Predictors of improved aerobic capacity in individuals with chronic stroke participating in cycling interventions

Abstract

Objective

To determine demographic and physiologic factors that predict improvement in aerobic capacity among individuals with chronic stroke participating in cycling interventions.

Design

Secondary analysis of data from two randomized clinical trials.

Setting

Research laboratory.

Participants

Individuals with chronic stroke (N=44).

Interventions

Participants were randomized to one of the following interventions: forced-rate aerobic exercise and upper extremity repetitive task practice (FE+RTP, n=16), voluntary-rate aerobic exercise and upper extremity repetitive task practice (VE+RTP, n=15), or a non-aerobic control group (control, n=13). All interventions were time-matched and occurred three times per week for eight weeks.

Main Outcome Measure

Aerobic capacity as measured by peak oxygen consumption (VO_2peak) during maximal cardiopulmonary exercise stress testing.

Results

Significant improvements in VO_{2 peak} were observed from baseline to post-intervention in the VE+RTP group (p<0.001). Considerable variability was observed among participants relating to post-intervention change in VO_{2 peak}. Among aerobic exercise participants, a multivariate regression analysis revealed that cycling cadence, baseline VO_{2 peak} and group allocation were significant predictors of change in VO_{2 peak}.

Conclusion

High exercise rate (cycling cadence) appears to be an important variable in improving aerobic capacity and should be considered when prescribing aerobic exercise for individuals with chronic stroke. Those with low VO_{2 peak} at baseline may benefit the most from aerobic interventions as it relates to cardiorespiratory fitness. Further investigation is warranted to understand the precise role of other exercise and demographic variables in the prescription of aerobic exercise for this population, and their effects on secondary stroke prevention and mortality.

It is well-accepted that aerobic capacity is compromised following stroke as peak oxygen uptake (VO_2peak) values of stroke survivors are roughly half those of age-matched healthy controls.¹ The American Heart Association/American Stroke Association acknowledged a high prevalence of physical inactivity post-stroke, recommending aerobic exercise training to improve aerobic capacity, functional mobility, and reduce risk factors for recurrent stroke.¹ Current exercise recommendations for stroke survivors include 20–60 minute aerobic exercise sessions at 40–70% of heart rate reserve (HRR), 3–5 days per week.¹ However, while aerobic capacity can improve with aerobic exercise, considerable variability in outcomes is reported among individuals, and the factors affecting VO_2peak responsiveness in individuals post-stroke have not been evaluated systematically.² A necessary step to developing evidence-based aerobic exercise recommendations for stroke survivors is understanding how specific exercise variables and participant demographics contribute to VO_2peak responsiveness.³ The aim of this study was to determine demographic and physiologic factors that predict improvement in aerobic capacity among individuals with chronic stroke participating in aerobic exercise cycling interventions.

METHODS

Cardiovascular data from two pilot randomized clinical trials (clinicaltrials.gov registration numbers NCT02076776 and NCT02494518) were analyzed for this project. Detailed inclusion/exclusion criteria in addition to a CONSORT diagram have been previously published.^4–6 The Cleveland Clinic Institutional Review Board approval was obtained and informed consent was completed by all participants.

Outcomes

At baseline and post-intervention, all participants completed cardiopulmonary exercise (CPX) testing using an incremental workload protocol administered by an exercise physiologist blinded to group allocation.^4–6 Gas analyses, including volume of oxygen consumption, were averaged over the final 30 seconds of each stage to capture steady state. Peak volume of oxygen (VO_2peak) was defined as the highest averaged sample obtained.

Intervention

Following baseline testing, participants were randomized to one of three groups: forced exercise and UE repetitive task practice (FE+RTP), voluntary exercise and UE RTP (VE+RTP), or a non-aerobic control group.^4,6 All sessions were ~90 minutes in length, with participants attending three days/week for eight weeks.

Participants in the VE+RTP and FE+RTP groups exercised for 45 minutes on semi-recumbent stationary cycles, and differed in cycling modes. The FE+RTP group exercised on a stationary bike custom-engineered with a motor that augmented pedaling rate by 30% of the participant’s self-selected cadence during the CPX test,^6,7 while the VE+RTP group exercised on a non-motorized bike at a self-selected cadence. Participants were instructed to remain in their target HR range, calculated at 60–80% of HRR using Karvonen’s formula based on results from their CPX. ^4,8 To measure exercise adherence, defined as a minimum of 60% of HRR, HR was monitored and recorded continuously in both groups using chest strap HR monitors synced to a visual display. While a minimum threshold of 60% of HRR was considered compliant, aerobic intensity was analyzed as a continuous variable using actual values of HRR as opposed to a dichotomous variable (compliant versus non-compliant). Along with the visual display of HR intended to provide continuous feedback to the participant, verbal encouragement to exercise within the target HR range was provided. While motivational strategies were not formally standardized, they were administered comparably to all participants by the same study personnel. Specific instructions on how to maintain aerobic intensity between 60–80% of HRR were not provided; rather, participants were encouraged to “work harder”, and typically utilized a combination of increasing cadence and workload on the cycle to achieve the target HR goal. The aerobic exercise protocol was not explicitly progressed during the 8-week intervention; however, in general, it was observed that participants were able to exercise within their target HR ranges for a greater percentage of time as they became increasingly fit. Following the cycling intervention, participants completed a 45-minute session of UE RTP.^4,6

