This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on http://www.researchprotocols.org, as well as this copyright and license information must be included.
Digital wearable devices provide a “real-world” assessment of physical activity and quantify intervention-related changes in clinical trials. However, the value of digital wearable device-recorded physical activity as a clinical trial outcome is unknown.
Because late sodium channel inhibition (ranolazine) improves stress laboratory exercise duration among angina patients, we proposed that this benefit could be quantified and translated during daily life by measuring digital wearable device-determined step count in a clinical trial.
We conducted a substudy in a randomized, double-blinded, placebo-controlled, crossover trial of participants with angina and coronary microvascular dysfunction (CMD) with no obstructive coronary artery disease to evaluate the value of digital wearable device monitoring. Ranolazine or placebo were administered (500-1000 mg twice a day) for 2 weeks with a subsequent 2-week washout followed by crossover to ranolazine or placebo (500-1000 mg twice a day) for an additional 2 weeks. The outcome of interest was within-subject difference in Fitbit Flex daily step count during week 2 of ranolazine versus placebo during each treatment period. Secondary outcomes included within-subject differences in angina, quality of life, myocardial perfusion reserve, and diastolic function.
A total of 43 participants were enrolled in the substudy and 30 successfully completed the substudy for analysis. Overall, late sodium channel inhibition reduced within-subject daily step count versus placebo (mean 5757 [SD 3076] vs mean 6593 [SD 339],
We report one of the first studies to use digital wearable device-determined step count as an outcome variable in a placebo-controlled crossover trial of late sodium channel inhibition in participants with CMD. Our substudy demonstrates that late sodium channel inhibition was associated with a decreased step count overall, although the subgroup with angina improvement had a step count increase. Our findings suggest digital wearable device technology may provide new insights in clinical trial research.
Clinicaltrials.gov NCT01342029; https://clinicaltrials.gov/ct2/show/NCT01342029 (Archived by WebCite at http://www.webcitation.org/6uyd6B2PO)
Mobile wireless devices and connected wearable biosensors have the potential to provide new insights into chronic medical conditions and help clinicians develop personalized treatment strategies. One potential application of digital wearable device technology is the real-time evaluation of change in daily activity as a clinical outcome. Patient activity level is traditionally assessed using surrogate measures such as exercise testing or patient questionnaires. However, these tools are limited to point-in-time observation and questionnaires provide subjective information that is retrospective. Furthermore, although patient questionnaires correlate with exercise stress testing [
One area for such potential is the management of angina pectoris, a chronic medical condition with symptoms that decrease exercise capacity, decrease quality of life (QoL), and lead patients to limit physical activities in daily life [
Late sodium channel inhibition (ranolazine) is an antiischemic pharmacologic therapy indicated for the treatment of angina due to CAD that has been demonstrated to improve physical activity, exercise duration, and QoL [
Angina clinical trials have used exercise laboratory testing, 6-minute walk, ECG monitoring, cardiac imaging, QoL questionnaires, and pedometers to determine if treatment of ischemia with antianginal therapy relieves symptoms and results in improved physical activity, functionality, and QoL. The totality of evidence suggests that treatment of angina can relieve symptoms and that no one therapy is more effective than the other. However, beyond larger modern studies that have used questionnaires or historical small studies that used pedometers, our understanding of how treatment with antianginal therapy affects daily activity is limited. As such, data generated by digital wearable devices has the potential to provide new insights into the efficacy of antianginal therapies as it relates to physical activity in daily life in patients with angina, particularly in the CMD population.
To further explore use of digital technology in a therapeutic cardiovascular trial, we conducted a substudy within a randomized, double-blinded, placebo-controlled, crossover trial of patients with angina and CMD and no obstructive CAD [
Participants were recruited from the two-site parent trial [
The parent trial was a double-blind, placebo-controlled, crossover trial with short-term (2 week treatment periods; week 1: ranolazine or placebo 500 mg twice a day, week 2: ranolazine or placebo 1000 mg twice a day) ranolazine-placebo exposure (order randomly assigned to a sequence of ranolazine or placebo first followed by a 2-week washout and subsequent crossover to placebo or ranolazine or vice versa) [
The duration of daily activity monitoring was designed for 1-week duration to account for daily activity that occurs in real life. In addition, it was previously demonstrated that ranolazine increases exercise performance in patients with angina-limited exercise after 1 week of treatment [
The wearable accelerometer was the Fitbit Flex, which continuously measures daily physical activity that is reported as step count [
Substudy design flow diagram, participant screening, enrollment, and randomization flow diagram. Treatment period 1 and 2 with daily activity monitoring during week 2 of each treatment period: randomized to sequence of ranolazine first followed by crossover to placebo or vice versa. BID: twice a day; DASI: Duke Activity Status Index; ECG: electrocardiogram; MRI: magnetic resonance imaging; SAQ: Seattle Angina Questionnaire; SF-36: 36-item Short Form Survey.
