Feasibility of Remote High-Intensity Interval Exercise Training in People with Spinal Cord Injury: A Pilot Study

Abstract

Purpose: Recent studies have shown that high-intensity interval training (HIIT) can improve cardiometabolic health in individuals with spinal cord injury (SCI); however, many barriers remain for individuals with spinal cord injury to participate in exercise such as lack of time, accessible equipment and facilities, and transportation. The use of telehealth interventions may be a form of exercise delivery that can ease the burden on the participant and lead to greater exercise participation. Thus, the purpose of this study was to determine the feasibility and efficacy of a home-based telehealth HIIT arm crank exercise training program for individuals with spinal cord injury. Methods: Participants were randomly assigned to 16 weeks of telehealth HIIT arm crank exercise training or a no-exercise control group. Body composition, resting energy expenditure (REE), blood lipids, insulin sensitivity, blood pressure, aerobic capacity (VO₂ max), and a qualitative interview were assessed at baseline and at 16 weeks post intervention. Results: Six individuals (four male and two female, mean age 52.7 ± 10.2 years) with spinal cord injury were recruited for this study. Four out of five HIIT participants showed improvements in aerobic capacity, insulin sensitivity, and resting energy expenditure. Three qualitative themes emerged: (1) convenience and perceived benefits were critical elements of engagement; (2) high-intensity exercise elicited time-sensitive responses; and (3) trainers played a key role in promoting strong program adherence. Conclusions: Overall, we found that this program could be easily implemented and per-formed at home in individuals with spinal cord injury. We also found that participants enjoyed the 1:1 training sessions with a telecoach and that the intervention was easy to adhere to, as demonstrated by participant attendance. There is a need for future randomized controlled trials to determine the efficacy of telehealth exercise training for improving cardiometabolic health in spinal cord injury.

1. Introduction

Spinal cord injury (SCI) affects over two million individuals worldwide, with most cases resulting from traumatic events such as falls, vehicular accidents, and work- or sports-related injuries [1]. Individuals with chronic SCI often experience severe muscle atrophy and increased adiposity, conditions that can be further exacerbated by low physical activity levels. Moreover, individuals with spinal cord injury often experience increased inflammation, dyslipidemia, insulin resistance, and reduced aerobic and muscular fitness, resulting in a threefold higher risk of heart disease, diabetes, cancer, and obesity compared to age- and weight-matched nondisabled individuals [2,3,4,5].

The high rate of physical inactivity among people with spinal cord injury (~50% sedentary) [1] represents a significant medical and public health concern that demands further study on identifying promising, dose-specific exercise strategies that can be readily achieved by this underactive/inactive population. Exercise training has been clinically proven to delay and, in many instances, reverse health conditions associated with cardiometabolic diseases in people without disabilities [6,7]. However, it remains difficult for exercise physiologists and healthcare professionals to convince people to adhere to current exercise recommendations of at least 150 min/week of moderate-intensity or 75 min/week of vigorous-intensity aerobic exercise [8,9].

Individuals with spinal cord injury experience significant disparities in physical activity participation compared to the general population. Key barriers include structural limitations (inaccessible equipment), time-intensive logistics (transportation and dressing), and a lack of specialized personnel [10]. Low-volume high-intensity interval training (HIIT) is a promising alternative that may mitigate these challenges while delivering comparable cardiometabolic benefits.

Recent reviews and a few smaller studies suggest that HIIT may lead to greater adherence, enjoyment, and overall health improvements in individuals with spinal cord injury [11,12,13,14]. Our previous research has demonstrated that just six weeks of only two HIIT sessions per week using an arm crank ergometer at a fitness facility significantly improved insulin sensitivity (9%), cardiovascular fitness (12.2%), and muscular strength (~15%) in individuals with spinal cord injury [13]. However, there were significant challenges with recruiting participants to attend in-person exercise sessions.

The use of telehealth interventions may be a form of exercise delivery that can ease the burden on the participant and lead to greater exercise participation. Advancements in smartphone technology and internet streaming have revolutionized home-based healthcare delivery [15]. For individuals with spinal cord injury, telehealth-based exercise could bypass common obstacles like transportation issues, inaccessible facilities, and the lack of specialized trainers [10]. Because low-volume HIIT is both time-efficient and effective for cardiometabolic health, delivering it via telehealth may be an ideal way to reduce environmental barriers. Consequently, this study aimed to evaluate the feasibility and effectiveness of a home-based telehealth HIIT program using arm crank ergometry for those with spinal cord injury.

