Abstract
Objective:
The aim of this study was to evaluate the association of oral menopausal hormone therapy (estrogen only [E-only] or estrogen and progestin [E+P]), as compared to placebo on postmenopausal headache severity among females with and without a history of migraine in the Women’s Health Initiative Hormone Therapy (WHI-HT) trials.
Background:
Migraine is three times more common among females than males and is a major source of pain and disability among women throughout the life course. Previous observational research found a positive association between menopausal hormone therapy (MHT) use and migraine prevalence, but the longitudinal association of MHT with postmenopausal migraine severity is unclear.
Methods:
We examined changes in self-reported headache severity between baseline and year 1 in 22,876 females (average age 64 years) across the United States enrolled in the WHI-HT trials between 1993 and 1998. We conducted a secondary, stratified analysis of data of two parallel, randomized, placebo-controlled clinical trials to estimate the association of MHT and headache severity overall and stratified by history of migraine at baseline. Additionally, we assessed the association between MHT and headache trajectory, based on whether headaches worsened over the follow-up period.
Results:
At baseline, the prevalence of a lifetime migraine diagnosis was 10%. Among participants with a history of migraine, randomization to E-only MHT was not associated with more severe headache at 1 year (adjusted odds ratio [aOR] = 1.14, 95% confidence interval [CI]: 0.90–1.44), or worsening headache trajectory (adjusted risk ratio [aRR] = 0.98, 95% CI: 0.71–1.34) after 1 year. Randomization to E+P MHT was modestly associated with more severe headache, although estimates did not reach statistical significance (aOR = 1.21, 95% CI: 0.98–1.50). E+P MHT was significantly associated with worsening headache trajectory (aRR= 1.53, 95% CI: 1.14–2.03). Among those without a migraine history, E-only MHT was not strongly associated with moderate-to-severe headache severity (aOR: 1.11, 95% CI: 1.01–1.21) or worsening headache trajectory (aRR = 1.08 95% CI: 0.96–1.22) after 1 year. However, E+P MHT was associated with modestly increased odds of moderate-to-severe headache severity (aOR = 1.14, 95% CI: 1.06–1.23) and a higher likelihood of worsening headache trajectory (aRR = 1.18, 95% CI: 1.07–1.30) after 1 year.
Conclusions:
In this study of MHT in the WHI hormone therapy trial participants, E-only MHT was not associated with increased odds of more severe postmenopausal headache severity. Among those with a history of migraine, E+P MHT was associated with more severe and worsening postmenopausal headache. Further research on the effect of newer MHT formulations on postmenopausal headache is warranted.
Keywords: epidemiology, headache, hormone therapy, migraine, randomized control trial
Plain Language Summary
Menopausal hormone therapy is widely used, but its impact on headache severity, especially among those with a history of migraine, is unclear. We analyzed data from randomized trials and found that estrogen-only therapy was not linked to worse headaches, while combined estrogen plus progestin therapy was associated with worsening headaches, particularly among women with prior migraine. These findings suggest that menopausal hormone therapy types differ in their association with headache outcomes.
INTRODUCTION
Approximately one-third of females will experience migraine in their lifetimes, resulting in significant pain, disability, and socioeconomic impact.1,2 There is evidence of a link between endogenous and exogenous hormones and migraine throughout the female life course, but the nature of these associations is not well established.3–5 Around the time of menarche, the prevalence of migraine begins to sharply diverge by sex, eventually becoming two to three times greater in females than males.3,6 Other reproductive events that cause fluctuation of ovarian hormone levels, such as menstruation, oral contraception use, and pregnancy have been shown to impact the clinical presentation of migraine.6 Substantial fluctuations in estrogen and progesterone production during perimenopause have been associated with a higher prevalence and intensity of migraine attacks.6
While some studies have shown a decline in past-year migraine prevalence in the decades following menopause, others have shown consistent migraine prevalence across pre-, peri-, and postmenopausal periods.7–10 In addition, the clinical presentation of migraine after menopause varies, with substantial proportions of individuals reporting improvement (8%–36%), worsening (9%–42%), or no change (37%–64%) in migraine.4 The association of menopause with migraine remains inconclusive, and is complicated by heterogeneity in migraine presentation, menopause type, and use of menopausal hormone therapy (MHT).3,7–9
Previous research on the association between MHT and headache or migraine remains limited and conflicting. A systematic review of migraine and MHT found evidence of differing patterns of migraine (improvement, no change, worsening) following menopause.7,11 Data from the Women’s Health Study showed an increased prevalence of migraine among MHT users compared to nonusers.11,12 However, as MHT was not randomized and the analysis was cross-sectional, the study was unable to account for confounding by indication, i.e., whether individuals with migraine were more or less likely to be prescribed or take MHT. A registry-based study in Korea found an increased risk of migraine with MHT, regardless of duration, but was limited to spontaneous menopause and migraine recorded in the medical record and potentially confounded by indication due to perimenopausal migraine experiences.6
Our research builds upon previous literature by examining a large cohort of postmenopausal females with and without a history of hysterectomy who were randomly assigned to MHT or placebo. We examined the association of MHT with headache severity after 1 year of MHT in postmenopausal individuals, overall and stratified by migraine history at baseline. We hypothesized that randomization to MHT would be associated with greater headache severity and higher odds of worsening migraine trajectory over the one-year follow-up.
METHODS
Study population
We conducted secondary, stratified analyses of data from two previous parallel, randomized, placebo-controlled clinical trials, although the specific hypothesis of this manuscript was developed a priori. We analyzed participants in the Women’s Health Initiative Hormone Therapy (WHI-HT) trials, which were designed to determine the risks and benefits of MHT for postmenopausal women. As this was a secondary analysis of existing data, no statistical power calculation for MHT and headache severity was completed. The sample size was based on the available data (eligible participants originally enrolled in the WHI-HT trials).
