The Relationship of Ovarian Steroids, Headache Activity and Menstrual Distress: A Pilot Study With Female Migraineurs
Jean C. Beckham,* Linda M. Krug,** Donald B. Penzien,** Cheryl A. Johnson,** Thomas H. Mosley, Jr.,** G. Rodney Meeks,*** Lori A. Pbert** and Rita C. Prather** *Duke **University
University Medical Center, Department of Psychiatry.
of Mississippi Medical Center Headache Clinic, Department of Psychiatry and Human Behavior.
***University
of Mississippi Medical Center, Department of Obstetrics and Gynecology.
Reprint requests to: Jean C. Beckham, Psychology Service,116B, Durham VAMC, Durham, NC 27705. Accepted for Publication: January 20, 1992. SYNOPSIS
Fourteen female volunteers who met diagnostic criteria for migraine headache monitored their headache activity and menstrual distress symptoms for one menstrual cycle. Serum estradiol and progesterone levels, and menstrual distress measures were collected at four points of the menstrual cycle: menstrual, ovulatory, luteal and premenstrual. Results indicated that one patient (7.1%) had menstrual migraine, 10 patients (71.4%) had menstrually-related headache and 3 (21.4%) had migraine headache unrelated to their menstrual cycle: subsequent analyses were conducted with the first two groups. Headache activity for the sample was highest during the premenstrual phase. Headache activity during the luteal and premenstrual phases was related to luteal phase progesterone levels. Menstrual distress was highest during the menstrual and premenstrual phases of the cycle, and these symptoms were related to higher estradiol levels, higher estradiol/progesterone ratios, and increased headache activity. These results indicated that for women with menstrual migraine or menstrually-related migraine, luteal progesterone levels may be significantly associated with headache activity, and estradiol and the estradiol/progesterone ratio may be significantly related to menstrual distress during the pre-menstrual phase of the cycle. The estradiol/progesterone ratio was not more related to headache or menstrual distress than either of these ovarian hormones alone. Suggestions for future research in this area are offered. Key words: progesterone, estradiol, ovarian steroids, migraine headache, menstrual distress Abbreviations: BBT basal body temperature, MDQ menstrual distress questionnaire (Headache 1992; 32:292-297) A long history of clinical observation and epidemiologic evidence exists to suggest an association between migraine and the menstrual cycle. The onset of migraine is likely to occur at puberty and the incidence of migraine after puberty is higher in women than men. In addition, a documented association exists between migraine headache and pregnancy, oral contraceptive, estrogen replacement therapy, and menopause. Many women also report increased migraines during the premenstrum, during menstruation and at ovulation. Several reviews have reported and summarized this research.1-6 Despite the wealth of information suggesting an association between the menstrual cycle and migraine, several important issues related to the documentation of this relationship remain. First, the prevalence of menstrual migraine is unclear. Frequency rates range widely from 8%7 to 70%.8 This large disparity may be due, at least in part, to retrospective reports and variable definitions of menstrual migraine.9 In a recent study, menstrual migraine was clearly defined as "a migraine attack which occurs regularly on or between days -2 to +3 of the menstrual cycle and at no other time.''9 Patients were labeled as having 'menstrually-related migraine' if they had an increased number of attacks at menstruation in addition to attacks at other times. This study minimizes the shortcomings of earlier reports, because it provides prospective headache recording, an explicit definition of menstrual migraine and inclusion of a related condition, menstrually-related migraine. Investigation of menstrual migraine requires explicit definitions of menstrual migraine, the use of established headache recording methods, and inclusion of possible subtypes of menstrual migraine (e.g., menstrually-related migraine) in order to provide valid results and comparability of findings across studies. A second issue related to the study of menstrual migraine concerns the covariation of hormone concentrations with headache activity. It has been hypothesized that migraines are associated with changing levels of estrogen and progesterone during the menstrual cycle. For example, headache activity has been postulated to increase because of a decline in serum estrogen concentration10 or to an increase in serum estrogen concentration.6 Alternatively, increased headache activity may be due to increased progesterone or to the change in estrogen to progesterone ratio across the menstrual cycle.4,6
