# Asymmetric U-shaped association between hypertension and wearable-derived sleep duration

## TL;DR

This study finds that both too little and too much sleep are linked to higher blood pressure and hypertension risk, with optimal sleep durations around 7.6-7.7 hours for sleep period time and 7.0-7.2 hours for total sleep time.

## Contribution

The novel finding is the asymmetric U-shaped relationship between sleep duration and hypertension, identified using wearable-derived sleep metrics.

## Key findings

- Short Sleep Period Time (<6 hours) is associated with a 2.05-fold higher hypertension prevalence.
- Blood pressure is lowest at 7.6–7.7 hours of Sleep Period Time and 7.0–7.2 hours of Total Sleep Time.
- Each additional hour of Sleep Period Time below optimal thresholds reduces systolic and diastolic blood pressure.

## Abstract

To investigate associations between hypertension and wearable-derived sleep duration, including Sleep Period Time (SPT) and Total Sleep Time (TST), in community-dwelling adults.

In this cross-sectional study, nocturnal SPT and TST were objectively quantified from single-lead electrocardiograms and analyzed using multimodal regression approaches including covariate-adjusted binary logistic regression, restricted cubic splines, and segmented linear regression.

Our analysis of 1,459 valid sleep records from 759 participants showed asymmetric U-shaped associations between objective sleep duration and both hypertension prevalence and blood pressure levels. After adjusting for confounding factors, participants with SPT < 6 h had a 2.05-fold higher prevalence of hypertension (95% CI: 1.18–3.54), while those with TST ≥ 8 h had a 1.82-fold increased odds (95% CI: 1.03–3.22). For SPT, blood pressure nadirs occurred at 7.6–7.7 h for both systolic blood pressure (SBP) and diastolic blood pressure (DBP). For TST, the optimal duration was 7.0–7.2 h in different models. Below optimal SPT thresholds, each 1-h increase was associated with a reduction of 2.94 mmHg in SBP (95% CI: −4.82 to −1.06) and 1.21 mmHg (95% CI: −2.05 to −0.36) in DBP. In partially adjusted models, each 1-h increase above the optimal SPT was associated with an increase in SBP of 2.27 mmHg (95% CI: 0.15 to 4.39). This effect was no longer significant in the fully adjusted analysis. For TST, similar dose–response patterns were significant only for DBP, in partially adjusted models.

Wearable-derived objective sleep duration demonstrates an asymmetric U-shaped association with both hypertension prevalence and blood pressure parameters. Blood pressure nadirs occurred at 7.6–7.7 h for SPT and 7.0–7.2 h for TST.