Prolongation of the Menstrual Cycle After Receipt of the Primary Series and Booster Doses of mRNA Coronavirus Disease 2019 (COVID-19) Vaccination

Compared with the primary doses of mRNA coronavirus disease 2019 (COVID-19) vaccination, the third booster shot is associated with a greater degree of menstrual cycle prolongation.

Abstract

OBJECTIVE:

To quantitatively evaluate the effect of a booster vaccination dose against coronavirus disease 2019 (COVID-19) on menstrual cycle in a large-scale retrospective cohort study using a menstrual cycle tracking smartphone application (app).

METHODS:

Prospectively or retrospectively recorded data, including the start and finish dates of menstrual cycles, were collected with the app. Detailed data on vaccinations, side effects, and participants' characteristics were retrospectively collected from a questionnaire on the app. For each COVID-19 vaccination shot (first, second, and third), within-individual changes in menstrual cycle length up to the fourth postvaccination cycle were evaluated.

RESULTS:

Among the 7,376 and 6,873 participants who had the first and second COVID-19 vaccine doses in different menstrual cycles, respectively, menstrual cycles immediately after the vaccination (first postvaccination cycles) were an average of 0.22 days (95% CI, 0.06–0.39) and 0.37 days (95% CI, 0.20–0.54) longer than the prevaccination cycle. In contrast, among the 1,672 participants who received the first and second doses in the same cycle, the first postvaccination cycle was an average of 4.21 days (95% CI, 3.69–4.72) longer. The second to fourth postvaccination cycles returned to the level of the prevaccination cycle. However, among the 4,768 participants who had the third COVID-19 vaccine dose, the menstrual cycle immediately after the vaccination was an average of 1.20 days (95% CI, 1.00–1.40) longer, with prolongation of cycles of 0.27 days (95% CI, 0.10–0.44) to 0.41 days (95% CI, 0.22–0.59) persisting from the second to the fourth postvaccination cycle.

CONCLUSION:

The booster shot against COVID-19 may have a greater and longer-lasting effect on menstrual cycles than the primary-series shots. Although the effect size was small, evidence on the side effects of immunization on menstruation should be accumulated.

Several vaccines have been developed in the fight against coronavirus disease 2019 (COVID-19). However, because the acquired immunity wanes over time, a third booster shot is recommended. So far, 84% of men and women in Japan aged 20–59 years have received at least two doses, and 66% have received a third dose.¹ The need for further periodic boosters is still being debated.²

In terms of possible adverse reactions, large-scale self-reported surveillance systems, such as the V-safe and Vaccine Adverse Event Reporting System in the United States^3,4 and the Medicines and Healthcare Products Regulatory Agency in the United Kingdom,⁵ show that a small number of women have reported menstrual irregularities after COVID-19 vaccination.^6,7 Since these reports, many studies have examined whether and to what extent COVID-19 vaccination influences the menstrual cycle.^8–16 Although most descriptive studies are based on self-reported data, several studies using applications (apps) with menstrual cycle data logs have quantitatively analyzed the effect of the primary-series (first and second) shots, revealing slight prolongation in the first menstrual cycle after vaccination, which resolved in subsequent cycles.^8,9,11,13,16 However, these larger studies have focused only on the primary series, and little is known about the effect of the booster shot.^17,18 Moreover, menstrual cyclicity varies among ethnicities,¹⁹ and there have been few reports in Asian women.²⁰

Therefore, we conducted a large-scale retrospective cohort study using the menstrual cycle tracking app Luna Luna to quantitatively investigate whether the booster shot affects menstrual cyclicity and the primary-series shots in the Japanese population.

METHODS

Luna Luna (https://www.mti.co.jp/eng/?page_id=2755) is a smartphone app that is freely available on Android and iOS platforms for providing services supporting menstrual schedule management, ovulation date prediction, and tracking of pill use.²¹ Users of this app can prospectively or retrospectively record the dates of their menstrual period. The app allows users to select one of the following three stages: normal stage, pregnant hoping stage, and pregnant stage. For this study, participants in the normal and pregnant hoping stages were requested to complete a questionnaire in Japanese regarding their baseline characteristics, details of COVID-19 vaccination, and vaccine-related side effects between July 12 and August 12, 2022. The questionnaire focused on age, height, body weight, parity, marital status, household income, educational level, occupation, smoking status, medical history, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection status, COVID-19 vaccination dates, vaccination types, side effects, severity, and use of any emergency contraceptives shortly after vaccination. For participants scheduled for the third vaccination but not vaccinated at the time of initial recruitment, a follow-up questionnaire on their third vaccination was administered between October 1 and October 31, 2022.