Despite these data, a need remains to comparatively investigate estrogen, progesterone and the estrogen/progesterone ratio in menstruating women with migraine headaches. A third issue regarding the investigation of menstrual migraine is that the associations of headache activity and hormone concentration have not been related to mood alterations and physical symptoms of menstrual distress. Several studies have examined menstrual distress symptoms and menstrual phase in normally cycling females without migraine headache, however findings are inconsistent. Two recent studies have found no relationship between cycle phase and either mood changes or performance levels,11-12 while other researchers have reported dramatic changes in the perimenstrual phases of the cycle.13-15 In a recent study examining menstrual distress, menstrual cycle and hormone activity, no relationship was found between hormone activity and distress, but a strong relationship existed between menstrual distress symptoms and cycle phases.16 For many women, headache is an important part of the constellation of somatic and psychological menstrual symptoms. Therefore, menstrual distress needs to be compared to headache activity, the menstrual cycle and hormone levels. The current study had three objectives: (1) to document prospectively the occurrence of menstrual migraine, menstrually-related migraine and non-menstrually-related migraine using daily headache logs in an outpatient setting, (2) to examine estrogen concentration, progesterone concentration, the estrogen/progesterone ratio, headache activity and menstrual distress during the menstrual cycle for patients with menstrual migraine or menstrually-related migraine, and (3) to determine how hormone levels, headache activity, and menstrual distress are related across the menstrual cycle. MATERIALS AND METHODS
Fourteen female volunteers, aged 26 to 46 years (M = 35) were recruited from patients seeking treatment at a university-based headache clinic. The Structured Diagnostic Interview for headache was conducted17 and a woman was invited to participate if she: (a) met diagnostic criteria for migraine headache,18 (b) reported experiencing at least one migraine episode per month, (c) had regular menstrual cycles of 25 to 35 days, (d) had not used oral contraceptives or other hormone preparations in the past 12 months, and (e) was in good health. All patients signed an informed consent form after the study procedures were fully explained. Patients were given basal thermometers and charts for recording basal body temperature (BBT), headache activity, changes in cervical mucous, day of menstrual cycle, and pertinent physical changes (e.g., illness, sleepless nights). BBT was recorded each morning before getting out of bed or engaging in physical activity. Patients received extensive verbal and written instructions for daily self-monitoring of headache activity and then charted daily headache activity for four consecutive weeks. Patients rated their headache pain on an 11-point scale (0 = no headache pain to 10 = extremely painful headache) four times per day (e.g., at mealtimes and before going to bed). Ambiguous responses or errors in daily self-monitoring were clarified or corrected with the patient. A headache index was calculated by averaging patients' headache ratings during a 2 to 5 day interval proximate to each of the four menstrual cycle phases described below. Blood samples were drawn at four points during the menstrual cycle: menstrual - days I to 3 beginning with onset of menses; ovulatory - 13 to 15 days prior to onset of next menses; luteal - 5 to 7 days prior to onset of menses, and premenstrual - 1 to 3 days prior to onset of menses. Onset of the next menses was estimated by patients' self-report of length of the previous two menstrual cycles, patient observation of a 0.3-0.5 degree rise in BBT and/or changes in cervical mucous (12-14 days following the change in viscosity). Serum levels of estradiol, progesterone, lutenizing hormone (LH) and follicle-stimulating hormone (FSH) were determined with commercial