2. Methods

2.1. Participants

A total of 6 individuals (4 male and 2 female, mean age 52.7 ± 10.2 years) with longstanding spinal cord injury were recruited for this study. Participants were eligible for the study if they met the following criteria: (1) a traumatic SCI at the C5–L2 level; (2) an AIS classification of A, B, C, or D (indicating motor/sensory complete or incomplete injuries); and (3) a minimum of three years post-injury. Exclusion criteria included cardiovascular or renal disease, orthopedic complications, or participation in a structured exercise program within the previous three months. Subjects were recruited via computer-generated lists from the UAB SCI Model System and the UAB/Lakeshore Research Collaborative databases.

2.2. Study Design

This study was a 16-week randomized controlled trial (Clinical Trials registration: Telehealth High-Intensity Interval Exercise and Cardiometabolic Health in Spinal Cord Injury, NCT04 940598, 16 June 2021). However, due to recruitment challenges during the COVID-19 pandemic and its aftermath, the number of participants enrolled was limited. Thus, this manuscript presents the feasibility and efficacy of a home-based telehealth HIIT arm crank exercise training program among a small cohort of individuals with spinal cord injury.

2.3. Randomization and Study Procedures

Six participants were recruited for this feasibility study. They were randomly assigned to either an immediate-start HIIT group or a waitlist control group. Permuted block randomization was used, and the randomization list was computer-generated. The randomization was not stratified. Participants and assessors were blinded to the group assignment. Three participants were assigned to the HIIT group and three were assigned to the waitlist control group. After completing the control arm, two participants from the control group chose to join the HIIT group. As a result, five participants completed the HIIT arm and three completed the control arm of the study. The duration for the intervention or wait period was 16 weeks. Thus, control group participants who participated in the HIIT group were in the study for 32 weeks. After participants completed the study, they were asked to participate in a one-on-one qualitative interview to describe their perceptions of completing the program.

2.4. Data Collection Protocol

At each data collection point (baseline and post), participants attended three visits: Visit 1, following an overnight fast, body composition, blood pressure, and resting energy expenditure (REE) were assessed; Visit 2, during which an oral glucose tolerance test (OGTT) was performed and baseline venipuncture blood samples were collected and stored at −80 °C until blood lipids, glucose, and insulin were analyzed; and Visit 3, where peak oxygen consumption (VO₂ peak) was assessed using indirect calorimetry during a graded arm cycle ergometer test, and peak anaerobic power was determined from a standard 30 s Wingate arm cycle test on a Lode Arm Crank Ergometer (Lode, Groningen, The Netherlands). Participants also completed the SF-36v2 survey during the preliminary testing.

2.5. Clinical Measures

2.5.1. Peak Oxygen Uptake (VO₂ Peak)

Aerobic capacity (VO₂ peak) was assessed on a Lode arm cycle ergometer. Participants were instructed to begin cycling at 10 W for the first 2 min, which increased by 10 W until voluntary fatigue. Due to the variability in heart rate across participants with different levels of SCI, VO₂ peak was determined by at least 2 of the following: (1) volitional exhaustion, (2) failure to maintain 60–65 revolutions per minute, (3) RER ≥ 1.10, and (4) rate of perceived exertion (RPE) >18 using the 6–20 Borg scale. Minute ventilation, oxygen uptake, and carbon dioxide production were continuously monitored and recorded via open circuit spirometry (ParvoMedics, Salt Lake City, UT, USA).

2.5.2. Peak Anaerobic Power Assessment

Peak power was determined by the standard 30 s Wingate protocol on a Lode arm cycle ergometer. Before testing, participants performed a 5 min warm-up at 25 W, incorporating three brief sprints, followed by a 5 min recovery period. For the assessment, subjects were instructed to hand-cycle at maximal velocity. Standardized verbal encouragement was provided throughout to ensure participants maintained their highest possible cadence. Post-test metrics included peak power, mean anaerobic power, fatigue rate, and relative peak power.

2.5.3. Oral Glucose Tolerance Test and Insulin Sensitivity

Whole-body insulin sensitivity was assessed via an oral glucose tolerance test (OGTT) at the UAB Clinical Research Unit (CRU). Following an overnight fast, participants arrived at the CRU between 7:00 and 9:00 AM; post-training assessments were conducted at least 24 h after the final exercise bout. A flexible intravenous catheter was placed in the antecubital vein for blood sampling. Two baseline samples were collected within the first 20 min to determine basal glucose and insulin levels. At time zero, participants consumed a 75 g glucose drink within a 5 min window. Subsequent blood samples were collected at 10, 20, 30, 60, 90, and 120 min. Following the test, samples were immediately centrifuged, and the serum was separated and stored at −80 °C. Analysis was performed at the UAB Center for Clinical and Translational Sciences Metabolism Core.