The study protocol was approved by the Ethics Committee of the National Center for Child Health and Development (ethical approval No. 1900; approved in May 2022). For deidentification, MTI Ltd anonymized and transferred the app and questionnaire data to the National Center for Child Health and Development for analysis. Participants were included if they confirmed the in-app notifications explaining the outline of the research plan and how data collected from the app would be used, as well as selecting the agreement button in the app, which equated to signing the informed consent form. MTI Ltd and the National Center for Child Health and Development published a press release on this study, accompanied by information on voluntary participation.

Participants who were pregnant or breastfeeding, were using hormonal drugs in the year before the first vaccination, had a history of malignant tumors or leukemia, or were younger than age 18 years were excluded from this analysis. In addition, participants who received vaccinations other than those from Pfizer–BioNTech or Moderna were excluded to investigate the adverse effects of mRNA vaccines on menstruation. Finally, we regarded menstrual cycles more than 90 days or less than 14 days as missing data because of the possibility of inaccurate data entry,²² and participants with missing data for at least two of the three cycles before vaccination were excluded.

The definitions of individual menstrual cycles with reference to vaccination timing are shown in Appendix 1, available online at http://links.lww.com/AOG/D517. Prevaccination cycle refers to the average length of two or three cycles occurring before the first or third dose; the postvaccination cycle of the first, second, or third dose refers to the cycle immediately after each dose. For participants who received the first and second doses in the same cycle, the term postvaccination cycle of the first and second doses was used instead of postvaccination cycle of the first dose and postvaccination cycle of the second dose. The term second to fourth postvaccination cycles refers to the cycles after the postvaccination cycles of either the primary-series shots or the third dose.

The severity of vaccine-related side effects was categorized according to the definitions provided by the Centers for Disease Control and Prevention²³: Fever was categorized as no fever, below 38.0°C, 38.0–38.9°C, 39.0–39.9°C, and 40.0°C or higher; systemic and local side effects were categorized as mild (does not interfere with activity), moderate (interferes with activity), and severe (prevents daily activity); and some local side effects, including swelling and redness at the vaccination site, were also classified into mild (larger than 2.0–5.0 cm), moderate (larger than 5.0–10.0 cm), and severe (larger than 10.0 cm). Age was categorized into 18–24, 25–34, 35–44, or 45 years or older; body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) as lower than 18, 18–24, 25–30, or higher than 30; household income as less than 2 million, 2–5 million, 6–9 million, or 10 million yen or more (for reference, the average household income in Japan is 5.45 million yen²⁴); and educational level as high school degree or less, junior college or vocational school, or college degree or higher. Participants were given the option to select “unwilling to answer” or “do not know” for every question. Vaccination timing was defined according to the inoculation phase in the menstrual cycle as follows: luteal phase (14 days before the first day of the subsequent menstrual cycle) and follicular phase (the remaining days between the first bleeding day and the last day before the luteal phase), as described in previous studies.^11,13 Missingness was considered a category for all variables in this study.

The primary outcomes were within-individual changes in the menstrual cycle length of each postvaccination cycle compared with the prevaccination cycle. The secondary outcomes were within-individual changes in the menstrual cycle length stratified by variables to investigate the background characteristics predisposing patients to menstrual abnormalities.

The sociodemographic characteristics and medical histories of participants who received the first, second, and third doses are described. For each vaccine, we observed within-individual changes in menstrual cycle length by comparing the mean and SD using paired t tests and reported estimated changes in length with 95% CIs. We used conditional logistic regression to estimate the odds ratios (ORs) of having a longer postvaccination cycle (more than 38 days) than a prevaccination cycle based on the International Federation of Gynecology and Obstetrics definition of infrequent menstruation.²⁵ Third, regression analyses adjusted for age, BMI, parity, marital status, education level, household income, occupation, and smoking status were used to calculate the adjusted within-individual changes in menstrual cycle length and how these changes differed with participant characteristics. In the sensitivity analyses, participants who used emergency contraceptives after COVID-19 vaccination and those with a medical history of menstrual irregularity were excluded to eliminate the influence of these conditions on our findings. An additional analysis of participants without a history of SARS-CoV-2 infection was performed to eliminate any bias attributable to previous SARS-CoV-2 infections. Stata SE 14 was used for statistical analysis, and P<.05 was considered statistically significant.

