This paper is in the following e-collection/theme issue:

Electronic/Mobile Data Capture, Internet-based Survey & Research Methodology (832) Cardiac Arrhythmia, Atrial Fibrillation (84) Wearable Devices and Sensors (888) Fitness Trackers and Smart Pedometers/Accelerometers (561) Innovations and Technology for Physical Activity Education (742)

Published on in Vol 25 (2023)

Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/43123, first published .
Step Count, Self-reported Physical Activity, and Predicted 5-Year Risk of Atrial Fibrillation: Cross-sectional Analysis

Step Count, Self-reported Physical Activity, and Predicted 5-Year Risk of Atrial Fibrillation: Cross-sectional Analysis

Step Count, Self-reported Physical Activity, and Predicted 5-Year Risk of Atrial Fibrillation: Cross-sectional Analysis

Authors of this article:

Ayelet Shapira-Daniels1 Author Orcid Image ;   Jelena Kornej2 Author Orcid Image ;   Nicole L Spartano3 Author Orcid Image ;   Xuzhi Wang4 Author Orcid Image ;   Yuankai Zhang4 Author Orcid Image ;   Chathurangi H Pathiravasan4 Author Orcid Image ;   Chunyu Liu4 Author Orcid Image ;   Ludovic Trinquart4 Author Orcid Image ;   Belinda Borrelli5 Author Orcid Image ;   David D McManus6, 7 Author Orcid Image ;   Joanne M Murabito1, 2 Author Orcid Image ;   Emelia J Benjamin2, 8, 9 Author Orcid Image ;   Honghuang Lin2, 6 Author Orcid Image
Article Authors Cited by Tweetations (18) Metrics

    Original Paper

    1Section of General Internal Medicine, Department of Medicine, Boston Medical Center, Boston University Chobanian and Avedisian School of Medicine, Boston, MA, United States

    2Boston University's Framingham Heart Study, National Heart, Lung, and Blood Institute, Framingham, MA, United States

    3Section of Endocrinology, Diabetes, Nutrition and Weight Management, Department of Medicine, Boston Medical Center, Boston University Chobanian and Avedisian School of Medicine, Boston, MA, United States

    4Department of Biostatistics, Boston University School of Public Health, Boston, MA, United States

    5Center for Behavioral Science Research, Henry M Goldman School of Dental Medicine, Boston University, Boston, MA, United States

    6Department of Medicine, University of Massachusetts Chan Medical School, Worcester, MA, United States

    7Department of Quantitative Health Sciences, University of Massachusetts Chan Medical School, Worcester, MA, United States

    8Department of Epidemiology, Boston University School of Public Health, Boston, MA, United States

    9Section of Cardiovascular Medicine, Department of Medicine, Boston Medical Center, Boston University Chobanian and Avedisian School of Medicine, Boston, MA, United States

    Corresponding Author:

    Honghuang Lin, PhD

    Department of Medicine

    University of Massachusetts Chan Medical School

    55 North Lake Avenue

    Worcester, MA, 01655

    United States

    Phone: 1 774 455 4881

    Email: Honghuang.Lin@umassmed.edu


    Background: Physical inactivity is a known risk factor for atrial fibrillation (AF). Wearable devices, such as smartwatches, present an opportunity to investigate the relation between daily step count and AF risk.

    Objective: The objective of this study was to investigate the association between daily step count and the predicted 5-year risk of AF.

    Methods: Participants from the electronic Framingham Heart Study used an Apple smartwatch. Individuals with diagnosed AF were excluded. Daily step count, watch wear time (hours and days), and self-reported physical activity data were collected. Individuals’ 5-year risk of AF was estimated, using the Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE)–AF score. The relation between daily step count and predicted 5-year AF risk was examined via linear regression, adjusting for age, sex, and wear time. Secondary analyses examined effect modification by sex and obesity (BMI≥30 kg/m2), as well as the relation between self-reported physical activity and predicted 5-year AF risk.

    Results: We examined 923 electronic Framingham Heart Study participants (age: mean 53, SD 9 years; female: n=563, 61%) who had a median daily step count of 7227 (IQR 5699-8970). Most participants (n=823, 89.2%) had a <2.5% CHARGE-AF risk. Every 1000 steps were associated with a 0.08% lower CHARGE-AF risk (P<.001). A stronger association was observed in men and individuals with obesity. In contrast, self-reported physical activity was not associated with CHARGE-AF risk.

    Conclusions: Higher daily step counts were associated with a lower predicted 5-year risk of AF, and this relation was stronger in men and participants with obesity. The utility of a wearable daily step counter for AF risk reduction merits further investigation.

    J Med Internet Res 2023;25:e43123

    doi:10.2196/43123

    Keywords

    atrial fibrillation; physical activity; fitness tracker; cardiovascular epidemiology; fitness; exercise; tracker; cardiology; heart; walk; step count; smartwatch; wearable; risk; cross-sectional analysis 


    Atrial fibrillation (AF) is the most common cardiac arrhythmia, and it is an important cause of stroke, heart failure, and death [1]. The Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE)–AF score is a tool that has been validated to estimate an individual’s 5-year risk of developing AF, using relevant clinical information and known risk factors, such as height, weight, blood pressure, and a history of heart failure and myocardial infarction [2-4].

