Inclusion/exclusion criteria
Main inclusion and exclusion criteria for HI-PACE are shown in Table 1. The HI-PACE study is designed to intervene in sedentary, overweight and obese AAs at high risk for T2D. Thus, we plan to enroll 60 sedentary, overweight and obese AA adults (body mass index [BMI]: 25.0-45.0 kg/m2, age 35-65 years). All participants will be sedentary/low active and not participating in exercise training at time of enrollment (<20 minutes, ≤2 days per week for the last 3 months). Major exclusions for the HI-PACE study include diagnosed type 1 or type 2 diabetes (or fasting glucose >125 mg/dL, or use of diabetes medication), known cardiovascular diseases (e.g. heart failure, serious arrhythmias, peripheral vascular disease), previous stroke or myocardial infarction, excessively high resting systolic (>180 mmHg) or diastolic (>100 mmHg) blood pressure, significant medical conditions, life-threatening conditions, pregnancy or plans to become pregnant, and other medical conditions that are contraindicated for exercise training. Additionally, individuals who plan to diet, engage in weight loss, or demonstrate non-compliance during screening visits will be excluded.
The study protocol has been approved by the East Carolina University (ECU) Institutional Review Board and is registered on Clinicaltrials.gov (NCT02892331). This study protocol was prepared on the basis of the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines. This SPIRIT Checklist is available as Additional file 1.
Recruitment and pre-screening
A detailed summary of the study visits is described in Table 2. Recruitment material will be disseminated through newspaper (general readership and an AA-specific newspaper), targeted social media advertisement (e.g. Facebook, Instagram), email sent through company employee listservs (i.e. ECU, Pitt Community College, Greenville government), and local organizational contacts in the Pitt County, North Carolina area (e.g. churches, physician offices, libraries, barbershops, etc.). In addition, a study website will be created to provide basic study information and to serve as a mechanism for web-screening potential participants. Web-screening will be performed using an online survey where basic inclusion/exclusion criteria questions can be completed and subsequently reviewed by study staff. This online survey is created using an online research database, REDCap (Nashville, TN) [40], which is connected to the main study database. Interested individuals can also contact HI-PACE staff by calling the study phone number or by directly emailing the research coordinator. After this, study staff will phone-screen individuals for major aspects of the inclusion/exclusion criteria and provide additional information about study participation. Individuals who are eligible and still interested following phone screening will progress to Screening visit 1.
Screening visits
During Screening visit 1, the research coordinator will describe all properties of study participation, answer questions from individuals, and obtain informed consent. Following consent, the research coordinator will screen participants for the full inclusion/exclusion criteria, collect contact/demographic information, and review prescribed medications (individuals will be required to bring in prescribed medications for verification). For inclusion/exclusion purposes, height and weight (without shoes) will be measured to calculate BMI (kg/m2) and seated resting blood pressure will be assessed via an automated blood pressure monitor (HEM-907XL, Omron Healthcare Co., Kyoto, Japan).
Individuals will be screened for ample time to participate in the exercise intervention by filling out an exercise calendar form, where they will be asked to identify specific days and times (and back-up times) they are available to exercise at our facility (See Appendix A). Staff will also conduct a standardized interview with potential participants in which: a) weekly time commitments, b) responsibilities for family care (i.e. child and elder), c) distances of home and work from our exercise facility, d) personal motivations for exercising, e) levels of familial support, and d) any other barrier(s) that would affect study adherence will be evaluated (See Appendix B). The exercise calendar and interview are intended to screen-out individuals depicting high risk for non-compliance and/or drop-out during the 24-week study. Previous studies utilizing similar methodologies exhibited high exercise training adherence and study retention [41,42].
