The High-Intensity Exercise Study to Attenuate Limitations and Train Habits in Older Adults With HIV (HEALTH): A Research Protocol

Abstract

The High-Intensity Exercise Study to Attenuate Limitations and Train Habits in Older Adults With HIV (HEALTH), which incorporates an exercise and biobehavioral coaching intervention, has the following overall goals: (a) to determine whether high-intensity interval training (HIIT) mitigates physical function impairments, fatigue, and impairments in mitochondrial bioenergetics of older people living with HIV (PLWH) to a greater extent than continuous moderate exercise (CME); and (b) to determine whether a biobehavioral coaching and mobile health (mHealth) text messaging intervention following HIIT or CME can promote long-term adherence to physical activity. The HEALTH study is a randomized trial of 100 older PLWH (≥ 50 years of age) who self-report fatigue and have a sedentary lifestyle. Enrolled participants will be randomized to 16 weeks of supervised HIIT or CME training, followed by a 12-week maintenance phase, involving a mHealth coaching intervention. Outcomes of the HEALTH study will inform the development of scalable, effective exercise recommendations tailored to the unique needs of aging PLWH.

With advances in treatment and care, a longer life expectancy is changing the demographics of the HIV epidemic. Nearly half of those living with HIV in the United States are now 50 years or older, and this population is estimated to grow to > 70% by 2030 (Smit et al., 2015). Yet even with effective antiretroviral therapy (ART), older people living with HIV (PLWH) often experience earlier onset of age-associated comorbidities (Marcus et al., 2020), poorer physical function (Oliveira et al., 2017; Schrack et al., 2016), and a disproportionately high symptom burden (Milanini et al., 2017). Fatigue is one of the most common symptoms, occurring in up to 88% of PLWH (Jong et al., 2010), and it is persistent and disruptive to daily life (Schreiner et al., 2020). As a multifactorial condition, fatigue can be challenging to treat, but without intervention, fatigue further contributes to impairments in key components of daily function (i.e., walking speed, balance, ability to rise from a chair) and diminished quality of life (Erlandson et al., 2014). Scalable strategies that can preserve high physical function and mitigate fatigue are urgently needed to maximize the healthspan (i.e., the part of one’s life in which they are in good health) of older PLWH.

Physical exercise is an important nonpharmacological strategy to counterbalance both the adverse effects of ART and chronic HIV infection. Across populations, including those living with or without HIV, regular exercise improves physical function and reduces fatigue, but physiological benefits are most apparent with at least a moderate intensity of exercise (Riebe et al., 2018). Most studies incorporate continuous moderate-intensity exercise (CME; Piercy et al., 2018) but, recently, high-intensity interval training (HIIT), repeated alternating short bouts of high-intensity (e.g., 80–95% of maximal heart rate [HR_max]) and lower-intensity aerobic exercise (e.g., 60% HR_max), has gained attention in improving health outcomes in people with varied health conditions (Dun et al., 2019; Ross et al., 2016). HIIT is demonstrated to be safe and to have superior efficacy in improving health outcomes compared to CME in those with chronic illnesses (e.g., coronary artery disease, diabetes; Pattyn et al., 2016). HIIT also reduces body fat, improves muscle mass, and enhances mitochondrial bioenergetics (i.e., mitochondrial respiration rate and electron transport chain activity), mechanisms proposed to underlie reductions in fatigue (S. Yang et al., 2019). Despite its benefits, little is known about HIIT for PLWH, although it has recently been shown to be a feasible and efficient strategy to improve cardiorespiratory fitness in older PLWH (Briggs et al., 2020).

Moderate-intensity aerobic and resistance exercise have been shown to improve physical function in older PLWH, with additional improvements in both physical function and fatigue from a higher-intensity exercise program (Erlandson et al., 2018). Despite these improvements, less than 50% of the participants continued to exercise at least once a week following the end of the supervised intervention, demonstrating that even when exercise is beneficial it can be difficult for PLWH to maintain long term. These data suggest that an innovative approach to exercise is needed to improve physical function, reduce fatigue, and maintain a self-directed exercise habit among older PLWH.

