Heart Rate Circadian Oscillations as Digital Biomarkers of Cardiometabolic Health Determinants
Preprint
Abstract
AIMS: Cardiometabolic risk factors may impair health by altering the autonomic modulation of the cardiovascular system, a physiological process described by heart rate (HR) circadian oscillations. However, the impact of cardiometabolic health determinants on HR circadian oscillations remains scarcely characterized in real-world, population-based settings. To address this, we applied digital health technologies to investigate how cardiometabolic health determinants shape HR circadian oscillations in a real-world cohort of individuals free of cardiometabolic diseases. METHODS: First, a 10-fold cross-validation of a model was performed, aiming at mitigating wearables’ measurement error caused by motion artifacts. This process was informed by 10,056 epochs of concurrent wearable-derived and polysomnographic HR assessment, yielding an average 1.3 bpm reduction in wearables measurement error. We subsequently applied this model to over 2 million 1-minute epochs of HR data, derived from 7-day continuous actigraphic recordings of 245 individuals free of cardiometabolic disorders. Functional-on-scalar regression modelling and both parametric and nonparametric analyses characterized HR circadian profiles and their relationships with demographics, lifestyle, chronotype, sleep health, and chronic insomnia diagnosis. A 6-dimension sleep health index was calculated. RESULTS: Sex, chronotype, and sleep health predominantly shaped HR circadian oscillations. In detail, females consistently showed higher HR across the 24 hours. Moreover, chronotype was associated to a phase shift in HR circadian profiles, with later timings corresponding to eveningness. Notably, sleep health impacted HR circadian oscillations in a dose-dependent fashion: each additional impaired sleep dimension was associated with a 1.2 bpm HR increase during nighttime, alongside reduced circadian robustness and delayed oscillation timings. Finally, the earlier occurrence of morning HR peaks served as a digital biomarker of insomnia (80% specificity, 74% sensitivity). CONCLUSIONS: This work provides a digital health framework to characterize HR circadian oscillations in free-living populations and supports its clinical utility in capturing the autonomic disruptions related to cardiometabolic health determinants. GRAPHICALKeywords
This publication has 39 references indexed in Scilit:
- Real-world findings of daridorexant in chronic insomnia: A pilot study on subjective and actigraphic sleep parametersSleep Medicine, 2026
- Predicting Acute Decompensated Heart Failure Using Circadian Markers From Heart Rate Time SeriesESC Heart Failure, 2025
- Day and night heart rate variability using 24-h ECG recordings: a systematic review with meta-analysis using a gender lensClinical Autonomic Research, 2023
- A method for characterizing daily physiology from widely used wearablesCell Reports Methods, 2021
- Acute electrical, autonomic and structural effects of binge drinking: Insights into the ‘holiday heart syndrome’International Journal of Cardiology, 2021
- ‘nparACT’ package for R: A free software tool for the non-parametric analysis of actigraphy dataMethodsX, 2016
- Adverse Effects of Cigarette and Noncigarette Smoke Exposure on the Autonomic Nervous SystemJACC, 2014
- Validity of the Italian Version of the Pittsburgh Sleep Quality Index (PSQI)Neurological Sciences, 2012
- Scaling regression inputs by dividing by two standard deviationsStatistics in Medicine, 2007
- Italian version of the Epworth sleepiness scale: external validityNeurological Sciences, 2003