Individuals who are still eligible at this point will wear a Fitbit Flex (Fitbit Inc., San Francisco, CA) and an activPAL accelerometer (PAL Technologies Ltd., Glasgow, UK) for 7 continuous days to assess baseline non-exercise physical activity level. The Fitbit Flex will be worn on the non-dominant wrist to obtain data on steps, miles, intensity, and calorie expenditure each day of wear (blinded to individual). Study staff will apply the activPAL by rolling a nitrile sleeve over the entire device and wrap an 8x10 cm sheet of transparent medical dressing completely around it to act as a waterproof barrier. Staff will rub an alcohol-based prep pad around the site, place the distal end of the activPAL towards the knee, and apply a separate sheet of medical dressing over the monitor to complete application to the leg (waterproofing method). The activPAL accelerometer will be worn on the individual’s mid-thigh, without removal for the entire 7 days. The activPAL measures postural aspects of time spent sitting/lying down, standing, and walking in hours per day, as well as energy expenditure (MET-hours per day), steps per day, and number of sit-to-stand transitions.
The HI-PACE study will utilize a REDCap database to store and track physical activity data during the physical activity assessment. During each day the devices are worn, REDCap surveys will be sent automatically to individuals’ email address to ask if the activPAL and Fitbit devices were worn on the previous day and if there were any extended periods of non-wear time. The purpose of the survey is to: 1) increase the accuracy of the physical activity assessment by being able to eliminate days affected by non-wear, and 2) prompt individuals to wear the devices consistently. Since changes in non-exercise physical activity can confound exercise-related changes in outcome measures [43,44], it is necessary to ensure that participants in the HI-PACE study can regularly wear the devices.
Following completion of the baseline physical activity assessment (7 days), individuals will return for Screening visit 2 in the morning, in the fasted state. The study nurse will perform a fasting blood draw and immediately send the sample to a clinical laboratory (LabCorp Inc., Burlington, NC) for complete metabolic panel, lipid panel, insulin level, and blood chemistries. Pre-menopausal women will also be required to complete a pregnancy test. The Fitbit and activPAL will be retrieved for accelerometer data to be downloaded and recorded in the study database. Upon completion of the screening visits, individuals will be scheduled for the baseline assessment visit.
Assessment visits (baseline, mid-intervention, and follow- up)
A flow chart of the present study is shown in Fig. 1. Primary (i.e. CRF) and secondary outcome measures (i.e. arterial stiffness, mitochondrial measures, insulin sensitivity, skeletal muscle oxidative capacity, quality of life, and food-frequency questionnaires [FFQ]) will be obtained at baseline and at follow-up (week 24). At mid-intervention (week 12), CRF, resting blood pressures, and anthropometry (i.e. body mass, waist circumference, and BMI) will be re-evaluated. Measurements of arterial stiffness, muscle biopsy, IVGTT, and NIRS will be obtained in this order, during the same visit (baseline and follow-up). Whereas, CRF, anthropometry, and body composition will be obtained during the same visit of a separate week. Randomization into a study group will occur upon completion of all baseline assessments. For an overview of the schedule of enrollment, allocation, intervention, and assessments, see Fig. 2 for the completed Standard Protocol Items: Recommendations for Interventional Trials (SPIRITS) figure.
Primary Outcome – change in maximal oxygen uptake (VO2 max)
The primary outcome measurement is CRF due to the well-established association between CRF levels and risk of T2D [8,10,11,27]. We will measure CRF as VO2 max (the gold standard) from a maximal exertion treadmill test under supervision of a physician. Maximal exercise testing will be conducted on a treadmill (Cardiac Science TM65, Davis Medical Electronics, Bothell, WA) under a modified Balke protocol. For the warm-up, participants will initially walk at a speed of 2.0 mph at 0% grade for 2 minutes. Subsequently, we will increase treadmill speed to 3.0 mph to begin the treadmill test. During the test, we will increase treadmill grade by 2.5% every 2 minutes until volitional exhaustion. Gas exchange (i.e. VO2, VCO2) and pulmonary ventilation will be measured continuously using a TrueOne 2400 Metabolic Measurement Cart (Parvo Medics, Salt Lake City, UT). Heart rate, blood pressure, rating of perceived exertion, and electrocardiogram will be monitored and recorded before, during, and after the exercise test. The electrocardiogram will be cleared by the study physician prior to participant randomization. A valid maximal exercise test will meet two of three end criteria: 1) elevated respiratory exchange ratio (RER) (≥1.10); 2) plateauing of VO2; or 3) within ±5 bpm of age-predicted maximal heart rate.