The WHI study design has been described in detail elsewhere.13,14 Briefly, the WHI-HT began in 1993 and enrolled 27,347 postmenopausal individuals between the ages of 50 and 79 years at 40 WHI clinical centers across the United States, 24,619 of whom had complete data on migraine status at baseline. Our analytic sample included 22,876 participants with complete data on headache severity at screening and 1-year follow-up (Figure 1). The WHI project was reviewed and approved by the Fred Hutchinson Cancer Research Center (Fred Hutch) Institutional Review Board (IRB) in accordance with the U.S. Department of Health and Human Services regulations at 45 CFR 46 (approval number: IR# 3467-EXT). Participants provided written informed consent to participate. Additional consent to review medical records was obtained through signed written consent. Fred Hutch has an approved Federalwide Assurance (FWA) on file with the Office for Human Research Protections (OHRP) under assurance number 0001920. The WHI ClinicalTrials.gov ID is NCT00000611.
FIGURE 1.
Women’s health Initiative Hormone Therapy trial enrollment flow chart.
Exposure assessment
At screening, individuals with intact uteri were randomly assigned with a 1:1 ratio to either estrogen plus progestin (E+P) oral hormone therapy (0.625 mg/day conjugated equine estrogens plus 2.5 mg/day medroxyprogesterone acetate) or placebo, while those who had undergone a hysterectomy were randomly assigned with a 1:1 ratio to estrogen only oral hormone therapy (E-only) or placebo. Due to baseline differences between the E+P and E-only trials, each MHT group was analyzed separately. E+P participants were compared to those receiving E+P placebo, and E-only participants were compared to those receiving E-only placebo.
Outcome assessment
Migraine history and headache severity were self-reported on WHI questionnaires. Participants reported their migraine history and headache status at baseline and 1-year follow-up. At baseline, participants were provided with a list of medical conditions and asked, “Has a doctor told you that you have any of the following conditions or have you had any of the following procedures? (Please mark all that apply)”. Participants marking migraine were considered to have a history of migraine status. At baseline and 1-year after follow-up, participants were provided with a list of symptoms and asked to mark how bothersome their symptoms were during the past 4 weeks. Participants indicated their level of “headaches or migraines” as “symptom did not occur,” “symptom was mild,” “symptom was moderate,” or “symptom was severe” at baseline and 1 year after randomization.
Covariate assessment
We stratified analyses by migraine status reported at baseline to compare a subgroup of individuals with and without migraine and isolate headaches which were more likely to be migraine. Participants self-reported whether they had ever been diagnosed with migraine by a healthcare provider on a form assessing chronic conditions and medical history. We also collected baseline information on education level, income, body mass index, region of residence, race and ethnicity, sleep quality, and smoking, as these factors may be related to headache severity.9,15,16
Statistical analysis
We first conducted descriptive analyses to characterize the population allocated to each type of MHT and placebo, compare baseline characteristics of individuals with and without a history of migraine, and examine headache severity at baseline across MHT arms and migraine history. As age at screening and age at menopause were found to be normally distributed (Kolmogorov–Smirnov p > .05), we calculated the mean and standard deviation for these variables. The distribution of categorical variables was examined using counts with proportions.
Using an “intent-to-treat” analysis, we fit separate four-level ordinal logistic regression models to estimate the association of MHT with higher headache severity category at baseline and year 1. While our data violated the proportional odds assumption, likely due to stronger effects in the higher severity categories, our findings were robust when we dichotomized the outcome to moderate-to-severe headache as compared to none vs. mild. We also fit binary regression models to estimate the association of MHT baseline to year 1 (headache worsened compared with remained unchanged or improved and headache improved compared with remained unchanged or worsened). We tested for multiplicative statistical interaction between baseline migraine and MHT on year 1 headache severity and trajectory using multiplicative interaction terms. Models converged without warnings and adequate cell counts were confirmed for all joint exposure categories used in interaction analyses. As each participant contributed one follow-up headache measurement to the data set, independence of observations was assumed. In analyses of headache trajectory, we adjusted for baseline headache severity to increase statistical power and provide a more precise estimate of the treatment effect.
Given previous findings, we were also interested in the association between baseline migraine history and headache trajectory over follow-up. We conducted bivariate analyses to describe headache trajectory among those with and without migraine and log-binomial models to estimate the association between baseline migraine status and risk of worsening headache over 1 year of follow-up. As migraine history was not randomly assigned, we adjusted for MHT arm, age category (50–54, 55–59, 60–69, 70–79 years), education (less than high school or equivalent, high school or equivalent, vocational/training, some college/associate’s, college graduate/baccalaureate, some post-graduate or professional, master’s degree, doctoral degree), geographic region (Northeast, South, Midwest, West), race (American Indian/Alaska Native, Asian, Native Hawaiian/other Pacific Islander, Black, white, more than one race, unknown/not reported), and sleep quality (very restless, restless, average, sound or restful, very sound or restful), as these factors may be associated with self-reported migraine diagnosis at baseline and headache severity at follow-up. Covariates were modeled categorically to avoid assumptions of linearity.
We analyzed data using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). A two-sided alpha level of 0.05 was considered statistically significant for all analyses. There were no observations with missing exposure or outcome information in the analytic sample. Missing covariate data was low, so participants with missing covariates were not included in final models (complete case analysis). The proportion of missing data for income was 3.1%, the highest of any covariate.
RESULTS
Descriptive results
WHI hormone therapy clinical trial participants (n = 22,876) were an average of 64 years at enrollment, when 38.6% of participants were randomized to E-only MHT or placebo and 61.4% were randomized to E+P MHT or placebo (Table 1) based on hysterectomy status at baseline. Participant characteristics were generally balanced between treatment and placebo groups. Participants in the E-only arm (with a history of hysterectomy) had a younger average age at menopause and tended to have a higher body mass index, live in the Southern United States, identify as Black, and report poorer sleep quality than participants in the E+P arm. These factors may also be associated with increased headache or migraine burden. Indeed, individuals in the E-only arm were slightly more likely than E+P participants to report an ever-diagnosis of migraine at baseline (11.4% vs. 8.5%).
TABLE 1.