Insulin was measured on a TOSOH Bioscience AIA900 (TOSOH Bioscience, San Francisco, CA, USA) using immunofluorescence (minimum sensitivity 0.5, intra-assay CV% 1.49, and inter-assay CV% 3.95). Glucose was measured with a SIRRUS analyzer (Stanbio Laboratory, Boerne, TX, USA) using glucose oxidase (minimum sensitivity 2, intra-assay CV% 1.28, and inter-assay CV% 3.95).

2.5.4. Blood Lipids

Biochemical analyses were conducted at the University of Alabama at Birmingham (UAB) within the Core Laboratories of the CRU, the Nutrition Obesity Research Center, and the Diabetes Research Center. Total cholesterol, high-density lipoprotein cholesterol (HDL-C), and triglycerides were measured using an automated analyzer (Sirrus; Stanbio Laboratory, Boerne, TX, USA). Low-density lipoprotein cholesterol (LDL-C) concentrations were subsequently derived using the Friedewald equation.

2.5.5. Body Composition

Total body composition, including fat mass, percent body fat, lean mass, and fat-free mass were measured by dual-energy X-ray absorptiometry (DEXA). The DEXA analysis was conducted following a 10 min wait period in which participants completed a series of questionnaires. Participants laid supine with arms at their sides on a padded table. Scans were analyzed using ADULT software version 1.35 (Lunar Radiation).

2.5.6. Resting Energy Expenditure (REE)

Resting energy expenditure (REE) was assessed between 7:00 and 9:00 AM following a 12 h overnight fast. After completing a five-minute dual-energy X-ray absorptiometry (DEXA) scan, participants remained in a supine position. Measurements were conducted in a quiet, dimly lit room while participants remained awake and stationary. Oxygen uptake (VO₂) and carbon dioxide production (CO₂) were monitored continuously using a computerized, open-circuit indirect calorimetry system with a ventilated hood (ParvoMedics, Murray, UT, USA). Following a 10 min equilibration period, data from a subsequent 20 min steady-state period were used for analysis.

2.6. Treatment Groups

2.6.1. Control Group

Participants randomly assigned to the control group (n = 3) underwent normal daily activities and were asked to refrain from participation in structured exercise. Control participants were also invited to participate in the HIIT group following their completion of the control arm of the study. Two of the three participants agreed to do so.

2.6.2. Exercise Group

The tele-exercise HIIT intervention was delivered through a custom, wireless internet-based system. The telehealth platform integrated several hardware and software components: a Samsung Galaxy Tab 2 10.1 tablet mounted on a Standzfree universal floor stand, a Zephyr Bioharness 3 wearable monitor (Medtronic, Boulder, CO, USA), and a custom web application. The Bioharness 3 transmitted real-time heart rate data to the tablet via Bluetooth^®, while the application streamed physiological data and video to a secure, dedicated server. This configuration enabled telecoaches to monitor participant data with a maximum of 5 s latency, allowing for simultaneous video conferencing and the delivery of written instructions. Telecoaches utilized this real-time feedback to adjust exercise intensity and correct movement quality during sessions conducted on a Hudson Fitness tabletop upper body ergometer (UBE-BDP, Fabrication Enterprises, White Plains, NY, USA). Safety procedures included the following: (1) live intensity monitoring through remotely measured heart rate and rating of perceived exertion; (2) video monitoring of contraindications to exercise (e.g., shortness of breath, excessive redness or soreness), losing consciousness or the ability respond verbally to coach conversation or prompts; and (3) coaches were trained to call emergency response for the participant in the case of an emergency.

The HIIT intervention consisted of 32 sessions performed twice weekly over 16 weeks, with at least 24 h between sessions to ensure adequate recovery. Participants scheduled sessions at their convenience. Training intensity was derived from peak anaerobic power achieved during the baseline arm crank Wingate test. Based on preliminary findings where 50% of peak power proved unsustainable, a more feasible intensity of 30% was utilized. This workload remained significantly higher than the power output at ensuring a high-intensity stimulus. Each 20 min session included four cycles of four minutes at 5% peak power followed by a 30 s interval at 30% peak power, concluding with a two-minute recovery at 5%.

2.7. Qualitative Interview

After completing the intervention, participants underwent a one-on-one semi-structured interview with a member of the research team (B.L.). The interview lasted no longer than 60 min and was conducted by phone. The interviews were guided by an interview script that included approximately 10 questions that pertained to the following: overall thoughts of the program, likes and dislikes, recommendations to improve the program, thoughts of the intervention characteristics (exercise intensity, duration, modality), and the rationale for their attendance to the program. The interviewer was an expert in qualitative research who had conducted more than 400 interviews pertaining to exercise and disability. Interviews were audio recorded so that they could be transcribed for analysis.