radioimmunoassay kits from Diagnostic Products Corporation. LH was and FSH were determined with double-antibody RIA procedures, with minimum detection levels of 2 mlU/ml and 1.2 mlU/ml, respectively. The kit for LH had an intra-assay range of 1.8 to 0.7% and an inter-assay range of 4.0 to 7.9%. The intra-assay range for the FSH kit was 3.1 to 6.5% and the inter-assay range was 4.2 to 7.7%. LH and FSH were included to determine whether subjects were cycling normally. Total estradiol levels were determined by a double-antibody RIA method. The minimum detectable dose is 8 pg/ml; intra-assay coefficient variations (CVs) have ranged from 4.0 to 7.0% and inter-assay CVs have ranged from 4.2 to 8.1%. Progesterone levels were determined with a solid-phase 125I-RIA. This procedure can detect as little as 0.05 ng/ml; intra-assay CVs ranged from 5.8 to 8.4% and inter-assay CVs from 6.6 to 10.0%. Patients also completed the Menstrual Distress Questionnaire (MDQ) following each phlebotomy. The MDQ is a widely used self-report measure sensitive to symptoms associated with menstrual cycle.19 Patients rate the presence of symptoms "today and one to two days ago" on a four-point scale according to severity of experience, from 1 ("not at all") to 4 ("to a high degree"). The MDQ is composed of two factors (somatic and psychological/behavioral) derived from seven stable and recurrent dimensions: pain, concentration, arousal, negative affect, behavioral change, water retention and control. The MDQ has been shown to be reliable and valid measure of menstrual distress symptoms.20-22 RESULTS
Study Objective 1. Sample Prevalence of Menstrual Migraine and Menstrually-Related Migraine. Patients were identified as having 'menstrual migraine' if they had headache activity which occurred on or between days -3 to +3 days of their menstrual cycle and at no other time. Patients were labeled as having 'menstrually-related migraine' if the mean of their total headache activity was greater during the premenstrual and menstrual phases compared to their ovulatory and luteal phases. Women with headache activity equal to or greater during the luteral and ovulatory phases as compared to premenstrual and menstrual phases were considered to have migraine unrelated to menstrual cycle. According to these definitions, one patient (7.1%) had menstrual migraine, 10 patients (71.4%) had menstrually-related headache, and 3 (21.4%) had migraine headache unrelated to menstrual cycle. Subsequent analyses were conducted with the 11 patients whose headache activity could be defined as 'menstrual migraine' or 'menstrually-related migraine'.
Study Objective 2: Hormonal Levels Across the Menstrual Cycle. Table 1 shows the means and standard errors for serum levels of estradiol, progesterone, LH and FSH during each cycle phase. Repeated measures analyses of variance (ANOVAS) with follow-up pairwise comparisons using paired t-tests showed as expected that estradiol varied significantly with the cycle (F = 6,51, p =.0019), and was at its zenith during the ovulatory phase. The mean progesterone also varied significantly (F = 20.59, p =.0001), and was elevated during the luteal phase. FSH and LH showed peak values around ovulation (F = 3.87, p = .02 and 4.52, p = .01 for overall tests, respectively). The estradiol/progesterone ratio also peaked during the ovulation phase of the cycle (overall test: F = 17.5, p = .0001). For each of the 11 subjects, ovulation was confirmed by progesterone greater than 3ng/ml during the luteal phase and by inspection of the basal body temperature graphs. All cycles were found to be ovulatory. Headache Activity, Hormones, and the Menstrual Cycle. Headache activity levels were observed to differ across the menstrual cycle phases (see Figure 1). Repeated measures ANOVA calculated for headache index across the four menstrual cycle phases was significant (F(3,30) = 7.34; p = .0008). Follow-up pairwise comparisons using paired t-tests revealed that patients experienced significantly greater headache activity during the premenstrual phase (M = 14.1, S. E. = 2.63) of their cycle than during the menstrual (M =9.14, S. E., 1.97), ovulatory (M = 7.95, S. E. = 2.43) and luteal (M = 5.86, S.E. = 1.55) phases of their cycle (all comparisons p < .05), and during the menstrual phase as compared to the luteal phase. Headache activity was also selectively related to hormone levels. Pearson product moment correlations between each hormone and headache activity