Baseline demographic, medical, and lifestyle characteristics of 22,876 Women’s Health Initiative Hormone Therapy trial participants by menopausal hormone therapy status.
| Parameter | Estrogen only users (N = 4408, 19.3%) | Estrogen only control (N = 4417, 19.3%) | Estrogen and progestin users (N = 7054, 30.8%) | Estrogen and progestin control (N = 6997, 30.6%) |
|---|---|---|---|---|
| Age in years, mean (SD) | 63.8 (7.2) | 63.9 (7.3) | 63.5 (7.1) | 63.5 (7.1) |
| Age category, N (%) | ||||
| 50–54 | 531 (12.1) | 533 (12.1) | 813 (11.5) | 780 (11.2) |
| 55–59 | 754 (17.1) | 758 (17.2) | 1442 (20.4) | 1446 (20.7) |
| 60–69 | 2018 (45.8) | 2019 (45.7) | 3225 (45.7) | 3198 (45.7) |
| 70–79 | 1105 (25.1) | 1107 (25.1) | 1574 (22.3) | 1573 (22.5) |
| Age at menopause in years, mean (SD) | 44.6 (7.6) | 44.4 (7.6) | 50.0 (4.7) | 50 (4.7) |
| Education level, N (%)a | ||||
| Less than high school or equivalent | 392 (9.1) | 472 (8.9) | 421 (6.0) | 431 (6.2) |
| High school or equivalent | 1033 (23.7) | 987 (22.5) | 1347 (19.2) | 1393 (20.1) |
| Vocational/training | 609 (14.0) | 609 (13.8) | 808 (11.5) | 786 (11.3) |
| Some college/Associate’s | 1289 (29.6) | 1303 (29.7) | 1993 (28.4) | 1875 (27.0) |
| College graduate/Baccalaureate | 313 (7.2) | 341 (7.8) | 736 (10.5) | 693 (10.0) |
| Some post-graduate or professional | 326 (7.5) | 347 (7.9) | 727 (10.4) | 785 (11.3) |
| Master’s degree | 338 (7.8) | 353 (8.1) | 870 (12.4) | 845 (12.2) |
| Doctoral degree | 57 (1.3) | 58 (1.3) | 122 (1.7) | 140 (2.0) |
| Family income, N (%)a | ||||
| <$10,000 | 321 (7.5) | 336 (7.9) | 337 (4.9) | 320 (4.7) |
| $10,000–$19,999 | 822 (19.3) | 808 (18.9) | 996 (14.5) | 958 (14.2) |
| $20,000–$34,999 | 1258 (29.5) | 1235 (28.9) | 1862 (27.2) | 1797 (26.5) |
| $35,000–$49,999 | 794 (18.6) | 830 (19.4) | 1428 (20.9) | 1442 (21.3) |
| $50,000–$74,999 | 615 (14.4) | 684 (13.7) | 1161 (17.0) | 1210 (17.8) |
| $75,000–$99,999 | 200 (4.7) | 218 (5.1) | 470 (6.9) | 509 (7.4) |
| $100,000–$149,999 | 114 (2.7) | 121 (2.8) | 307 (4.5) | 259 (3.8) |
| ≥$150,000 | 38 (0.9) | 36 (0.8) | 121 (0.8) | 119 (1.) |
| Don’t know | 105 (2.5) | 105 (2.5) | 167 (2.4) | 157 (2.3) |
| Body mass index, kg/m2, N (%)a | ||||
| Underweight (<18.5) | 16 (0.4) | 18 (0.4) | 49 (0.7) | 51 (0.7) |
| Normal (18.5–24.9) | 904 (20.6) | 895 (20.4) | 2104 (30.0) | 2119 (30.5) |
| Overweight (25.0–29.9) | 1495 (34.1) | 1557 (35.5) | 2443 (34.8) | 2429 (34.9) |
| Obesity I (30.0–34.9) | 1108 (25.3) | 1106 (25.2) | 1524 (21.7) | 1422 (20.5) |
| Obesity II (35.0–39.9) | 559 (12.8) | 518 (11.8) | 630 (9.0) | 607 (8.7) |
| Obesity III (≥40) | 302 (6.9) | 294 (6.7) | 271 (3.9) | 327 (3.9) |
| Region, N (%) | ||||
| Northeast | 850 (19.3) | 872 (19.7) | 1628 (23.1) | 1612 (23.0) |
| South | 1228 (27.9) | 1230 (27.9) | 1617 (22.9) | 1618 (23.1) |
| Midwest | 1095 (24.8) | 1084 (24.5) | 1845 (26.2) | 1809 (25.9) |
| West | 1235 (28.0) | 1231 (27.9) | 1964 (27.8) | 1958 (28.0) |
| Race, N (%) | ||||
| American Indian/Alaska Native | 19 (0.4) | 17 (0.4) | 19 (0.3) | 18 (0.3) |
| Asian | 65 (1.5) | 52 (1.2) | 171 (2.4) | 149 (2.1) |
| Native Hawaiian/Other Pacific Islander | <5 (<1.0) | 15 (0.3) | 11 (0.2) | 9 (0.1) |
| Black | 625 (14.2) | 635 (14.4) | 428 (6.1) | 462 (6.6) |
| White | 3552 (80.4) | 3552 (80.4) | 6240 (88.5) | 6187 (88.4) |
| More than one race | 78 (1.8) | 80 (1.8) | 77 (1.1) | 80 (1.1) |
| Unknown/not reported | 65 (1.5) | 66 (1.5) | 108 (1.5) | 92 (1.3) |
| Ethnicity, N % | ||||
| Not Hispanic/Latino | 4110 (93.2) | 4138 (93.7) | 6619 (93.8) | 6607 (94.4) |
| Hispanic/Latino | 273 (6.2) | 262 (5.9) | 391 (5.5) | 358 (5.1) |
| Unknown/not reported | 25 (0.6) | 17 (0.4) | 44 (0.6) | 32 (0.5) |
| Sleep quality, N % | ||||
| Very restless | 131 (3.0) | 120 (2.7) | 162 (2.3) | 132 (1.9) |
| Restless | 764 (17.4) | 775 (17.6) | 974 (13.8) | 984 (14.1) |
| Average | 1950 (44.4) | 1968 (44.7) | 2994 (45.5) | 3006 (43.1) |
| Sound or restful | 1146 (26.1) | 1116 (25.3) | 2040 (29.0) | 2005 (28.8) |
| Very sound or restful | 403 (9.2) | 428 (9.7) | 871 (12.4) | 846 (12.1) |
| Smoking, N % | ||||
| Never smoked | 2256 (51.6) | 2198 (50.3) | 3465 (49.6) | 3470 (50.2) |
| Past smoker | 1681 (38.5) | 1714 (39.2) | 2823 (40.4) | 2738 (39.6) |
| Current smoker | 432 (9.9) | 461 (10.5) | 704 (10.1) | 712 (10.2) |
| Ever migraine diagnosis at baseline, N % | 526 (11.9) | 480 (10.9) | 579 (8.2) | 613 (8.8) |
Abbreviation: BMI, body mass index; SD, standard deviation.