    In recent years, researchers have investigated AF risk modification via lifestyle changes, such as decreased alcohol consumption, weight loss, smoking cessation, and other factors [5-11]. In particular, physical activity has been examined as a means to decrease the risk of cardiovascular disease (CVD) and AF [12-20]. Studies that examine objectively measured physical activity and AF have mostly used research-grade accelerometers or implantable loop recorders, which limit applicability to daily life [16-20]. Although research-grade accelerometers provide superior validity and precision in their ability to reflect an individual’s physical activity, they are not as widely available to the general public as commercially available wearable devices [21]. Additionally, in studies that measured physical activity with research-grade accelerometers, physical activity data were collected for only a brief period of time (4-7 days) [17,18]. The sharp rise in the prevalence of wearable devices with step counters has resulted in a unique opportunity for health management and optimization, with direct applicability to individuals’ lifestyles. To our knowledge, a direct relation between the long-term tracking of daily step counts from commercially available wearable devices and the risk of AF has yet to be investigated.

    We hypothesized that a higher daily step count, as measured by wearable devices, is associated with a lower 5-year risk of AF, as predicted by the CHARGE-AF score.


    Study Sample

    The Framingham Heart Study (FHS), which originated in 1948 to investigate CVD, is a community-based cohort study that spans 3 generations of families [22,23]. In recent years, participants from the Third Generation Cohort, multiethnic Omni Group 2 Cohort, and New Offspring Spouse Cohort were invited to enroll in the electronic FHS (eFHS) at the time of their third research examination (2016-2019) [24]. The use of the Apple Watch (Series 0; Apple Inc)—a smartwatch that allows for the tracking of daily steps and heart rate—was incorporated into the eFHS in 2016. Participants were required to be English speakers and own an iPhone (Apple Inc) to be eligible for the study.

    Ethics Approval

    This study was approved by the Institutional Review Board of Boston University Medical Center (approval number: H-36586). All participants provided written informed consent.

    Collection of Objective Physical Activity Data

    In this analysis, we selected eFHS participants, as displayed in Figure S1 in Multimedia Appendix 1. Of the 3521 participants examined in the research center, 1948 had a compatible iPhone, provided consent, and were ultimately enrolled in our eFHS sample. Only 1185 used the Apple Watch and returned step data. We excluded 244 participants who did not wear the watch for >30 days and 18 participants with prevalent AF. Participants were encouraged to wear their Apple Watch daily.

    The total number of hours of watch wear time per day and the number of days participants wore the watch were recorded. A wear hour was defined as an hour with at least one heart rate measure or the time when at least 30 steps were accumulated. An active day was defined as a day with at least 5 watch wear hours. Average daily step count and watch wear time were defined as participants’ mean daily step counts and their total wear hours, respectively, for active days.

    Estimation of AF Risk

    The primary dependent measure was participants’ 5-year risk of AF, which was estimated based on the CHARGE-AF risk scores that were calculated by using the clinical risk factors assessed when participants were examined at the FHS research center [24]. This previously validated prediction model is a Cox proportional hazard regression model. The prediction model uses an individual’s age (years), height (cm), weight (kg), self-reported race and ethnicity, systolic and diastolic blood pressure (mm Hg), current smoking status, antihypertensive medication use, history of diabetes mellitus, history of heart failure, and history of myocardial infarction to predict the individual’s 5-year risk of AF.

    Self-reported Physical Activity

    Participants were asked to complete a questionnaire during their FHS research examinations to determine their physical activity levels. This questionnaire, which has been used in other FHS studies, asked participants to estimate the number of hours in a typical day that they spent performing varying levels of physical activity over the past year [25,26]. We then calculated a physical activity index (PAI) score as a weighted composite of hours per day spent performing activities of varying physical intensity. For example, sleeping was weighted at 1.0; slight activity, such as standing and walking, was weighted at 1.5; and high-intensity activity, such as jogging and swimming, was weighted at 5. As such, PAI scores could hypothetically range from 24 (eg, 24 hours per day of sleeping) to 120 (eg, 24 hours per day of high-intensity exercise), with higher PAI scores indicating a participant’s perception of higher daily physical activity levels.

    Statistical Methods

    Daily step count, watch wear time, and clinical variables were reported as means with SDs for continuous variables and as n values with percentages for dichotomous variables. When continuous variable distribution was skewed, medians with IQRs were reported.

    The primary analysis examined the association between daily step count (independent measure) and the CHARGE-AF risk score (dependent measure) via linear regression, adjusting for age, sex, and average wear time per day. Both daily step count and the CHARGE-AF score were treated as continuous variables.