at each phase of the cycle were calculated. Headache index and progesterone were positively correlated (r = .65, p = .03) during the luteal phase. In order to evaluate the hypothesis that headache activity during the premenstrual phase may be due to the surge in progesterone during the luteal phase, the correlation between luteal phase progesterone and pre-menstrual phase headache activity was calculated. Although failing to reach statistical significance (owing to the small sample size), the correlation was still moderate to high (4 = .41 p = .2). Headache activity was not otherwise significantly correlated with hormone levels, but there were several noteworthy trends. Estradiol and the estradiol/progesterone ratio were related to headache activity in the luteal and premenstrual phases with correlations ranging from .32 to .38, although these correlations were not statistically significant. Menstrual Distress and the Menstrual Cycle. Means for both the somatic and psychological / behavioral factors are shown in Figure 2. Separate repeated measures ANOVAS performed on the two MDQ factors showed that there were significant cycle-related changes for the somatic factor of the MDQ (F (3,30) = 8.43; p = .0005). Post hoc pairwise comparisons using paired t-tests indicated somatic symptoms were highest at the menstrual and premenstrual times of the cycle as compared to the ovulatory and luteal phases of the cycle. Although psychological/behavioral symptoms also varied in the same pattern across the cycle, mean differences were not statistically significant. Study Objective 3: Menstrual Distress, Hormonal Levels and Headache Activity. Pearson product moment correlations between MDQ factors, hormone measures and headache activity across the menstrual cycle are presented in Table 2. Estradiol, the estradiol/progesterone ratio and headache activity were strongly related to the somatic factor scores during the menstrual and premenstrual phases, with higher levels associated with increased somatic symptoms. For the MDQ psychological/behavioral Table 1 Means and Standard Errors for Serum Levels of Estradiol, Progesterone, Estradiol/Progesterone Ratio, FSH and LH in the Menstrual, Ovulatory, Luteal and Premenstrual Phase of One Menstrual Cycle in Women with Migraine. Menstrual Ovulatory Luteal Premenstrual Fvalue Hormone Mean SE Mean SE Mean SE Mean SE Estradiol 43.7(12.8) 144.4(21.0) 94.7(14.3) 99.7(14.3) 6.5** (pmol/I) Progest. 1.0(.05) 1.8(.6) 9.0(1.3) 2.7(.5) 20.6*** (nmol/I) E/P Ratio 43.0(13.0) 130.00(23.7) 11.7(2.6) 39.0(11.6) 17.5'*** FSH 15.4(.93) 18.0(2.8) 9.3(1.0) 12.3(2.4) 2.8* (mU/1) LH 6.9(.60) 39.0(13.5) 6.3(.8) 8.0(1.6) 4.5** (mU/1) *p < 0.05 ** p < 0.01 ***p < 0.001
factor, these patterns were similar. During the menstrual and premenstrual phases, higher estradiol/progesterone ratio and headache activity were associated with more psychological/behavioral symptoms. During the menstrual phase, higher estradiol was related to increased psychological/behavioral symptoms. Headache activity was related to the psychological/behavioral factor during the luteal phase. Table 2 Correlations Between MDQ Factors, Hormone Measures and Headache Across the Menstrual Cycle Phase Menstrual Cycle Phase MDQ Factor Menstrual Ovulatory Luteal Premenstrual Somatic Estradiol .60** .15 .18 .72** Progesterone -.28 .35 .44 .48 E/P ratio1 .61 ** .14 -.06 .72** Headache Index .50* .3 .38 .58* Psychological/Behavioral Estradiol .59** .40 -.23 .23 Progesterone -.36 .02 -.06 -.15 E/P ratio .61** .40 .36 .74** Headache Index .54* .26 .50* .70** *p < .10 **p < .05 1E/P ratio = estradiol/progesterone ratio DISCUSSION
In this study, headache activity was found to be highest during the premenstrual and menstrual phases of the cycle for women with menstrual migraine or menstrually-related migraine. Headache activity during the luteal and premenstrual phases was moderately to strongly related to luteal phase progesterone levels, and there was a trend toward association between premenstrual estradiol levels and headache activity during this period, Contrary to previous studies,4,5 the results of this study suggest that serum progesterone concentration as wall as serum estradiol concentration may be important in understanding the occurrence of menstrual migraine and menstrually-related headache. Menstrual distress also varied across the cycles for this sample. Somatic symptoms were highest at the premenstrual and menstrual phases of the