Missingness of covariates is as follows: education (n = 159, 0.7%), income (n = 716, 3.1%), BMI (n = 128, 0.6%), sleep quality (n = 61, 0.3%), smoking (n = 232, 1.0%).
Participants reporting a baseline migraine diagnosis (n = 2198, 9.6%) were similar to those without a history of migraine with respect to most measured characteristics. However, participants with a history of migraine were slightly younger, more likely to identify as white (87.9% vs. 85.1%), and report poorer sleep quality (Table 2) than those without a history of migraine.
TABLE 2.
Baseline demographic, medical, and lifestyle characteristics of 22,876 Women’s Health Initiative Hormone Therapy trial participants by ever migraine diagnosis at baseline.
| Parameter | Ever migraine diagnosis (n = 2198, 9.6%) |
No migraine diagnosis (n = 20,678, 90.4%) |
|---|---|---|
| Age in years, mean (SD) | 62 (7) | 64 (7) |
| Age category, N (%) | ||
| 50–54 | 343 (15.6) | 2314 (11.2) |
| 55–59 | 485 (22.1) | 3915 (18.9) |
| 60–69 | 966 (44.0) | 9494 (45.9) |
| 70–79 | 404 (18.4) | 4955 (24.0) |
| Age at menopause in years, mean (SD) | 48 (7) | 47 (7) |
| Education level, N (%) | ||
| Less than High School or equivalent | 173 (7.9) | 1465 (7.1) |
| High school or GED | 438 (20.0) | 4322 (21.1) |
| Vocational/training | 248 (11.3) | 2560 (12.5) |
| Some college/Associate’s | 691 (31.5) | 5769 (28.1) |
| College graduate/Baccalaureate | 184 (8.4) | 1899 (9.3) |
| Some post-graduate or professional | 202 (9.2) | 1983 (9.7) |
| Master’s degree | 226 (10.3) | 2180 (10.3) |
| Doctoral degree | 29 (1.3) | 348 (1.7) |
| Income, N (%) | ||
| <$10,000 | 145 (6.8) | 1169 (5.8) |
| $10,000-$19,999 | 333 (15.7) | 3251 (16.2) |
| $20,000-$34,999 | 594 (28.0) | 5558 (27.7) |
| $35,000-$49,999 | 425 (20.0) | 4069 (20.3) |
| $50,000-$74,999 | 344 (16.2) | 3226 (16.1) |
| $75,000-$99,999 | 123 (5.8) | 1274 (6.4) |
| $100,000-$149,999 | 76 (3.6) | 725 (3.6) |
| ≥$150,000 | 28 (1.3) | 286 (1.4) |
| Don’t know | 55 (2.6) | 479 (2.4) |
| Body mass index, kg/m2, N (%) | ||
| Underweight (<18.5) | 17 (0.8) | 117 (0.6) |
| Normal (18.5–24.9) | 562 (25.7) | 5460 (26.6) |
| Overweight (25.0–29.9) | 774 (35.4) | 7150 (34.8) |
| Obesity I (30.0–34.9) | 508 (23.2) | 4652 (22.6) |
| Obesity II (35.0–39.9) | 216 (9.9) | 2098 (10.2) |
| Obesity III (≥40) | 111 (5.1) | 1083 (5.3) |
| Region, N (%) | ||
| Northeast | 476 (21.7) | 4486 (21.7) |
| South | 528 (24.0) | 5165 (25.0) |
| Midwest | 536 (24.4) | 5297 (25.6) |
| West | 658 (29.9) | 5730 (27.7) |
| Race, N (%) | ||
| American Indian/Alaska Native | 7 (0.3) | 66 (0.3) |
| Asian | 22 (1.0) | 415 (2.0) |
| Native Hawaiian/Other Pacific Islander | <5 (<1) | 39 (0.2) |
| Black | 153 (7.1) | 1997 (9.7) |
| White | 1931 (87.9) | 17,600 (85.1) |
| More than one race | 44 (2.0) | 271 (1.3) |
| Unknown/not reported | 41 (1.9) | 290 (1.4) |
| Ethnicity, N (%) | ||
| Not Hispanic/Latino | 2039 (92.3) | 19435 (94.0) |
| Hispanic/Latino | 152 (6.9) | 1132 (5.5) |
| Unknown/not reported | 7 (0.3) | 111 (0.5) |
| Sleep quality, N (%) | ||
| Very restless | 82 (3.7) | 463 (2.2) |
| Restless | 470 (21.5) | 3027 (14.7) |
| Average | 994 (45.4) | 8924 (43.3) |
| Sound or restful | 500 (22.8) | 5807 (28.2) |
| Very sound or restful | 145 (6.6) | 2403 (11.7) |
| Smoking, N (%) | ||
| Never smoked | 1106 (50.7) | 10283 (50.3) |
| Past smoker | 848 (38.9) | 8108 (39.6) |
| Current smoker | 226 (10.4) | 2073 (10.1) |
| HRT arm, N (%) | ||
| Estrogen only | 526 (23.9) | 3882 (18.8) |
| Estrogen only control | 480 (21.8) | 3937 (19.0) |
| Estrogen and progestin | 579 (26.3) | 6475 (31.3) |
| Estrogen and progestin control | 613 (27.9) | 6384 (30.9) |
Abbreviation: GED, general education development; HRT, hormone replacement therapy; SD, standard deviation.