RESULTS

Overall, 22,509 users participated in this study (Fig. 1): 9,048 who had the first dose of the COVID-vaccine, 8,545 who had the primary series (first and second doses), and 4,768 who had the booster shot were eligible. Most participants who completed the booster vaccine were aged 25–34 years (42.0%), had normal-range BMI (70.2%), were nulliparous (80.3%), were unmarried (63.4%), had an average household income of 2–5 million yen (39.3%), had college degrees or higher (41.4%), were current workers (77.6%), were nonsmokers (87.3%), and had no medical history of menstrual irregularity (87.4%) and no previous history of SARS-CoV-2 infection at recruitment (87.1%) (Table 1). Approximately two-thirds of the participants received the Pfizer–BioNTech vaccine for the first and second doses (68.0% and 67.4%), and more than half received the Moderna vaccine as the booster (53.0%). The median intervaccine interval was 24 days between the first and second doses and 212 days between the second and third doses. Regarding individual adverse reactions, fewer than half of participants (47.7%) had a fever after the first dose, whereas about two-thirds of participants had a fever after the second and third doses (68.6% and 68.9%, respectively). In addition, more severe symptoms were observed with the second and third doses (14.6% and 13.5%, respectively) than with the first dose (9.4%). Detailed systemic side effects are shown in Appendix 2, available online at http://links.lww.com/AOG/D517. Throughout all vaccinations, almost half of the participants were vaccinated in their follicular phase. Fewer than 5% of the participants used emergency contraceptives after each vaccination. Regarding the per-dose analysis, among participants with primary-series shots in the same cycle, most (81.8%) received the Pfizer–BioNTech vaccine, and the median intervaccine interval was shorter than in those who received a single dose per cycle (21 and 28 days, respectively), possibly because of the difference in inoculation interval between Pfizer–BioNTech (3 weeks) and Moderna (4 weeks) (Appendix 3, available online at http://links.lww.com/AOG/D517).

Fig. 1. Flowchart for eligibility. COVID-19, coronavirus disease 2019.

Hosoya. COVID-19 Vaccination and Menstrual Cycle Length. Obstet Gynecol 2024.

Table 1.

Characteristic and Vaccination Details of the Participants

For the primary-series shots, the average change in the length of the postvaccination cycle for the first dose and second dose was 0.22 days (95% CI, 0.06–0.39) and 0.37 days (95% CI, 0.20–0.54) (unadjusted change in cycle length), respectively, for participants who received the shots in different cycles. Participants who received the primary-series doses within the same cycle had an average change of 4.21 days (95% CI, 3.69–4.72) (Table 2). Overall, the first postvaccination cycle of the primary series showed a 1.12-day (95% CI, 0.95–1.29) increase compared with the length of prevaccination cycles; however, the increases were resolved in the second, third, and fourth postvaccination cycles (−0.04 days [95% CI, −0.19 to 0.12], −0.06 days [95% CI, −0.21 to 0.09], and −0.15 days [95% CI, −0.30 to 0.00], respectively) (Table 3).

Table 2.

Within-Individual Changes in Menstrual Cycle Length of the First Postvaccination Cycle After Each Dose

Table 3.

Within-Individual Changes in Menstrual Cycle Length of the Postvaccination Cycles After the Primary Series and Third Dose Compared With the Prevaccination Cycle

For the third dose, the vaccinated individuals experienced an average of 1.20-day (95% CI, 1.00–1.40) increase in the length of the postvaccination cycle of the third dose (Table 2). Although the prolongation weaned over time, statistically significant increases in length compared with the prevaccination cycles persisted in the second, third, and fourth postvaccination cycles (0.27 days [95% CI, 0.10–0.44], 0.27 days [95% CI, 0.10–0.45], and 0.41 days [95% CI, 0.22–0.59], respectively) (Table 3).

