Daily Heart Rate per Step: A Wearables Metric Associated With Cardiovascular Disease in a Cross‐Sectional Study of the All of Us Research Program
Abstract
Background
Simple biometrics such as peak heart rate and exercise duration remain core predictors of cardiovascular disease (CVD). Commercial wearable devices track physical and cardiac electrical activity. Detailed, longitudinal data collection from wearables presents a valuable opportunity to identify new factors associated with CVD.
Methods and Results
This cross‐sectional study analyzed 6947 participants in the Fitbit Bring‐Your‐Own‐Device Project, a subset of the All of Us Research Program. The primary exposure daily heart rate per step (DHRPS) was defined as the average daily heart rate divided by steps per day. Our analysis correlated DHRPS with established CVD factors (type 2 diabetes, hypertension, stroke, heart failure, coronary atherosclerosis, myocardial infarction) as primary outcomes. We also performed a DHRPS‐based phenome‐wide association study on the spectrum of human disease traits for all 1789 disease codes across 17 disease categories. Secondary outcomes included maximum metabolic equivalents achieved on cardiovascular treadmill exercise stress testing. We examined 5.8 million person‐days and 51 billion total steps of individual‐level Fitbit data paired with electronic health record data. Elevated DHRPS was associated with type 2 diabetes (odds ratio [OR], 2.03 [95% CI, 1.70–2.42]), hypertension (OR, 1.63 [95% CI, 1.32–2.02]), heart failure (OR, 1.77 [95% CI, 1.00–3.14]), and coronary atherosclerosis (OR, 1.44 [95% CI, 1.14–1.82]), even after adjusting for daily heart rate (DHR) and step count. DHRPS also had stronger correlations with maximum metabolic equivalents achieved on exercise stress testing compared with steps per day (∆ρ=0.04, P<0.001) and heart rate (∆ρ=0.31, P<0.001). Lastly, DHRPS‐based phenome‐wide association study demonstrated stronger associations with CVD factors (P<1×10−55) compared with daily heart rate or step count.
Conclusions
In the All of Us Research Program Fitbit Bring‐Your‐Own‐Device Project, DHRPS was an easy‐to‐calculate wearables metric and was more strongly associated with cardiovascular fitness and CVD outcomes than DHR and step count.
Nonstandard Abbreviations and Acronyms
- DHRPS
- daily heart rate per step
- PheWAS
- phenome‐wide association study
Clinical Perspective
What Is New?
•
We introduce daily heart rate per step as a new wearable‐derived biomarker for cardiovascular disease that combines heart rate and step count.
•
Daily heart rate per step showed stronger correlations with cardiovascular fitness, disease outcomes, and patient self‐reported fitness than either metric alone.
What Are the Clinical Implications?
•
Daily heart rate per step may identify individuals at higher risk of cardiovascular morbidity and mortality, facilitating early detection using data from widely available wearable devices.
Wearable devices, such as the Fitbit and Apple Watch, are widely used to monitor and track physical activity and cardiac metrics.1, 2 Consumer devices have become increasingly popular for their ability to provide real‐time feedback on health‐related behaviors.3 These features have raised public awareness about the importance of physical activity; however, the mechanisms underlying the relationship between daily steps and cardiovascular health remain unclear. Specifically, daily step count reflects motor activity driven by conscious effort, serving as a proxy for exercise duration and an indirect measure of exercise capacity. Understanding these relationships are essential for improving how wearable data can be used to assess and manage cardiovascular health.
Cardiac stress tests directly evaluate the heart's response to external stress in a controlled clinical environment.4 Low cardiorespiratory fitness on a cardiac stress test predicts increased cardiovascular disease (CVD) incidence and mortality.5 Previous studies have advocated for more widespread use of exercise capacity in clinical medicine given the strong correlations between cardiovascular and all‐cause mortality.6, 7, 8, 9 However, adherence to prescriptions for outpatient cardiac stress tests in the United States has been less than 60%.10, 11 In contrast, consumers spend money to manage their well‐being with Fitbit devices, which track physical activity, heart rate (HR), sleep patterns, and other health metrics. Wearable devices have been voluntarily adopted by the public, and the availability of granular data from wearable devices on the scale of minutes and hours presents an opportunity to identify new biomarkers associated with an increase in CVD, potentially improving on the rough estimate of step count, and with better compliance than intermittent cardiac stress tests.12
The purpose of this study was to use a cross‐sectional data set to examine the association between daily heart rate per step (DHRPS) and CVD, then to compare these to previously‐studied metrics like average daily heart rate (DHR) and step count.13, 14, 15 In total, we analyzed individual‐level Fitbit device data paired with electronic health records (EHR) through the All of Us Research Program Fitbit Bring‐Your‐Own‐Device Project. Based on previous literature, we hypothesized that an elevated DHRPS could be associated with increased CVD prevalence.16, 17 This new wearable biomarker could potentially provide better individual stratification for CVD as an effective, low‐cost, high‐compliance screening tool.
