The prevention of cardiovascular disease requires identification of risk before the development of clinical events. Conventional cardiovascular-risk scores use established variables such as age, blood pressure, smoking, diabetes and serum lipids. [1,2] Although these factors are fundamental to preventive cardiology, they provide limited information regarding an individual's everyday physiological and behavioral patterns.

The development of wearable sensors has created an opportunity for continuous cardiovascular phenotyping.[3] Smartwatches and other wearable devices can measure heart rate, inter-beat intervals, physical activity, sedentary behavior, sleep and, in selected devices, cardiac rhythm. Such objectively measured signals can be transformed into digital biomarkers reflecting physiological or behavioral processes relevant to cardiovascular health. [4,5]

Physical activity is a particularly important wearable-derived biomarker. Device-based research has demonstrated an inverse relationship between daily step count and cardiovascular events and mortality. [6,7] Accelerometer-derived activity measurements may provide more objective assessment than self-reported physical-activity questionnaires and can quantify both active behavior and prolonged sedentary exposure.[8]

Sleep represents another important cardiovascular domain. Both insufficient sleep and sleep irregularity have been associated with cardiometabolic abnormalities and cardiovascular events. [9,10] Wearable devices allow repeated estimation of sleep duration, efficiency and regularity across multiple nights and may therefore capture habitual sleep patterns that cannot be represented adequately by a single questionnaire or laboratory assessment.

Wearable sensors may also detect previously unrecognized rhythm abnormalities. The Apple Heart Study demonstrated the feasibility of smartwatch-based irregular-pulse screening in a very large free-living population.[11] Similarly, the Fitbit Heart Study and other wearable-monitoring investigations have shown that prolonged digital monitoring can identify atrial fibrillation that might not be captured by a single resting ECG. [12,13]

No individual wearable biomarker is likely to represent the full complexity of cardiovascular risk. Resting heart rate reflects autonomic and metabolic status; HRV reflects autonomic regulation; step count reflects habitual activity; sedentary time captures inactivity; sleep parameters reflect recovery and circadian behavior; and rhythm monitoring may detect electrical abnormalities. Combining these domains could potentially provide a more informative risk phenotype than examining a single variable alone.[14]

The present study therefore evaluated multimodal wearable-derived digital biomarkers across conventional cardiovascular-risk categories and assessed whether an exploratory composite digital-biomarker score could discriminate high cardiovascular risk more effectively than individual wearable parameters.

This hospital-based prospective observational analytical study included a cross-sectional cardiovascular-risk assessment combined with a short longitudinal wearable-monitoring component.

 Participants

A total of 150 adults aged 30–70 years without previously established cardiovascular disease were enrolled. Participants with known coronary artery disease, prior myocardial infarction, stroke, heart failure, clinically significant valvular disease or documented sustained arrhythmias were excluded.

 Conventional cardiovascular-risk assessment

Participants underwent structured clinical history, anthropometric measurements, resting blood pressure, 12-lead ECG, fasting glucose and/or HbA1c and fasting lipid profile. An established cardiovascular-risk assessment approach was used to classify participants as low-, moderate- or high-risk.

The resulting groups comprised 68 low-risk, 52 moderate-risk and 30 high-risk participants.

 Wearable assessment

A validated wrist-worn wearable device was used continuously for at least seven consecutive days, with 14 days as the target monitoring period.

The following digital biomarkers were extracted:

• Resting heart rate and night-time heart rate
• SDNN and RMSSD
• Mean daily step count
• Active minutes per day
• Sedentary time
• Sleep duration
• Sleep efficiency
• Night-to-night sleep variability
• Irregular-rhythm notifications
• Single-lead ECG classification when available

An exploratory composite digital-biomarker score was generated from standardized individual wearable biomarkers, as defined in the parent study protocol.

 Outcome measures

The primary outcome for the present manuscript was the ability of the composite digital-biomarker score to discriminate participants categorized as having high cardiovascular risk.

Secondary outcomes included relationships of physical activity, sedentary behavior, sleep and rhythm abnormalities with conventional cardiovascular-risk categories.

 Statistical analysis

Continuous variables were summarized as mean ± SD and categorical variables as frequency and percentage. Comparisons across cardiovascular-risk groups were made using appropriate group-comparison tests.

Associations between wearable biomarkers and conventional risk scores were evaluated using correlation coefficients. Categorical rhythm outcomes were compared using chi-square or Fisher's exact testing as appropriate.

ROC curves were generated for resting heart rate, RMSSD, SDNN, daily steps, sedentary time, sleep efficiency and the composite digital-biomarker score. AUC, 95% confidence intervals, sensitivity and specificity were reported. Statistical significance was defined as p<0.05.

This hospital-based prospective observational analytical study included a cross-sectional cardiovascular-risk assessment combined with a short longitudinal wearable-monitoring component.

 Participants

A total of 150 adults aged 30–70 years without previously established cardiovascular disease were enrolled. Participants with known coronary artery disease, prior myocardial infarction, stroke, heart failure, clinically significant valvular disease or documented sustained arrhythmias were excluded.