2.8. Analysis

For quantitative outcomes, descriptive statistics such as means and standard deviations were obtained for baseline and 6-month study values and the pre-to-post change. Due to the very small sample size of each group, between-group and within-group comparisons were not performed. SAS software (version 9.4; SAS Institute, Cary, NC, USA) was used to conduct descriptive analyses for quantitative outcomes. Qualitative outcomes were analyzed using thematic analysis [16], specifically the 6 steps proposed by Braun and Clarke [16]. Two members of the research team (B.L. and J.A.) analyzed the transcriptions and generated codes, sub-themes, and themes. To enhance the quality of the results, the analysts acted as “critical friends” [17] and we paid careful attention to ensure that codes, sub-themes, and themes were saturated as best as possible (relevantly important and strong enough to warrant report) given the small sample size. The analysis was underpinned by an interpretivist approach, that is, the analysts acknowledged that the participants could have multiple explanations for a phenomenon that could be shaped by their backgrounds and interactions with others, including the research team and interviewer. The research team’s philosophical assumptions aligned with dialectical pluralism [18], meaning the research team had separate belief systems for the quantitative and qualitative methods (positivism and interpretivism, respectively). The lead analyst (B.L.) was a qualitative expert, and the second (J.A.) was a graduate student who was trained by the lead analyst.

3. Results

A total of six participants enrolled in the study: three were randomized to the immediate start HIIT group and three to the waitlist control group. In total, two participants in the control group opted to participate in the HIIT group following their completion of the control arm of the study. This resulted in n = 5 for HIIT and n = 3 for the control. All six participants underwent assessment at baseline. The HIIT group performed sixteen weeks of tele-exercise followed by post-testing. The waitlist control group began their training immediately following their post-test evaluation, which served as their pre-test score for their training component. Characteristics of the participants are displayed in

Table A1. Descriptive characteristics of study participants.

Subject	Gender	Age	Injury Level	Height (cm)	Weight (kg)	Race
Control (n = 3)
2	M	65.4	T8 Complete	186.44	110.1	1
4	F	49.7	T11/T12 Incomplete	168.4	58.6	2
5	F	54.1	L1/T12 Incomplete	153.67	114.1	1
Mean (SD)		56.4 (8.1)		169.5 (16.4)	94.3 (31)
HITT (n = 5)
1	M	42.2	C6 Incomplete	180.34	106.5	2
2	M	66.8	T8 Complete	186.44	111.5	1
3	M	56	T5	179.83	93.6	1
4	F	50.6	T11/T12 Incomplete	168.4	63.6	2
6	M	37.4	L1	170.18	61	1
Mean (SD)		50.6 (11.6)		177.0 (7.6)	87.2 (23.7)

Descriptive characteristics of study participants. Race: 1 = Caucasian, 2 = African American.

.

3.1. VO₂ Peak

There were no pre–post changes in VO₂ peak in the three participants who were in the waitlist control group (mean pre–post change (SD) = −0.1 mL/kg/min (0.1 mL/kg/min). However, four of the five HIIT participants improved their VO₂ peak by an average of 16%. Improvements ranged from as high as 84% in participant 1 to 6.9% in participant 6. The mean pre–post change was not different (mean pre–post change (SD) = 1.9 mL/kg/min (2.8 mL/kg/min). Descriptive statistics are shown in

Table A2. Baseline and post-exercise changes in the HIIT group.

Variable	Baseline	16 Week	Pre-to-Post Change (Post—Pre)
(n = 5)	Mean (SD)	Mean (SD)	Mean (SD)
VO2 Peak (mL/kg/min)	11.6 (3.0)	13.6 (2.2)	1.9 (2.8)
Peak Power (Watts)	313.6 (44.5)	315.0 (67.5)	1.4 (38.1)
Matsuda	3.5 (1.8)	5.3 (3.5)	1.8 (2.4)
HOMA	2.5 (1.1)	1.9 (0.9)	−0.6 (1.1)
Fasting Glucose (mg/dL)	95.2 (3.5)	95.6 (5.6)	0.4 (2.6)
Fasting Insulin (mg/dL)	10.7 (4.3)	7.9 (3.3)	−2.8 (4.4)
Cholesterol (mg/dL)	179.2 (36.1)	193.8 (49.5)	14.6 (21.5)
Triglycerides (mg/dL)	132.6 (46.4)	122.8 (39.1)	−9.8 (27.3)
HDL (mg/dL)	45.6 (6.9)	52.0 (9.6)	6.4 (5.0)
LDL (mg/dL)	107.1 (30.4)	117.5 (48.1)	10.4 (23.5)
Body Weight (kg)	87.2 (23.7)	86.5 (24.1)	−0.7 (5.2)
Lean Mass (kg)	48.8 (15.2)	48.8 (14.8)	−0.02 (0.99)
Fat Mass (kg)	35.7 (10.4)	34.9 (9.9)	−0.8 (2.8)
Arm Fat (%)	30.5 (3.0)	29.3 (3.0)	−1.1 (1.3)
Resting Energy Expenditure (kcal/day)	1444.8 (250.2)	1611.2 (386.7)	166.4 (189.0)
Systolic Blood Pressure (mmHg)	149.6 (9.2)	146.2 (12.5)	−3.4 (5.7)
Diastolic Blood Pressure (mm Hg)	92.4 (9.2)	94.4 (5.9)	2.0 (8.5)

and

Table A3. Baseline and post-exercise changes in the control group.