Successful exercise maintenance depends on factors that are different from those leading to exercise initiation and therefore requires different interventions (Maula et al., 2019). Personal coaching is helpful in exercise maintenance, but it can be difficult to scale; mobile health (mHealth) interventions are increasingly popular to support behavior change but are usually either nontailored or include one-time tailoring using participants’ baseline demographic or clinical characteristics only (Oikonomidi et al., 2019). Even though tailoring to basic demographics is likely to be helpful, additional gains can be achieved by theory-based tailoring (Noar et al., 2007), particularly when messages are tied to participants’ immediate experiences based on ecological momentary assessments or sensor data collected in real time (Cook et al., 2018). Highly tailored, theory-based mHealth interventions have been rarely tested and never for physical activity (PA) maintenance among PLWH (Oikonomidi et al., 2019).

To address these gaps in the literature, we are currently conducting the “High-Intensity Exercise Study to Attenuate Limitations and Train Habits in Older Adults with HIV (HEALTH),” which incorporates an exercise and a biobehavioral coaching intervention. The overall goals of the HEALTH study are to:

• Determine whether HIIT mitigates physical function impairments, fatigue, and impairments in mitochondrial bioenergetics of older PLWH to a greater extent than CME;

• Determine whether a biobehavioral coaching and mHealth text messaging intervention following HIIT or CME can promote long-term adherence to PA, which is a crucial component of the sustainability of the intervention.

Methods

The HEALTH Study is a randomized, parallel group, superiority trial of 100 older PLWH (≥ 50 years of age) who self-report fatigue and have a sedentary lifestyle. The study is being conducted at two sites (University of Colorado Anschutz Medical Campus and University of Washington) that provide HIV care for diverse PLWH. The timeline includes 4.5 years of research procedures, with data collection expected to start in April 2021. Enrollment and baseline procedures are expected to continue through January 2024, and all study procedures should be completed by July 2025. The study is approved by the Colorado Multiple Institutional Review Board, which is the IRB of record for this study. Participants will sign a written informed consent and will be reminded of their ability to withdraw from the study at any time. Participants will receive compensation for testing and procedures but not for the exercise sessions. Parking vouchers or bus passes may be provided, as needed, to attend the study sessions. The investigators will provide progress reports at 6-month intervals to the independent Data and Safety Monitoring Board assembled for the study. The HEALTH study is registered at ClinicalTrials.gov #NCT04550676.

Study Overview

Before enrolling in the study, participants will receive explanation about the HEALTH study procedures, potential risks and benefits, and give written informed consent. Participant demographics and medical history (i.e., clinical diagnoses, HIV lab values, current and pertinent prior medications, and comorbid health conditions) will be collected in REDCap® (Harris et al., 2009). Medical history variables will be obtained through review of the participant’s medical record, and accuracy of the information will be verified during baseline visits. A brief and targeted physical examination will focus on any difficulties with following instructions that may prevent the intervention, including vital signs, weight, lung and cardiac examination, and assessment of gait and balance.

As part of their eligibility screening and baseline assessments, participants will complete the following: medical history, physical exam, graded exercise test (GXT) with peak aerobic power (VO_2peak), blood draw, muscle biopsy, systemic and skeletal muscle mitochondrial assays, PA level, physical function (Short Physical Performance Battery [SPPB], 400-m walk test [400-MWT], and frailty), fatigue, dual-energy absorptiometry scan (DXA) for body composition, and one-repetition maximum test (1-RM).

After baseline assessments, participants will be randomized 1:1 to 16 weeks of HIIT or CME training (exercise intervention). After completing the 16-week exercise intervention, participants will repeat baseline assessments. The supervised exercise training period will be followed by a 12-week maintenance phase (biobehavioral intervention), wherein participants will be randomized to an mHealth coaching intervention or to a control condition. At the end of the biobehavioral intervention, participants will complete a final assessment of blood drawn, PA level, physical function and fatigue (Figures 1 and 2).

Figure 1.

HEALTH study procedures flowchart.