Overall, 82.9% of participants were considered adherent to the trial protocol (took 80% of MHT pills) for the 1-year observation period. Adherence was 86.0% among the intervention group and 79.8% among the control group. Adherence did not differ by migraine status at baseline (83.0% of participants with migraine and 81.7% of participants without migraine). The results below reflect intention-to-treat (ITT) analyses. Results of per-protocol (PP) analyses are reported in the supplemental material (Tables S1–S3). There were no substantial differences in results across these analysis types.
MHT and headache severity
Overall
Moderate-to-severe headache (compared to no or mild headache) at 1 year of follow-up was more common among individuals with history of migraine than those without (34.6% vs. 7.9%), regardless of treatment group.
Migraine at baseline
When analyzing the 9.6% of participants who reported a history of migraine at baseline, we observed no association of MHT with increasing headache severity among the E-only group (adjusted odds ratio [aOR] = 1.14, 95% confidence interval [CI]: 0.90–1.44). Among the E+P group we noted a modest positive association between MHT and increasing headache severity (aOR = 1.21, 95% CI: 0.98–1.50), adjusted for baseline headache severity, although estimates were imprecise and did not reach statistical significance (Table 3).
TABLE 3.
Postmenopausal headache severity after 1 year of follow-up, by menopausal hormone therapy use and baseline migraine status.
| Parameter | No headache, (n, %) | Mild headache, (n, %) | Moderate headache, (n, %) | Severe headache (n, %) | Crude odds of more severe headache | Adjusted odds of more severe headachea |
|---|---|---|---|---|---|---|
| Overall | ||||||
| E-Only | 2179 (49.4%) | 1635 (37.1%) | 466 (10.6%) | 128 (2.9%) | 1.13 (1.04–1.23) | 1.12 (1.02–1.22) |
| E-Only control | 2328 (52.7%) | 1560 (35.3%) | 427 (9.7%) | 102 (2.3%) | ref | ref |
| E+P | 3985 (56.5%) | 2382 (33.8%) | 564 (8.0%) | 123 (1.7%) | 1.10 (1.03–1.18) | 1.14 (1.06–1.22) |
| E+P control | 4100 (58.6%) | 2317 (33.1%) | 480 (6.9%) | 100 (1.4%) | ref | ref |
| Reported migraine at baseline | ||||||
| E-Only | 106 (20.2%) | 209 (39.7%) | 152 (28.9%) | 59 (11.2%) | 1.14 (0.91–1.43) | 1.14 (0.90–1.44) |
| E-Only control | 105 (21.9%) | 200 (41.7%) | 126 (26.3%) | 49 (10.2%) | ref | ref |
| E+P | 131 (22.6%) | 249 (43.0%) | 141 (24.4%) | 59 (10.2%) | 1.36 (1.10–1.67) | 1.21 (0.98–1.50) |
| E+P control | 179 (29.2%) | 259 (42.3%) | 128 (20.9%) | 47 (7.7%) | ref | ref |
| Did not report migraine at baseline | ||||||
| E-Only | 2073 (53.4%) | 1426 (36.7%) | 314 (8.1%) | 69 (1.8%) | 1.13 (1.03–1.23) | 1.11 (1.01–1.21) |
| E-Only control | 2223 (56.5%) | 1360 (34.5%) | 301 (7.7%) | 53 (1.4%) | ref | ref |
| E+P | 3854 (59.5%) | 2133 (32.9%) | 423 (6.5%) | 65 (1.0%) | 1.09 (1.02–1.17) | 1.14 (1.06–1.23) |
| E+P control | 3921 (61.4%) | 2058 (32.2%) | 352 (5.5%) | 53 (0.8%) | ref | ref |
Abbreviations: E-only, estrogen only; E+P, estrogen plus progestin.
Adjusted for baseline headache severity.
No migraine at baseline
Among those without a history of migraine at baseline, there was also a modest association of E-only MHT with odds of more severe headache (aOR = 1.11, 95% CI: 1.01–1.21). Those assigned to E+P exhibited 14% increased odds of more severe headache (aOR = 1.14, 95% CI: 1.06–1.23), compared with E+P controls, adjusted for baseline headache severity.
MHT and headache trajectory
Overall
Overall, 34.5% of the cohort experienced a change in headache trajectory between baseline and 1-year follow up. Across migraine history and treatment groups, 17.6%–27.7% of participants reported an improvement in their headaches, 48.5%–68.7% reported their headaches remained the same, and 14.1%–25.6% reported their headaches worsened (Table 4).
TABLE 4.
Headache trajectory, baseline headache and headache after 1 year of follow-up, by menopausal hormone therapy status.
| Parameter | Headache improved (n, %) | Headache remained unchanged (n, %) | Headache worsened (n, %) | Risk ratio (worsened)a | Risk ratio (improved)a |
|---|---|---|---|---|---|
| Overall | |||||
| E-Only | 808 (18.3%) | 2793 (63.4%) | 807 (18.3%) | 1.07 (0.96–1.20) | 0.84 (0.74–0.95) |
| E-Only control | 860 (19.5%) | 2781 (63.0%) | 776 (17.6%) | ref | ref |
| E+P | 1188 (16.8%) | 4669 (66.2%) | 1197 (17.0%) | 1.18 (1.08–1.30) | 0.90 (0.81–1.00) |
| E+P control | 1232 (17.6%) | 4731 (67.6%) | 1034 (14.8%) | ref | ref |
| Reported migraine at baseline | |||||
| E-Only | 118 (22.1%) | 287 (54.6%) | 121 (23.0%) | 0.98 (0.71–1.34) | 0.71 (0.52–0.98) |
| E-Only control | 133 (27.7%) | 233 (48.5%) | 114 (23.8%) | ref | ref |
| E+P | 148 (25.6%) | 283 (48.9%) | 148 (25.6%) | 1.53 (1.14–2.03) | 1.03 (0.76–1.39) |
| E+P control | 135 (22.0%) | 346 (56.4%) | 132 (21.5%) | ref | ref |
| Did not report migraine at baseline | |||||
| E-Only | 690 (17.8%) | 2506 (64.5%) | 686 (17.7%) | 1.08 (0.96–1.22) | 0.86 (0.75–0.99) |
| E-Only control | 727 (18.5%) | 2548 (64.7%) | 662 (16.8%) | ref | ref |
| E+P | 1040 (16.1%) | 4386 (67.7%) | 1049 (16.2%) | 1.18 (1.07–1.30) | 0.89 (0.80–1.00) |
| E+P control | 1097 (17.2%) | 4385 (68.7%) | 902 (14.1%) | ref | ref |
Abbreviations: E-only, estrogen only; E+P, estrogen plus progestin.
