peak performance[51].