 Conventional cardiovascular-risk assessment

Participants underwent structured clinical history, anthropometric measurements, resting blood pressure, 12-lead ECG, fasting glucose and/or HbA1c and fasting lipid profile. An established cardiovascular-risk assessment approach was used to classify participants as low-, moderate- or high-risk.

The resulting groups comprised 68 low-risk, 52 moderate-risk and 30 high-risk participants.

 Wearable assessment

A validated wrist-worn wearable device was used continuously for at least seven consecutive days, with 14 days as the target monitoring period.

The following digital biomarkers were extracted:

• Resting heart rate and night-time heart rate
• SDNN and RMSSD
• Mean daily step count
• Active minutes per day
• Sedentary time
• Sleep duration
• Sleep efficiency
• Night-to-night sleep variability
• Irregular-rhythm notifications
• Single-lead ECG classification when available

An exploratory composite digital-biomarker score was generated from standardized individual wearable biomarkers, as defined in the parent study protocol.

 Outcome measures

The primary outcome for the present manuscript was the ability of the composite digital-biomarker score to discriminate participants categorized as having high cardiovascular risk.

Secondary outcomes included relationships of physical activity, sedentary behavior, sleep and rhythm abnormalities with conventional cardiovascular-risk categories.

 Statistical analysis

Continuous variables were summarized as mean ± SD and categorical variables as frequency and percentage. Comparisons across cardiovascular-risk groups were made using appropriate group-comparison tests.

Associations between wearable biomarkers and conventional risk scores were evaluated using correlation coefficients. Categorical rhythm outcomes were compared using chi-square or Fisher's exact testing as appropriate.

ROC curves were generated for resting heart rate, RMSSD, SDNN, daily steps, sedentary time, sleep efficiency and the composite digital-biomarker score. AUC, 95% confidence intervals, sensitivity and specificity were reported. Statistical significance was defined as p<0.05.

The present study demonstrated that cardiovascular-risk status was associated with a multidimensional wearable phenotype involving autonomic function, physical activity, sedentary behavior, sleep and cardiac rhythm. Importantly, combining information across these domains resulted in better discrimination of high cardiovascular risk than any individual wearable-derived variable.

Daily step count decreased progressively across cardiovascular-risk groups, from 8,426 steps/day among low-risk participants to 5,018 steps/day in high-risk participants. This is consistent with the growing evidence that objectively measured physical activity is associated with cardiovascular outcomes. Banach et al., in a large meta-analysis, reported a graded inverse association between daily step count and both all-cause and cardiovascular mortality.[6] Del Pozo Cruz et al. similarly demonstrated associations between daily steps and cardiovascular incidence and mortality.[7]

Sedentary time showed the opposite relationship, increasing from 7.1 to 9.4 hours/day. Device-based cohort studies have demonstrated that accelerometer-measured physical activity and sedentary behavior are associated with incident cardiovascular disease.[8,15] These observations support the use of activity and sedentary behavior as complementary digital cardiovascular biomarkers.

Sleep parameters also differed significantly. High-risk participants slept approximately one hour less per night than low-risk participants and demonstrated poorer sleep efficiency and greater night-to-night variability. Sleep duration and regularity have increasingly been recognized as relevant components of cardiovascular health. St-Onge et al. described associations between sleep duration, cardiometabolic health and cardiovascular risk.[9] Huang et al. reported that sleep irregularity was associated with cardiovascular events in the Multi-Ethnic Study of Atherosclerosis.[10]

The rhythm-monitoring findings provide another clinically relevant dimension. Irregular-rhythm notifications increased from 2.9% in low-risk participants to 20.0% among high-risk participants. Five participants had recordings suggestive of atrial fibrillation. Large-scale studies have established that wearable photoplethysmography can identify irregular pulse patterns compatible with AF.[11,12] The Apple Heart Study enrolled more than 400,000 participants and demonstrated the feasibility of smartwatch-based irregular-pulse notification.[11]

However, wearable-detected pulse irregularity should be regarded as a screening signal rather than a definitive diagnosis. Confirmation using diagnostic-quality ECG is required before clinical decisions are made.[13,16]

The principal finding of this analysis was the performance of the composite digital-biomarker score. Individual AUC values ranged from 0.69 for sleep efficiency to 0.78 for RMSSD. The multimodal composite increased the AUC to 0.86, with sensitivity of 83.3% and specificity of 79.2%.

This observation is conceptually important because cardiovascular risk is multifactorial. A single wearable parameter captures only one physiological domain. HRV primarily represents autonomic regulation, activity measures behavior, sleep represents recovery and circadian function, and rhythm monitoring captures electrical abnormalities. Multimodal integration can therefore potentially increase the signal-to-noise ratio by combining complementary information.