Instead of aerobic exercise, the control group completed either 90-minutes of UE RTP, or 45-minutes of stroke-related education followed by 45-minutes of UE RTP.^4–6 For purposes of examining VO_2peak responsiveness to aerobic exercise, both non-aerobic groups were combined into one control group.

Statistical Analysis

Descriptive statistics were computed to describe group characteristics. A linear mixed effects model was performed to determine the interaction of group and time (baseline to post-intervention) on VO_2peak. If a significant interaction was found, post-hoc pairwise comparisons were performed with Bonferroni correction. Next, a multivariate linear regression was performed among exercisers to determine which participant characteristics and exercise variables were associated with improved VO_2peak. Participant characteristics included in the model were: age, group allocation, body mass index, sex, baseline six-minute walk test distance, and baseline VO_2peak. Exercise variables included in the model were aerobic exercise intensity (% HRR), exercise rate (pedaling cadence), and power. To check for collinearity, variance inflation factors were examined. Significance was set at p<0.05 and statistical analyses were performed using R software version 3.4.0 (https://cran.r-proiect.org).

RESULTS

Participant demographics, baseline characteristics, exercise variables, and the primary outcomes are summarized in Table 1. Exercise variables were similar across groups except cadence, which was significantly higher for the FE+RTP group compared to VE+RTP. The linear mixed effects model revealed a significant group by time interaction (p=0.035). Post-hoc analysis revealed a significant improvement in VO_2peak for the VE group from 16.6±4.0 at baseline to 18.1±5.0 at post-intervention (p<0.001). No other significant post-hoc comparisons were found among groups or time points.

Table 1.

Participant demographics by randomized group.

Participant Characteristics	FE+RTP (N=16)		VE+RTP (N=15)		Control (N=13)		p-value
Age	51±12		60±14		58±11		0.07
Male sex (versus female)	12 (75%)		10 (67%)		12 (92%)		0.26
Body Mass Index	31.6 ± 6.6		28.6 ± 6.0		32.4 ± 8.5		0.84
Race:							0.16
African American	8 (50%)		2 (13%)		5 (38%)
White	5 (31%)		12 (80%)		7 (54%)
Asian	1 (6%)		1 (7%)		0 (0%)
Other	2 (12%)		0 (0%)		1 (8%)
Hispanic ethnicity	1 (6%)		0 (0%)		1 (8%)		0.67
Dominant side affected (%)	10 (62%)		7 (47%)		6 (46%)		0.59
Months since stroke	12 [7,16]		16 [11,25]		12 [9,20]		0.50
Baseline UE Fugl Meyer score	37±8		33±11		30±9		0.15
Baseline Peak VO2 (mL/kg/min)	18.0±6.1		16.6±4.0		15.7±3.2		0.42
Baseline Six Minute Walk Test (m)	127 [106, 139]		86 [59, 116]		102 [57, 116]		0.01
Respiratory Exchange Ratio (Baseline)	1.15±0.07		1.18±0.08		1.13±0.10		0.76
Respiratory Exchange Ratio (Post-Int)	1.15±0.08		1.16±0.09		1.18±0.08		0.36
Exercise Variables
Average cadence (RPM)	74 [68, 82]		59 [53, 68]		--		0.01
Percentage of HRR (%)	0.60 [0.48, 0.65]		0.57 [0.46, 0.61]		--		0.32
Power (watts)	32.4 [24.1, 41.8]		34.75[30.1, 59.3]		--		0.25
Aerobic exercise time (min/session)	41.89±8.14		41.79±7.24		--		0.95
Outcome	FE + RTP		VE + RTP		Control		p-value for
	Baseline	Post-Int	Baseline	Post-Int	Baseline	Post-Int	Interaction
Peak VO2	18.0±6.1	18.5±5.6	16.6±4.0	18.1±5.0	15.7±3.2	15.8±3.5	0.03

Summary statistics presented as mean ± standard deviation for normally distributed data, median [Q1, Q3] for skew data, or N (%) for categorical data.