Note. *Given the potential for dropout, the second randomization will be conducted at week 16 and will also be balanced by initial randomization (2×2 factorial design). 1-RM = one-repetition maximum test; DXA = dual-energy absorptiometry scan; GXT = graded exercise test; PA = physical activity; VO2 = difference between oxygen inspired and oxygen expired in a unit of time.

Figure 2.

Visual scheme of HEALTH study procedures.

Note. HIIT and CME will be matched for caloric expenditure. The first weeks of exercise training (weeks 1 to 8) will focus on familiarization; intensity and duration of exercises will progress over the first 8 weeks to reach the target goals, and this prescription will be maintained until the end of the intervention. From weeks 8 to 16, the aerobic exercise duration will be 42 minutes for HIIT and 50 minutes for CME (main part plus warm-up and cool down). Given the potential for dropout, the second randomization will be conducted at week 16 and will also be balanced by initial randomization (2×2 factorial design). 1-RM = one-repetition maximum test; CME = continuous moderate exercise; DXA = dual-energy absorptiometry scan; GXT = graded exercise test; HIIT = high-intensity interval training; HRR= heart rate reserve; PA = physical activity; VO2 = difference between oxygen inspired and oxygen expired in a unit of time.

Population Characteristics and Recruitment

The study population will be composed of older PLWH (aged ≥ 50 years) who are virally suppressed, currently report experiencing fatigue (≥ 2.0 on either of the first two screening items on the HIV-Related Fatigue Scale), and are sedentary (defined by self-reported PA that breaks a sweat < 3 days/week, with no regular resistance exercise for 3 months preceding the study). Complete eligibility criteria are described in Table 1. The study enrollment goal is 100 PLWH. This sample size was based on power calculations assuming two-sided, two-sample/paired t-tests for between/within group comparisons of the change from baseline with a 0.05 significance level. We assume conservative within subject correlations (rho) of 0.7 or 0.5 to estimate the standard deviation (SD) of the outcome change, whereas larger correlations and repeated measures will result in additional power. For change in our primary outcome, 400-MWT, we assume an effect size comparable to Toohey (2018). Converting Toohey’s 6-minute walk to 400-MWT, assuming rho = 0.7 and a sample size of 40/group, we will have > 88% power to detect comparable mean (SD) between group differences of 45 (38) and 20 (38) second improvement in 400-MWT in HIIT and CME, respectively. For morning and evening fatigue, we assume baseline means (SD) of 19 (8) and 27 (8) points, respectively (Webel, unpublished data). With rho = 0.5, and a sample size of 40/group, we will have > 80% power to detect changes of ≥ 27% and ≥ 20%, respectively. Finally, for the change in state three mitochondrial respiration, we estimate a baseline mean (SD) of 580 (180) pmol O₂/min/mL using rho = 0.5, based on Robinson (2017). This provides > 81% power to detect a difference in the changes of 50% (HIIT) and 30% (CME).

Table 1.

Inclusion and Exclusion Criteria of the Study

Inclusion criteria

• Age ≥ 50 years; • Sedentary lifestyle, defined by self-reported physical activity that breaks a sweat <3 days/week, with no regular resistance exercise for 3 months preceding study; • Fatigued (≥ 2.0 on either of the first two screening items on the HIV-Related Fatigue Scale); • Prescribed HIV antiretroviral therapy for ≥ 12 months, with no current use (within 1 year) of older drugs with established mitochondrial toxicity (i.e., stavudine, didanosine, zidovudine); • HIV-1 RNA level < 200 copies/mL, for a minimum of 12 months prior to enrollment, with an allowed blip to 500 copies/mL presuming repeat assessments are below 200 copies; • For women, postmenopausal status due to confounding of menstrual status in mitochondrial response to exercise; • Cell phone with ability to receive text messaging in order to participate in the biobehavioral maintenance intervention; • Able to speak, read, and write in English.