METHODS
The All of Us Research Program is a comprehensive health data repository supported by the National Institutes of Health. Details of the All of Us recruitment methods have been described separately.18 Briefly, health provider organizations and community partners recruited participants, with more than 75% of participants from groups historically underrepresented in biomedical research. All of Us gathered self‐reported surveys, physical measurements, and encoded EHRs, with ethics approval from the All of Us Institutional Revie Board to ensure ethical compliance in recruitment, data collection, and use. All terms referring to race were defined by All of Us. The categories of American Indian or Alaskan Native, More than one category, Other, and Prefer not to say were collapsed into a single category, “Not otherwise listed”, to reflect the All of Us policy to not publish cell values <20.
The data are available in the All of Us Research Program public repository, accessible after completing the required training (https://workbench.researchallofus.org/). Informed consent was obtained from each participant by the All of Us Research Program (https://allofus.nih.gov/article/all‐us‐consent‐process).
Study Design and Eligibility Criteria
Data from the All of Us version 7 (C2022Q4R9 CDR) Fitbit Bring‐Your‐Own‐Device Project were used to establish a cross‐sectional, observational study population with recruitment dates from summer 2017 to July 1, 2022. The study uses wearable data to derive metrics that serve as exposures, linking them to modifiable cardiovascular risk factors (eg, hypertension, type 2 diabetes) and specific major adverse cardiovascular events (eg, myocardial infarction, stroke). The primary cohort included participants with a minimum of 10 days of Fitbit device data, excluding individuals missing data on age, sex at birth, body mass index (BMI), enrollment survey responses, or EHR linkage. Additionally, a subgroup analysis was conducted on participants with Fitbit data and cardiovascular treadmill exercise testing to compare wearable‐derived metrics with established clinical measures of cardiovascular fitness.
Wearable Data Sampling and Exposure Variables
Wearable data included minute‐by‐minute physical activity and HR measurements. The data were obtained from participants with a minimum of 10 days of recording, with most extending beyond this period; however, the wearable data were retrospective and cross‐sectional in nature. For each participant, average DHR was calculated by averaging the HR across all minutes in a day (midnight to midnight). Steps per day were also counted. A derived metric, DHRPS, was calculated by dividing DHR by steps per day for each available day, then averaging across all days with data. Three wearable‐derived metrics were analyzed as exposures: DHR, steps per day, and DHRPS.
Outcomes of Interest
Disease phecodes were mapped from International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM) and Tenth Revision, Clinical Modification (ICD‐10‐CM) billing codes.19 We tested associations of our exposure variable (DHRPS) against 2 types of outcome variables. The first was the prevalence of modifiable cardiovascular risk factors: hypertension (401.1), type 2 diabetes (250.2), and coronary atherosclerosis (411.4). The second was a subset of major adverse cardiovascular events: myocardial infarction (411.2), stroke (433.3), and heart failure (428.2). The cross‐sectional data structure implies that relationships between exposure and outcome variables establish associations, not causality.
Statistical Analysis
Primary Analysis
To evaluate the clinical association between DHRPS and CVD, a baseline model was constructed using 6 logistic regressions. These regression models estimated odds ratios (OR) and 95% CI for the association between DHRPS and each of the 6 primary outcomes, adjusting for age, sex, race, ethnicity, and BMI. We then created an adjusted multivariate model to demonstrate the incremental discriminative value of DHRPS beyond existing wearable‐derived metrics. The adjusted model included the same 6 logistic regressions, but with 2 additional independent variables—steps per day and DHR—added to the baseline independent variables. The change in ORs before and after adjusting for these metrics quantified the incremental discriminative information contributed by DHRPS. The prevalence ratio was calculated by dividing the prevalence of a condition in the exposed group by its prevalence in the unexposed group, with a corresponding 95% CI.
Additionally, the concordance index was used to assess the improvement in model discrimination provided by including DHRPS. Two models were constructed for this purpose: Model 1 included age, sex, race, ethnicity, BMI, steps per day, and DHR. Model 2 incorporated the same variables as Model 1 along with DHRPS. The 95% CIs for each concordance index were calculated using bootstrapping with 10 000 resamples, and the incremental improvements in discrimination between the 2 models were tested using the Wilcoxon signed‐rank test across the bootstrapped samples.