Secondary outcome measures
Secondary outcome measures include: change in insulin action, arterial stiffness, mitochondrial function, body fat, C-reactive protein, and psychological surveys.
Insulin sensitivity
Insulin sensitivity will be assessed at baseline and follow-up (24 hours following the last exercise session for MOD-INT and HIGH-INT groups) via an intravenous glucose tolerance test (IVGTT). After collection of fasting blood samples, glucose (Dextrose 50%) will be injected into a catheter placed in the antecubital vein, at a dose of 0.3 g/kg body weight. Subsequently, blood samples will be obtained at the following time points: 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 19, 22, 25, 30, 40, 50, 60, 70, 80, 90, 100, 120, 140, 160, and 180 minutes. Insulin will be injected at minute-20 of the test, at a dose of 0.025 U/kg body weight. Blood samples will be centrifuged and stored at -80°C until sample analysis for glucose and insulin. Insulin sensitivity index will be determined through a minimal model [45]. Follow-up IVGTT will be assessed within 18-24 hours of the MOD-INT and HIGH-INT participants’ last exercise session.
Arterial stiffness
Carotid-to-femoral PWV and aortic blood pressure parameters will be measured using a SphygmoCor XCEL (AtCor Medical, Itasca, IL). Carotid-femoral PWV, an index of the degree of arterial stiffness, is the gold standard measurement of arterial stiffness [46]. Arterial stiffness measurements will occur during the morning in a quiet, temperature-controlled room at baseline and follow-up. Prior to each measure, participants will refrain from vigorous exercise, tobacco, caffeine, and alcohol for at least 12 hours, as well as large meals for at least 6 hours. Participants will take their prescribed medications which will be logged and repeated at follow-up. The methodology for arterial stiffness measurements will adhere to a position stand by the American Heart Association [47].
For aortic blood pressure and stiffness measurements, participants will be in the seated position for a 5-minute rest period. Following rest, aortic blood pressure (e.g. brachial blood pressures, aortic blood pressures, etc.) and stiffness (e.g. augmentation index, wave reflection, etc.) parameters will be obtained on the basis of acquisition of brachial artery pressure waveforms with application of a generalized transfer function to derive the central aortic pressure waveform, from which estimates of aortic blood pressures are generated. Three measurements will be performed for aortic blood pressure parameters with a 1-minute rest period between each measurement.
Subsequently, PWV will be obtained in the supine position following a 15-minute rest. During rest, body surface measurements will be measured via a Gulick tape measure (Baseline, Fabrication Enterprises, White Plains, NY) in triplicate to determine the distance traveled by the pulse wave between the carotid and femoral artery sites. Study staff will palpate and mark the carotid artery pulse (between the larynx and sternocleidomastoid muscle in the neck), the sternal notch (superficial landmark of aortic arch), and femoral artery pulse (over the ventral thigh halfway between the pubic symphysis and anterior superior iliac spine) [48]. The mean distance for each site will be used for PWV calculation. Pressure waveforms at the carotid arterial site will be acquired via applanation tonometry and electrocardiographic gating, while the femoral arterial site will be simultaneously acquired using the oscillometric device within the SphygmoCor XCEL. The PWV measurement will be conducted in duplicate, and the mean of these measurements will be the reported value. Both measurements must be within 0.5 meters per second to be considered acceptable for data purposes. If the two measurements differ by more than 0.5 meters per second, a third measurement will be obtained, and the reported value will be the median of the three measurements.