The OR of the increase in the proportion of participants with menstrual cycle length longer than 38 days in the postvaccination cycle compared with the prevaccination cycle showed similar differences by dose (Appendix 4, available online at http://links.lww.com/AOG/D517). For the primary series, the first postvaccination cycle showed an OR of 1.79 (95% CI, 1.59–2.02); however, no significant differences were observed for the second, third, and fourth postvaccination cycles (OR 1.09 [95% CI, 0.97–1.24], 1.04 [95% CI, 0.92–1.17], and 1.03 [95% CI, 0.91–1.17], respectively). For the third dose, the first postvaccination cycle showed an OR of 2.85 (95% CI, 2.35–3.47), and significant differences persisted for the second, third, and fourth postvaccination cycles (OR 1.62 [95% CI, 1.34–1.96], 1.68 [95% CI, 1.38–2.04], and 1.78 [95% CI, 1.45–2.19], respectively). The distribution of menstrual cycle lengths before and after vaccination with the primary series and the third dose is shown in Appendices 5 and 6, available online at http://links.lww.com/AOG/D517.

Restricted analysis of participants without emergency contraceptive use (Appendix 7, available online at http://links.lww.com/AOG/D517), medical history causing menstrual irregularity (Appendix 8, available online at http://links.lww.com/AOG/D517), or past SARS-CoV-2 infection at recruitment (Appendix 9, available online at http://links.lww.com/AOG/D517) generated similar results of a 1-day increase in the postvaccination cycle of the primary series and third dose, with a statistically significant increase in the subsequent postvaccination cycles persisting only after the third dose. No change was observed in the duration of menses between the primary series and the third dose (Appendix 10, available online at http://links.lww.com/AOG/D517).

According to the participant's characteristics (Table 4), perimenopausal participants aged 45–57 years experienced longer increases than participants aged 25–34 years for both the primary-series shots and third doses (adjusted difference in mean of change in menstrual cycle length: 1.00 day [95% CI, 0.27–1.73] and 1.01 days [95% CI, 0.20–1.83], respectively). No statistically significant changes were observed for BMI, parity, and smoking status. Similarly, no statistically significant differences were observed in the association between changes in menstrual cycle length and detailed medical comorbidities, except in participants with thyroid diseases and longer cycles in the third dose cycle (Appendix 11, available online at http://links.lww.com/AOG/D517). Regarding vaccination-related factors, the average changes were longer for participants who experienced a higher body temperature for the primary series, although a significant difference was observed only in those with a fever of 39.0–39.9°C compared with those with no fever (adjusted difference in mean of change in cycle length 0.67 days, 95% CI, 0.03–1.31).

Table 4.

Multivariate Analysis of the Within-Individual Changes in Menstrual Cycle Length of the First Postvaccination Cycles After the Primary Series and Third Dose Compared With the Prevaccination Cycle

Moreover, regarding vaccination timing, participants vaccinated during the follicular phase showed statistically significant prolongation after the second and third doses (adjusted mean of change in cycle length 2.24 days [95% CI, 2.00–2.48] and 2.24 days [95% CI, 1.97–2.51], respectively), whereas those vaccinated during the luteal phase showed no menstrual prolongation. However, from the perspective of the per-dose analysis, participants who received the second vaccination of one dose per cycle in the luteal phase showed statistically significant menstrual cycle shortening (adjusted mean of change in cycle length −1.39 days, 95% CI, −1.66 to −1.12) (Appendix 12, available online at http://links.lww.com/AOG/D517). Among participants who received their primary vaccines in the same cycle, those who received two doses during the follicular phase (n=199) had the longest prolongation (adjusted mean of change in cycle length: 17.40 days, 95% CI, 16.06–18.74). In addition, their mean prevaccination cycle length for the first dose was also longer (mean±SD 36.62±8.79 days) compared with those of participants vaccinated during different phases of their cycle (n=1,456) and those vaccinated during the luteal phase (n=17) (31.98±5.75 and 30.26±3.67 days, respectively) (data not shown in table). No statistically significant changes in menstrual cycle length were observed in the severity of side effects or vaccination type for either the primary series or the third dose (Table 4).