Phenome‐Wide Association Study Analysis
For the phenome‐wide association study (PheWAS), multivariate regression models were constructed to estimate the associations between wearable‐derived factors versus all 1789 disease phecodes across 17 disease categories while adjusting for age, sex, race, ethnicity, and BMI. To prevent false positives due to 1‐time miscoding, a case was counted if a participant had been coded for the same phecode in 2 or more encounters, occurring on separate dates, according to standard methodology.20 Multiple testing was adjusted with Bonferroni correction. The Z score and P value of the regression coefficient for wearable‐derived factors were used for comparison.
Subgroup Analyses
For the subgroup of participants with exercise stress test data available, we used Spearman's rank correlation to correlate wearable‐derived factors with maximum metabolic equivalents achieved. The difference in correlation values (DHRPS versus steps per day, DHRPS versus HR) were compared using Wilcoxon signed‐ranked test on 10 000 bootstrapped samples drawn with replacement. As a participant‐reported measure, we also examined the correlation between wearable‐derived factors and self‐reported physical health quality from the All of Us enrollment health survey. Likert‐based self‐reported physical health (1 “Poor” through 5 “Excellent”) was extracted from the survey question “In general, how would you rate your physical health?”
The participants were also divided into 3 groups based on DHRPS quartiles: low quartile (≤25 percentile or raw DHRPS value ≤0.0081), medium quartile (>25 percentile and <75 percentile), high quartile (≥75 percentile or raw DHRPS value ≥0.0147). Polynomial regression was used to estimate potentially nonlinear central tendency and uncertainty of CVD as it pertains to age as a continuous variable. According to our hypothesis, a high DHRPS value should be associated with elevated disease associations, having a higher ratio of average HR over lower number of steps. In contrast, a lower DHRPS should be associated with decreased disease associations, with the cardioprotective effects from a lower average HR over higher number of steps. After dividing participants into 3 DHRPS groups (low, medium, high), a second‐order polynomial regression model estimated the effects of each DHRPS group on primary outcome associations across the age spectrum, with the goal of uncovering curvilinear relationships that may exist and shedding light on critical age thresholds in CVD associations.
All analyses were performed on the All of Us Research Workbench in Python using the statsmodels package.
RESULTS
A total of 6947 individuals (mean±SD age, 54.6±15.1 years) were included in analysis, including 5.8 million person‐days and 50 billion total steps of individual‐level Fitbit device data. The median number of days with Fitbit data available was 795 days (interquartile range, 365–1342), ranging from 10 to 2710 days. The characteristics of the participants in each group are shown in Table 1.
| Overall | DHRPS profiles | |||
|---|---|---|---|---|
| Low (<25%) | Medium (25%–75%) | High (>75%) | ||
| No. | 6947 | 1737 | 3473 | 1737 |
| Age, y (SD) | 56.2 (15.4) | 58.2 (14.5) | 55.3 (15.5) | 56.1 (15.8) |
| Sex, n (%) | ||||
| Female | 5047 (72.7) | 1064 (61.3) | 2567 (73.9) | 1416 (81.5) |
| Male | 1900 (27.3) | 673 (38.7) | 906 (26.1) | 321 (18.5) |
| Race or ethnicity, n (%) | ||||
| Asian | 175 (2.5) | 36 (2.1) | 108 (3.1) | 31 (1.8) |
| Black, African‐American, or African | 377 (5.4) | 61 (3.5) | 185 (5.3) | 131 (7.5) |
| Hispanic, Latino, or Spanish | 389 (5.6) | 86 (5.0) | 200 (5.8) | 103 (5.9) |
| Not otherwise listed | 258 (3.7) | 52 (3.0) | 127 (3.7) | 79 (4.5) |
| White | 5748 (82.7) | 1502 (86.5) | 2853 (82.1) | 1393 (80.2) |
| Body mass index (SD) | 29.7 (6.9) | 26.9 (5.2) | 29.3 (6.2) | 33.1 (8.2) |
| Heart rate (SD) | 76.8 (8.5) | 72.9 (7.6) | 77.2 (8.0) | 80.0 (9.0) |
| Steps per day (SD) | 7597.6 (5376.0) | 11916.0 (8878.1) | 7242.8 (1311.5) | 3988.7 (1198.0) |
| DHRPS* (SD) | 12.9 (10.0) | 6.5 (1.1) | 10.9 (1.8) | 23.2 (15.3) |
| Diagnosis, n (%) | ||||
| Coronary atherosclerosis | 593 (8.5) | 151 (8.7) | 248 (7.1) | 194 (11.2) |
| Myocardial infarction | 192 (2.8) | 41 (2.4) | 78 (2.2) | 73 (4.2) |
| Heart failure | 257 (3.7) | 38 (2.2) | 93 (2.7) | 126 (7.3) |
| Stroke | 123 (1.8) | 24 (1.4) | 51 (1.5) | 48 (2.8) |
| Type 2 diabetes | 937 (13.5) | 139 (8.0) | 398 (11.5) | 400 (23.0) |
| Hypertension | 2691 (38.7) | 577 (33.2) | 1250 (36.0) | 864 (49.7) |
“High” DHRPS profiles have a large heart rate/step ratio and are associated with higher disease risks, whereas “low” DHRPS profiles have a small ratio. DHRPS indicates daily heart rate per step.
