Women who sleep less may face more irregular menstrual cycles

By Dr. Priyom Bose, Ph.D.Reviewed by Lauren HardakerMay 27 2026

Tracking 42,759 menstrual cycles with wearable devices, researchers uncovered how sleep patterns, age, and physiology intertwine, revealing that even modest sleep disruptions may be linked to measurable changes in menstrual health. 

Study: The menstrual cycle through the lens of a wearable device: insights into physiology, sleep, and cycle variability. Image credit: New Africa/Shutterstock.com

A recent NPJ Digital Medicine study investigated how wearable device data can reveal relationships among menstrual cycle phases, sleep patterns, and physiological variability in women.

Current understanding and unresolved questions in menstrual cycle physiology

The menstrual cycle involves recurring hormonal, metabolic, and behavioral fluctuations in individuals of reproductive age. Cycle length and variability are key indicators of female health, with increased variability and irregularity linked to negative menstrual symptoms and greater risk for long-term health conditions such as cancer, diabetes, cardiovascular disease, fractures, and premature mortality. Additionally, decreased motivation to train, increased negative mood, and reduced sleep quality are frequently reported before or during menstruation.

Physiological, sleep, and performance metrics are known to fluctuate across the menstrual cycle. Digital tools and wearable devices have facilitated large-scale monitoring of menstruation, physiological biometrics, and behaviors such as sleep and physical activity. Biometrics, such as skin temperature and resting heart rate, show systematic variation across the cycle and are influenced by age and reproductive status.

Despite these advances, significant research gaps persist. Most studies lack daily-resolution biometric data across diverse ages and cycle lengths, limiting understanding of individual variability. The relationship between behavioral factors, particularly sleep duration, and menstrual cycle characteristics remains inadequately defined, especially in real-world settings. These gaps highlight the need for comprehensive, longitudinal research to clarify how behaviors and physiological patterns interact across the menstrual cycle.

Assessing the effects of the menstrual cycle on physiological biometrics and sleep patterns

The current study analyzed data from regular WHOOP device users. The devices collected nightly biometric data, including resting heart rate, respiratory rate, heart rate variability (HRV), skin temperature, and blood oxygen saturation levels, as well as sleep and workout metrics.

Inclusion criteria required consistent device wear and regular cycles (median length 21–35 days); women using hormonal contraceptives, pregnant, or with menopause symptoms were excluded. The final cohort comprised 2,596 participants, 42,759 cycles, and over 1.29 million days of data. The authors noted that the cohort likely represented individuals who were more active and health-conscious than the general population and may not fully reflect broader demographic groups.

Daily time series integrated menstrual, behavioral, and physiological data, with cycles segmented into premenstrual, menstrual, postmenstrual, and other phases. Sleep and workout metrics were summarized for each day. Some biometric data were unavailable for certain participants due to device upgrades.

Biometric time series were interpolated, filtered, and normalized to enable cross-individual comparisons. Missing data of ≤7 days were linearly interpolated; longer gaps were filled with the participant’s mean for that metric.

Cycle length was defined as the interval between cycle starts, with deviations classified as cycles differing by ±3 days from the participant’s median cycle length. Sleep duration variability was assessed as variance within each cycle. Generalized Estimating Equations (GEE) were used for inference, with cycles as the unit of analysis and covariates including age, BMI, seasonality, sleep, and workout metrics. Behavioral changes were identified based on stable sleep averages over three weeks.

Biometric modeling employed Generalized Additive Models (GAM) to disentangle the effects of age and cycle length, with covariates for sleep, exercise, BMI, seasonality, and weekend effects. Temporal relationships between biometrics were explored using Vector Autoregression (VAR).

Menstrual cycle stability is influenced by sleep and physiological variation

Average menstrual cycle length was 28.4 days, decreasing from 29.1 days at age 24 to 26.9 days at age 44. Cycle length variability followed a U-shaped pattern, with the least variability near age 33.

Shorter and inconsistent sleep were strongly associated with greater menstrual cycle variability, while average cycle length remained largely unchanged. Sleeping less than 7.3 hours or having irregular sleep patterns was associated with greater cycle irregularity, suggesting a potential role for regular sleep in menstrual cycle stability. A within-participant analysis of 813 individuals confirmed that greater sleep variability within an individual was associated with more variable cycle lengths.

High-resolution biometric data revealed clear physiological rhythms across the menstrual cycle. Most biometrics dipped during menstruation and peaked before the next cycle, except HRV, which showed the opposite pattern. Resting heart rate, HRV, and respiratory rate were closely linked throughout the cycle.

With age, fluctuations in HRV and resting heart rate lessened, while changes in other biometrics were minor. Longer cycles were associated with a greater range of cardiorespiratory biometrics. Blood oxygen saturation showed little cyclicality. Thus, age and cycle length are key in shaping menstrual physiology.

Population-level biometrics captured main trends, but individual cycles showed much greater variability. HRV, for example, often fluctuated by nearly half of a participant’s mean value within a single cycle. This underscores substantial within-person variation beyond population averages.

Although population waveforms are stable, individual cycles reveal marked variability. Shorter sleep, especially in the premenstrual week, was associated with increased resting heart rate, reduced HRV, and changes in other physiological metrics. This pattern was similar across menstrual phases, underscoring that sleep loss was consistently linked to physiological changes throughout the cycle.

Conclusions

This study shows that regular, sufficient sleep is closely linked to menstrual cycle stability, with age and cycle length also playing significant roles in shaping physiological rhythms. Substantial individual variability in biometric patterns highlights the need for personalized approaches to menstrual health. Future research should clarify the mechanisms underlying these associations and assess whether interventions that promote sleep regularity could help improve cycle stability and overall well-being.