Before randomization, each participant was provided with a Fitbit Flex, a dongle for wireless synchronization, a USB cord, and a wall charger. Participants were instructed regarding the device placement and the daily physical activity monitoring protocol, which included scheduled monitoring dates, device charging, wrist placement, and device synchronization. Participants were encouraged not to modify their daily activity or achieve any specific step count goal or daily activity level. The device was worn by the participant on the nondominant wrist continuously each day during week 2 of both the active treatment and placebo periods. All days with step count data were considered a valid day of wear as long as the participant did not report any violations of the protocol in that they continuously wore the Fitbit per protocol during the entirety of the monitoring period. Research coordinators verbally confirmed that participants complied with the protocol after completion of each monitoring period. Per protocol, the device could only be removed temporarily for bathing or swimming, but otherwise removal was not allowed unless a participant was instructed to remove the device by the research coordinator (ie, adverse device reaction, device failure, or device charging while sleeping if indicated).
Digital wearable device data were uploaded regularly by participants via computer synchronization using the “Sync Now” function in the Fitbit Connect software, which was installed on each participant’s personal computer or device and aggregated by research coordinators for analysis. Each device was accompanied by a unique account that was password protected and accessible only to research coordinators; therefore, participants were not able to view their own step count data. They were able to view the Fitbit Flex indicator lights, although they were not instructed as to what the specific meaning of the light patterns indicated. During the clinical trial, participants received reminder telephone calls regarding device charging, placement, and synchronization. In the case that a research coordinator determined that a device needed to be charged, the participant was instructed to remove the device during sleep and place the device on the nondominant wrist immediately on waking. Research coordinators monitored device synchronization and intervened if the device failed to upload data during monitoring periods. To assure participant safety and equipment integrity, participants were also instructed about allergic reactions, exposure to liquids, and device cleaning. In total, 15 Fitbit Flex’s purchased between 2013 and 2015 were used throughout the duration of the study.
Patient questionnaires were administered as described [
The cardiac MRI protocol was performed as previously described and was conducted under identical conditions and timing, dosing, and settings, approximately 4 hours after the morning study drug dose [
Clinically indicated invasive coronary reactivity testing [
This substudy was planned to enroll 30 participants to achieve 80% power to detect a mean difference in daily step count between a patient’s ranolazine and placebo treatments of 1000 using a paired
The analytic approach was a within-subject comparison (paired) of the difference between a participant’s mean daily step count during 1 week on ranolazine and their mean daily step count during 1 week on placebo. The main endpoint of the parent trial was the difference between change in SAQ for ranolazine versus placebo. There were two treatment periods, ranolazine and placebo, giving two measurements per participant for mean step count and cardiac MRI variables. There were a total of four measurements per participant, including treatment baselines and posttreatment values, for SAQ and QoL giving two treatment changes per participants. The primary approach was a standard paired
The study was an investigator-initiated, intramurally funded substudy embedded within a parent trial ancillary to the National Heart, Lung, and Blood Institute (NHLBI)-sponsored Women’s Ischemia Syndrome Evaluation (WISE) study. The parent trial was funded, in part, by Gilead Sciences and was overseen by the WISE Data Safety Monitoring Committee. Statistical analysis was performed by investigators independent of NHLBI and Gilead, masked to treatment assignment. The decision to submit for publication was made by investigators who had access to all data after the last participant completed the study.