cycle. Psychological/behavioral symptoms, albeit not statistically significantly, followed the same pattern. Hormonal concentrations and headache activity were associated with menstrual distress. Menstrual and premenstrual somatic symptoms were positively related to estradiol, estradiol/progesterone ratio and headache activity. Increased menstrual and premenstrual psychological/behavioral symptoms were related to increased estradiol/progesterone ratio levels and headache activity. Higher estradiol levels were associated with increased psychological/ behavioral symptoms during the menstrual phase of the cycle. The results regarding hormonal levels and menstrual distress are consistent with a study investigating menstrual distress and hormonal levels in women with high levels of menstrual distress, but who did not experience migraine headaches.23 This study provides a preliminary investigation into the covariation of headache, menstrual distress symptoms and ovarian hormones in women with menstrual migraine or menstrually-related headache. However, results need to be evaluated with a larger sample across several menstrual cycles in order to assess the stability of findings. Following a larger group of women across several menstrual cycles would allow for both predictive and pattern analysis of the relationship between headache activity, hormonal levels and menstrual distress. Inclusion of menstrual migraine, menstrually-related migraine and unrelated migraine would allow for comparative analyses. One would predict for example, that migraineurs with headache unrelated to menstrual cycle may report increased menstrual distress, but hormonal levels and headache activity would be unrelated in these women. Comparison to other related disorders, such as late luteal phase dysphoric disorder, could assist in identifying the unique hormonal and symptom complex of menstrual migraine and menstrually-related migraine. The authors would like to thank John C. Barefoot, Thomas L. Haney, Robin M. French, William B. Saun-
ders and John Feaganes for their comments of earlier drafts of this paper. This research was supported part by grant ;1 M01 RR02303-05 awarded to Dr. Penzien by the General Clinical Research Centers Program, Division of Research Resources, National Institutes of Health. REFERENCES
1.
Edelson RN: Menstrual migraine and other hormonal aspects of migraine. Headache 1985; 25:376-379.
2.
Martignoni E, Sances G, Nappi G: Significance of hormonalchanges in migraine and cluster headache Gynecol Endocrinol 1987; 1:295-319.
3.
Nattero G: Menstrual headache. Adv Neurol 1982; 33:215-226.
4.
Silberstein SD, Merrian GR: Estrogens, progestins and headache. Neurology 1991; 41:786-793.
5.
Uknis A, Silberstein SD: Review article: Migraine and pregnancy. Headache 1991; 31:372-374.
6.
Welch KMA, Darnely D, Simkins RT: The role of estrogen in migraine: A review and hypothesis. Cephalalgia 1984; 4:227-236.
7.
Digre K, Damasio H: Menstrual migraine: Differential diagnosis, evaluation, and treatment. Clin Obs Gyn 1987; 30:417-430.
8.
Diamond S, Dalessio DJ: The practicing physician's approach to headache. Baltimore: Williams and Wilkins, 1982.
9.
MacGregor EA, Chia H, Vohrah RC, Wilkinson M: Migraine and menstruation: A pilot study. Cephalalgia 1990; 10:305-310.
10.
Somerville BW: The role of estradiol withdrawal in the etiology of menstrual migraine. Neurology 1972; 22:355-365.
11.
Cockrill CL, Appel MA, Carson RR: The effect of menstrual cycle on physiological and psychological function. Paper presented at the meeting of Society of Behavioral Medicine, Boston, April 1988.
12.
Nattero G, Allais G, De Lorenzo D, Torre E, Ancona M, Benedetto C, Massobrio M: Menstrual migraine: New biochemical and psychological aspects. Headache 1988; 28:103-107.
13.
Brahams D: Pre-menstrual syndrome: A disease of the mind? Lancet 1981; 2:710-718.
14.
Glass GS, Heninger GR, Lansky M, Talan K: Psychiatric emergency related to the menstrual cycle Am J Psychiatry 1971; 128:705.
15.
Dalton K: Cyclical criminal acts in premenstrual syndrome. Lancet 2:1070, 1980.
16.
Collins A, Eneroth P, Langren B: Psychoneuroendocrine stress responses and mood as related to the menstrual cycle. Psychosom Med 1985; 47: 512-527.
17.
Penzien DB, Andrew ME, Knowlton GE, McAnulty RD, Rains JC, Johnson CA, Hursey KG, Jacks SD: Computer-aided system for headache diagnosis with the IHS headache diagnostic criteria: Development and validation. Cephalalgia 1991; 11(Suppl 11):325-326.