Mitochondrial Function
A percutaneous muscle biopsy (100-200 mg of tissue) will be obtained using sterile techniques at baseline and follow-up from the vastus lateralis with a 5-mm Bergström muscle biopsy cannula with suction (Stille Surgical Instruments, Eskilstuna, Sweden), as previously described [49]. Briefly, participants will lie supine with legs extended (0º flexion) and two operators will spray ethyl chloride on the biopsy site and administer 1% lidocaine at each level of subcutaneous tissue, stopping superficial to the fascia. Following 2-3 minutes to allow for the local anesthetic effects, a 1-cm incision will be made through the skin and subcutaneous tissues, parallel to the femur, until an incision is made through the muscle fascia. The operator will utilize the biopsy cannula to locate the fascia incision site and advance the needle past the fascia, angled downward towards the floor to rapidly clip and collect the muscle sample with suction by the second operator. Following the biopsy, the sample will be trimmed of visible adipose tissue, weighed on a scale (AL54, Mettler Toledo, Columbus, OH), and snap-frozen in liquid nitrogen to be stored at -80ºC until analysis at study completion.
Peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC1-α), COX IV, GLUT-4, and CPT-1 content will be determined; these proteins were selected as they are downstream of PGC1-α and represent distinct steps in oxidative metabolism [50]. Approximately 50 mg of muscle tissue will be homogenized (T8 Ultra Turrax; IKA, Wilmington, NC) in 20 volumes of cell lysis buffer (50 mM HEPES, 12 mM sodium pyrophosphate, 100 mM sodium fluoride, 100 mM ethylenediaminetetraacetic acid, 10 mM sodium orthovanate, 1% Triton X-100) supplemented with a protease and phosphatase inhibitor cocktails (Sigma-Aldrich, St. Louis, MO). Lysates will be sonicated for 5 seconds, rotated for ~1 hour at 4°C, and centrifuged at 13,500 x g for 15 minutes at 4°C. Protein concentration for each sample homogenate will be determined via a commercially available bicinchoninic acid protein assay kit (Pierce, Rockford, IL). Aliquots containing 30 µg of total protein will be diluted in 4x Laemmli Buffer (BioRad, Hercules, CA) with 5% ß-mercaptoethanol (ßME) at a 3:1 ratio, prior to heating at 70°C for 10 minutes. Denatured samples will be brought to room temperature, loaded onto a 10% polyacrylamide gel, separated by SDS-PAGE, and transferred to nitrocellulose membranes. Membranes will be blocked with Odyssey Blocking Buffer (OBB; Li-Cor, Lincoln, NE) for 1 hour and incubated with primary antibodies. Membranes will be washed with TBST and incubated with an anti-rabbit or anti-mouse fluorophore-conjugated secondary antibody (1:20,000, Li-Cor) in OBB supplemented with 0.1% Tween-20 for 1 hour. The membranes will then be washed with TBST followed by TBS prior to being scanned on the Odyssey CLx Imaging System (Li-Cor) and quantified on Image Studio software (V4.0.21; Li-Cor). GAPDH will be used as a loading control.
Citrate synthase activity will be determined with a colorimetric reaction using reagents in a commercial kit (Sigma CD0720, St. Louis, MO), as in a previous study [51]. A 10- to 15-mg piece of muscle will be diluted 20-fold in a buffer containing 100 mM KH2PO4 and 0.05% bovine serum albumin and homogenized at 4°C using the Ultra Turrax. Homogenates will undergo four freeze-thaw cycles before experimentation. Protein content will be measured using the bicinchoninic acid assay and citrate synthase activity will be assessed with reagents provided in the commercial kit (Sigma CS0720), which uses a colorimetric reaction to measure the reaction rate of acetyl coenzyme A and oxaloacetic acid.