Between February 12, 2014 and June 1, 2016, 43 participants entered the Fitbit pilot study, 31 of whom were also enrolled in the parent trial (29 included in the parent trial analyses), and 12 who were enrolled in the substudy after the parent trial closed enrollment. Among the 43 participants who entered the substudy, 13 participants (30%) could not be analyzed for within-subject change in step count because of dropout before protocol initiation (n=3), incomplete data (n=6), Fitbit Flex not returned (n=1), protocol nonadherence (n=1), and washout dropout (n=2), leaving 30 participants (70%) who adhered to the protocol sufficient to provide adequate data for analysis as an outcome variable (
There was mean 6.5 (SD 1.0) days of digital wearable device monitoring during treatment with ranolazine (independent of randomization sequence) with 7.6% (16/210) days missing data and mean 6.6 (SD 0.9) days of activity monitoring during treatment with placebo (independent of randomization sequence) with 6.2% (13/210) days missing data. Pertinent baseline demographics and clinical variables are summarized in
Compliance to the intervention, measured by returned pill counts, was available for 56 of 60 sessions and was 100% for those 56 sessions overall. Ranolazine and placebo interventions were well tolerated with three (ranolazine) and three (placebo) participants reduced to 500 mg twice daily dosing for adverse effects, as per protocol. No serious adverse events during the ranolazine period occurred. Nonserious adverse events during the ranolazine period occurred in two participants—nausea and lightheadedness (n=1) and excessive sweating (n=1)—and in one participant during placebo—throat swelling (n=1)—as previously reported [
Substudy participant screening, enrollment, randomization, and completion flow diagram. Participants were randomized to sequence of ranolazine first followed by crossover to placebo (treatment period 1) or vice versa (treatment period 2). For both treatment periods 1 and 2, daily activity monitoring occurred during week 2.
Baseline demographics and clinical variables (N=30).
Variables | Analyzed participants | ||
Age in years, mean (SD) | 54.03 (10.59) | ||
Female, n (%) | 29 (97) | ||
Postmenopausal (n=29), n (%) | 20 (69) | ||
27.71 (6.74) | |||
>30 kg/m2, n (%) | 10 (35) | ||
Race (non-Caucasian), n (%) | 25 (83) | ||
Current | 1 (3) | ||
Former | 8 (27) | ||
Never | 21 (70) | ||
Hypertension, n (%) | 16 (53) | ||
Hyperlipidemia, n (%) | 13 (43) | ||
Family history of premature coronary artery disease, n (%) | 19 (63) | ||
Qualifying coronary flow reserve (n=13) | 2.38 (0.63) | ||
Qualifying coronary blood flow (n=6) | 90.82 (71.18) | ||
Qualifying acetylcholine response (n=10) | –0.08 (12.25) | ||
Qualifying nitroglycerin response (n=11) | 19.31 (19.55) | ||
Qualifying cold pressor testing response (n=8) | 7.94 (12.95) | ||
Left ventricular end-diastolic pressure (n=9) | 13.33 (3.57) | ||
Baseline systolic blood pressure (n=28) | 123.61 (21.03) | ||
Baseline diastolic blood pressure (n=28) | 68.79 (12.38) | ||
Baseline heart rate (n=29) | 70.59 (11.38) | ||
Baseline rate pressure product (n=28) | 8804 (2284) | ||
Global myocardial perfusion reserve index (n=11) | 1.69 (0.20) | ||
Subendocardial MPRI (n=11) | 1.52 (0.14) | ||
Subepicardial MPRI (n=11) | 1.75 (0.27) | ||
Typical angina | 12 (40) | ||
Shortness of breath | 24 (80) | ||
Palpitations | 15 (50) | ||
Nausea | 8 (27) | ||
Beta blockers | 12 (40) | ||
Calcium channel blockers | 7 (23) | ||
Angiotensin-converting enzyme inhibitors | 2 (7) | ||
Angiotensin receptor blockers | 6 (20) | ||
Nitrates | 10 (33) | ||
Prior ranolazine | 4 (13) | ||
Statins | 13 (43) | ||
Aspirin | 15 (50) | ||
Diuretic | 3 (10) | ||
Hormone replacement therapy (n=19) | 4 (21) | ||
Vitamin D | 10 (33) |
Step counts on individual days ranged from 7 to 19,879 during ranolazine and 25 to 18,110 during placebo. The individual mean across days within the treatment period ranged from 68 to 14,465 during ranolazine and 228 to 13,311 during placebo. During the ranolazine period, participants had significantly lower mean daily step counts compared to placebo (
Mean daily step count, Seattle Angina Questionnaire (SAQ), hemodynamic, myocardial perfusion reserve, and diastolic filling treatment effect.