*
DHRPS was calculated as heart rate/steps per day, as noted in the Methods section. For convenience of display in the table, it was scaled by multiplying by 103.
High DHRPS Was Associated With Increased Prevalence of CVD
Overall, participants in the high DHRPS quartile group experienced at least a 34% increase in the prevalence ratio for all 6 primary outcomes compared with the medium DHRPS group (Table 2 P<0.001 for each among type II diabetes, hypertension, heart failure, stroke, coronary atherosclerosis, and myocardial infarction). Further, the high DHRPS quartile, with a high ratio of DHR to steps, was associated with at least a 56% increase in the prevalence ratio for all 6 primary outcomes compared with the low DHRPS quartile, a larger disparity compared with the medium DHRPS group.
| Outcome | Prevalence ratio (high vs medium) | P value | Prevalence ratio (high vs low) | P value |
|---|---|---|---|---|
| Coronary atherosclerosis | 1.35 [1.09–1.68] | <0.001 | 1.56 [1.30–1.87] | <0.001 |
| Myocardial infarction | 1.34 [1.10–1.66] | <0.001 | 1.90 [1.38–2.59] | <0.001 |
| Heart failure | 2.81 [2.15–3.66] | <0.001 | 3.28 [2.30–4.77] | <0.001 |
| Stroke | 2.04 [1.38–3.01] | <0.001 | 2.08 [1.28–3.37] | <0.001 |
| Type 2 diabetes | 2.03 [1.78–2.31] | <0.001 | 3.13 [2.60–3.79] | <0.001 |
| Hypertension | 1.42 [1.32–1.52] | <0.001 | 1.62 [1.49–1.77] | <0.001 |
High quartile (>75 percentile); medium group (>25 and ≤75 percentile); low quartile (≤25 percentile).
High DHRPS values were also associated with individual risk factors in multivariate regression. When including DHRPS in statistical models with both modifiable clinical risk factors and major adverse cardiovascular events, multivariate logistic regression showed that higher DHRPS values (achieved through increasing of DHR, decreasing of steps per day, or a combination of both) were associated with >40% increases in odds for all primary outcomes (P<0.001) while controlling for age, sex, race, ethnicity, and BMI (Figure 1). To ensure that DHRPS has discriminative value above the isolated wearable data components of steps per day and DHR, we evaluated DHRPS in an additional multivariable model which included steps per day and average HR as covariates. In this model, DHRPS was still associated with a statistically significant increased OR in type 2 diabetes (OR, 2.03 [95% CI, 1.70–2.42]; P<0.001), hypertension (OR, 1.63 [95% CI, 1.32–2.02]; P<0.001), heart failure (OR, 1.77 [95% CI, 1.00–3.14]; P=0.05), and coronary atherosclerosis (OR, 1.44 [95% CI, 1.14–1.82]; P<0.001). As further evidence that DHRPS provides incremental discriminative value, we calculated the concordance index for the 2 sets of logistic regression models established above. Compared with Model 1 with demographic and wearable factors, incorporating DHRPS in Model 2 showed a statistically significant increase in the concordance index for primary outcomes, indicating that DHRPS added incremental information to CVD risk discrimination (Table S1).
Figure 1. Incremental odds ratio and 95% CI for the DHRPS variable in logistic regression models on each primary outcome.