DISCUSSION

This study included about 10,000 participants and provided several novel insights, notably a greater degree of menstrual cycle prolongation after the booster dose than after the first and second doses when administered in different cycles and slight prolongation persisting in consecutive cycles after the booster shot. In addition, the small transient menstrual prolongation after COVID-19 vaccinations is similar in our Japanese population compared with prior studies in other populations.

Several large-scale app-based studies have also shown that the primary series of COVID-19 mRNA vaccinations cause slight transient menstrual prolongation.^8,9,11,16 Furthermore, participants who received the primary series in the same cycle also had a greater degree of prolongation than those who received them separately,^8,9 and perimenopausal age has been identified as a risk factor for such prolongation.²⁶ Moreover, in two previous studies, vaccination during the follicular phase caused longer menstrual prolongation than during the luteal phase, similar to our findings.^11,13 Thus, COVID-19 vaccinations can cause menstrual prolongation, possibly attributable to a delay in ovulation, when administered during the follicular phase. In addition, its negative effects may be maximized with the primary dose administration in the same follicular phase, although such a population originally showed a longer prevaccination cycle and might have underlying backgrounds with ovulation delay. Our study and a previous report¹¹ indicated that administering the second dose in the luteal phase induced menstrual cycle shortening, suggesting luteal function insufficiency. Although the third dose did not affect menstrual cyclicity when administered during the luteal phase in our study, future studies using more accurate ovulation and luteal function data, such as basal body temperature, should be considered.

Because mRNA vaccines induce a robust immune response²⁷ and women may generally have a higher level of immune response to mRNA vaccines than men,²⁸ various inflammatory cytokines might affect ovulation and hormonal production through the disruption of the hypothalamus–pituitary–ovarian axis, resulting in menstrual irregularity.^29–31 Similarly, because some non-mRNA vaccines, including virus–vector or recombinant protein vaccines, induce menstrual abnormalities,^11,32 the immune response may be associated with irregular menstruation. In this study, participants with higher body temperatures had pronounced cycle prolongation after the primary doses. Because a higher body temperature implies a stronger inflammatory response,³³ this observation is consistent with the above hypothesis.

It is interesting that, although the general adverse reactions after the third vaccination were not different from those in the second vaccination, the degree of menstrual cycle prolongation was greater, and some prolongation persisted in the subsequent cycles. Because the prolongation differences attributable to side effects in the primary series were not found in the booster dose, a booster dose may have a different immunization mechanism, causing menstrual irregularities. Although the World Health Organization withdrew its recommendation for booster vaccination for low-risk healthy adults in March 2023, it is still suggested for health care workers and groups at high risk²; thus, future studies on how and why these shots disturb menstrual cycles are required. Moreover, although the COVID-19 vaccination effect on fertility has been a clinical concern, its negative effect on fertility outcomes has been denied.^34–36 However, if vaccinations could slightly delay ovulation or disrupt luteal function, clinics should consider vaccination timing when scheduling infertility treatments.

This study had several limitations. First, self-reported app data may be inaccurate when participants retrospectively entered their menstrual dates or erroneously identified abnormal bleeding as a normal menstrual cycle. However, given the purpose of this app, because most users use it to manage and predict their menstrual periods, there is no advantage for the user to enter a false menstrual period date. Moreover, considering that logs collected through smartphone apps can be the most reliable big data on menstruation in the general population³¹ and because we excluded participants with unreliable logs, the influence of inaccuracy is negligible. Second, the retrospective data collection using questionnaire induces a possible recall bias; however, the proportions of vaccination-related side effects were almost similar to those of the official Japanese data.³⁷ Third, the seasonal effect on menstruation was not considered because a previous study using the Luna Luna showed that seasons had negligible effects on menstrual cycles.³⁸ However, other confounders may exist, such as menstrual natural cyclicity^19,39 or self-selection bias attributable to participants with a longer prevaccination cycle and menstrual cycle prolongation after the primary-series doses skipping the booster dose (Appendix 13, available online at http://links.lww.com/AOG/D517). Finally, the result of menstrual prolongation after third dose results may not be generalizable to other populations.

In conclusion, the booster shot against COVID-19 may have a greater and longer-lasting effect on menstrual cycle than the primary-series shots. Although the effect size was small and the number of booster shots required for COVID-19 prevention is debatable, evidence on the side effects of immunization on menstruation should be accumulated.