Variablesa | Ranolazine | Placebo | Treatment change | |||||||
Mean (SD) | N | Mean (SD) | N | Mean (95% CI) | N | |||||
Mean daily step count | 5756.49 (3075.51) | 30 | 6593.29 (3393.44) | 30 | –836.8 (–1464.74, –208.86) | 30 | .01 | |||
Physical limitation | 69.35 (26.10) | 28 | 65.18 (21.44) | 28 | 2.01 (–4.04, 8.06) | 27 | .50 | |||
Angina stability | 55.17 (27.04) | 29 | 54.17 (26.33) | 30 | –6.03 (–22.63, 10.56) | 29 | .46 | |||
Angina frequency | 55.86 (27.32) | 29 | 56.33 (23.56) | 30 | –4.14 (–13.76, 5.48) | 29 | .39 | |||
Treatment satisfaction | 66.31 (23.61) | 29 | 66.81 (23.72) | 30 | –3.09 (–12.62, 6.44) | 29 | .51 | |||
Quality of life | 50.29 (21.76) | 29 | 45.56 (19.54) | 30 | 1.72 (–4.80, 8.25) | 29 | .59 | |||
SAQ-7 | 58.19 (24.43) | 29 | 55.86 (19.85) | 30 | –0.64 (–6.74, 5.46) | 29 | .83 | |||
Heart rate (bpm) | 97.6 (14.1) | 30 | 104.4 (16.34) | 30 | –6.8 (–9.8, –3.7) | 30 | <.001 | |||
SBP (mmHg) | 126.37 (24.55) | 30 | 128.23 (20.42) | 30 | –1.87 (–8.68, 4.95) | 30 | .58 | |||
DBP (mmHg) | 58.83 (15.00) | 30 | 60.93 (15.47) | 30 | –2.1 (–7.5, 3.3) | 30 | .43 | |||
Stress RPP | 12275.97 (2745.26) | 30 | 13292.47 (2543.46) | 30 | –1016.5 (–1860.2, –172.8) | 30 | .02 | |||
Global MPRI | 2.09 (0.75) | 30 | 2.16 (0.71) | 30 | –0.08 (–0.43, 0.28) | 30 | .67 | |||
MPRI Subendocardial midventricular | 1.9 (0.9) | 30 | 1.9 (0.7) | 30 | –0.03 (–0.43, 0.36) | 30 | .86 | |||
PFR (mL/s) | 322.8 (128.9) | 30 | 321.35 (116.98) | 30 | 1.5 (–27.6, 30.5) | 30 | .92 | |||
tPFR (ms) | 171.75 (43.31) | 30 | 158.41 (39.85) | 30 | 13.34 (0.91, 25.76) | 30 | .04 |
aDBP: diastolic blood pressure; MPRI: myocardial perfusion reserve index; PFR: peak filling rate; RPP: rate pressure product; SAQ: Seattle Angina Questionnaire; SBP: systolic blood pressure; tPFR: time to peak filling rate.
bPharmacologic Stress=adenosine or regadenoson infusion.
Correlation between change in angina and change in mean daily step count (n=29,
There were no significant differences in baseline variables among patients who experienced a within-subject mean daily step count increase of 500 steps or greater (n=6), a decrease of 500 steps or greater (n=17), or a change of less than 500 steps (n=7). The change in 500 steps was informally based on the distribution of the step count data and a visual median
There were no significant correlations for DASI, SF-36, diary-reported angina frequency and nitroglycerin usage, MPRI, and diastolic function with mean daily step count.
Subgroup analysis among the participants with typical angina (n=12) showed within-subject mean daily step count was lower during ranolazine compared to placebo (mean 5522 [SD 2891] vs mean 7150 [SD 3490],
Due to the relatively small sample size of the substudy and potential confounding variables, linear regression models used individual’s change in step count as the outcome between ranolazine and placebo and adjusted for various factors that may be associated with it, including age, body mass index, and typical angina versus nontypical angina. The model adjusting for age, body mass index, and typical angina versus nontypical angina did not identify any of these three factors to be significantly associated with change in step count.