Variable	Baseline	16 Week	Pre-to-Post Change (Post—Pre)
(n = 3)	Mean (SD)	Mean (SD)	Mean (SD)
VO2 Peak (mL/kg/min)	10.1 (1.0)	10.0 (0.9)	−0.1 (0.1)
Peak Power (Watts)	281.7 (55.5)	244.7 (101.9)	−37.0 (114.1)
Matsuda	6.1 (4.2)	6.5 (6.0)	0.4 (3.8)
HOMA	2.8 (1.5)	5.8 (6.6)	3.1 (5.3)
Fasting Glucose (mg/dL)	148.3 (85.6)	84.0 (25.7)	−64.3 (110.6)
Fasting Insulin (mg/dL)	7.6 (4.1)	41.7 (57.6)	34.1 (58.0)
Cholesterol (mg/dL)	178.0 (4.4)	188.0 (25.4)	10.0 (23.4)
Triglycerides (mg/dL)	126.3 (30.1)	102.3 (44.3)	−24.0 (58.7)
HDL (mg/dL)	51.7 (0.6)	57.7 (8.0)	6.0 (8.0)
LDL (mg/dL)	101.1 (8.0)	109.7 (22.0)	8.6 (14.8)
Body Weight (kg)	94.3 (31.0)	98.7 (30.7)	4.4 (2.7)
Lean Mass (kg)	48.8 (13.9)	49.6 (14.3)	0.8 (1.0)
Fat Mass (kg)	45.5 (18.0)	45.0 (17.6)	−0.5 (0.5)
Arm Fat (%)	38.0 (12.7)	37.4 (11.8)	−0.5 (0.9)
Resting Energy Expenditure (kcal/day)	1737.3 (321.7)	1830.7 (294.5)	93.3 (133.7)
Systolic Blood Pressure (mmHg)	149.33 (6.66)	132.0 (18.0)	−17.3 (23.8)
Diastolic Blood Pressure (mm Hg)	94.3 (0.6)	92.3 (0.6)	−2.0 (0)

. Individual VO₂ peak responses for HIIT and control are shown in

Table A4. Individual Pre- and Post-Treatment Data Responses.

Group	Sub.	Power/Wgt/Fat	Glucose/Ins/HOMA	SBP/DBP/REE	Lipids (TC/TG/HDL/LDL)
HIIT	1	320 → 375/107 → 98/41 → 35	98 → 102/15 → 10/3.8 → 2.5	153 → 150/106 → 94/1570 → 1880	142 → 132/78 → 88/44 → 46/81 → 69
	2	375 → 365/118 → 118/49 → 50	96 → 100/11.5 → 13/2.6 → 3.0	149 → 137/85 → 86/1830 → 2060	140 → 160/131 → 150/45 → 56/68 → 73
	3	300 → 320/95 → 95/39 → 38	92 → 92/5 → 7.5/1.0 → 1.5	151 → 148/93 → 97/1350 → 1690	212 → 258/195 → 175/38 → 47/135 → 173
	4	250 → 205/62 → 64/26 → 27	97 → 97/7.5 → 4/1.5 → 0.8	160 → 164/95 → 102/1200 → 1240	200 → 195/99 → 88/57 → 67/123 → 113
	6	335 → 320/59 → 59/24 → 26	91 → 89/14.5 → 5/3.2 → 1.1	135 → 132/83 → 93/1280 → 1190	203 → 226/159 → 109/44 → 44/129 → 161
CTRL	2	300 → 365/113 → 114/47 → 46	104 → 104/12 → 13.5/2.9 → 3.3	145 → 132/94 → 92/1890 → 1830	182 → 205/110 → 152/52 → 50/109 → 126
	4	225 → 205/63 → 64/27 → 27	93 → 93/4 → 4/1.0 → 0.7	146 → 150/94 → 92/1370 → 1540	175 → 200/106 → 66/52 → 66/102 → 120
	5	330 → 175/115 → 115/63 → 62	245 → 55/6.5 → 108/4.1 → 13.2	157 → 114/95 → 93/1960 → 2130	177 → 158/161 → 87/51 → 57/93 → 85

Key (Pre → Post): Column 1: Power (Watts), Weight (kg), Fat (%); Column 2: Glucose (mg/dL), Insulin (mg/dL), HOMA; Column 3: Systolic BP (mmHg), Diastolic BP (mmHg), Resting energy expenditure; Column 4: Chol, Trig, HDL, LDL (all mg/dL).