Exclusion criteria

• Body mass index < 18 or > 48 kg/m2 or weight over 450 pounds (due to limitations of the dual-energy absorptiometry machine); • Use of sex hormone therapy or other hormone replacement, if on for ≤ 3 months (stable doses for > 3 months will be permitted); • Anemia (hemoglobin ≤ 9 g/dL for women or ≤ 10 g/dL for men) due to contribution to fatigue; • Diagnosis of mitochondrial disease; • Active substance abuse or other factors that could prevent compliance or safety with study visits, at the discretion of the site investigator; • For participants undergoing the muscle biopsy only, use of anticoagulant therapy other than low-dose aspirin that cannot be held for at least 7 days for the muscle biopsy. Aspirin and nonsteroidal use will be permitted but will be held for 7 days prior to the muscle biopsy and can be resumed following the biopsy; • Due to the expected fatigue associated with COVID-19 and potential infection risk, anyone with a diagnosis of COVID will not be eligible for enrollment until at least 30 days after symptom resolution and return to baseline level of function as determined by provider evaluation; • Reasons for medical exclusion, as determined by study site physician:   ○ Uncontrolled hypertension, defined as resting systolic blood pressure > 150 mm Hg or diastolic blood pressure > 90 mm Hg;   ○ Unstable ischemic heart disease (e.g., angina, ST-segment depression) or serious arrhythmias at rest or during the graded exercise test without negative follow-up evaluation will be cause for exclusion;   ○ New York Heart Association class III or IV congestive heart failure, clinically significant aortic stenosis, uncontrolled angina, or uncontrolled arrhythmia;   ○ Pulmonary disease requiring the use of supplemental oxygen at rest or with physical exertion;   ○ Malignancy requiring chemotherapy or radiation therapy within 24 weeks prior to enrollment;   ○ Poorly controlled diabetes, as evidenced by hemoglobin A1c > 8.5, documented within 6 months of study visit or current use of insulin;   ○ Surgery/trauma/injury/fracture within 24 weeks prior to enrollment that may impact a subject’s baseline functional testing and ability to exercise;   ○ Balance impairments that may impact ability to safely exercise as reported by the participant or in their medical record;   ○ Orthopedic problems (e.g., severe osteoarthritis, rheumatoid arthritis) that greatly limit the ability to perform moderate intensity resistance exercise (e.g., unable to be properly positioned in exercise equipment or to have severely restricted range of motion even after modifications have been made).

We expect to enroll a highly diverse population of older PLWH. Recruitment strategies include: a) invitation during clinic visits, b) advertisement of the study to individuals who have participated in or expressed interest in previous studies, c) presentation of the study at local HIV Community Advisory Boards, d) distribution of pamphlets in the HIV clinics, local Community Advisory Boards, local AIDS Service Organizations, and relevant local events, and e) providing information to HIV providers who may refer interested volunteers. Retention will be facilitated by regular contact with participants during the first part of the study, and by daily tailored digital messages during the second part of the study, as well as by participant incentives.

Exercise Intervention

Participants will complete 16 weeks of in-person supervised aerobic and resistance training on 3 nonconsecutive days per week, for a total of 48 exercise sessions. Strategies to maximize adherence include submaximal exercise training loads calculated from the participants’ pretraining capacity, modifying resistance exercises if needed to accommodate preexisting movement restrictions, progressive exercise prescriptions that provide time for cardiovascular and musculoskeletal adaptations, rescheduling missed exercise sessions, and parking vouchers. The exercise intervention was designed and will be coordinated and monitored by exercise physiologists experienced with older adults with HIV. Participants will be supervised by experienced exercise trainers during every exercise session. Aerobic training will be composed of CME or HIIT plus the same resistance training exercises for both groups. During the exercise intervention, the physical function, fatigue, and 1-RM assessments will be repeated every 4 weeks (at weeks 4, 8, and 12). The first weeks of training will focus on familiarization with the equipment and exercises at lower volume and intensity. The intensity and duration of exercises will progress over the first 8 weeks to reach the target goals, and then this prescription will be maintained until the end of the intervention (Figure 2).

Aerobic training.