Abbreviations: FE+RTP, forced-rate exercise and repetitive task practice; VE+RTP, voluntary-rate exercise and repetitive task practice; UE, upper extremity; VO₂, volume of oxygen consumption; mL/kg/min, milliliters per kilogram per minute; m, meters; Post-Int, Post-Intervention

As shown in Figure 1, considerable inter-individual variability in change in VO_2peak was observed among participants. The multivariate regression analysis found that the overall model including all exercise and participant characteristics was significant in predicting change in VO_2peak (F₉,₂₁ = 2.66, p = 0.03, Multiple R squared = 0.53). Of all variables, cadence (p=0.01), baseline VO_2peak (p=0.02), and group allocation (p=0.02) contributed significantly to the model with higher cadence, lower baseline VO_2peak and VE group allocation predictive of improved VO_2peak. Variance inflation factors ranged from 1.2 to 4.2.

Figure 1 :

Plot depicting percent change in Peak VO₂ from baseline to post-intervention among all participants, revealing considerable variability among individuals. Color coding represents average cycling cadence across all exercise sessions or participants who were randomized to the control group (red), revealing that in general, individuals who exercised at a cadence ≥60 revolutions per minute were more likely to demonstrate improved Peak VO₂. The aerobic intensity at which each participant exercised (measured as percent of heart rate reserve) for individuals in the FE or VE groups is indicated at the end of each bar. This variable measuring aerobic intensity did not appear to markedly impact change in Peak VO₂. Lastly, group assignment to FE or VE is indicated along the x-axis.

DISCUSSION

The 8-week intervention resulted in considerable variability in change in aerobic capacity across exercise participants, ranging from a 13% worsening to a 30% improvement in VO_2peak. Given the variability of cardiopulmonary response to aerobic exercise training^9,10, it was critical to identify factors that contribute to improvements in cardiopulmonary fitness. Understanding predictive variables is particularly important in the stroke population, as diminished motor control due to hemiparesis and other stroke-related neurological sequelae may influence exercise performance.

The negative correlation between baseline VO_2peak and change in aerobic capacity suggests that individuals with the lowest cardiopulmonary function have the most to gain from the aerobic exercise intervention. Others have reported similar findings in populations with stroke¹⁰ and heart disease.⁹ Interestingly, while baseline VO_2peak was a predictor of change in aerobic capacity, baseline six-minute walk test performance was not a predictor in our cohort. This implies that lower extremity function did not influence the participants’ ability to improve aerobic capacity,¹¹ and individuals with varying levels of lower limb function can harness the cardiopulmonary benefits of AE.

Pedaling rate was also predictive of VO_2peak response, with higher cadence predicting greater improvement. While group allocation was also significant with VE participants demonstrating greater increases in VO_2peak, nine of the top ten responders exercised at ≥60 RPM. Of these nine individuals, five were in the VE group, but were able to maintain relatively high cycling cadence without the mechanical augmentation utilized in the FE mode. Macko and colleagues have reported that increased velocity during treadmill training predicted improved aerobic capacity post-stroke.¹² Thus, increased pedaling rate may have a comparable training effect as treadmill velocity. Interestingly, aerobic intensity measured as percent HRR was not predictive of change in aerobic capacity. Thus, in a stroke population, exercising at a high percent HRR may not be sufficient to elicit change in VO_2peak if the individual is exercising at a low cadence. While percent HRR should be considered in the exercise prescription, our results indicate that exercise prescriptions for survivors of stroke should include exercise at higher cadences, even at the cost of lower aerobic intensity. We have recently reported that mechanically augmenting pedaling cadence improved UE motor recovery, when FE was combined with UE RTP.^4,6 The current analysis suggests that increased cadence provided through FE or voluntarily exercising at high cadences also optimizes aerobic capacity.

Study Limitations

Our findings were based on of a relatively small, heterogeneous sample of persons with chronic ischemic and hemorrhagic stroke, and may not be generalizable to all persons post-stroke. Aerobic interventions longer than 8 weeks may be necessary to induce changes in cardiorespiratory fitness. Also, overall levels of free living physical activity or measures of motivation to be physically active were not measured, yet may provide insight into levels of fitness.

Conclusions

Our findings suggest that to maximize the cardiopulmonary benefits of aerobic exercise post-stroke, exercise prescriptions should include a cycling cadence greater than 60 RPM and 60–80% HRR. Additional studies are underway to further delineate the effect of patient characteristics and training variables on cardiopulmonary outcomes to enable clinicians to prescribe patient-specific aerobic exercise interventions to optimize health and reduce disability.

Highlights.

• Change in peak VO₂ was highly variable following an 8-week aerobic cycling program

• Higher pedaling cadence during cycling was predictive of improvements in peak VO₂

• Low baseline peak VO₂ was also predictive of improvements in peak VO₂ post-exercise

• Aerobic intensity did not predict change in peak VO₂ in persons with chronic stroke

Abbreviations

VO_2peak

peak oxygen uptake

AE

aerobic exercise

HRR

heart rate reserve

CPX

cardiopulmonary exercise

UE

upper extremity

FE+RTP

forced exercise and upper extremity repetitive task practice

VE+RTP

voluntary exercise and upper extremity repetitive task practice

6MWT

six-minute walk test

RPM

revolutions per minute