Adjusted for baseline headache severity.
Migraine at baseline
We found that 48.7% of individuals with migraine experienced either a worsening or improvement in their headache trajectory over year one of follow-up. E-only MHT was not associated with worsening headaches, regardless of migraine history (adjusted risk ratio [aRR] = 1.07, 95% CI: 0.96–1.20). E+P MHT was associated with worsening headache overall (aRR = 1.18, 95% CI: 1.08–1.30), and particularly among those with migraine history at baseline (aRR = 1.53, 95% CI: 1.14–2.03).
No migraine at baseline
Thirty-three percent of individuals without a history of migraine experienced a change in their trajectory of headache severity over follow-up. Among those without a history of migraine at baseline, E-only MHT was not associated with worsening headache (aRR = 1.08, 95% CI: 0.96–1.22). The association of E+P with worsening headache was attenuated among individuals without a history of migraine at baseline (aRR = 1.18, 95% CI: 1.07–1.30), compared with those with a history of migraine. However, there was no statistical evidence of interaction by history of migraine diagnosis in the association between E+P MHT and worsened headaches (p > .05).
Migraine at baseline and headache trajectory
Independent of MHT, and adjusting for potential confounders (age, income, education, region, race, and sleep quality) individuals with a history of migraine at baseline had a 42% greater risk of worsening headache (aRR = 1.42, 95% CI: 1.30–1.54) and a 32% greater risk of improved headache (aRR = 1.32, 95% CI: 1.22–1.43) compared to those without a history of migraine at baseline (Table 5).
TABLE 5.
Headache trajectory, baseline headache and headache after 1 year of follow-up, by ever diagnosis of migraine at baseline.
| Baseline migraine history | Headache improved | Headache remained unchanged | Headache worsened | Adj risk ratioa
(worsened) |
Adj risk ratiob
(improved) |
|---|---|---|---|---|---|
| Migraine diagnosis | 534 (24.3%) | 1149 (52.3%) | 515 (23.4%) | 1.42 (1.30–1.54) | 1.32 (1.22–1.43) |
| No migraine diagnosis | 3554 (17.2%) | 13,825 (66.9%) | 3299 (16.0%) | ref | ref |
Abbreviation: MHT, menopause hormone therapy.
Adjusted for baseline headache severity.
Adjusted for MHT arm, age category, income, education, region, race, and sleep quality.
DISCUSSION
In this study of the association of randomized hormone therapy with migraine and headache severity, we found modest associations between E+P MHT and postmenopausal headache severity regardless of migraine history. E+P MHT was particularly associated with higher risk of worsening headache in individuals with a history of migraine. Among individuals without a history of migraine at baseline, E+P was also associated with increased risk of worsening headache, although to a lesser extent.
While the increase in risk of worsening headache with E+P MHT was qualitatively stronger among individuals with a history of migraine, the difference in estimates did not reach statistical significance. This may be due to a low statistical power to detect differences among subgroups, or a true lack of heterogeneity of the degree to which E+P is associated with worsening headache across migraine history subgroups. Individuals with a history of migraine were more likely to experience a change in their headache severity between baseline and 1 year of follow-up. This may be due to underlying headache susceptibility and/or fewer individuals with migraine reporting no headaches at baseline and is an interesting observation in the context of the general assumption that migraine tends to resolve in the postmenopausal period.
Previous cross-sectional and observational studies have been methodologically inconsistent and resulted in positive, negative, and null findings for the association between MHT and postmenopausal migraine.3,8,10–12,17,18 Interpreting the association of MHT with headaches and migraine is further complicated by headache heterogeneity, different MHT regimens, and methods of administration.17–19 A previous randomized control trial that tested the effect of low-dose E+P MHT versus tibolone (a selective estrogen receptor modulator) on headaches in postmenopausal females found that E+P MHT resulted in more headache days and increased pain associated with migraine without aura, but fewer headache days and decreased pain associated with tension-type headache. However, the use of tibolone instead of placebo in the comparator group limits comparability to our findings.17 Our findings are consistent with previous studies suggesting divergent headache trajectories in the postmenopausal period among MHT users and non-users, with headaches or migraine worsening for some and improving for others, although the exact mechanisms or phenotypes at increased risk are unknown.7–10
The “estrogen-withdrawal” hypothesis posits that fluctuations in serum estradiol levels trigger migraine in hormonally susceptible individuals. However, the exact role of exogenous hormones in preventing or exacerbating migraine is unknown.3–5,7,8 Similarly, progestogens may be protective against migraine attacks by modulating pain perception but have also been associated with an increase in headache among those using progestin-only pills for birth control, though there is variability across specific progestins.19,20 These complex biological pathways may have influenced our differing findings among women receiving E+P or E-only.
Limitations
Our study was limited to the specific MHT regimens and doses tested in the WHI, limiting generalizability to the other existing formulations, doses, and routes of estrogen and progestogens. Results cannot be generalized to oral contraceptive use, which have different formulations and doses, or non-oral MHT (e.g., hormone patch), which may have less impact on migraine. In addition, our study was conducted in postmenopausal individuals. As those with headache disorders tend to report improvement in symptoms after menopause, these findings may not be generalizable to pre- and perimenopausal individuals. In addition, this study recruited participants between 1993 and 1998. Therefore, generalization to contemporary patient populations is limited by significant shifts in population demographic, lifestyle, and health characteristics, and shifts in clinical practice of headache.
