Blood glucose variance measured by continuous glucose monitors across the menstrual cycle

Georgianna Lin, Rumsha Siddiqui, Zixiong Lin, Joanna M. Blodgett, Shwetak N. Patel, Khai N. Truong, Alex MariakakisNpj Digit. Medicine 2023 · DOI

Summary

By leveraging continuous glucose monitors (CGMs) and daily hormone testing, researchers have mapped a distinct biphasic pattern of blood glucose levels across the menstrual cycle in non-diabetic individuals. The study reveals that median glucose levels peak during the luteal phase and hit their lowest point during the late-follicular phase, a trend that remains significant even when accounting for exercise, diet, and sleep.

Addressing Data Gaps in Menstrual Research

While it is well-established that hormones like estrogen and progesterone influence insulin sensitivity, previous research on how blood glucose fluctuates across the menstrual cycle has been notoriously inconsistent. Some studies suggested glucose rises in the luteal phase, while others found no change at all.

The authors identify two primary culprits for this inconsistency: sparse data and inaccurate phase mapping. Most prior studies relied on infrequent clinical blood draws and used self-reported "calendar methods" to estimate cycle phases, which fail to account for the high degree of inter- and intra-individual variability in cycle length.

This paper addresses these gaps by using high-frequency, longitudinal data from CGMs and objective daily hormone measurements to provide a high-resolution look at glucose dynamics in healthy, non-diabetic participants.

Methodology and Data Collection

The study employed a multimodal data collection strategy over an average of 79 days per participant, involving 49 non-diabetic individuals (ages 18–30) in Canada. The resulting dataset includes 149 complete menstrual cycles and 554 individual cycle phases.

  • Hardware: Participants wore a Dexcom G6 CGM (measuring glucose every 5 minutes) and a Fitbit Sense (tracking steps and heart rate).
  • Hormone Tracking: To ensure precise labeling, the study used the Mira Plus Starter Kit to measure daily urinary levels of Luteinizing Hormone (LH) and Estrone-3-glucuronide (E3G, a metabolite of estrogen). This allowed for the accurate identification of four phases: Menstrual, Late-follicular, Ovulation, and Luteal.
Distribution of menstrual cycle phases relative to cycle progress.
  • Statistical Modeling: The team used periodic restricted cubic splines to model the non-linear, cyclic relationship between time-in-cycle and glucose. They also employed linear mixed-effects models to account for both fixed effects (like age and BMI) and random effects (individual baseline variations).

Key Insights: The Biphasic Wave

The study confirmed that blood glucose follows a predictable wave rather than remaining static. It is lowest in the late-follicular phase ($5.8 \pm 0.7$ mmol/L) and highest in the luteal phase ($6.1 \pm 0.8$ mmol/L).

Median daily glucose levels alongside LH and estrogen (E3G) concentrations across the menstrual cycle.

Estrogen as a Regulator

The study found a statistically significant negative association between estrogen levels and glucose. As estrogen peaks (around ovulation), glucose levels tend to drop, supporting the theory that estrogen enhances insulin sensitivity. Relative to the late-follicular phase (the baseline), glucose levels were significantly higher in the ovulation ($\beta = 0.08, p < 0.05$) and luteal ($\beta = 0.16, p < 0.001$) phases.

Lifestyle and Behavioral Factors

Interestingly, higher food cravings were positively associated with higher median glucose levels. This suggests that the "luteal peak" in glucose might be driven both by hormonal insulin resistance and behavioral changes, such as increased caloric intake. However, the cyclic pattern of glucose persisted even after the models were adjusted for age, BMI, step count, and self-reported fatigue, suggesting the menstrual cycle is an independent driver of glucose variance.

Implications and Future Directions

The researchers suggest that blood glucose could eventually serve as a "digital biomarker" for fertility tracking. Furthermore, they highlight the need for similar high-resolution studies in Type 1 diabetics, who may need phase-specific insulin dosing strategies to avoid hyperglycemia during the luteal phase.

While the study provides robust data, the authors acknowledge the homogeneity of the cohort (young, healthy, low BMI). Future work is needed to determine if these trends hold for individuals with PCOS, endometriosis, or those in perimenopause.

Summary updated

BibTeX
@article{lin2023blood,
  author = {Georgianna Lin and Rumsha Siddiqui and Zixiong Lin and Joanna M. Blodgett and Shwetak N. Patel and Khai N. Truong and Alex Mariakakis},
  title = {{Blood glucose variance measured by continuous glucose monitors across the menstrual cycle}},
  journal = {Npj Digit. Medicine},
  year = {2023},
  doi = {10.1038/s41746-023-00884-x},
}
Figure