For the aerobic training component, participants will be randomized to either CME or HIIT, which will be matched for caloric expenditure (i.e., isocaloric) assuming 5 kcal/L O₂ and 50 min/session duration for CME to meet the Department of Health and Human Services PA guidelines (at least 150 minutes of moderate aerobic exercise per week; Piercy et al., 2018). For the final goals of HIIT group, following a warm-up at 50% of heart rate reserve (HRR), higher- and lower-intensity exercise bouts will alternate: five bouts of 4-minute vigorous-intensity exercise (up to 90% HRR) with four 3-minute bouts of moderate-intensity aerobic exercise (50% HRR). For the CME group, following a warm-up at 50% HRR, the participant will walk/jog continuously at 60% HRR (Riebe et al., 2018). HHR will be calculated using the supine resting heart rate obtained before the GXT and maximal heart rate achieved during the GXT. Both groups will complete a comparable cool-down period. From weeks 8 to 16, the aerobic exercise duration will be 42 minutes for HIIT and 50 minutes for CME.

Resistance training.

Resistance training will be the same for HIIT and CME. After an initial familiarization goal of two sets of 8 to 10 repetitions at moderate intensity (50% of 1-RM) in the first weeks, participants will gradually progress to perform three sets of 8 to 10 repetitions at vigorous intensity (75–80% of 1-RM) for the target goals. Load adjustment will occur every 4 weeks as guided by 1-RM testing, to provide a progressive stimulus as participants become more conditioned. The exercises consist of lateral pulldown, leg press, and chest press, all on weight-stack equipment.

Biobehavioral Intervention

After completing 16 weeks of supervised exercise, participants will be randomized to a 12-week coaching intervention (i.e., monthly telephone coaching and daily tailored text messages) vs. a control condition (i.e., daily generic text messages with information about exercise). The goal of this biobehavioral intervention is to promote long-term adherence to physical activity; it was designed by a clinical psychologist and will be supervised by licensed mental health professionals. During the biobehavioral intervention period, all participants will complete physical function and fatigue assessments every 4 weeks (weeks 20 and 24 of the HEALTH study, Figure 2).

Coaching component.

Three telephone coaching sessions will be provided at weeks 17, 20, and 24. The exercise trainers will use motivational interviewing (MI), a well-established counseling intervention that has been used to support PA among healthy adults and people living with chronic diseases (Lundahl et al., 2013). During each session, the trainer will ask about current PA, barriers to adhering to PA, and any factors that help to maintain PA over time. The trainer will use standard MI strategies such as reflective listening and open-ended questions to explore the participant’s intentions to exercise as well as factors that interfere with exercise. The trainer may provide additional education based on the participant’s stated concerns, using materials from the structured exercise phase to reinforce earlier messages. However, no preset educational content will be delivered. The focus of MI is, instead, to explore and resolve the participant’s ambivalence about PA using a guiding and empowering counseling style.

mHealth component.

Monthly telephone MI coaching will be augmented with daily text messages tailored to the individual participant’s self-reported symptom experiences and barriers to exercise on that specific day. The text-messaging mHealth component will begin with daily survey items assessing common barriers to PA maintenance identified in prior research, such as boredom, competing priorities, negative mood, environmental barriers, low self-efficacy, low motivation for exercise, fatigue, or other physical symptoms (van Stralen et al., 2009). Tailored messages will address the top barrier identified based on the individual participant’s survey responses that day and will be randomly selected from a list of barrier-matched items that are further varied based on a range of effective daily behavior-change strategies proposed by two minds theory (Cook et al., 2018). Two minds theory is an interdisciplinary approach to understanding health behavior that explains the gap between the intentions and behaviors of an individual by postulating that intentions are one of the multiple inputs considered for in-the-moment decisions, along with habits, attention, and in-the-moment perceptions and emotional reactions (Cook et al., 2018). Additional messages will be developed based on participants’ feedback about PA barriers and facilitators and specific suggestions on message wording or content elicited during qualitative interviews as part of the study protocol. Because of the random element in message selection and the plan for ongoing message development, participants who continue to report the same barriers will receive different text messages; this strategy was identified in prior research as key to prevent boredom and maintain participants’ attention (Cook et al., 2015). To further enhance motivation, tailored messages will be generated automatically by an algorithm but “signed” by the participant’s exercise trainer to enhance the interpersonal connection in between scheduled calls.