.

3.2. Peak Power

Peak anaerobic power improved in 2/5 participants following HIIT and 1/3 participants in the control group. The mean pre–post change in both groups were similar (HIIT: mean pre–post change (SD) = 1.4 W (38.1 W); control: mean pre–post change (SD) = −37.0 W (114.1 W). Descriptive statistics are shown in Table A2 and Table A3. Thus, HIIT did not appear to lead to improvements in peak anaerobic power output. Individual peak power responses for HIIT and the control are shown in Table A4.

3.3. Oral Glucose Tolerance Test

Descriptive statistics for the Matsuda Index and homeostatic model assessment for insulin resistance (HOMA) are shown in Table A2 and Table A3, and individual responses are shown in Table A4. Insulin sensitivity assessed by the Matsuda Index improved in 4/5 participants following HIIT. There was a 50% improvement in the HIIT group and a 4% improvement in the control group. HOMA improved in 3/5 participants following HIIT, while there was an increase in 2/3 participants in the control group. The mean pre–post change in both groups were similar for the Matsuda Index (HIIT: mean pre–post change (SD) = 1.8 (2.4); control: mean pre–post change (SD) = 0.4 (3.8)) and for HOMA (HIIT: mean pre–post change (SD) = −0.6 (1.1); control: mean pre–post change (SD) = 3.1 (5.3)).

3.4. Blood Lipids

Descriptive statistics for fasting cholesterol, triglycerides, HDL, and LDL are shown in Table A2 and Table A3, and individual responses are shown in Table A4. The mean pre–post change in both groups were similar for fasting cholesterol (HIIT: mean pre–post change (SD) = −14.6 mg/dL (21.5 mg/dL); control: mean pre–post change (SD) = 10.0 mg/dL (23.4 mg/dL)); for triglycerides (HIIT: mean pre–post change (SD) = −9.8 mg/dL (27.3 mg/dL); control: mean pre–post change (SD) = −24.0 mg/dL (58.7 mg/dL)); for HDL (HIIT: mean pre–post change (SD) = 6.4 mg/dL (5.0 mg/dL); control: mean pre–post change (SD) = 6.0 mg/dL (8.0 mg/dL)); or for LDL (HIIT: mean pre–post change (SD) = 10.4 mg/dL (23.5 mg/dL); control: mean pre–post change (SD) = 8.6 mg/dL (14.8 mg/dL)). Responses were variable throughout. Cholesterol decreased in 2/5 participants following HIIT and 1/3 participants in the control group, whereas triglycerides decreased in 3/5 participants following HIIT and 2/3 participants in the control group. Additionally, LDL decreased in 2/5 participants following HIIT and 1/3 in the control group, whereas HDL increased in 4/6 participants following HIIT and 2/3 in the control group.

3.5. Body Composition

Descriptive statistics for body weight, lean mass, fat mass, and % arm fat are shown in Table A2 and Table A3, and individual responses are shown in Table A4. The mean pre–post change in both groups were similar for body weight (HIIT: mean pre–post change (SD) = −0.7 kg (5.2 kg); control: mean pre–post change (SD) = 4.4 kg (2.7 kg)); for lean mass (HIIT: mean pre–post change (SD) = −0.02 kg (0.1 kg); control: mean pre–post change (SD) = 0.8 kg (1.0 kg)); for fat mass (HIIT: mean pre–post change (SD) = −0.8 kg (2.8 kg); control: mean pre–post change (SD) = −0.5 kg (0.5 kg)); or for % arm fat (HIIT: mean pre–post change (SD) = −1.1% (1.3%); control: mean pre–post change (SD) = −0.5% (0.9%)). Participant 1 showed a reduction in body weight following HIIT, which was due to a 5 kg reduction in fat mass, with no major changes to lean mass. Participant 4 increased body weight by 2 kg following HIIT, which was due almost entirely to increased lean mass with no changes in fat mass. % arm fat decreased in 4/5 participants in the HIIT group and 1/3 participants in the control group. These data demonstrate significant variability in body composition responses following HIIT; however, there were apparent improvements in both lean mass and % arm fat in several participants following HIIT.