In vivo skeletal muscle mitochondrial oxidative capacity
As an additional measure of mitochondrial function, in vivo skeletal muscle mitochondrial oxidative capacity will be measured non-invasively via near-infrared spectroscopy (NIRS) at baseline and follow-up. This NIRS approach measures the recovery kinetics of skeletal muscle oxygen consumption (mVO2) following brief exercise and has demonstrated strong correlations with current in vivo and ex vivo gold standard measurements of mitochondrial function (i.e. magnetic resonance spectroscopy and muscle biopsy) [52,53]. We will implement a NIRS testing protocol similar to Ryan et al. [53]. NIRS data will be obtained using an OxiplexTS (ISS, Champaign, IL), a frequency-domain tissue oximeter. Briefly, the OxiplexTS is equipped with 2 independent data acquisition channels, with 8 infrared diode lasers (four emitting at 691 nm and four at 830 nm) and a detector within each (emitter-detector distances 2.0-4.0 cm). The absolute values of oxygenated hemoglobin (O2Hb) and deoxygenated hemoglobin (HHb) will be calculated in micromoles according to manufacturer instructions. Data will be collected at 4 Hz. Both NIRS probes will be calibrated prior to each test using a phantom with known optical properties, once the device warms-up for at least 20 minutes.
For each NIRS measurement, participants will be supine on a padded table with both legs extended (0º flexion). A skinfold caliper (Lange, Beta Technology, Santa Cruz, CA) will be used to measure subcutaneous adipose tissue thickness at the probe site (~10 cm above the patella). The NIRS probe will be secured to the skin at the vastus lateralis site with double-sided adhesive tape and Velcro straps. Additionally, a blood pressure cuff (Hokanson SC-10D or SC-10L, D.E. Hokanson, Inc., Bellevue, WA) will be placed proximal to the NIRS probe, as high as anatomically possible to prevent unwanted signal noise from cuff inflation. A 15-gallon air compressor (Model D55168, Dewalt, Baltimore, MD) set to 30 psi will power a rapid-inflation system (Hokanson E20, D.E. Hokanson) to control the blood pressure cuff.
Upon securement of the probe and cuff, participants will complete a short-duration (~10-30 seconds), submaximal, repeated knee extension and/or isometric quadricep exercise to increase mVO2. Following exercise, the recovery kinetics of mVO2 will be determined from a series of repeated arterial occlusions (275-300 mmHg) for ~5-7 minutes in duration using the following inflation/deflation timing: 5 seconds inflated/5 seconds deflated for ~90 seconds, then 10 seconds inflated/10 seconds deflated for remainder of test. The beginning and end of each occlusion will be marked for calculations of mVO2 from the deoxygenated hemoglobin/myoglobin signal (i.e. slope during occlusion). The post-exercise mVO2 data will then be fit to a mono-exponential function to calculate the rate constant, which is directly related to the mitochondrial respiratory capacity [52]. Three trials of exercise and occlusion procedures will be performed, and results will be averaged. Baseline and follow-up NIRS data will be analyzed via custom-written routines in MATLAB R2017b (MathWorks, Natick, MA).
Blood sample collection
A venous blood sample will be drawn with a 21-gauge needle with the participant in the fasted state at baseline and follow-up. A total of 21 mL of blood will be drawn by the study nurse and will immediately be sent to a clinical laboratory (LabCorp Inc., Burlington, NC) for a complete metabolic panel, lipid panel, insulin level, and blood chemistries. Prior to glucose injection during the baseline and follow-up IVGTT, we will collect vials of archive plasma, serum, and red blood cells to be stored at -80°C for future analysis. We also will send an additional serum separator tube to LabCorp for measurement of C-reactive protein.
Anthropometry and body composition
Body weight will be measured in the fasted state via a calibrated scale (DigiTol 8510, Mettler Toledo, Columbus, OH) (recorded to the nearest 1/10 of a kg). Dual Energy X-ray Absorptiometry (DEXA) (GE Lunar Prodigy Advance, Fairfield, CT) will be used to measure body composition (fat and fat-free mass) at baseline and follow-up. Waist circumference will be measured via a Gulick tape measure at the natural waist (halfway point from the inferior border of the rib cage and the superior point of the iliac crest). Participants will be instructed to stand straight and upright, with their feet together, and arms to their side. Study staff will mark each landmark and measure the distance to determine the proper measurement site. For each measure, study staff will confirm: 1) the tape is parallel to the floor; 2) the tape touches the entire circumference of the participant; 3) the tape is not compressing any abdominal tissue; 4) the tape is not within abdominal folds; and 5) the measurement is recorded following a normal exhalation by the participant. Duplicate waist circumference measurements will be obtained. If measurements are ±0.5 cm, the reported value will be the average of the two. If measurements differ >0.5 cm, a third measurement will be assessed, and the reported value will be the average of the three measurements. Waist circumference will be evaluated at baseline, mid-intervention, and follow-up.