Biobehavioral control.

The control group will complete the same daily survey delivered to the coaching group but will only receive texts from the study team with generic information about exercise (e.g., “It’s important to exercise for at least 30 minutes 5 times a week.”) or generic encouragement (e.g., “Keep exercising. Physical activity helps you the most when itś a long-term habit!”). These text messages will be primarily social or informational in content and serve to maintain involvement and enhance retention of the control group, while also blinding the participant to the biobehavioral intervention. In the control group, the messages will not be signed by the coach, and the content will not be tied to daily survey responses.

Outcome Measures

Physical function.

Physical function will be measured by a 400-MWT, SPPB in its original and modified (mSPPB) forms, and Fried Frailty Criteria. The 400-MWT will take place along an indoor corridor with a marked 25-meter course. Participants will be instructed to walk at a brisk pace that they can maintain for the full eight laps. Participants will be allowed to rest while standing for up to 60 seconds at a time, if necessary. Time to complete the test, or distance completed and time stopped for those unable to complete the test, will be recorded. The SPPB is a brief measure of physical performance that includes a 4-meter timed walk, five times sit-to-stand, and balance testing of increasing difficulty (Guralnik et al., 1994). The SPPB is a well-regarded, valid (test-rest reliability = 0.87), objective assessment of physical function (particularly lower extremity function), that is associated with short-term mortality, disability, hospitalizations, and nursing home admission. The mSPPB improves discrimination of physical function in people with less severe impairments by using 10 times sit-to-stand, increasing each balance test to 30 seconds instead of 10, and adding a single leg stand (Simonsick et al., 2001). The Fried Frailty Criteria considers measures of weakness (handgrip strength), slowness (4-meter timed walk), low level of PA, self-reported exhaustion, and unintentional weight loss (predictive validity for clinical outcomes ranging from hazard ratio = 1.11–2.24; Fried et al., 2001). Each criterion receives a score of 1 when they are present, and the sum score of these five criteria classifies people into one of three frailty stages: not frail (score 0), pre-frail (score 1–2), and frail (score 3–5).

Fatigue.

Diurnal changes in fatigue will be assessed using the Lee Fatigue Scale (LFS, internal reliability ranging from α = 0.91–0.96; Lee et al., 1991). This validated, 13-item visual analog scale asks participants to rate, on a scale of 0 to 10 (higher scores indicate higher fatigue severity), how much fatigue they are feeling “right now.” Participants will be instructed to complete the LFS within 30 minutes of waking on seven consecutive mornings and within 30 minutes before going to bed on the same seven consecutive evenings. We will use the recommended LFS scoring in which all items are averaged into morning and evening fatigue scores. The 7-day averages will be our primary fatigue outcomes: total morning fatigue and total evening fatigue scores. We will also administer the Patient-Reported Outcomes Measurement Information System fatigue measure (PROMIS F-SF, internal consistency reliability α = 0.84; M. Yang et al., 2019), a global measure of fatigue. PROMIS F-SF includes seven items with response options on a five-point Likert scale, ranging from 1 = “never” to 5 = “always” (just one item is reverse scored so that higher scores indicate greater fatigue).

Blood draw and muscle biopsy.

A 12-hour fasting blood draw will be collected during the morning. Following the blood draw, specimens of the vastus lateralis muscle will be obtained by percutaneous biopsy from participants who have refrained from exercise for ≥ 48 hours using the technique described by Evans et al. (1982). After cleansing the thigh with chlorhexidine, 1% lidocaine without epinephrine will be injected subcutaneously. A 3- to 5-mm incision will be made in the skin and fascia over the belly of the vastus lateralis, and 100 to 150 mg of muscle tissue will be removed using a Bergström side-cutting biopsy needle (Pelomi, Denmark). Participants will be contacted 48 to 72 hours post-procedure to ensure no complications and to be reminded on wound care to ensure healing. The muscle specimens will be divided for immediate mitochondrial bioenergetic measures and the remaining specimen will be frozen in liquid nitrogen and stored at −80 C.

