Can heart rate patterns measured by wearable devices help diagnose premenstrual disorders?

Background: Screening premenstrual disorders (PMDs) is time-consuming and challenging, resulting in delays in detection and treatment. To characterize the temporal dynamics of portable digital-marker across the menstrual cycle and its biological basis in PMDs, we investigated the menstrual fluctuation of wearable device-based real-time heart rate variability (HRV) through menstrual cycles and its associations with premenstrual disorders (PMDs) symptoms. Methods: A prospective study of female participants nested from the Care of Premenstrual Emotion Cohort was conducted. Daily HRV metrics were collected by the Huawei Fitness Tracker over 1-2 menstrual cycles. PMDs symptoms were assessed with the Daily Record of Severity of Problems on a daily basis. HRV variability across cycles was described using descriptive statistics and splines, while associations between HRV metrics and PMDs symptoms were estimated using a mixed-effect model. Results: In total, 193 participants (with 68 prospectively confirmed PMDs) were included, with measures from 293 menstrual cycles. In both women with and without PMDs, SDNN, rMSSD, and HF decreased before menses and increased afterwards; the increase trends were more pronounced in women without PMDs. During the week before or after menses, levels of these HRV metrics were inversely associated with PMDs symptoms among women with PMDs (e.g., rMSSD, postmenstrual week, β = -0·036 per SD, 95% CI: -0·065 to -0·048), whereas null association was noted for those without PMDs (β = -0·001, 95% CI -0·011 to 0·009; P-for-difference < 0·001). Conclusions: Wearable device-estimated HRV fluctuate across menstrual cycles, with varying strengths of association with PMDs symptoms between individuals with and without PMDs, which may facilitate the development of digital biomarkers for future PMDs diagnostics.

Reproduced under CC-BY.