3.6. Resting Energy Expenditure and Systolic and Diastolic Blood Pressure

Descriptive statistics for resting energy expenditure, systolic blood pressure, and diastolic blood pressure are shown in Table A2 and Table A3, and individual responses are shown in Table A4. The mean pre–post change in both groups were similar for resting energy expenditure (HIIT: mean pre–post change (SD) = 166.4 kcal/day (189.0 kcal/day); control: mean pre–post change (SD) = 93.3 kcal/day (133.7 kg)); for systolic blood pressure (HIIT: mean pre–post change (SD) = −3.4 mmHg (5.7 mmHg); control: mean pre–post change (SD) = −17.3 mmHg (23.8 mmHg)); or for diastolic blood pressure (HIIT: mean pre–post change (SD) = 2.0 mmHg (8.5 mmHg); control: mean pre–post change (SD) = −2.0 mmHg (0 mmHg)). We found that 4/5 participants improved their resting energy expenditure by an average of 11% following 16 weeks of HIIT. Improvements ranged from as high as 25% in participant 3 to 3% in participant 4. In total, 2/3 participants showed modest improvements in the control group, with a mean change of 5%. There were no significant changes in either SBP or DBP following HIIT, as both mean SBP and DBP were nearly identical pre- and following HIIT.

3.7. Qualitative Themes

Considering the small sample size, we developed three themes that we thought were highly saturated. First, we found that convenience and perceived benefits were critical elements of engagement. Second, high-intensity exercise elicits time-sensitive responses. Third, telecoaches facilitated strong attendance to the program.

3.8. Convenience and Perceived Benefits

Engagement in the program was strong and was due to four factors. First, participants noted that exercise participation seemed to improve their physical health in the form of strength, energy, and body composition. Observing and anticipating health benefits across the program made participants enjoy their time exercising. One participant reported that exercise reversed the physiological deconditioning that occurred from being ill with COVID-19. Combined with a convenient telehealth communication tower at home, these benefits provided participants with a sense of accomplishment and fostered strong adherence to the program.

3.9. High-Intensity Exercise Elicits Time-Sensitive Responses

The exercise regime was physiologically demanding for the participants. The intensity and structured nature of the program was new for all participants, who reported that they had never been in a program that matched the intensity and duration that was prescribed. There were two notable time points where internal motivation to attend the sessions wavered. The early stages of the intervention (e.g., first week) were identified as a difficult time point where the physiological demand of the high-intensity exercise highly stressed participants’ work output capability. Time was needed for participants’ bodies to physiologically adapt to the early stages of training, which caused internal motivational challenges. In contrast, the mid-to-late stages of the training also caused motivational challenges due to the stacking physiologic workload that accumulated throughout the program. Unsupervised training likely would not have had strong adherence, given that the telecoaches were identified as the primary agents that overcame early and mid-to-late motivational challenges.

3.10. Trainers Facilitated Strong Attendance

Real-time supervision during exercise by a telecoach was instrumental for keeping participants engaged in their exercise prescription. The telecoach motivated participants in two ways: (1) by providing them with a sense of accountability to show up to the scheduled sessions and (2) by providing a sense of engagement during exercise sessions. These influences seemingly required telecoaches who were personable with the participants. The telecoaches were highly praised for their professionalism, positivity, and passionate and personable communication skills. Notably, participants reported that they likely would not have adhered well to the exercises without support from the telecoach because of the physiological stressors involved with high-intensity exercise training (discussed in the next section).

4. Discussion

This study was designed to examine the feasibility and efficacy of a remote high-intensity exercise training program for individuals with spinal cord injury. Specifically, we wanted to determine if this form of exercise training could be effectively delivered via telehealth technology using arm crank ergometry while also enhancing study recruitment, improving exercise adherence, and promoting cardiometabolic health outcomes in individuals with spinal cord injury. Overall, our findings indicate that the telehealth HIIT program was both feasible and accessible for individuals with spinal cord injury to perform at home. Participants reported enjoying the 1:1 training sessions with a telecoach and, despite the physiological demands, found the intervention manageable, as reflected by strong attendance and adherence.

A key highlight reported by participants during the qualitative interview was the convenience of exercising from home at their preferred time of day. The transportable equipment was easy to set up and use, no adverse health events occurred, and the flexibility of the weekly training schedule contributed to 100% adherence and no dropouts among enrolled participants.

In terms of efficacy, HIIT participants showed improvements in aerobic capacity, insulin sensitivity, and resting energy expenditure. While we did not have a large enough sample to examine statistical differences, we did see improvements in these cardiometabolic outcomes in four of the five participants in the intervention group compared to one out of three in the control group. The individual data reveals that HIIT participant 1 was the most successful responder, showing a simultaneous 55 W increase in power, a 6% reduction in body fat, and improved cholesterol and blood pressure. HIIT participant 6 also saw significant metabolic gains, specifically a dramatic drop in insulin and HOMA levels. In contrast, the HIIT group had outliers like participant 3, whose cholesterol and LDL levels spiked sharply despite gains in power. Within the control group, participant 5 was an outlier, showing a large loss in power and fluctuations in glucose and insulin responses. Additionally, it would be interesting to observe differences between complete and incomplete injury; however, we only had one participant (participant 2) with a complete injury in this study. This participant did not improve following HIIT for most outcomes other than an increase in REE.