Non-exercise physical activity levels
Non-exercise physical activity data will be monitored in all randomization groups using a Fitbit Flex activity tracker throughout the intervention period. Each group will be blinded to the number of steps accrued and instructed to not change their non-exercise physical activity levels from baseline. Prior to each exercise session, the Fitbit device will be removed from the participant (to not mix exercise and non-exercise physical activity data) and synced to the Fitbit software to upload their non-exercise physical activity levels. Participants in the CON group will sync their data at home using the Fitbit software and be monitored by study staff to assure compliance. Automated REDCap surveys will be emailed 3 times per week to all participants to inquire about Fitbit wear. Participants will be instructed to fill-out these surveys to validate consistent device wear as this will allow staff to input non-exercise physical activity data on a weekly basis. This process helps to ensure consistent daily wearing of the device and to determine if the participant did not wear the Fitbit for extended periods of time.
Study staff will use a database program (Fitabase, Small Steps Labs, San Diego, CA) to centralize all non-exercise physical activity data (i.e. total daily steps, minutes of light, moderate and vigorous physical activity, miles traveled, and estimated kilocalorie energy expenditure). All non-exercise physical activity data synced to Fitabase will be stored in a custom-made REDCap database. Study staff will compile and inp daily steps, intensity, miles, and energy expenditure from each participant on a weekly basis.
Dietary composition
Dietary intake will be tracked at baseline and follow-up via the Block food frequency questionnaire (FFQ) [54]. The FFQ consists of 105 categorized items and assesses both frequency of consumption and portion size selections, in which participants will recall their typical eating habits within the previous 3 months at baseline and follow-up timepoints. The questionnaire estimates daily intake values of kilocalories, select macronutrients and micronutrients, and also calculates servings by food group. The research coordinator will instruct participants at screening visit 1 to maintain current dietary habits and not to begin intentional dieting for the entirety of the study. Additionally, study staff will remind all participants on a weekly basis to not change their eating habits during the intervention. The FFQ serves as a semi-quantitative measure to ensure dietary habits are not changed throughout the intervention.
Psychological parameters
The short form health survey (SF-36) [55] will be utilized to measure quality of life at both baseline and follow-up. Exercise enjoyment will be assessed via the Physical Activity Enjoyment (PACE) Scale [56] and the Feeling Scale [57]. As secondary outcomes, the impact of exercise intensity on these affective responses (i.e. feelings of overall pleasure/displeasure and enjoyment) play an important role in physical activity participation and adherence [58–60]. The PACE Scale is comprised of 8 items rated on a 7-point semantic differential scale with “4” representing a neutral position. The PACE Scale will be collected every 4 weeks during the exercise intervention. Affective responses to exercise will be assessed by having participants complete the Feeling Scale. The Feeling Scale is a single-item, 11-point scale which assesses how individuals feel at a specific moment in time, ranging from -5 (very bad) to +5 (very good) with 0 representing neutral feelings. Study staff will collect Feeling Scale data every 5 minutes during the first exercise session of every week of the exercise intervention.
Randomization
Participants will be randomized to either the non-exercise control (CON), moderate-intensity (MOD-INT), or high-intensity (HIGH-INT) group upon completion of all baseline assessments and approval by the study physician. The study biostatistician will generate a randomization list to allocate participants in a 1:1:1 ratio to study groups. The randomization process will be performed by an individual separate from the research team, who has no interaction with the study participants or access to HI-PACE study data. All other research staff will not have access to the randomization list (including the principal investigator). Upon randomization, the intervention period will begin the following week. A study flow chart is shown in Fig. 1.