Our study is also limited by the absence of detailed data on the exact type of headache types. Although individuals with a history of migraine likely reported migraine symptoms, the data did not allow us to conclusively categorize headache types, and thus cannot conclude that MHT influenced migraine headaches specifically. Therefore, the findings should be interpreted as an effect on general headache severity rather than an effect specific to migraine. Non-adherence to study protocols would lead to unpredictable bias given the polytomous nature of our headache severity variable, however adherence to the trial protocol in the first year was high (82.9%), and there were no substantial differences between ITT and PP analyses, indicating that this was not a likely source of significant bias.21
Our study relied on self-report of baseline physician-diagnosis of migraine and subjective headache severity, rather than diary-based symptom tracking. Previous studies have shown >80% agreement between self-reported migraine symptoms or self-reported physician diagnosed migraine and International Classification of Headache Disorder diagnostic criteria.22,23 However, our older population may have tended to under-report migraine diagnosis that occurred earlier in life, and migraine diagnoses was less common in the decades prior to this study. This may partially explain the low overall migraine prevalence in our sample. Our outcome focused on the degree to which headaches were recently bothersome, a relevant measure for patients. Assessing headache on a four-point scale may have reduced our ability to determine subtle changes in headache status particularly in those consistently reporting severe headaches. However, the risk of moderate-to-severe headaches or lack of headache improvement remain important to patients.
Not all participants reported their migraine status at baseline or headache severity at both time points, slightly reducing the number of participants in our analytic sample, although differential missing data are unlikely in this randomized control trial. While randomization of MHT would be expected to balance both measured and unmeasured covariates in the overall sample, this balance may not hold in the stratified analyses, as migraine history was not included in the original randomization scheme. We did not account for pre-enrollment reproductive conditions which may be associated with headache history or burden and MHT type. For example, overall access to health care or comorbid underlying conditions may lead to hysterectomy (and therefore E-only group assignment) and increased headache. However, parallel trials of E+P placebo and E-only MHT would balance these factors across treatment and placebo arms. In addition, measured covariates appear to be well balanced across migraine history and treatment subgroups (Table 2), decreasing concerns of substantial confounding.
Strengths of our study include a large sample, assessment of headache severity, randomized allocation to MHT in the overall sample, and assessment of headache at etiologically relevant time points (baseline and 1 year after initiation of MHT).
CONCLUSION
We found that E+P was associated with modestly increased odds of more severe postmenopausal headache, regardless of migraine history. E+P, but not E-only MHT, was associated with worsening headaches after 1 year, and this increase was stronger among individuals with a history of migraine at baseline. Our findings also suggest that postmenopausal headache trajectories are not uniform. Our study adds to the limited body of literature on the association between MHT and headache or migraine trajectories in postmenopausal individuals by using data from a large randomized controlled trial of MHT use. Future studies to explore predictors of worsening or improving headache in the postmenopausal period are necessary for improved risk prediction and treatment of headaches and migraine in the postmenopausal period.
Supplementary Material
Additional supporting information can be found online in the Supporting Information section at the end of this article.
ACKNOWLEDGMENTS
The authors thank the following WHI investigators and staff for their dedication, and the study participants for making the program possible. Program Office: (National Heart, Lung, and Blood Institute, Bethesda, Maryland) Jacques Rossouw, Jared Reis, and Candice Price Clinical Coordinating Center: (Fred Hutchinson Cancer Center, Seattle, WA) Garnet Anderson, Ross Prentice, Andrea LaCroix, and Charles Kooperberg Steering Committee and Academic Centers: (University of Alabama at Birmingham) Gretchen Wells; (Albert Einstein College of Medicine) Yasmin Mossavar-Rahmani; (University at Buffalo) Amy Millen; (University at Buffalo) Jean Wactawski-Wende; (Fred Hutchinson Cancer Center) Marian Neuhouser; (Fred Hutchinson Cancer Center) Holly Harris; (University of Massachusetts) Brian Silver; (University of North Carolina) Nora Franceschini; (Stanford Prevention Research Center) Marcia L. Stefanick; (The Ohio State University) Electra Paskett; (Wake Forest University) Mara Vitolins For a list of all the investigators who have contributed to WHI Science, please visit:: https://s3-us-west-2.amazonaws.com/www-whi-org/wp-content/uploads/WHI-Investigator-Long-List.pdf.
FUNDING INFORMATION
This project was supported by NIH Research Grant U54AG062322 funded by The National Institute on Aging (NIA) and Office of Research on Women’s Health (ORWH) and NIH Research Grant K01MH136417 funded by the National Institute on Mental Health. The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts (75N92021D00001, 75N92021D00002, 75N92021D00003, 75N92021D00004, 75N92021D00005).
Abbreviations:
- aOR
adjusted odds ratio
- aRR
adjusted risk ratio
- CI
confidence interval
- E-only
estrogen only
- E+P
estrogen plus progestin
- Fred Hutch
Fred Hutchinson Cancer Research Center
- FWA
Federalwide Assurance
- GED
general education development
- HT
hormone therapy
- IRB
institutional review board
- ITT
intention-to-treat
- MHT
menopausal hormone therapy
- OHRP
Office for Human Research Protections
- OR
odds ratio
- PP
per protocol
- RR
risk ratio
- SD
standard deviation
- WHI
Women’s Health Initiative
- WHI-HT
Women’s Health Initiative Hormone Therapy
Footnotes
CONFLICT OF INTEREST STATEMENT
Holly M. Crowe, Tracy E. Madsen, Janet W. Rich-Edwards, Aladdin H. Shadyab, Longjian Liu, Leslie V. Farland, Peter F. Schnatz, JoAnn E. Manson, Kathryn M. Rexrode, and Marcia L. Stefanick have no conflicts of interest to declare.