    Secondary analyses tested for interactions between daily step count and sex and between daily step count and obesity (BMI≥30 kg/m2) in their association with the CHARGE-AF risk score, given that male sex and obesity are established independent risk factors for AF. For graphic purposes, we performed an analysis that examined the CHARGE-AF score, as a dependent variable, in high versus low physical activity groups (<7500 vs ≥7500 daily steps). High and low physical activity groups were determined by using the average step count of the study sample as the cutoff. The analysis was performed by using a Wilcoxon rank-sum test and adjusted for age.

    Additional analyses examined the association between self-reported physical activity (PAI score) and CHARGE-AF risk via linear regression, adjusting for age and sex. Sensitivity analyses were performed by using different thresholds for watch wear time (5 vs 10 hours/day) and number of active days (30 vs 60 vs 90 days). A 2-sided P value of <.05 was considered statistically significant. All of the analyses were performed by using R software package version 4.0.3 (R Foundation for Statistical Computing).


    Participant Characteristics

    We included 923 participants in this study. The mean age of participants was 53 (SD 9) years, 563 (61%) participants were female, and 838 (90.8%) participants identified as White. The median daily step count was 7227 (IQR 5699-8970), and the median watch wear time was 13.6 (IQR 12.4-14.7) hours per day for 324 (IQR 137-563) active days. The median CHARGE-AF risk score was 1% for men and 0.5% for women (Figure 1). Participants’ mean PAI score, which we calculated based on their self-reported physical activity levels, was 33.4 (SD 4.7; range 24-120; lower scores indicate less self-reported physical activity). Additional demographic characteristics can be seen in Table 1.

    ‎
    Figure 1. Distribution of 5-year AF risk among participants. Most participants (823/923, 89.2%) had a 5-year AF risk of <2.5%, as determined by the CHARGE-AF score. AF: atrial fibrillation; CHARGE: Cohorts for Heart and Aging Research in Genomic Epidemiology.
    Table 1. Characteristics of the study participants.
    CharacteristicsaParticipants (N=923)
    Age (years), mean (SD)53 (9)
    Sex (female), n (%)563 (61)
    BMI (kg/m2), mean (SD)28.2 (5.5)
    Height (cm), mean (SD)169 (9)
    Weight (kg), mean (SD)81 (18)
    Systolic blood pressure (mm Hg), mean (SD)118 (14)
    Diastolic blood pressure (mm Hg), mean (SD)76 (9)
    History of heart failure, n (%)6 (0.7)
    History of myocardial infarction, n (%)12 (1.3)
    Current smoking, n (%)44 (4.8)
    Diabetes mellitus, n (%)53 (5.7)
    Antihypertensive medication use, n (%)193 (20.9)
    Self-reported physical activity index score, mean (SD)33.4 (4.7)
    Self-reported race and ethnicity, n (%)

    		White838 (90.8)

    		Black21 (2.2)

    		Asian17 (1.8)

    		Hispanic27 (2.9)

    		Other20 (2.2)
    Daily step count, median (IQR)7227 (5699-8970)
    Active hours per day, mean (SD)13.4 (1.9)
    Active days, median (IQR)324 (137-562)

    aPresented are means with SDs for continuous traits with a normal distribution, medians with IQRs for continuous traits with a skewed distribution, and n values with percentages for dichotomous traits.

    Association Between Daily Step Count and 5-Year Risk of AF

    After adjusting for age, sex, and watch wear time, our primary analysis, in which linear regression was performed, showed that daily step count and the CHARGE-AF score were inversely associated. Every 1000 steps were associated with a 0.08% lower CHARGE-AF risk score (Table 2).

    Figure 2 shows an age-adjusted analysis of high versus low physical activity levels. Participants who took ≥7500 steps daily had lower estimated CHARGE-AF risk scores than those of participants who took <7500 steps daily (mean 0.90% vs mean 1.3%; P<.001).

    We observed significant interactions by sex and obesity; the association between the CHARGE-AF score and daily step count was stronger in men and participants with obesity. The CHARGE-AF risk score was 0.14% and 0.05% lower for every 1000 steps in men and women, respectively (interaction: P<.001). Similarly, the CHARGE-AF score was 0.10% and 0.03% lower for every 1000 steps in individuals with obesity and individuals without obesity, respectively (interaction: P<.001).

    Participants’ self-reported physical activity levels, as determined by the PAI score, were not associated with the CHARGE-AF score.

    Sensitivity analyses were conducted to investigate the association between daily step count and the CHARGE-AF score, using different device wear thresholds. The watch wear cutoff was increased to ≥60 days and ≥90 days from the cutoff of ≥30 days used in primary analysis. In a separate analysis, the wear time threshold was increased to 10 hours from the threshold of 5 hours used in the primary model. The inverse association between daily step count and the CHARGE-AF risk score remained significant in these subgroups (Tables S1 and S2 in Multimedia Appendix 1).

    Table 2. Association between average daily step count and the predicted 5-year atrial fibrillation (AF) riska.

