Y-ECCO Literature Review: Lina Welz

The timely identification of Inflammatory Bowel Disease (IBD) activity remains a key challenge. Current tools such as blood tests, stool biomarkers, imaging, and endoscopy are all invasive to some extent, can be burdensome and are limited to single time-point assessments, while patient-reported symptoms carry a high risk of subjective bias. These gaps in monitoring can result in substantial delays to effective intervention and cause many patients with IBD to have suboptimal disease control in the interim.

One potential solution is the use of increasingly available wearable technologies to capture physiological signals both passively and continuously. Devices such as the Apple Watch, Fitbit and the Oura Ring can measure heart rate (HR), resting heart rate (RHR), heart rate variability (HRV), step count and oxygen saturation. These measures reflect autonomic nervous system function, which is known to be impaired in IBD, causing, for example, predominance of sympathetic states, particularly during active inflammation [1–3]. To date, however, only a limited number of studies have investigated the utility of wearables for IBD management [4–7]. Against this background, Hirten and colleagues performed the IBD Forecast Study, the largest prospective investigation on the topic to date, to determine whether wearable-derived physiological data may be of value in identifying and predicting IBD flares.

The study enrolled 309 participants across 36 states in the United States. The mean age of participants was 40 years and two-thirds were female. Both Crohn’s Disease (CD) and Ulcerative Colitis (UC) were represented, and participants were followed for an average of 213 days.

Inflammatory flares were defined by elevated biomarkers, specifically C-reactive protein >5 mg/dL, erythrocyte sedimentation rate >30 mm/h or faecal calprotectin >150 mg/g. Values were assigned to a ±7-day window around collection.

Symptomatic flares were identified using daily patient-reported outcomes (PRO-2): for CD, flares were defined as a PRO-2 score of ≥8 based on abdominal pain and stool frequency, while for UC, PRO-2 ≥1 with rectal bleeding >0 or stool frequency ≥1 was used. Statistical methods comprised linear and cosinor mixed-effects models as well as mixed-effects logistic regression.

The results revealed significant physiological changes during IBD flares. Both HR and RHR were elevated in periods of inflammatory and symptomatic flares as compared with phases of remission. HRV circadian rhythms exhibited relevant disruptions during inflammatory flares, allowing differentiation from inflammatory remission and reflecting elevated sympathetic activity. Similar changes were observed during symptomatic flares, although they were less predominant. Step counts declined significantly during inflammatory flares but did not fall during mere symptom-driven episodes.

Crucially, these physiological signals distinguished between symptomatic flares accompanied by underlying inflammation and those without such inflammation. Furthermore, oxygen saturation enabled differentiation between periods of symptomatic remission with and without concurrent presence of inflammation, thereby uncovering inflammatory states that might escape conventional monitoring.

Equally important, wearable-derived data revealed predictive potential. Alterations in HRV, HR, RHR, physical activity and oxygenation were detectable as early as seven weeks before flare onset. Predictive models that integrated these parameters achieved area under the curve values exceeding 0.95 and F1 scores between 0.81 and 0.90 up to 49 days in advance.