Mitochondrial assays.

Using the vastus lateralis tissue specimens and the blood drawn, we will measure skeletal muscle and systemic peripheral blood mononuclear cells (PBMCs) mitochondrial bioenergetics (i.e., mitochondrial respiration rate and electron transportation chain complex activity) using the Oroboros Oxygraph-2k (O2k) respirometer with modified techniques described by Pesta and Gnaiger (2012). For mitochondrial content analysis, frozen tissue will be homogenized using a MagNA Lyser (Roche Molecular Systems, Branchburg, NJ). Total RNA will be extracted from the supernatant using the Trizol Plus RNA Purification Kit (Life Technologies, Carlsbad, CA). Expression of citrate synthase and other mitochondrial markers of signaling, function, and dynamics (AMPK, eNOS, PGC-1alpha, mitochondrial complexes I-V, Mnf1, OPA1) will be measured by established enzymatic assay (Spinazzi et al., 2012) and Western blot techniques using commercially available antibodies (Cell Signaling Technology, Inc., Beverly, MA).

Physical activity level.

To understand the effectiveness of our biobehavioral intervention, each participant will wear an ActiGraph® GT3X/1 monitor (ActiGraph, LLC, Fort Walton Beach, FL) during approximately 10 consecutive days on their nondominant hip, starting two days after the muscle biopsy. PA endpoints are time spent in moderate-to-vigorous PA and steps per day. These endpoints will be set with the adult cut-points for tri-axial accelerometers (Sasaki et al., 2011). A valid wear cycle will have data recorded for ≥ 10 hours/day for ≥ 4 days, including at least 1 weekend day. Non-wear time will be defined as 0 counts/minute for ≥60 minutes. We will require consecutive epochs outside of the activity threshold for wear time to resume. Participants not meeting standards will be asked to repeat the ActiGraph® monitoring period. The data meeting wear time criteria will be analyzed. Data will be sampled at 30 Hz, using 60-second epochs and the normal filter (Migueles et al., 2017), and all the calculations will be conducted using the ActiLife software. Participants will also maintain exercise logs describing the PA they complete during the biobehavioral intervention phase (weeks 17–28).

Graded exercise test and VO_2peak.

Supine resting heart rate (beats/min) will be measured after a 30-minute rest, prior to the start of the GXT. The treadmill GXT will begin at a comfortable walking speed and 0% elevation. Speed will be maintained and the grade increased by 2% every 2 minutes, until the participant reaches volitional exhaustion or the test is otherwise terminated (Erlandson et al., 2018). A 12-lead electrocardiogram will be monitored throughout. VO_2peak will be used to determine the effects of training (difference from baseline to week 16). The O₂ and CO₂ content of expired air will be measured continuously by open circuit spirometry and averaged every 30 seconds using an automated online system. In the absence of evidence that VO_2max has been attained (≥ 2 of: a plateau in VO₂, respiratory exchange ratio ≥ 1.10, maximal heart rate within 10 beats of age-predicted) the maximum measured VO₂ will be deemed VO_2peak. Increases in VO_2peak in response to training will verify that the exercise was of sufficient intensity. Maximal heart rate will be determined in the same interval that VO_2max (or VO_2peak) is reached.

Dual-energy absorptiometry scan.

Total body DXA scans will be performed by trained technicians on Hologic instruments (Marlborough, MA) for estimation of total body mass, lean body mass, and fat mass.

One-repetition maximum test.

The 1-RM (i.e., maximal weight that can be lifted one time using correct form through the full range of motion) will be used to prepare resistance exercise prescriptions and to determine the effects of resistance training. Participants will be familiarized with each of the resistance machines at the study sites by performing 8 to 10 repetitions of a light load (~50% of predicted 1-RM). After 1 minute of rest, participants will perform a greater load (~80% of estimated 1-RM) through the full range of motion for the first exercise. After each successful repetition, the weight will increase until a failed attempt occurs. Participants will rest for 1 minute between each attempt; the 1-RM will be attained within five attempts and approximately 3 minutes of rest between each exercise. To facilitate recovery and reduce the effect of fatigue on performance, exercises will alternate between the upper and lower body.