    		Change in AF risk score per 1000 steps (%), βb (SE)P value
    All participants−0.08 (0.01)<.001
    Men−0.14 (0.02)<.001
    Women−0.05 (0.01).001
    No obesity−0.03 (0.01)<.001
    Obesityc−0.10 (0.03)<.001

    aThe model was adjusted for age, sex (for the model including all participants), and wear time.

    bβ represents the change in 5-year AF risk for every 1000-step increase in daily step count.

    bObesity was defined as a BMI of ≥30 kg/m2.

    ‎
    Figure 2. Participants' 5-year AF risk by high versus low daily step count. Participants with lower daily step counts had higher CHARGE-AF scores (mean 1.3% for low daily step count vs mean 0.9% for high daily step count; P<.001). The model was adjusted for age. AF: atrial fibrillation; CHARGE: Cohorts for Heart and Aging Research in Genomic Epidemiology.

    Main Findings

    In our observational cross-sectional analysis of eFHS participants, the 5-year AF risk, as predicted by the CHARGE-AF score, was low (0.5% in women and 1% in men), and the score was 0.08% lower for every 1000 steps in a model adjusted for age, sex, and wear time. Interactions were significant for sex (P<.001) and obesity (P<.001); in men and participants with obesity, the association between the CHARGE-AF score and daily step count was stronger. Finally, self-reported physical activity, as determined by the PAI score, was not associated with the CHARGE-AF score. A graphic abstract that presents our main findings is presented in Multimedia Appendix 2.

    Relation Between Physical Activity and AF

    The overall benefits of increased physical activity have been well documented, including reductions in all-cause mortality, CVD-related mortality, and overall CVD risk [12-15,27,28]. The relation between physical activity and the risk of AF is more complex. Studies that examined incident AF and physical activity suggested an increased AF risk with vigorous physical activity but a protective benefit of low to moderate physical activity against AF [29-33]. These studies however relied on self-reported physical activity [29-33].

    The association between future AF risk and objectively measured physical activity has been investigated in both short-term accelerometer use and the long-term use of implantable devices (≥25 months). AF risk was either estimated with the CHARGE-AF score or calculated after monitoring for incident AF [17-20]. All of these methods detected an association between lower physical activity levels and a higher risk of AF, but these studies have limitations. In studies that used research-grade accelerometers, which are not commercially available, the device wear time was ≤7 days, and it may not have reflected an individual’s habitual, long-term physical activity lifestyle [17,18]. Additionally, these studies categorized physical activity into levels, such as “light,” “moderate,” or “vigorous.” This methodology aligns with the recommendations from public health organizations, which currently recommend ≥150 minutes of moderate physical activity per week or ≥75 minutes of vigorous physical activity per week. Although the use of step counts may result in the overestimation or underestimation of intentional physical activity, step count targets may be more understandable to and practical for the general population than the abstract concepts of moderate and vigorous levels of physical activity. Additionally, evidence is evolving to suggest that ≥7000 daily steps in adults may be equivalent to the current public health recommendation of ≥150 minutes of moderate physical activity per week or ≥75 minutes of vigorous physical activity per week [34]. Studies that use invasive devices, such as implantable loop recorders or implantable cardioverter-defibrillators, are less applicable to the general population, and such devices are not regularly used [19,20]. Our study may offer a more accurate window into a participant’s overall physical activity lifestyle and routine. As such, monitoring daily step count for a longer period of time with a commercially available noninvasive device may provide a more real-life reflection of the relation between physical activity and the estimated risk of AF. Additionally, our use of daily step count instead of a physical activity level (eg, moderate to vigorous physical activity) may provide a clearer and more practical target for AF risk reduction in the future.

    Our study also demonstrated that sex and obesity may modify the association between AF risk and daily step count, as a stronger association between AF risk and daily step count was noted in men and participants with obesity. The independently increased risk of AF in men and individuals with obesity may have contributed to the stronger association that we observed between step count and the CHARGE-AF score [35].

    Finally, our study did not find evidence of an association between self-reported physical activity and predicted AF risk. This lack of association may be related to the inconsistent validity of self-reported physical activity [36,37]. Our findings are consistent with the lack of a significant association between self-reported physical activity and the predicted risk of CVD [15].

    The biological mechanisms behind the inverse association between step count, as an objective measure of physical activity, and estimated AF risk may be embedded within the shared risk factors between CVD and AF. On a cellular level, increased step counts have been associated with lower chronic inflammation and cardiac biomarkers, such as lipoproteins, white blood cell count, troponin, and N-terminal pro–brain natriuretic peptide [38-40]. As such, physical activity has been associated with lower atherosclerosis burden, as evidenced by decreased carotid artery plaque, thickness, and stiffness, as well as lower coronary artery calcification scores [41-43]. Finally, higher step counts and physical activity levels have also been associated with lower blood pressure and lower rates of obesity, coronary heart disease, and heart failure, of which all are major risk factors for AF [27,44,45].

