This study was designed as a population-based retrospective cohort study using the KNHIS database. We investigated the relationship between reproductive factors and the risk of migraine in premenopausal women. Our study’s main finding was that a higher parity number, a longer duration of breastfeeding, and the use of OCs were related to a higher migraine risk in premenopausal women.
Parity and migraine
During pregnancy, elevated serum estrogen levels that rarely fluctuate may be associated with migraine improvement 4. The serum estrogen level rises during the sixth and seventh weeks of pregnancy and gradually increases to its highest level during the third trimester 4. Previous studies have reported that migraine activity tended to improve during pregnancy, particularly in the second and third trimesters 12–14. A prospective study found that migraine was improved by 46.8% during the first trimester, 83.0% during the second trimester, and 87.2% during the third trimester 15.
The effect of parity on the course of headaches during pregnancy among migraineur women is still inconclusive. Two large prospective studies have reported that parity did not influence migraine activity during pregnancy 16,17. Another prospective study found a lower migraine prevalence among pregnant women that was significant for nulliparous but not for primiparous or multiparous women 7. However, the effects of childbirth on migraine in nonpregnant women have rarely been studied.
Although the mechanism of the association between migraine and parity remains unclear, it is thought that a patient’s hormonal state may play an important role. A longer lifetime number of years of menstruation (LNYM) increases the risk of breast and endometrial cancer, while more experiences of childbirth reduce this risk 18–21. The LNYM is the total duration of menstruation (excluding pregnancy and breastfeeding), which reflects the cumulative exposure to endogenous estrogen in women 18. There is a significant inverse relationship between LNYM and sex hormone-binding globulin (SHBG). Low SHBG induces serum estrogen to leave the circulation and enter cells, resulting in increased endogenous estrogen exposure. SHBG is increased according to a woman’s parity number 19. A meta-analysis reported a significant inverse association between migraine and breast cancer risk 22. Previously, we concluded that a shorter LNYM increases the migraine prevalence in postmenopausal women 11. Our present study indicated that higher migraine risk was associated with the experience of childbirth, which is related to a shorter length of endogenous estrogen exposure. These results suggest that shorter endogenous estrogen exposure might have an impact on migraine risk.
Our results indicate that the effects of parity on migraine were not significant after 55 years, and there were no differences between primiparous and multiparous women. These results suggest that the effect of childbirth on migraine may decrease significantly with increasing age. Considering that the average age at menopause in Korea was reported as 49.3 ± 3.5 years, women older than 55 years might experience a relatively longer menstruation period over their lifetimes 23. This longer endogenous estrogen exposure may influence the outcome, but the exact mechanism is unclear. In a previous study, the incidence of migraine in nonpregnant women was lower in nulliparous women than in women with children, but there was no statistical significance after the age of 40, which was similar to the finding of our study 7.
Women with children may have more parenting stress than nulliparous women, which can aggravate migraine 24,25. However, our study did not investigate the relationship between stress factors and migraine risk. Therefore, a cautious interpretation of our results is required, and further studies on other factors are needed to draw more definitive conclusions.
Breastfeeding and migraine
Breastfeeding inhibits ovulation and maintains low estrogen levels with little fluctuation 4. Although breastfeeding affects migraine activity positively, a link has not yet been clearly established. Some studies have reported that breastfeeding impedes the recurrence of migraine 15,17. Migraine reportedly recurred within the first postpartum month in 100% of women who did not breastfeed and in 43.2% of those who were breastfeeding 15. In contrast, a large population-based study of women who experienced migraine during pregnancy and puerperium reported that breastfeeding did not influence the occurrence of migraine 12. Our result shows that breastfeeding for > 6 months was related to a higher migraine risk, which was significant in patients < 60 years of age. The present study was unique in that we investigated the long-term effect of breastfeeding following childbirth on migraine prevalence in premenopausal women. One partial explanation for this result is that as the lactation period is prolonged, women often experience a decrease in sleep quality due to the need for night feeding and the delay in returning to their regular daily activities. In the 45-year-old group, breastfeeding for < 6 months showed a significant protective effect after adjusting for parity. In Korea, 7.1% of all childbirths occur in women between 40 and 45 years of age, so we were unable to exclude women who had experienced recent childbirth or were currently breastfeeding in this age group 26.
Oral contraceptives and migraine
OCs are frequently used by women for birth control and may exacerbate migraine symptoms 27. A randomized controlled study reported that 70% of migraineurs experience a worsening in migraine during treatment with OCs 28. In migraine patients, headache patterns did not change in 44–67%, became worse in 24–36%, and improved in 5–8% with the use of OC 29–31. A large cross-sectional population study that investigated the effect of OCs on the prevalence of migraine in 13,994 premenopausal women reported that women currently using OCs had a higher prevalence of migraine (OR: 1.4; 95% CI: 1.2–1.7) than the group that had never used OCs 9. Previously, we reported that HRT increases the risk of migraine in postmenopausal women 11. These results suggest that the hormonal factors that cause an interruption in the normal menstruation cycle may also increase the migraine risk. The present study showed that all OC use groups had a higher risk than the non-OC use group, and there was a tendency for the risk to increase according to the period of use. Considering the limited data available on the exact type, dose, and persistence of OC usage, the results of this study need to be carefully interpreted.
This large population-based study analyzed information from the linked KNHIS and KHE databases. As a result, there were some limitations to our study. We enrolled only migraine participants who were currently taking migraine-specific abortive and preventive medications and had ICD-10 codes in the KNHIS database. Therefore, it is possible that our study did not include migraineurs who only take simple analgesics to treat their migraines. We obtained the patients’ clinical information from the KHE database, which has some limitations. KHE data is limited to people 40 years of age or older, so our participants may have included not only premenopausal but also perimenopausal women. The KHE questionnaire also did not account for recent childbirth, current breastfeeding, parenting stress, and current use of OCs. In addition, we could not determine the exact type of OCs that were being used. However, only combined OCs are available in Korea, so our result reflects the influence of combined OCs.
In conclusion, our study showed that the reproductive factors that induce an interruption in the normal menstrual cycle may also influence migraine risk in premenopausal women.