18.
Horowitz M, Wilren N, Alvarez W: Impact of event scale: A measure of subjective stress. Psychosom Med 1979; 41:209-218.
19.
Moos RH: The development of a Menstrual Distress Questionnaire. Psychosom Med 1968; 30:853-867.
20.
Moos RH: Typology of menstrual cycle symptoms. Am J Obstet Gynecol 1969; 103:390-402.
21.
Moos RH, Leiderman DB: Toward a menstrual cycle symptom typology. J Psychosom Res 1978; 22:31-40.
22.
Moos RH: Perimenstrual symptoms: A manual and overview of research with the Menstrual Distress Questionnaire, Social ecology Lab. Sanford University School of Medicine and V.A. Medical centers, Palo Alto, CA. 1985.
23.
Dennerstein L, Spencer-Gardner C, Brown JB, Smith MA, Burrows GD: Premenstrual tension-hormonal profiles. J Psychosom Obstet Gynaecol 1984; 3:27-51.
Jean C. Beckham,* Linda M. Krug,** Donald B. Penzien,** Cheryl A. Johnson,** Thomas H. Mosley, Jr.,** G. Rodney Meeks,*** Lori A. Pbert** and Rita C. Prather** *Duke **University
University Medical Center, Department of Psychiatry.
of Mississippi Medical Center Headache Clinic, Department of Psychiatry and Human Behavior.
***University
of Mississippi Medical Center, Department of Obstetrics and Gynecology.
Reprint requests to: Jean C. Beckham, Psychology Service,116B, Durham VAMC, Durham, NC 27705. Accepted for Publication: January 20, 1992. SYNOPSIS
Fourteen female volunteers who met diagnostic criteria for migraine headache monitored their headache activity and menstrual distress symptoms for one menstrual cycle. Serum estradiol and progesterone levels, and menstrual distress measures were collected at four points of the menstrual cycle: menstrual, ovulatory, luteal and premenstrual. Results indicated that one patient (7.1%) had menstrual migraine, 10 patients (71.4%) had menstrually-related headache and 3 (21.4%) had migraine headache unrelated to their menstrual cycle: subsequent analyses were conducted with the first two groups. Headache activity for the sample was highest during the premenstrual phase. Headache activity during the luteal and premenstrual phases was related to luteal phase progesterone levels. Menstrual distress was highest during the menstrual and premenstrual phases of the cycle, and these symptoms were related to higher estradiol levels, higher estradiol/progesterone ratios, and increased headache activity. These results indicated that for women with menstrual migraine or menstrually-related migraine, luteal progesterone levels may be significantly associated with headache activity, and estradiol and the estradiol/progesterone ratio may be significantly related to menstrual distress during the pre-menstrual phase of the cycle. The estradiol/progesterone ratio was not more related to headache or menstrual distress than either of these ovarian hormones alone. Suggestions for future research in this area are offered. Key words: progesterone, estradiol, ovarian steroids, migraine headache, menstrual distress Abbreviations: BBT basal body temperature, MDQ menstrual distress questionnaire (Headache 1992; 32:292-297) A long history of clinical observation and epidemiologic evidence exists to suggest an association between migraine and the menstrual cycle. The onset of migraine is likely to occur at puberty and the incidence of migraine after puberty is higher in women than men. In addition, a documented association exists between migraine headache and pregnancy, oral contraceptive, estrogen replacement therapy, and menopause. Many women also report increased migraines during the premenstrum, during menstruation and at ovulation. Several reviews have reported and summarized this research.1-6 Despite the wealth of information suggesting an association between the menstrual cycle and migraine, several important issues related to the documentation of this relationship remain. First, the prevalence of menstrual migraine is unclear. Frequency rates range widely from 8%7 to 70%.8 This large disparity may be due, at least in part, to retrospective reports and variable definitions of menstrual migraine.9 In a recent study, menstrual migraine was clearly defined as "a migraine attack which occurs regularly on or between days -2 to +3 of the menstrual cycle and at no other time.''9 Patients were labeled as having 'menstrually-related migraine' if they had an increased number of attacks at menstruation in addition to attacks at other times. This study minimizes the shortcomings of earlier reports, because it provides prospective headache recording, an explicit definition of menstrual migraine and inclusion of a related condition, menstrually-related