    Study Limitations

    Our study has several limitations. First, this study is limited by its observational and cross-sectional design, precluding the ability to establish causality, establish temporality, or rule out residual confounding. Additionally, the cross-sectional design precludes prospective follow-ups for the occurrence of incident AF. As such, the use of the estimated CHARGE-AF risk (as opposed to incident AF) as the primary outcome limits interpretations of clinical significance. However, CHARGE-AF risk has been extensively validated in large data sets with good predictive performance for incident AF [46-48]. A second limitation is the absence of heart rate as a variable in our regression model; heart rate data were unavailable to be contemporarily correlated with step count at the time of analysis. The lack of concurrent heart rate data may limit the inference of the association between physical activity and AF risk, but it does not affect the association between objectively measured step count and estimated AF risk. A third limitation is that the majority of participants in this study (n=838/923 (90.8%)) were White, middle-aged or older (age: mean 52 years) who were presumably living in Massachusetts. Additionally, compared to the FHS participants who did not enroll in the eFHS, the participants in the eFHS were younger, had higher levels of education, and had fewer CVD risk factors [24]. Further, the accuracy of wearable devices among individuals with darker skin tones is unclear [49,50]. Given these limitations, the applicability of this study to the general population is limited. Additionally, we anticipate participation and observation bias due to the language and smartphone eligibility requirements, as well as inaccuracies in measures derived from wearable devices, as previously reported [51]. The inference of physical activity level by using step counts may result in the overestimation and underestimation of vigorous physical activity.

    Conclusions

    Increasing daily steps can be a practical, lifestyle-modifying method for reducing an individual’s AF risk. Future studies could investigate a dose-dependent relation between step count and AF risk and examine this relation in more ethnically diverse, racially diverse, and age-diverse populations. Given the emerging relation between AF risk, objectively measured physical activity, and daily step count, assessing commercially available wearable devices for AF preventative risk reduction merits further investigation.

    Acknowledgments

    We would like to warmly acknowledge Emily S Manders of Boston University’s Framingham Heart Study and the National Heart, Lung, and Blood Institute for her contribution to this project.

    This study was supported by an award from the Robert Wood Johnson Foundation (number 74624) and a grant from the National Heart, Lung, and Blood Institute (R01HL141434). The Framingham Heart Study was supported by a contract from the National Heart, Lung, and Blood Institute (principal investigator [PI]: Vasan S Ramachandran 75N92019D00031). The investigators were supported by the following National Institutes of Health grants: R01HL126911 (EJB), R01HL092577 (EJB), R01 R01AG066010 (EJB), 18SFRN34110082 (EJB), U54HL120163 (EJB), R01HL126911 (DDM), R01HL137734 (DDM), R01HL137794 (DDM), R01HL13660 (DDM), U54HL143541 (DDM), and U01AG068221 (HL). This work was also supported by the American Heart Association (18SFRN34150007; PI: LT). JK received funding from the Marie Sklodowska-Curie Actions under the European Union’s Horizon 2020 research and innovation program (agreement number: 838259). The Apple watches were provided to Boston University by Apple Inc at no cost to the study.

    Conflicts of Interest

    NLS receives research support from Novo Nordisk for an investigator-initiated grant unrelated to this study. JMM was a guest lecturer at Merck, unrelated to this work. DDM has received honoraria, speaking, consulting, or editorial fees from the Heart Rhythm Society, Bristol-Myers Squibb, Pfizer, Samsung, Flexcon, Philips, Fitbit, Avania, and Venturewell; Data Safety Monitoring Boards including the Boston Biomedical Associates, Avania, and NAMSA. DDM also reports receiving research grants or support from Boehringer Ingelheim, Bristol-Myers Squibb, Philips, Samsung, Apple, Biotronik, Boehringer Ingelheim, Pfizer, Sanofi, Flexcon, Fitbit, and National Institutes of Health (NIH). EJB reports receiving grants from NIH.

    ‎
    Multimedia Appendix 1

    Supplementary figure and tables.

    PDF File (Adobe PDF File), 261 KB
    ‎
    Multimedia Appendix 2

    Graphic abstract.