Models of primary cardiovascular outcomes vs DHRPS. Using 6 logistic regressions, one with each cardiovascular risk factor as the dependent variable, the odds ratio and 95% CI for DHRPS were calculated, controlling for the risk factors age, sex, race, ethnicity, and body mass index. Primary outcome with asterisk (*) remain statistically significant after controlling for the same baseline risk factors plus steps per day and daily heart rate. Dots are point estimates, plus 95% CIs. The red dotted line represents an odds ratio of 1 (ie, no association).
DHRPS Was Relevant Across the Lifespan
We also evaluated the relationship between DHRPS and CVD across the age spectrum (Figure S1). As expected, for all 6 primary outcomes, low DHRPS scores were protective for primary outcomes across the age spectrum; however, the effects were nonlinear. In middle aged and older participants (starting at age 35), high DHRPS had a statistically significant increase in CVD in type 2 diabetes, hypertension, stroke, and coronary atherosclerosis, compared with middle and low scores. The difference in the prevalence of heart failure between high DHRPS group versus other groups also increased with age.
DHRPS‐Based PheWAS
We then assessed the association between DHRPS and all EHR‐coded diagnoses. DHRPS‐based PheWAS identified 518 statistically significant phecodes, including previously established CVD risk factors such as obesity, type 2 diabetes, sleep apnea, and hypertension (Figure 2, Table S2). PheWAS based on DHRPS had higher statistical significance and at least 1.3 times higher regression coefficient Z score for primary outcomes, indicating stronger associations with disease status.21 In particular, the coefficient Z scores for DHRPS were stronger than DHR and step count combined, suggesting that DHRPS encodes additional clinical information beyond DHR and step count (Figure 3).
Figure 2. DHRPS‐based phenome‐wide association study.
Manhattan plot of DHRPS‐based PheWAS. PheWAS analyzes the association between many phenotypes compared with a single attribute, DHRPS, while adjusting for age, sex, race, ethnicity, and body mass index. The 1789 disease codes across 17 disease categories phenotypes were derived from preexisting All of Us classifications and the dashed line denotes phenome‐wide significance after Bonferroni corrections for multiple comparisons, equivalent to a P value of 2.9×10−5. Larger log(P value) represents smaller P‐values for the individual phenotype‐DHRPS comparisons. DHRPS indicates daily heart rate per step; GERD gastroesophageal reflux disease; and PheWAS, phenome‐wide association study.
Figure 3. Comparison of correlations between wearable risk factors and medical diagnoses: DHRPS, steps per day and heart rate.
Bar plots indicate DHRPS, steps per day, and heart rate PheWAS coefficient Z scores for the 6 primary outcomes. A higher coefficient Z score represents a stronger association with each primary outcome. DHRPS indicates daily heart rate per step; and PheWAS, phenome‐wide association study.
Exercise Stress Test Subgroup Analysis
In the small subgroup with Fitbit data and exercise stress testing (n=21), DHRPS showed stronger correlations with maximum metabolic equivalents achieved on exercise stress testing (ρ=0.47 [95% CI, 0.37–0.56]) compared with steps per day (∆ρ=0.04, P<0.001) and HR (∆ρ=0.31, P<0.001). Additionally, data from enrollment surveys for participants with Fitbit data revealed that DHRPS showed stronger correlations with self‐reported physical health than steps per day or HR (Figure 4).
Figure 4. Correlation between wearable risk factors and self‐reported physical health.
Bivariate correlation and 95% CI between self‐reported physical health at the time of recruitment and wearable risk factors. *** represents P<0.001. DHRPS indicates daily heart rate per step.
DISCUSSION
We report the development and initial validation of a novel wearables biomarker, DHRPS, for CVD risk discrimination. In this cross‐sectional study, we demonstrated that DHRPS was strongly associated with EHR coding for modifiable cardiovascular risk factors and major adverse cardiovascular events. These data were supported by the PheWAS, which evaluated the entire spectrum of clinical diagnoses associated with a continuous variable without selecting candidate variables for the model a priori. In the PheWAS, we found that the strongest association with DHRPS occurred in established CVD risk factors, including obesity, type 2 diabetes, sleep apnea, and hypertension. Thus, our data establish a relationship between elevated DHRPS and CVD both when we prespecify cardiovascular risk factors and adverse outcomes and when we use a hypothesis‐free PheWAS approach.
Further supporting the central link between elevated DHRPS and CVD, in a 21‐participant subgroup, DHRPS had stronger correlations than steps per day or DHR with established clinical measures of cardiovascular fitness on cardiovascular treadmill stress tests. The significance of these findings is that a favorable DHRPS score is a sensitive marker of overall health in the population. Equally important, DHRPS is a better model discriminator of CVD risk than either HR or steps‐per‐day.