Aerobic exercise training
All exercise sessions will be supervised by study staff and performed on a treadmill (Precor TRM 885, Precor Inc., Woodinville, WA) to sustain control of energy expenditure from exercise. Participants in the MOD-INT group will exercise at a target heart rate associated with 45-55% VO2 max, and participants in the HIGH-INT group will exercise at a target heart rate associated with 70-80% VO2 max. The heart rate range for each participant will be determined based on the maximal exercise test (baseline and mid-intervention). The full exercise dose for both groups will be 600 MET-minutes per week, which is consistent with current public health guidelines [5]. Since participants will be sedentary at baseline, we will increase exercise dose incrementally throughout the study to avoid potential adverse events during exercise. Initially, the exercise dose will be 300 MET-minutes during week 1 and will increase by 50 MET-minutes per week until the maximum exercise volume of 600 MET-minutes is reached at week 9. The exercise dose will remain at 600 MET-minutes until conclusion of the intervention (see Fig. 3). We will calculate the number of MET-minutes exercised based on treadmill speed/grade and the participants’ weight, using the standard American College of Sports Medicine (ACSM) walking equation [61]. Custom-made Excel spreadsheets will be utilized to determine exercise time for each session, based on 1) the required weekly MET-minutes, 2) the participants’ weight, 3) exercise speed/grade, and 4) the amount of expected sessions per week (3-4 sessions per week).
At the first exercise session of each week, study staff will weigh participants (without shoes) on a calibrated scale and remind participants to not alter their diet or engage in an exercise program outside of the study. Additionally, we will ask participants about any changes to their prescribed medications on a weekly basis.
Prior to starting exercise, participants will rest for 5 minutes in the seated position, after which study staff will measure systolic/diastolic blood pressures using a mercury sphygmomanometer and record resting heart rate via a Zephyr Bioharness 3 monitor (Medtronic, Annapolis, MD). Each participant will be instructed to complete a 5-minute warm-up on the treadmill at a low speed (~2.0 miles per hour) at 0% grade. Following completion of the warm-up, participants will begin their prescribed exercise by adjusting speed and/or grade of the treadmill. During the supervised exercise, heart rate will be monitored continuously using the Bioharness monitor to confirm exercise intensity and participants will be required to remain within their target heart range (MOD-INT: heart rate associated with 45-55% VO2 max; HIGH-INT: heart rate associated with 70-80% VO2 max). Heart rate will be recorded every 5 minutes, along with participants’ subjective rating of perceived exertion (RPE) via the Borg scale [62]. The Feeling Scale will also be recorded every 5 minutes on the first exercise session of each week. Study staff will keep mobile laptop carts nearby the exercising participant(s) and will use custom-made Excel spreadsheets to: 1) quantify the number of MET-minutes accumulated during exercise and number of MET-minutes remaining in the session, 2) mean heart rate and RPE of the session, and 3) calculate amount of time remaining in the current session. The spreadsheet calculates these variables in real-time and can compensate for potential adjustments during the exercise session, such as increasing or decreasing treadmill speed and/or grade.
Once completed with the exercise session, participants will perform a 5-minute cool-down at a similar intensity as the warm-up. Subsequently, participants will rest for 5 minutes in the seated position for recording of post-exercise heart rate and systolic/diastolic blood pressures. Lastly, study staff will input all exercise session data (i.e. total exercise duration, MET-minutes, caloric expenditure, miles traveled, average speed, grade, heart rate, and RPE, and percentage of time participant exercised within target heart rate range) into the database.
For each exercise session the mean heart rate will be calculated using Omnisense Analysis version 5.0 software (Medtronic, Annapolis, MD). The Bioharness monitors continuously record heart rate data on a second-by-second basis. Thus, to collect mean heart rate for each session, study staff will analyze heart rate data only during exercise by creating a time-specific sub-session annotation within the Omnisense software to exclude non-exercise heart rate data from calculation. This process will more accurately calculate exercise intensities during training sessions since heart rate will be measured continuously as opposed to intervals (e.g. every 5 minutes).