    PDF File (Adobe PDF File), 309 KB
    1. Benjamin EJ, Wolf PA, D'Agostino RB, Silbershatz H, Kannel WB, Levy D. Impact of atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation 1998 Sep 08;98(10):946-952. [CrossRef] [Medline]
    2. Alonso A, Krijthe BP, Aspelund T, Stepas KA, Pencina MJ, Moser CB, et al. Simple risk model predicts incidence of atrial fibrillation in a racially and geographically diverse population: the CHARGE-AF consortium. J Am Heart Assoc 2013 Mar 18;2(2):e000102 [FREE Full text] [CrossRef] [Medline]
    3. Himmelreich JCL, Veelers L, Lucassen WAM, Schnabel RB, Rienstra M, van Weert HCPM, et al. Prediction models for atrial fibrillation applicable in the community: a systematic review and meta-analysis. Europace 2020 May 01;22(5):684-694 [FREE Full text] [CrossRef] [Medline]
    4. Poorthuis MHF, Jones NR, Sherliker P, Clack R, de Borst GJ, Clarke R, et al. Utility of risk prediction models to detect atrial fibrillation in screened participants. Eur J Prev Cardiol 2021 May 22;28(6):586-595 [FREE Full text] [CrossRef] [Medline]
    5. Abed HS, Wittert GA, Leong DP, Shirazi MG, Bahrami B, Middeldorp ME, et al. Effect of weight reduction and cardiometabolic risk factor management on symptom burden and severity in patients with atrial fibrillation: a randomized clinical trial. JAMA 2013 Nov 20;310(19):2050-2060. [CrossRef] [Medline]
    6. Aune D, Schlesinger S, Norat T, Riboli E. Tobacco smoking and the risk of atrial fibrillation: A systematic review and meta-analysis of prospective studies. Eur J Prev Cardiol 2018 Sep;25(13):1437-1451 [FREE Full text] [CrossRef] [Medline]
    7. Lau DH, Nattel S, Kalman JM, Sanders P. Modifiable risk factors and atrial fibrillation. Circulation 2017 Aug 08;136(6):583-596. [CrossRef] [Medline]
    8. Middeldorp ME, Ariyaratnam J, Lau D, Sanders P. Lifestyle modifications for treatment of atrial fibrillation. Heart 2020 Mar;106(5):325-332. [CrossRef] [Medline]
    9. Schnabel RB, Yin X, Gona P, Larson MG, Beiser AS, McManus DD, et al. 50 year trends in atrial fibrillation prevalence, incidence, risk factors, and mortality in the Framingham Heart Study: a cohort study. Lancet 2015 Jul 11;386(9989):154-162 [FREE Full text] [CrossRef] [Medline]
    10. Voskoboinik A, Kalman JM, De Silva A, Nicholls T, Costello B, Nanayakkara S, et al. Alcohol abstinence in drinkers with atrial fibrillation. N Engl J Med 2020 Jan 02;382(1):20-28. [CrossRef] [Medline]
    11. Conner SC, Lodi S, Lunetta KL, Casas JP, Lubitz SA, Ellinor PT, et al. Refining the association between body mass index and atrial fibrillation: G-formula and restricted mean survival times. J Am Heart Assoc 2019 Aug 20;8(16):e013011 [FREE Full text] [CrossRef] [Medline]
    12. Kannel WB, Belanger A, D'Agostino R, Israel I. Physical activity and physical demand on the job and risk of cardiovascular disease and death: the Framingham Study. Am Heart J 1986 Oct;112(4):820-825. [CrossRef] [Medline]
    13. Shortreed SM, Peeters A, Forbes AB. Estimating the effect of long-term physical activity on cardiovascular disease and mortality: evidence from the Framingham Heart Study. Heart 2013 May;99(9):649-654. [CrossRef] [Medline]
    14. Kubota Y, Evenson KR, Maclehose RF, Roetker NS, Joshu CE, Folsom AR. Physical activity and lifetime risk of cardiovascular disease and cancer. Med Sci Sports Exerc 2017 Aug;49(8):1599-1605 [FREE Full text] [CrossRef] [Medline]
    15. Lin H, Sardana M, Zhang Y, Liu C, Trinquart L, Benjamin EJ, et al. Association of habitual physical activity with cardiovascular disease risk. Circ Res 2020 Oct 23;127(10):1253-1260 [FREE Full text] [CrossRef] [Medline]
    16. Semaan S, Dewland TA, Tison GH, Nah G, Vittinghoff E, Pletcher MJ, et al. Physical activity and atrial fibrillation: Data from wearable fitness trackers. Heart Rhythm 2020 May;17(5 Pt B):842-846 [FREE Full text] [CrossRef] [Medline]
    17. O'Neal WT, Bennett A, Singleton MJ, Judd SE, Howard G, Howard VJ, et al. Objectively measured physical activity and the risk of atrial fibrillation (from the REGARDS study). Am J Cardiol 2020 Aug 01;128:107-112. [CrossRef] [Medline]
    18. Khurshid S, Weng LC, Al-Alusi MA, Halford JL, Haimovich JS, Benjamin EJ, et al. Accelerometer-derived physical activity and risk of atrial fibrillation. Eur Heart J 2021 Jul 01;42(25):2472-2483 [FREE Full text] [CrossRef] [Medline]
    19. Bonnesen MP, Frodi DM, Haugan KJ, Kronborg C, Graff C, Højberg S, et al. Day-to-day measurement of physical activity and risk of atrial fibrillation. Eur Heart J 2021 Oct 07;42(38):3979-3988 [FREE Full text] [CrossRef] [Medline]