To put these data in context, the component metrics of DHRPS are strong independent predictors of CVD risk. More steps per day is strongly associated with decreased risk of CVD and all‐cause mortality.13, 15, 22 The measurement of steps per day is a practical and accessible proxy for assessing an individual's physical activity level, reflecting lower‐level ambulatory activity and sedentary time.23 Separately, elevated resting HR has been positively associated with increased incidence of hypertension,24, 25 coronary atherosclerosis,26, 27 and heart failure.28 The parasympathetic nervous system also plays a key role in coordinating exercise capacity, with higher exercise capacity associated with lower resting HR through increased cardiac vagal activity.29
Given the pathogenic mechanism behind these behavioral biomarkers, we sought to develop DHRPS, a new wearables metric that quantifies the average HR per step taken throughout the day, by integrating both the frequency of physical activity (steps) and the cardiovascular response (HR) into a single measure. As an individual's activity levels vary naturally throughout the day, the HR also varies according to physiological demand and in response to the patient's underlying fitness level and autonomic state. DHRPS may offer insights into the heart's responsiveness to increased activity demands, reflecting its capacity to adjust under stress or exertion through established physiological mechanisms such as coronary flow reserve, which represents the heart's ability to increase blood flow to meet heightened oxygen demands.30 In this study, DHRPS had stronger associations with CVD than either HR or steps per day alone and was correlated with self‐reported physical health, suggesting that this easy‐to‐calculate metric may be a superior exposure variable in population health studies.
If validated in future studies, DHRPS may serve as a useful complement to clinical risk scores for CVD risk prediction. Additional prospective research, including studies that incorporate cardiopulmonary exercise testing with detailed metabolic data, will be needed to verify which components of the cardiovascular and autonomic system are being evaluated by DHRPS and to delineate the physiologic mechanisms that provide incremental value above steps per‐day and DHR. With the expanding availability of commercially available wearable devices, this new biomarker is a potential source of high‐quality data that encode both activity and cardiovascular fitness.
The cross‐sectional nature of All of Us is the most important limitation of this study. The design prevents establishment of causation or inference of longitudinal relationships. Although we have many days of DHRPS data for each participant, these are still fundamentally cross‐sectional data, gathered as a convenience sample in a population‐health protocol. Therefore, the wearable metrics were defined as exposure variables for disease prevalence and not chronological risk factors. Further, the subset of population who choose to participate in the All of Us Fitbit Bring‐Your‐Own‐Device Project likely introduces selection bias by including a predominance of women, self‐reported White race, and individuals who are interested in tracking their health status in the study population. Although Fitbit devices provide convenient and noninvasive data collection, factors such as device calibration, sensor accuracy, and participant compliance can influence the precision of data collected.1 Strengths of this study include the large number of participants with wearable device as well as the paired EHR.
CONCLUSIONS
In the All of Us Research Program Fitbit Bring‐Your‐Own‐Device Project, DHRPS had strong associations with CVD and provided stronger clinical associations with exercise capacity and aerobic fitness than isolated measures of HR or steps per day. With the expanding availability of commercially available wearable devices, this new biomarker could potentially enable clinicians to identify individuals at higher risk for poor CVD outcomes.
Sources of Funding
Partial support for this work was provided by National Institutes of Health R01 HL164773. The All of Us Research Program is supported by the National Institutes of Health, Office of the Director: Regional Medical Centers: 1 OT2 OD026549; 1 OT2 OD026554; 1 OT2 OD026557; 1 OT2 OD026556; 1 OT2 OD026550; 1 OT2 OD 026552; 1 OT2 OD026553; 1 OT2 OD026548; 1 OT2 OD026551; 1 OT2 OD026555; IAA #: AOD 16037; Federally Qualified Health Centers: HHSN 263201600085U; Data and Research Center: 5 U2C OD023196; Biobank: 1 U24 OD023121; The Participant Center: U24 OD023176; Participant Technology Systems Center: 1 U24 OD023163; Communications and Engagement: 3 OT2 OD023205; 3 OT2 OD023206; and Community Partners: 1 OT2 OD025277; 3 OT2 OD025315; 1 OT2 OD025337; 1 OT2 OD025276. In addition, the All of Us Research Program would not be possible without the partnership of its participants.
Footnotes
This article was sent to Tazeen H. Jafar, MD, MPH, Associate Editor, for review by expert referees, editorial decision, and final disposition.
