(M-075) Leveraging Digital Biomarkers to Enhance Clinical Pharmacology and Pharmacometrics Informed Drug Development: A Case Study

Corinna Maier, na – Sr. Pharmacometrician, AbbVie Deutschland GmbH & Co.; Jie Shen, na – VP, Local Delivery Translational Sciences, AbbVie; Dan Webster, na – Director, Digital Science, AbbVie; Insa Winzenborg, na – Director, Pharmacometrics, AbbVie Deutschland GmbH & Co.; Ronilda D'Cunha, na – Associate Director, Clinical Pharmacology, AbbVie; Sven Mensing, na – Head of Pharmacometrics, AbbVie Deutschland GmbH & Co.

Integrating digital health technologies (DHTs), such as wearable sensors and mobile device applications, into model-informed drug development offers unprecedented opportunities by collecting continuous patient data remotely and objectively assessing patient behavior. This study aims to showcase the benefits and challenges of integrating digital biomarkers to enhance the evaluation of drug efficacy and safety, using a case study on sleep parameter monitoring in a Phase 2 clinical trial.

In this study, sleep monitoring was collected from 35 glucocorticoid-dependent patients over 36 weeks including a 24 week glucocorticoid taper. Therapeutic glucocorticoids have been found to disrupt the normal cortisol circadian rhythm, impairing sleep quality.1 Our analysis investigated the potential improvement in sleep and restoration of the normal cortisol circadian rhythm when tapering off synthetic glucocorticoids and introduction of drug X in a glucocorticoid-dependent patient population.

The wrist-wearing actigraphy watch was used to collect continuous physical activity and sleep data for a week after each clinical visit. The sleep parameters were derived from the minute-level sleep/awake epoch data generated by the Cole-Kripke algorithm 2 from the raw accelerometer data. By integrating the sleep monitoring data with data on drug exposure, we aimed to determine a potential return to the normal circadian rhythm and an improvement in sleep quality. A model for normal cortisol circadian rhythm was developed for placebo subjects and pre-dose measurements of cortisol2 to characterize deviations from the normal circadian rhythm.

This case study did not reveal clear exposure-response trends on aggregated sleep quality data nor clear effects of glucocorticoid tapering on sleep. Individual analysis of time profiles shows strong variability within and between subjects. The case study identified several challenges, including the management of vast volumes of raw data requiring advanced preprocessing methods, e.g. signal processing, dimension reduction and feature engineering. Limited availability of baseline measurements due to constrained study procedures and limited sample size from voluntary participation, posed a challenge to robust longitudinal PKPD analysis. Despite these challenges, the integration of DHTs into pharmacometric modeling presents considerable advantages: It offers more objective and precise outcomes than traditional patient-reported sleep questionnaires, facilitates early detection of drug efficacy and safety markers, and can provide a deeper understanding of drug action mechanisms. This can eventually further inform response time-courses enhancing the overall understanding of drug efficacy and disease progression.

Citations: [1] Szmyd et al. (2021), “The impact of glucocorticoids and statins on sleep quality” Sleep Medicine Reviews. https://doi.org/10.1016/j.smrv.2020.101380.
[2] Cole RJ et al (1992). Automatic sleep/wake identification from wrist activity. Sleep. Oct;15(5):461-9. doi: 10.1093/sleep/15.5.461. PMID: 1455130.
[3] Scherholz et al. (2019) “Chronopharmacology of glucocorticoids” Advanced Drug Delivery Reviews. https://doi.org/10.1016/j.addr.2019.02.004.