    20. Palmisano P, Guerra F, Ammendola E, Ziacchi M, Pisanò ECL, Dell'Era G, Italian Association of Arrhythmology and Cardiac Pacing (AIAC). Physical activity measured by implanted devices predicts atrial arrhythmias and patient outcome: Results of IMPLANTED (Italian Multicentre Observational Registry on Patients With Implantable Devices Remotely Monitored). J Am Heart Assoc 2018 Feb 24;7(5):e008146 [FREE Full text] [CrossRef] [Medline]
    21. Strath SJ, Kaminsky LA, Ainsworth BE, Ekelund U, Freedson PS, Gary RA, American Heart Association Physical Activity Committee of the Council on Lifestyle and Cardiometabolic Health and Cardiovascular, Exercise, Cardiac Rehabilitation and Prevention Committee of the Council on Clinical Cardiology, and Council. Guide to the assessment of physical activity: Clinical and research applications: a scientific statement from the American Heart Association. Circulation 2013 Nov 12;128(20):2259-2279. [CrossRef] [Medline]
    22. McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: the Framingham study. N Engl J Med 1971 Dec 23;285(26):1441-1446. [CrossRef] [Medline]
    23. Andersson C, Johnson AD, Benjamin EJ, Levy D, Vasan RS. 70-year legacy of the Framingham Heart Study. Nat Rev Cardiol 2019 Nov;16(11):687-698. [CrossRef] [Medline]
    24. McManus DD, Trinquart L, Benjamin EJ, Manders ES, Fusco K, Jung LS, et al. Design and preliminary findings from a new electronic cohort embedded in the Framingham Heart Study. J Med Internet Res 2019 Mar 01;21(3):e12143 [FREE Full text] [CrossRef] [Medline]
    25. Spartano NL, Davis-Plourde KL, Himali JJ, Murabito JM, Vasan RS, Beiser AS, et al. Self-reported physical activity and relations to growth and neurotrophic factors in diabetes mellitus: The Framingham Offspring Study. J Diabetes Res 2019 Jan 09;2019:2718465 [FREE Full text] [CrossRef] [Medline]
    26. Tan ZS, Spartano NL, Beiser AS, DeCarli C, Auerbach SH, Vasan RS, et al. Physical activity, brain volume, and dementia risk: The Framingham Study. J Gerontol A Biol Sci Med Sci 2017 Jun 01;72(6):789-795 [FREE Full text] [CrossRef] [Medline]
    27. Rothenbacher D, Koenig W, Brenner H. Lifetime physical activity patterns and risk of coronary heart disease. Heart 2006 Sep;92(9):1319-1320 [FREE Full text] [CrossRef] [Medline]
    28. Saint-Maurice PF, Troiano RP, Bassett DRJ, Graubard BI, Carlson SA, Shiroma EJ, et al. Association of daily step count and step intensity with mortality among US adults. JAMA 2020 Mar 24;323(12):1151-1160 [FREE Full text] [CrossRef] [Medline]
    29. Ricci C, Gervasi F, Gaeta M, Smuts CM, Schutte AE, Leitzmann MF. Physical activity volume in relation to risk of atrial fibrillation. A non-linear meta-regression analysis. Eur J Prev Cardiol 2018 May;25(8):857-866. [CrossRef] [Medline]
    30. Mozaffarian D, Furberg CD, Psaty BM, Siscovick D. Physical activity and incidence of atrial fibrillation in older adults: the cardiovascular health study. Circulation 2008 Aug 19;118(8):800-807 [FREE Full text] [CrossRef] [Medline]
    31. Aizer A, Gaziano JM, Cook NR, Manson JE, Buring JE, Albert CM. Relation of vigorous exercise to risk of atrial fibrillation. Am J Cardiol 2009 Jun 01;103(11):1572-1577 [FREE Full text] [CrossRef] [Medline]
    32. Mohanty S, Mohanty P, Tamaki M, Natale V, Gianni C, Trivedi C, et al. Differential association of exercise intensity with risk of atrial fibrillation in men and women: Evidence from a meta-analysis. J Cardiovasc Electrophysiol 2016 Sep;27(9):1021-1029. [CrossRef] [Medline]
    33. Kunutsor SK, Seidu S, Mäkikallio TH, Dey RS, Laukkanen JA. Physical activity and risk of atrial fibrillation in the general population: meta-analysis of 23 cohort studies involving about 2 million participants. Eur J Epidemiol 2021 Mar;36(3):259-274 [FREE Full text] [CrossRef] [Medline]
    34. Lane-Cordova AD, Jerome GJ, Paluch AE, Bustamante EE, LaMonte MJ, Pate RR, Committee on Physical Activity of the American Heart Association Council on Lifestyle and Cardiometabolic Health. Supporting physical activity in patients and populations during life events and transitions: A scientific statement from the American Heart Association. Circulation 2022 Jan 25;145(4):e117-e128 [FREE Full text] [CrossRef] [Medline]
    35. Staerk L, Sherer JA, Ko D, Benjamin EJ, Helm RH. Atrial fibrillation: Epidemiology, pathophysiology, and clinical outcomes. Circ Res 2017 Apr 28;120(9):1501-1517 [FREE Full text] [CrossRef] [Medline]
    36. Ferrari P, Friedenreich C, Matthews CE. The role of measurement error in estimating levels of physical activity. Am J Epidemiol 2007 Oct 01;166(7):832-840. [CrossRef] [Medline]