migraine. Investigation of menstrual migraine requires explicit definitions of menstrual migraine, the use of established headache recording methods, and inclusion of possible subtypes of menstrual migraine (e.g., menstrually-related migraine) in order to provide valid results and comparability of findings across studies. A second issue related to the study of menstrual migraine concerns the covariation of hormone concentrations with headache activity. It has been hypothesized that migraines are associated with changing levels of estrogen and progesterone during the menstrual cycle. For example, headache activity has been postulated to increase because of a decline in serum estrogen concentration10 or to an increase in serum estrogen concentration.6 Alternatively, increased headache activity may be due to increased progesterone or to the change in estrogen to progesterone ratio across the menstrual cycle.4,6
Despite these data, a need remains to comparatively investigate estrogen, progesterone and the estrogen/progesterone ratio in menstruating women with migraine headaches. A third issue regarding the investigation of menstrual migraine is that the associations of headache activity and hormone concentration have not been related to mood alterations and physical symptoms of menstrual distress. Several studies have examined menstrual distress symptoms and menstrual phase in normally cycling females without migraine headache, however findings are inconsistent. Two recent studies have found no relationship between cycle phase and either mood changes or performance levels,11-12 while other researchers have reported dramatic changes in the perimenstrual phases of the cycle.13-15 In a recent study examining menstrual distress, menstrual cycle and hormone activity, no relationship was found between hormone activity and distress, but a strong relationship existed between menstrual distress symptoms and cycle phases.16 For many women, headache is an important part of the constellation of somatic and psychological menstrual symptoms. Therefore, menstrual distress needs to be compared to headache activity, the menstrual cycle and hormone levels. The current study had three objectives: (1) to document prospectively the occurrence of menstrual migraine, menstrually-related migraine and non-menstrually-related migraine using daily headache logs in an outpatient setting, (2) to examine estrogen concentration, progesterone concentration, the estrogen/progesterone ratio, headache activity and menstrual distress during the menstrual cycle for patients with menstrual migraine or menstrually-related migraine, and (3) to determine how hormone levels, headache activity, and menstrual distress are related across the menstrual cycle. MATERIALS AND METHODS
Fourteen female volunteers, aged 26 to 46 years (M = 35) were recruited from patients seeking treatment at a university-based headache clinic. The Structured Diagnostic Interview for headache was conducted17 and a woman was invited to participate if she: (a) met diagnostic criteria for migraine headache,18 (b) reported experiencing at least one migraine episode per month, (c) had regular menstrual cycles of 25 to 35 days, (d) had not used oral contraceptives or other hormone preparations in the past 12 months, and (e) was in good health. All patients signed an informed consent form after the study procedures were fully explained. Patients were given basal thermometers and charts for recording basal body temperature (BBT), headache activity, changes in cervical mucous, day of menstrual cycle, and pertinent physical changes (e.g., illness, sleepless nights). BBT was recorded each morning before getting out of bed or engaging in physical activity. Patients received extensive verbal and written instructions for daily self-monitoring of headache activity and then charted daily headache activity for four consecutive weeks. Patients rated their headache pain on an 11-point scale (0 = no headache pain to 10 = extremely painful headache) four times per day (e.g., at mealtimes and before going to bed). Ambiguous responses or errors in daily self-monitoring were clarified or corrected with the patient. A headache index was calculated by averaging patients' headache ratings during a 2 to 5 day interval proximate to each of the four menstrual cycle phases described below. Blood samples were drawn at four points during the menstrual cycle: menstrual - days I to 3 beginning with onset of menses; ovulatory - 13 to 15 days prior to onset of next menses; luteal - 5 to 7 days prior to onset of