REFERENCES
- 1.Mattsson P, Svärdsudd K, Lundberg PO, Westerberg CE. The prevalence of migraine in women aged 40–74 years: a population-based study. Cephalalgia. 2000;20(10):893–899. doi: 10.1046/J.1468-2982.2000.00133.X/ASSET/41853BA8-296A-44FD-A032-5C809C8F96C3/ASSETS/IMAGES/LARGE/10.1046_J.1468-2982.2000.00133.X-FIG3.JPG [DOI] [PubMed] [Google Scholar]
- 2.Smitherman TA, Burch R, Sheikh H, Loder E. The prevalence, impact, and treatment of migraine and severe headaches in the United States: a review of statistics from national surveillance studies. Headache. 2013;53(3):427–436. doi: 10.1111/HEAD.12074 [DOI] [PubMed] [Google Scholar]
- 3.Hipolito Rodrigues MA, Maitrot-Mantelet L, Plu-Bureau G, Gompel A. Migraine, hormones and the menopausal transition. Climacteric. 2018;21(3):256–266. doi: 10.1080/13697137.2018.1439914 [DOI] [PubMed] [Google Scholar]
- 4.Martin VT, Behbehani M. Ovarian hormones and migraine headache: understanding mechanisms and pathogenesis – part 2. Headache. 2006;46(3):365–386. doi: 10.1111/J.1526-4610.2006.00370.X [DOI] [PubMed] [Google Scholar]
- 5.MacGregor EA. Female sex hormones and migraine. Revue Neurologique. Vol 161. Elsevier Masson SAS; 2005:677–678. doi: 10.1016/S0035-3787(05)85114-8 [DOI] [PubMed] [Google Scholar]
- 6.Kim S, Park JW. Migraines in women: a focus on reproductive events and hormonal milestones. Headache Pain Res. 2024;25(1):3–15. doi: 10.62087/HPR.2024.0003 [DOI] [Google Scholar]
- 7.MacGregor EA. Migraine, menopause and hormone replacement therapy. Post Reprod Health. 2018;24(1):11–18. doi: 10.1177/2053369117731172 [DOI] [PubMed] [Google Scholar]
- 8.MacGregor EA. Migraine, the menopause and hormone replacement therapy: a clinical review the City of London migraine clinic. J Fam Plann Reprod Health Care. 2007;33(4):33. Accessed March 18, 2024. http://srh.bmj.comDownloadedfrom [DOI] [PubMed] [Google Scholar]
- 9.Martin VT, Pavlovic J, Fanning KM, Buse DC, Reed ML, Lipton RB. Perimenopause and menopause are associated with high frequency headache in women with migraine: results of the American migraine prevalence and prevention study. Headache. 2016;56(2):292–305. doi: 10.1111/head.12763 [DOI] [PubMed] [Google Scholar]
- 10.Pavlović JM. Evaluation and management of migraine in mid-life women. Menopause. 2018;25(8):927–929. doi: 10.1097/GME.0000000000001104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ornello R, Caponnetto V, Frattale I, Sacco S. Patterns of migraine in postmenopausal women: a systematic review. Neuropsychiatr Dis Treat. 2021;17:859–871. doi: 10.2147/NDT.S285863 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Misakian AL, Langer RD, Bensenor IM, et al. Postmenopausal hormone therapy and migraine. Headache. 2004;12(10):1027–1036. doi: 10.1089/154099903322643956 [DOI] [PubMed] [Google Scholar]
- 13.Anderson G, Cummings S, Freedman LS, et al. Design of the Women’s Health initiative clinical trial and observational study. Control Clin Trials. 1998;19(1):61–109. doi: 10.1016/S0197-2456(97)00078-0 [DOI] [PubMed] [Google Scholar]
- 14.Manson JAE, Chlebowski RT, Stefanick ML, et al. The Women’s Health initiative hormone therapy trials: update and overview of Health outcomes during the intervention and post-stopping phases. JAMA. 2013;310(13):1353. doi: 10.1001/JAMA.2013.278040 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Schumacher LM, Farris SG, Thomas JG, et al. Interrelationships of sleep quality, obesity severity, and clinical headache features among women with comorbid migraine and obesity. J Clin Med. 2023;12(5):1742. doi: 10.3390/jcm12051742 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Braden JB, Young A, Sullivan MD, Walitt B, Lacroix AZ, Martin L. Predictors of change in pain and physical functioning among postmenopausal women with recurrent pain conditions in the women’s health initiative observational cohort. J Pain. 2012;13(1):64–72. doi: 10.1016/j.jpain.2011.10.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Nappi RE, Sances G, Sommacal A, et al. Different effects of tibolone and low-dose EPT in the management of postmenopausal women with primary headaches. Menopause. 2006;13(5):818–825. doi: 10.1097/01.GME.0000227399.53192.F5 [DOI] [PubMed] [Google Scholar]
- 18.Nappi G, Costa A, Tassorelli C, Santorelli FM. Migraine as a complex disease: heterogeneity, comorbidity and genotype-phenotype interactions. Funct Neurol. 2000;15(2):87–93. [PubMed] [Google Scholar]
- 19.van Lohuizen R, Paungarttner J, Lampl C, MaassenVanDenBrink A, Al-Hassany L. Considerations for hormonal therapy in migraine patients: a critical review of current practice. Expert Rev Neurother. 2024;24(1):55–75. doi: 10.1080/14737175.2023.2296610/ASSET/61FFA51C-320E-45A1-A959-EC8CD31845F0/ASSETS/GRAPHIC/IERN_A_2296610_F0002_OC.JPG [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Nappi RE, Merki-Feld GS, Terreno E, Pellegrinelli A, Viana M. Hormonal contraception in women with migraine: is progestogen-only contraception a better choice? J Headache Pain. 2013;14(1):66. doi: 10.1186/1129-2377-14-66 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Jackson M, Berman N, Huber M, et al. Research staff turnover and participant adherence in the Women’s Health initiative. Control Clin Trials. 2003;24(4):422–435. doi: 10.1016/s0197-2456(03)00027-8 [DOI] [PubMed] [Google Scholar]
- 22.Qiu C, Williams MA, Aurora SK, et al. Agreement of self-reported physician diagnosis of migraine with international classification of headache disorders-II migraine diagnostic criteria in a cross-sectional study of pregnant women. BMC Womens Health. 2013;13(1):50. doi: 10.1186/1472-6874-13-50 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Schürks M, Buring JE, Kurth T. Agreement of self-reported migraine with ICHD-II criteria in the Women’s Health study. Cephalalgia. 2009;29(10):1086–1090. doi: 10.1111/j.1468-2982.2008.01835.x [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.