One of the objectives in this study was to implement a longer training period. This involved training participants 2 days a week for 16 weeks. We found that two days per week of training may provide sufficient stimulus to improve cardiorespiratory fitness when performed at high intensities (30% of peak anaerobic power). This protocol was determined based on our previous work in which we performed the same training protocol in a lab setting [13] and showed significant improvements in the VO₂ peak. This previous investigation was the first study to show that low-volume HIIT could lead to improvements in VO₂ peak in individuals with spinal cord injury. A number of previous studies [13,19,20,21,22] have shown that various forms of exercise, including arm crank electrical stimulation hybrid, HIIT, and moderate-intensity training (MIT), can all improve aerobic capacity.

Similar to our individual VO₂ peak responses, we also found improvements in glucose tolerance in the majority of our participants. We have previously shown, using the same HIIT protocol in the lab, that we could improve insulin sensitivity after only 6 weeks of exercise training [13]. The current study found a 50% mean improvement in insulin sensitivity following 16 weeks of HIIT. While not all studies in individuals with spinal cord injury have shown improvements in glucose tolerance or insulin sensitivity following exercise training, [23] there are several that have shown benefits following different types of exercise training, including HIIT, moderate-intensity training (MIT), hybrid exercise, combined aerobic and resistance training, and locomotor training [24,25,26,27,28]. Thus, there is convincing data demonstrating that various forms of exercise that involve muscle contractile activity can be beneficial for improving glucose tolerance and insulin sensitivity in individuals with spinal cord injury.

Resting energy expenditure has been shown to be significantly lower in individuals with spinal cord injury [29]. This has been attributed to reduced muscle mass and sympathetic nervous system activity [30,31]. Thus, identifying ways to increase muscle mass or the activation of existing muscle mass and the activation of the sympathetic nervous system is important for increasing resting energy expenditure in SCI. In four of our five HIIT participants, we found a mean increase in REE of nearly 170 kcal following HIIT. This increase in REE is clinically meaningful, as it equates to nearly a 1200 kcal increase REE each week. In our previous study [13], we found no change in REE following six weeks of HIIT and about a 100 kcal decrease in REE following MIT. This reduction in REE following exercise training has been shown in a previous study [32] that assessed REE following 16 weeks of either functional electrical stimulation cycling or arm crank ergometry for 16 weeks. They showed a decreased REE of nearly 200 kcal following their exercise training. Thus, it is possible that performing high-intensity exercise may be important for overcoming this compensatory reduction in REE that has been seen following lower intensity exercise programs in individuals with spinal cord injury.

4.1. Limitations

The study was limited by its small sample size, with only five participants in the HIIT group and three in the control group. This was primarily due to the COVID-19 pandemic but may also reflect the high time commitment required for onsite data collection. Furthermore, while a major limitation in SCI research is the lack of inclusion of more sophisticated outcomes, such as body composition, aerobic capacity, and blood lipid and glucose measures, it may be more feasible to focus on more convenient approaches that can be collected from home when designing exercise training studies in SCI. The burden of multiple on-site visits may be too burdensome on participants and restrict recruitment. Thus, the data obtained by including these physiological and biological outcomes may not be translatable due to limitations in sample sizes.

4.2. Future Studies

There is a need for randomized controlled trials that are well powered to determine the efficacy of telehealth exercise training in individuals with spinal cord injury. Additionally, while studies are needed that include physiological and biological health outcomes to better understand the impact of exercise on cardiometabolic health in individuals with spinal cord injury, it may be more feasible to focus on home-based physiological data outcomes, such as wearable technologies, remote monitoring of blood pressure, glucose, body weight, and other clinical outcomes. Alternatively, while the use of telehealth to implement exercise in this population is promising, in order to address the small sample size limitations when incorporating site visits for outcome measures, the use of well-controlled multisite studies may be more effective for recruitment barriers.

5. Conclusions

The study found that a home-based, telehealth-delivered HIIT program is feasible, safe, and well-accepted by individuals with chronic spinal cord injury. Participants achieved 100% adherence and emphasized the convenience of exercising at home with real-time telecoach support. Physiologically, most HIIT participants improved aerobic capacity, insulin sensitivity, and resting energy expenditure, whereas changes in the control group were minimal. Recruitment challenges and the small sample size limited statistical comparisons and highlight the need for larger, multisite studies with remote outcome assessments. Overall, the findings suggest that telehealth HIIT is a promising strategy to reduce exercise barriers and improve cardiometabolic health in people with spinal cord injury.