Physiological Data Collected From Wearable Devices Identify and Predict Inflammatory Bowel Disease Flares

Abstract

Background & Aims

Wearable devices capture physiological signals noninvasively and passively. Many of these parameters have been linked to inflammatory bowel disease (IBD) activity. We evaluated the associative ability of several physiological metrics with IBD flares and how they change before the development of flare.

Methods

Participants throughout the United States answered daily disease activity surveys and wore an Apple Watch (Apple), Fitbit (Google), or Oura Ring (Oura Health). These devices collected longitudinal heart rate (HR), resting heart rate (RHR), heart rate variability (HRV), steps, and oxygenation. C-reactive protein, erythrocyte sedimentation rate, and fecal calprotectin were collected as standard of care. Linear mixed-effect models were implemented to analyze HR, RHR, steps, and oxygenation, and cosinor mixed-effect models were applied to HRV circadian features. Mixed-effect logistic regression was used to determine the predictive ability of physiological metrics.

Results

Three hundred and nine participants were enrolled across 36 states. Circadian patterns of HRV differed significantly between periods of inflammatory flare and remission and symptomatic flare and remission. Marginal means for HR and RHR were higher during periods of inflammatory flare and symptomatic flare. There were fewer daily steps during inflammatory flares. HRV, HR, and RHR differentiated whether participants with symptoms had inflammation. HRV, HR, RHR, steps, and oxygenation were significantly altered up to 7 weeks before inflammatory and symptomatic flares.

Conclusions

Longitudinally collected physiological metrics from wearable devices can identify and change before IBD flares, suggesting their feasibility to monitor and predict IBD activity.