Presented at the American College of Cardiology Scientific Sessions in Chicago, Illinois, March 29–31, 2025.
Supplemental Material is available at Supplemental Material
For Sources of Funding and Disclosures, see page 8.
Supplemental Material
Tables S1–S2
Figure S1
- Download
- 1.22 MB
REFERENCES
1.
Feehan LM, Geldman J, Sayre EC, Park C, Ezzat AM, Yoo JY, Hamilton CB, Li LC. Accuracy of Fitbit devices: systematic review and narrative syntheses of quantitative data. JMIR Mhealth Uhealth. 2018;6:e10527.
2.
Strain T, Wijndaele K, Dempsey PC, Sharp SJ, Pearce M, Jeon J, Lindsay T, Wareham N, Brage S. Wearable‐device‐measured physical activity and future health risk. Nat Med. 2020;26:1385–1391.
3.
Diaz KM, Krupka DJ, Chang MJ, Peacock J, Ma Y, Goldsmith J, Schwartz JE, Davidson KW. Fitbit®: an accurate and reliable device for wireless physical activity tracking. Int J Cardiol. 2015;185:138–140.
4.
Ladapo JA, Blecker S, Douglas PS. Physician decision making and trends in the use of cardiac stress testing in the United States: an analysis of repeated cross‐sectional data. Ann Intern Med. 2014;161:482–490.
5.
Carnethon MR, Gulati M, Greenland P. Prevalence and cardiovascular disease correlates of low cardiorespiratory fitness in adolescents and adults. JAMA. 2005;294:2981–2988.
6.
Salokari E, Laukkanen JA, Lehtimaki T, Kurl S, Kunutsor S, Zaccardi F, Viik J, Lehtinen R, Nikus K, Kööbi T, et al. The Duke treadmill score with bicycle ergometer: exercise capacity is the most important predictor of cardiovascular mortality. Eur J Prev Cardiol. 2019;26:199–207.
7.
Ross R, Blair SN, Arena R, Church TS, Després J‐P, Franklin BA, Haskell WL, Kaminsky LA, Levine BD, Lavie CJ, et al. Importance of assessing cardiorespiratory fitness in clinical practice: a case for fitness as a clinical vital sign: a scientific statement from the American Heart Association. Circulation. 2016;134:e653–e699.
8.
Mora S, Redberg RF, Cui Y, Whiteman MK, Flaws JA, Sharrett AR, Blumenthal RS. Ability of exercise testing to predict cardiovascular and all‐cause death in asymptomatic women: a 20‐year follow‐up of the lipid research clinics prevalence study. JAMA. 2003;290:1600–1607.
9.
Lauer M, Froelicher ES, Williams M, Kligfield P. Exercise testing in asymptomatic adults: a statement for professionals from the American Heart Association Council on Clinical Cardiology, Subcommittee on Exercise, Cardiac Rehabilitation, and Prevention. Circulation. 2005;112:771–776.
10.
Baugh CW, Blankstein R, Ganguli I, Januzzi JL, Morrow DA, Joseph JW, Jordan C, Donohoe G, Fofi J, McKinley K, et al. Frequency, compliance, and yield of cardiac testing after high‐sensitivity troponin accelerated diagnostic protocol implementation. Am J Emerg Med. 2023;72:64–71.
11.
Milano P, Carden DL, Jackman KM, Rongkavilit A, Groves K, Tyndall J, Moll J. Compliance with outpatient stress testing in low‐risk patients presenting to the emergency department with chest pain. Crit Pathw Cardiol. 2011;10:35–40.
12.
Xu Z, Zahradka N, Ip S, Koneshloo A, Roemmich RT, Sehgal S, Highland KB, Searson PC. Evaluation of physical health status beyond daily step count using a wearable activity sensor. NPJ Digit Med. 2022;5:164.
13.
Hall KS, Hyde ET, Bassett DR, Carlson SA, Carnethon MR, Ekelund U, Evenson KR, Galuska DA, Kraus WE, Lee I‐M, et al. Systematic review of the prospective association of daily step counts with risk of mortality, cardiovascular disease, and dysglycemia. Int J Behav Nutr Phys Act. 2020;17:1–14.
14.
Ayabe M, Brubaker PH, Dobrosielski D, Miller HS, Kiyonaga A, Shindo M, Tanaka H. Target step count for the secondary prevention of cardiovascular disease. Circ J. 2008;72:299–303.
15.
Sheng M, Yang J, Bao M, Chen T, Cai R, Zhang N, Chen H, Liu M, Wu X, Zhang B, et al. The relationships between step count and all‐cause mortality and cardiovascular events: a dose–response meta‐analysis. J Sport Health Sci. 2021;10:620–628.