Exercise economy
To address potential variability in exercise economy at a given workload, study staff will directly measure energy expenditure (EE) rate via indirect calorimetry (TrueOne 2400) at the participant’s prescribed exercise speed and grade on a treadmill [42,63,64]. This exercise economy test will be performed on week-1, -3, -5, and then monthly until the conclusion of the intervention. The rate of EE determined through indirect calorimetry will be divided by estimated EE determined from the ACSM walking equation to develop a correction factor (i.e. actual EE rate/predicted EE rate). This correction factor will be used to: 1) adjust the EE calculated from the ACSM equation to more accurately implement the exercise prescription which corresponds to 600 MET-minutes per week, 2) adjust participants’ exercise session time according to potential changes in metabolic and/or biomechanical efficiency from exercise training, and (3) verify required MET-minutes exercised by increasing or decreasing exercise session time accordingly.
Training data management
Exercise volume adherence will be defined as MET-minutes exercised divided by required MET-minutes. Exercise intensity adherence will be quantified as time within the required target heart rate range divided by total exercise time. Exercise compliance will be defined as the amount of sessions attended divided by the amount of sessions required. The research team will actively monitor exercise volume/intensity adherence, compliance, and other indicators of intervention fidelity (e.g. target heart rate compliance, wear rate of accelerometer, participant morale, progression rate of speed/grade) on a weekly basis in study meetings. In all randomization groups, Fitbit wear compliance will be monitored throughout the 6-month intervention. Weekly reports will be compiled from the study databases to monitor and review the compliance and adherence rates of all participants.
Statistical considerations
The results of the current pilot study will be used to advise the design (effect size/statistical power) of a larger prospective intervention. The response variable for the primary outcome is change in VO2 max. The three treatment groups, CON, MOD-INT, and HIGH-INT will be compared in terms of baseline VO2 max using side-by-side boxplots and the corresponding numeric summaries along with mean and standard deviation. This will be repeated for post treatment values of VO2 max as well as for VO2 max differences of post-treatment and baseline. If there are distributional concerns, log transformation of VO2 max will be considered. Unless there are extreme outliers or severe heteroscedasticity, one-way ANOVA will be used for inference regarding the primary outcome. The two sample t-test (without assuming equal variances) along with the associated confidence intervals will be used for differences in group means; confidence intervals for differences in group means obtained from the one-way ANOVA will also be reported.
The above steps used for VO2 max (the primary outcome) will be repeated for each of the numeric variables used for secondary outcomes. These variables are change in insulin sensitivity, mitochondrial protein content, citrate synthase activity, skeletal muscle mitochondrial oxidative capacity, arterial stiffness parameters, body fat %, and C-reactive protein.
The ordinal variables, exercise enjoyment and quality of life, will be dichotomized into “no improvement” and “improvement.” Fisher’s exact test will be used to obtain an overall p-value and restriction to two of the treatment groups will provide estimated odds ratios and the associated confidence intervals.
The three groups will be compared using the following demographic and other variables that may be related to exercise, change in VO2 max, or one of the other response variables: age, sex, body weight, BMI, waist circumference, body fat %, fat mass, fat-free mass, cholesterol, triglycerides, and blood pressures. If differences among the treatment groups in terms of one or more of these variables are deemed important, adjustments to the above comparisons will be made using higher order ANOVA, ANCOVA, and/or linear regression.
Power ranges from .85 for 15 participants per group to .96 for 20 participants per group using -.112, .124, and .200 L/min as the means for the three groups based on previous data [66], a common standard deviation of .250 L/min, and significance level a =.05. We expect attrition to be approximately 10-15% so that enrollment of 60 participants (20 per group) will be sufficient for the primary outcome.
All statistical analyses will be performed using statistical software in R version 3.5.1. [65]. The resultant mean change and standard deviation of the change of outcome measures (if indicative of enhance cardiometabolic improvements in the HIGH-INT compared to the MOD-INT and CON groups) will be used for power calculations to determine the necessary sample size of a larger study.