    37. Colley RC, Butler G, Garriguet D, Prince SA, Roberts KC. Comparison of self-reported and accelerometer-measured physical activity in Canadian adults. Health Rep 2018 Dec 19;29(12):3-15 [FREE Full text] [Medline]
    38. Kotani K, Taniguchi N. Pedometer step counts and oxidized low-density lipoprotein levels among asymptomatic subjects. Ann Clin Lab Sci 2012;42(4):435-438. [Medline]
    39. Parsons TJ, Sartini C, Welsh P, Sattar N, Ash S, Lennon LT, et al. Objectively measured physical activity and cardiac biomarkers: A cross sectional population based study in older men. Int J Cardiol 2018 Mar 01;254:322-327 [FREE Full text] [CrossRef] [Medline]
    40. Klenk J, Denkinger M, Nikolaus T, Peter R, Rothenbacher D, Koenig W, ActiFE Study Group. Association of objectively measured physical activity with established and novel cardiovascular biomarkers in elderly subjects: every step counts. J Epidemiol Community Health 2013 Feb;67(2):194-197. [CrossRef] [Medline]
    41. Chen L, Bi Y, Su J, Cui L, Han R, Tao R, et al. Physical activity and carotid atherosclerosis risk reduction in population with high risk for cardiovascular diseases: a cross-sectional study. BMC Public Health 2022 Feb 07;22(1):250 [FREE Full text] [CrossRef] [Medline]
    42. Sung KC, Hong YS, Lee JY, Lee SJ, Chang Y, Ryu S, et al. Physical activity and the progression of coronary artery calcification. Heart 2021 Nov;107(21):1710-1716. [CrossRef] [Medline]
    43. Tanaka H, Palta P, Folsom AR, Meyer ML, Matsushita K, Evenson KR, et al. Habitual physical activity and central artery stiffening in older adults: the Atherosclerosis Risk in Communities study. J Hypertens 2018 Sep;36(9):1889-1894 [FREE Full text] [CrossRef] [Medline]
    44. Kelley GA, Kelley KS, Tran ZV. Walking and resting blood pressure in adults: a meta-analysis. Prev Med 2001 Aug;33(2 Pt 1):120-127. [CrossRef] [Medline]
    45. Igarashi Y, Akazawa N, Maeda S. The required step count for a reduction in blood pressure: a systematic review and meta-analysis. J Hum Hypertens 2018 Dec;32(12):814-824. [CrossRef] [Medline]
    46. Kartoun U, Khurshid S, Kwon BC, Patel AP, Batra P, Philippakis A, et al. Prediction performance and fairness heterogeneity in cardiovascular risk models. Sci Rep 2022 Jul 22;12(1):12542 [FREE Full text] [CrossRef] [Medline]
    47. Khurshid S, Kartoun U, Ashburner JM, Trinquart L, Philippakis A, Khera AV, et al. Performance of atrial fibrillation risk prediction models in over 4 million individuals. Circ Arrhythm Electrophysiol 2021 Jan;14(1):e008997 [FREE Full text] [CrossRef] [Medline]
    48. Hulme OL, Khurshid S, Weng LC, Anderson CD, Wang EY, Ashburner JM, et al. Development and validation of a prediction model for atrial fibrillation using electronic health records. JACC Clin Electrophysiol 2019 Nov;5(11):1331-1341 [FREE Full text] [CrossRef] [Medline]
    49. Sañudo B, De Hoyo M, Muñoz-López A, Perry J, Abt G. Pilot study assessing the influence of skin type on the heart rate measurements obtained by photoplethysmography with the Apple Watch. J Med Syst 2019 May 22;43(7):195. [CrossRef] [Medline]
    50. Shcherbina A, Mattsson CM, Waggott D, Salisbury H, Christle JW, Hastie T, et al. Accuracy in wrist-worn, sensor-based measurements of heart rate and energy expenditure in a diverse cohort. J Pers Med 2017 May 24;7(2):3 [FREE Full text] [CrossRef] [Medline]
    51. Germini F, Noronha N, Debono VB, Philip BA, Pete D, Navarro T, et al. Accuracy and acceptability of wrist-wearable activity-tracking devices: Systematic review of the literature. J Med Internet Res 2022 Jan 21;24(1):e30791 [FREE Full text] [CrossRef] [Medline]

    ‎
    AF: atrial fibrillation
    CHARGE: Cohorts for Heart and Aging Research in Genomic Epidemiology
    CVD: cardiovascular disease
    eFHS: electronic Framingham Heart Study
    FHS: Framingham Heart Study
    PAI: physical activity index

    Edited by R Kukafka, G Eysenbach; submitted 30.09.22; peer-reviewed by N Freene; comments to author 17.11.22; revised version received 24.11.22; accepted 30.11.22; published 06.03.23

    Copyright

    ©Ayelet Shapira-Daniels, Jelena Kornej, Nicole L Spartano, Xuzhi Wang, Yuankai Zhang, Chathurangi H Pathiravasan, Chunyu Liu, Ludovic Trinquart, Belinda Borrelli, David D McManus, Joanne M Murabito, Emelia J Benjamin, Honghuang Lin. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 06.03.2023.

    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 the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on https://www.jmir.org/, as well as this copyright and license information must be included.