Randomization

After screening, participants will be assigned with equal probability to one of the two exercise intervention arms (HIIT or CME). Randomization procedures will be site-specific and will use permuted blocks with stratification by sex and age, to ensure similar distribution in each arm. After completion of 16 weeks of exercise, participants will be randomized a second time to one of two biobehavioral interventions (mHealth coaching or control). Given the potential for dropout during the exercise intervention, the second randomization will also be balanced by sex, age, and exercise arm (2×2 factorial design). The randomization sequences will be created by the study statistician, who will develop the code and program the randomization instrument in REDCap®. The allocation to study arm will remain blinded to laboratory personnel.

Analysis

Baseline characteristics will be compared between groups prior to comparisons of endpoints, with log transformation for normality, as appropriate. Age, sex, smoking, exercise intensity, and body mass index will be considered for inclusion into regression models, although if matching and/or balance by age and sex is achieved, these will be precision variables, rather than confounders. Differences in response by ART class; prior ART exposure to zidovudine, stavudine, didanosine; use of statin; psychiatric medication burden; and use of hormone replacement therapy will also be explored. Given repeated measurements, a linear mixed effects model will be used with change from baseline as the primary outcome.

We will implement efforts to minimize loss to follow-up. To address missing data, we will conduct a thorough investigation of the mechanisms for missing data. For outcomes measured repeatedly over time, mixed models, sensitivity analysis will include multiple imputation (Schafer, 1999). Once missing values have been imputed, each dataset will be analyzed using standard complete data. We will calculate final parameter estimates and their standard errors using Rubin’s formula for combining results from multiply imputed data sets (Rubin, 1996; Schafer, 1999). We will analyze data and report final study results with and without employing the multiple imputation strategy and will carefully examine and describe discrepancies found. Finally, given the potential for nonignorable dropout, we will accommodate dropout using a semiparametric varying-coefficient approach (Forster et al., 2012; Moore et al., 2017). The primary analyses will assume that the slope beyond dropout is linear, with sensitivity analyses utilizing a slope that is attenuated by 50% and a zero slope beyond dropout, similar to a last value carried forward. A sensitivity analysis using only data from completers will be conducted.

Expected Outcomes

We expect to observe greater improvements in physical function and fatigue with HIIT, driven in part by changes in mitochondrial bioenergetics. We also expect that HIIT will result in greater exercise satisfaction and likelihood of long-term continuation of exercise. Using an mHealth coaching intervention for self-directed exercise that combines MI and tailored text messages during the maintenance phase, we seek to develop the ideal PA “cocktail” to promote healthspan among older PLWH in the current era of ART.

Conclusion

As older PLWH experience more comorbidities and complications, nonpharmacological approaches to improve physical function and alleviate symptoms of the growing population are urgently needed. The HEALTH study will generate scientifically rigorous knowledge on physical function and fatigue responses to exercise, the associated mitochondrial adaptations, and investigate strategies to instill sustained, self-directed exercise behavior. These data will inform the development of scalable, effective exercise recommendations tailored to the unique needs of aging PLWH.

Key Considerations.

• Older people with HIV often experience earlier onset of age-associated comorbidities, poorer physical function, and a disproportionately high symptom burden. Fatigue is one of the most common symptoms, is persistent and disruptive to daily life, and contributes to impairments in key components of daily function.

• High-intensity interval training is demonstrated to be safe and to have superior efficacy in improving health outcomes compared to continuous moderate exercise in those with chronic illnesses, also reducing body fat, improving muscle mass, and enhancing mitochondrial bioenergetics (i.e., mitochondrial respiration rate and electron transport chain activity), mechanisms proposed to underlie reductions in fatigue.

• The High-Intensity Exercise Study to Attenuate Limitations and Train Habits in Older Adults With HIV (HEALTH) is expecting to generate greater improvements in physical function and fatigue with high-intensity interval training, and to result in greater exercise satisfaction and likelihood of long-term continuation of exercise. This would promote healthspan among older PLWH in the current era of ART.