16.
Rossen JD, Winniford MD. Effect of increases in heart rate and arterial pressure on coronary flow reserve in humans. J Am Coll Cardiol. 1993;21:343–348.
17.
McGinn AL, White CW, Wilson RF. Interstudy variability of coronary flow reserve. Influence of heart rate, arterial pressure, and ventricular preload. Circulation. 1990;81:1319–1330.
18.
Ramirez AH, Sulieman L, Schlueter DJ, Halvorson A, Qian J, Ratsimbazafy F, Loperena R, Mayo K, Basford M, Deflaux N, et al. The All of Us Research Program: data quality, utility, and diversity. Patterns. 2022;3:100570.
19.
Wu P, Gifford A, Meng X, Li X, Campbell H, Varley T, Zhao J, Carroll R, Bastarache L, Denny JC, et al. Mapping ICD‐10 and ICD‐10‐CM codes to phecodes: workflow development and initial evaluation. JMIR Med Inform. 2019;7:e14325.
20.
Verma A, Ritchie MD. Current scope and challenges in phenome‐wide association studies. Curr Epidemiol Rep. 2017;4:321–329.
21.
Master H, Annis J, Huang S, Beckman JA, Ratsimbazafy F, Marginean K, Carroll R, Natarajan K, Harrell FE, Roden DM, et al. Association of step counts over time with the risk of chronic disease in the All of Us Research Program. Nat Med. 2022;28:2301–2308.
22.
Paluch AE, Bajpai S, Ballin M, Bassett DR, Buford TW, Carnethon MR, Chernofsky A, Dooley EE, Ekelund U, Evenson KR, et al. Prospective association of daily steps with cardiovascular disease: a harmonized meta‐analysis. Circulation. 2023;147:122–131.
23.
Nayor M, Chernofsky A, Spartano NL, Tanguay M, Blodgett JB, Murthy VL, Malhotra R, Houstis NE, Velagaleti RS, Murabito JM, et al. Physical activity and fitness in the community: the Framingham Heart Study. Eur Heart J. 2021;42:4565–4575.
24.
Palatini P, Julius S. Heart rate and the cardiovascular risk. J Hypertens. 1997;15:3–17.
25.
Palatini P, Benetos A, Julius S. Impact of increased heart rate on clinical outcomes in hypertension: implications for antihypertensive drug therapy. Drugs. 2006;66:133–144.
26.
Fox K, Ford I, Steg PG, Tendera M, Robertson M, Ferrari R. Heart rate as a prognostic risk factor in patients with coronary artery disease and left‐ventricular systolic dysfunction (BEAUTIFUL): a subgroup analysis of a randomised controlled trial. Lancet. 2008;372:817–821.
27.
Huikuri HV, Jokinen V, Syvänne M, Nieminen MS, Airaksinen KJ, Ikäheimo MJ, Koistinen JM, Kauma H, Kesäniemi AY, Majahalme S, et al. Heart rate variability and progression of coronary atherosclerosis. Arterioscler Thromb Vasc Biol. 1999;19:1979–1985.
28.
Böhm M, Swedberg K, Komajda M, Borer JS, Ford I, Dubost‐Brama A, Lerebours G, Tavazzi L. Heart rate as a risk factor in chronic heart failure (SHIFT): the association between heart rate and outcomes in a randomised placebo‐controlled trial. Lancet. 2010;376:886–894.
29.
Gourine AV, Ackland GL. Cardiac vagus and exercise. Physiology. 2019;34:71–80.
30.
Kato S, Saito N, Nakachi T, Fukui K, Iwasawa T, Taguri M, Kosuge M, Kimura K. Stress perfusion coronary flow reserve versus cardiac magnetic resonance for known or suspected CAD. J Am Coll Cardiol. 2017;70:869–879.
Information & Authors
Information
Published In
Copyright
© 2025 The Author(s). Published on behalf of the American Heart Association, Inc., by Wiley. This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
Versions
You are viewing the most recent version of this article.
History
Received: 3 September 2024
Accepted: 20 March 2025
Published ahead of production: 29 March 2025
Published in print: 6 May 2025
Published online: 7 May 2025
Keywords
Subjects
Authors
Disclosures
None.
Metrics & Citations
Metrics
Citations
Download Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Select your manager software from the list below and click Download.
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginPurchase Options
Purchase this article to access the full text.
eLetters(0)
eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.
Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.