Kańtoch and Kańtoch have previously demonstrated the feasibility of combining multimodal smartwatch-derived biomarkers for cardiovascular and pre-frailty risk assessment.[14] More broadly, advances in machine learning and digital phenotyping increasingly support the integration of heterogeneous wearable signals into clinically interpretable risk models.[17]

Nevertheless, the present composite score should be considered exploratory. It was developed and assessed within the same dataset. As acknowledged in the thesis, this can produce optimistic estimates of discriminatory performance and requires validation in an independent cohort.

Wearable-derived digital biomarkers spanning autonomic regulation, physical activity, sedentary behavior, sleep and cardiac rhythm were significantly associated with conventional cardiovascular-risk status. Individuals at higher risk demonstrated lower HRV, reduced daily activity, greater sedentary exposure, shorter and less efficient sleep and more frequent irregular-rhythm notifications. The composite digital-biomarker score achieved an AUC of 0.86 with 83.3% sensitivity and 79.2% specificity, exceeding the discriminatory performance of the individual wearable parameters examined. Multimodal wearable monitoring may therefore provide a useful additional layer of information for early cardiovascular-risk assessment. It should presently complement, rather than replace, established clinical risk tools, pending prospective outcome-based and external validation.

1. Prabhakaran D, Jeemon P, Roy A. Cardiovascular diseases in India: current epidemiology and future directions. Circulation. 2016;133:1605-1620. doi:10.1161/CIRCULATIONAHA.114.008729.
2. Arnett DK, Blumenthal RS, Albert MA, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease. Circulation. 2019;140:e596-e646. doi:10.1161/CIR.0000000000000678.
3. Bayoumy K, Gaber M, Elshafeey A, et al. Smart wearable devices in cardiovascular care: where we are and how to move forward. Nat Rev Cardiol. 2021;18:581-599. doi:10.1038/s41569-021-00522-7.
4. Coravos A, Goldsack JC, Karlin DR, et al. Digital medicine: a primer on measurement. Digit Biomark. 2019;3:31-71. doi:10.1159/000500413.
5. Coravos A, Khozin S, Mandl KD. Developing and adopting safe and effective digital biomarkers to improve patient outcomes. NPJ Digit Med. 2019;2:14. doi:10.1038/s41746-019-0090-4.
6. Banach M, Lewek J, Surma S, et al. The association between daily step count and all-cause and cardiovascular mortality: a meta-analysis. Eur J Prev Cardiol. 2023;30:1975-1985. doi:10.1093/eurjpc/zwad229.
7. del Pozo Cruz B, Ahmadi MN, Lee IM, Stamatakis E. Prospective associations of daily step counts and intensity with cardiovascular disease incidence and mortality and all-cause mortality. JAMA Intern Med. 2022;182:1139-1148. doi:10.1001/jamainternmed.2022.4000.
8. Ramakrishnan R, Doherty A, Smith-Byrne K, et al. Accelerometer measured physical activity and the incidence of cardiovascular disease: evidence from the UK Biobank cohort study. PLoS Med. 2021;18:e1003487.
9. St-Onge MP, Grandner MA, Brown D, et al. Sleep duration and quality: impact on lifestyle behaviors and cardiometabolic health. Circulation. 2016;134:e367-e386.
10. Huang T, Mariani S, Redline S. Sleep irregularity and risk of cardiovascular events: the Multi-Ethnic Study of Atherosclerosis. J Am Coll Cardiol. 2020;75:991-999.
11. Perez MV, Mahaffey KW, Hedlin H, et al. Large-scale assessment of a smartwatch to identify atrial fibrillation. N Engl J Med. 2019;381:1909-1917. doi:10.1056/NEJMoa1901183.
12. Lubitz SA, Faranesh AZ, Selvaggi C, et al. Detection of atrial fibrillation in a large population using wearable devices: the Fitbit Heart Study. Circulation. 2022;146:1415-1424. doi:10.1161/CIRCULATIONAHA.122.060291.
13. Guo Y, Wang H, Zhang H, et al. Mobile photoplethysmographic technology to detect atrial fibrillation. J Am Coll Cardiol. 2019;74:2365-2375. doi:10.1016/j.jacc.2019.08.019.
14. Kańtoch E, Kańtoch A. Cardiovascular and pre-frailty risk assessment based on multimodal biomarkers collected from smart telemedical wearables. J Clin Med. 2021;10:1997. doi:10.3390/jcm10091997.
15. Dempsey PC, Strain T, Khaw KT, Wareham NJ, Brage S, Wijndaele K. Prospective associations of accelerometer-measured physical activity and sedentary time with incident cardiovascular disease. Circulation. 2020;141:1113-1115.
16. Varma N, Cygankiewicz I, Turakhia M, et al. 2021 ISHNE/HRS/EHRA/APHRS collaborative statement on mHealth in arrhythmia management. Eur Heart J Digit Health. 2021;2:7-48. doi:10.1093/ehjdh/ztab001.
17. Krittanawong C, Rogers AJ, Johnson KW, et al. Integration of novel monitoring devices with machine learning technology for scalable cardiovascular management. Nat Rev Cardiol. 2021;18:75-91.