When participants were categorized as “sedentary” daily physical activity [
We found in this substudy that late sodium channel inhibition was associated with a decreased step count overall during daily life in these participants, although the subgroup with angina improvement had a step count increase. This suggests that digital wearable device technology may be useful as a clinical variable and outcome in clinical trials research.
These substudy results are consistent with primary findings of the parent trial (eg, late sodium current blockade with ranolazine did not improve overall group angina in participants with CMD). Surprisingly, we observed a significantly lower mean daily step count during late sodium current blockade compared to placebo in our substudy. This finding is new and differs from the previous studies of ranolazine in CMD [
A variety of explanations may have contributed to the finding of reduced step count. The result could be a chance finding; however, it was present in the overall and subgroup analyses. It is possible that in CMD patients, those who are more active may have become less active due to unreported ranolazine side effects (ie, dizziness or gastrointestinal intolerance). In fact, two of the three reported side effects occurred during treatment with ranolazine. It is also possible that the digital wearable device was unable to detect the various types of daily activities participants may have performed, such as swimming or bicycling versus walking, or changes in daily activity intensity. Additionally, our results failed to detect a relationship between change in SAQ angina or QoL and change in step count, although our substudy was underpowered to evaluate this.
The unique attributes of our population and description of their daily activity should be noted. Women comprised 97% of our population, precluding conclusions regarding response in men, who typically are more represented in pivotal chronic angina trials with ranolazine (75%) and suffered from obstructive CAD [
The strengths of our substudy include the crossover design, which allowed participants to serve as his or her own control thus negating the effects of daily activity intersubject variability and reducing confounding effects of other antianginal therapies. Participant compliance with study medication is also a notable strength of our substudy.
The limitations of our substudy include a small sample size and a mean “low active” [
Digital wearable devices have the potential for an expanded role in research as a clinical trial outcome and should be explored in a variety of health conditions.
We report one of the first studies to use digital wearable device-determined step count as an outcome variable in a placebo-controlled crossover trial of late sodium channel inhibition in patients with CMD. Our substudy demonstrates that late sodium channel inhibition was associated with a decreased step count overall, although the subgroup with angina improvement had a step count increase. Our findings suggest digital wearable device technology may provide new insights in clinical trial research.
coronary artery disease
coronary flow reserve
coronary microvascular dysfunction
Duke Activity Status Index
electrocardiogram
myocardial perfusion reserve index
magnetic resonance imaging
National Heart, Lung, and Blood Institute
peak filling rate
quality of life
Seattle Angina Questionnaire
time to peak filling rate
Women’s Ischemia Syndrome Evaluation
The parent trial was supported by an unrestricted research grant from Gilead and by contracts from the National Heart, Lung, and Blood Institutes (N01-HV-68161, N01-HV-68162, N01-HV-68163, N01-HV-68164, K23HL105787, T32-HL-116273); a GCRC grant (MO1-RR00425) from the National Center for Research Resources; grant UL1RR033176 the NIH/National Center for Advancing Translational Sciences (NCATS); UCLA CTSI grant UL1TR000124; and UF CTSI grant UL1TR001427, grant R01 HL089765; and grants from the Gustavus and Louis Pfeiffer Research Foundation, Denville, NJ; the Women’s Guild of Cedars-Sinai Medical Center, Los Angeles, CA; the Edythe L Broad Women’s Heart Research Fellowship, Cedars-Sinai Medical Center, Los Angeles, CA; the Constance Austin Women’s Heart Research Fellowship; the Barbra Streisand Women’s Cardiovascular Research and Education Program, Cedars-Sinai Medical Center, Los Angeles, CA; the Erika Glazer Women’s Heart Health Project, Cedars-Sinai Medical Center, Los Angeles, CA. This pilot study was intramurally funded by the Cedars-Sinai Clinical and Translational Science Institute Clinical Scholars Program, Cedars-Sinai Medical Center, Los Angeles, CA. CJP was supported by NHLBI grants HL33610, HL56921, UM1 HL087366, and the Gatorade Trust through funds distributed by the University of Florida, Department of Medicine; the NIH NCATS-University of Florida Clinical and Translational Science UL1TR001427; and the PCORI-OneFlorida Clinical Research Consortium CDRN-1501-26692.
CLS, PKM, LEJT, and CNBM received an investigator-initiated research grant from Gilead, the manufacturer of Ranexa (ranolazine), to support the clinical trial.