Your BMI could be affecting your menstrual cycle—researchers explain how

By Tarun Sai LomteReviewed by Susha Cheriyedath, M.Sc.Mar 2 2025

New data from 8,700 women shows how underweight and obesity disrupt cycles—and why hitting a BMI sweet spot of 20 could boost fertility and ovulation.

Study: Body mass index and menstrual irregularity in a prospective cohort study of smartphone application users. Image Credit: FotoDuets / Shutterstock

In a recent study published in the journal npj Women’s Health, researchers investigated the relationship between body mass index (BMI) and menstrual irregularities.

The menstrual cycle characteristics are among the most accessible indicators for female health assessment. Regular menstruation depends on a functional hypothalamic-pituitary-ovarian (HPO) axis, whose disruption could lead to abnormalities, such as anovulation, amenorrhea, and irregular menstrual cycles. Research suggests that individuals with extreme BMIs are at a higher risk of menstrual irregularities.

Further, high BMI and obesity are frequently associated with infertility, primarily due to hormonal and metabolic disruptions affecting ovulation. Besides, underweight individuals are reportedly at a higher risk of ovulatory infertility, albeit studies have been limited to specific populations, such as female athletes and those with anorexia nervosa. Moreover, existing studies on menstrual irregularities and BMI extremities are inconsistent.

About the Study

The present study investigated the relationship between menstrual cycle irregularities and BMI. The study used a period-tracking mobile application (LunaLuna) in Japan. Menstrual data were collected from January 2019 to March 2021 from app users who agreed to participate. Participants completed bimonthly questionnaires, which collected information on health conditions, lifestyle, education, and employment.

The first questionnaire (wave 1) was administered from January 23 to March 25, 2020, and the second (wave 2) was administered from May 14 to June 14, 2020. Respondents with data from both waves were included in the analyses. Participants were excluded if they had a BMI <15 kg/m² or >35 kg/m², were pregnant, undergoing infertility treatment, using hormonal contraceptives or IUD/IUS, or logged fewer than three menstrual cycles. Additionally, extreme outlier cycles (>4 standard deviations from the mean cycle length) were removed to ensure data quality.

The primary outcomes were cycle length (CL), the proportion of persons with menstrual irregularities, and the proportion of biphasic cycles.

Menstrual irregularity indicators were infrequent menstrual bleeding (IMB, defined as CL ≥ 39 days but < 90 days) and absent menstrual bleeding (AMB, defined as CL ≥ 90 days, per FIGO guidelines). For basal body temperature (BBT) analysis, the luteal phase was defined as the 10 days up to the day before the next menstruation, while the follicular phase was defined as the first 10 days from the start of the menstrual cycle.

A cycle was considered biphasic if the difference in the mean luteal phase BBT from follicular phase BBT was greater than 0.3°C, an indicator of ovulatory function. The relationship between BMI and outcomes was assessed using a restricted cubic spline model, with linear regression for continuous outcomes, logistic regression for binary outcomes, and Poisson regression for proportions (of biphasic cycles).

Findings

In total, 10,465 individuals completed both questionnaires (waves 1 and 2). Each subject logged 21 cycles on average. They were excluded if they had a BMI < 15 kg/m², fewer than three cycles, or were pregnant, under infertility treatment, or on hormonal contraceptives. Following exclusions, 8,745 participants with 191,426 menstrual cycles were included. BBT data were collected from 3,221 participants, with 15,883 cycles.

Overall, participants were in their late 20s or early 30s and were unmarried, employed, and non-smokers. Using Asian-specific BMI categories, around 14% of participants were underweight (15–18.4 kg/m²), 59% had normal BMI (18.5–22.9 kg/m²), 13% were overweight (23–24.9 kg/m²), and 14% were obese (25–35 kg/m²). The average CL was 31.5 days. Further, 7% of subjects had IMB, and 4% had AMB. Participants with a BMI of 20 kg/m² had the lowest mean CL, while those with a BMI of ≤16 kg/m² or ≥30 kg/m² had significantly longer cycles (+1.03 and +1.06 days, respectively).

Overweight and obese individuals had a higher IMB risk than those with normal BMI (Overweight: OR 1.56; Obese: OR 2.63). Further, underweight and obese individuals showed greater AMB risks than those with normal BMI (Underweight: OR 1.78; Obese: OR 1.94). A J-shaped relationship between BMI with IMB and AMB indicated that higher and lower BMIs increased the prevalence of IMB and AMB.

In addition, individuals with a BMI ≤18 kg/m² or ≥21 kg/m² had a higher IMB risk compared to those with a BMI of 20 kg/m². Similarly, AMB risk was significantly higher in individuals with a BMI ≤19 kg/m² or ≥26 kg/m².

Moreover, there was an inverted J-shaped relationship between BMI and the proportion of biphasic cycles, suggesting that ovulatory function was optimal at BMI 20 kg/m² but decreased at both higher and lower BMI values. This finding highlights that both overweight and underweight individuals may be at an increased risk of anovulation.

Conclusions

The findings illustrate a nonlinear relationship between BMI and menstrual cycle characteristics. Individuals with a low or high BMI showed elevated risks of longer, irregular menstrual cycles. Further, an inverted J-shaped relationship for the proportion of biphasic cycles confirmed that both high and low BMI increased the risk of non-ovulatory cycles, potentially leading to ovulatory infertility.

Overall, individuals with normal BMI (particularly those around 20 kg/m²) had the least risk of irregular, non-ovulatory menstrual cycles, highlighting the benefits of maintaining a normal BMI for reproductive health.

However, the study was conducted exclusively in Japan, and the authors note that BMI-adiposity relationships differ across ethnic groups. These findings may not be directly generalizable to non-Asian populations. Additionally, potential selection bias should be considered, as app users may not fully represent the general population. The reliance on self-reported BMI data is another limitation that should be acknowledged.