menses, and premenstrual - 1 to 3 days prior to onset of menses. Onset of the next menses was estimated by patients' self-report of length of the previous two menstrual cycles, patient observation of a 0.3-0.5 degree rise in BBT and/or changes in cervical mucous (12-14 days following the change in viscosity). Serum levels of estradiol, progesterone, lutenizing hormone (LH) and follicle-stimulating hormone (FSH) were determined with commercial radioimmunoassay kits from Diagnostic Products Corporation. LH was and FSH were determined with double-antibody RIA procedures, with minimum detection levels of 2 mlU/ml and 1.2 mlU/ml, respectively. The kit for LH had an intra-assay range of 1.8 to 0.7% and an inter-assay range of 4.0 to 7.9%. The intra-assay range for the FSH kit was 3.1 to 6.5% and the inter-assay range was 4.2 to 7.7%. LH and FSH were included to determine whether subjects were cycling normally. Total estradiol levels were determined by a double-antibody RIA method. The minimum detectable dose is 8 pg/ml; intra-assay coefficient variations (CVs) have ranged from 4.0 to 7.0% and inter-assay CVs have ranged from 4.2 to 8.1%. Progesterone levels were determined with a solid-phase 125I-RIA. This procedure can detect as little as 0.05 ng/ml; intra-assay CVs ranged from 5.8 to 8.4% and inter-assay CVs from 6.6 to 10.0%. Patients also completed the Menstrual Distress Questionnaire (MDQ) following each phlebotomy. The MDQ is a widely used self-report measure sensitive to symptoms associated with menstrual cycle.19 Patients rate the presence of symptoms "today and one to two days ago" on a four-point scale according to severity of experience, from 1 ("not at all") to 4 ("to a high degree"). The MDQ is composed of two factors (somatic and psychological/behavioral) derived from seven stable and recurrent dimensions: pain, concentration, arousal, negative affect, behavioral change, water retention and control. The MDQ has been shown to be reliable and valid measure of menstrual distress symptoms.20-22 RESULTS
Study Objective 1. Sample Prevalence of Menstrual Migraine and Menstrually-Related Migraine. Patients were identified as having 'menstrual migraine' if they had headache activity which occurred on or between days -3 to +3 days of their menstrual cycle and at no other time. Patients were labeled as having 'menstrually-related migraine' if the mean of their total headache activity was greater during the premenstrual and menstrual phases compared to their ovulatory and luteal phases. Women with headache activity equal to or greater during the luteral and ovulatory phases as compared to premenstrual and menstrual phases were considered to have migraine unrelated to menstrual cycle. According to these definitions, one patient (7.1%) had menstrual migraine, 10 patients (71.4%) had menstrually-related headache, and 3 (21.4%) had migraine headache unrelated to menstrual cycle. Subsequent analyses were conducted with the 11 patients whose headache activity could be defined as 'menstrual migraine' or 'menstrually-related migraine'.
Study Objective 2: Hormonal Levels Across the Menstrual Cycle. Table 1 shows the means and standard errors for serum levels of estradiol, progesterone, LH and FSH during each cycle phase. Repeated measures analyses of variance (ANOVAS) with follow-up pairwise comparisons using paired t-tests showed as expected that estradiol varied significantly with the cycle (F = 6,51, p =.0019), and was at its zenith during the ovulatory phase. The mean progesterone also varied significantly (F = 20.59, p =.0001), and was elevated during the luteal phase. FSH and LH showed peak values around ovulation (F = 3.87, p = .02 and 4.52, p = .01 for overall tests, respectively). The estradiol/progesterone ratio also peaked during the ovulation phase of the cycle (overall test: F = 17.5, p = .0001). For each of the 11 subjects, ovulation was confirmed by progesterone greater than 3ng/ml during the luteal phase and by inspection of the basal body temperature graphs. All cycles were found to be ovulatory. Headache Activity, Hormones, and the Menstrual Cycle. Headache activity levels were observed to differ across the menstrual cycle phases (see Figure 1). Repeated measures ANOVA calculated for headache index across the four menstrual cycle phases was significant (F(3,30) = 7.34; p = .0008). Follow-up pairwise comparisons using paired t-tests revealed that patients experienced significantly greater headache activity during the premenstrual phase (M = 14.1, S. E. = 2.63) of their cycle than during the menstrual (M =9.14, S. E., 1.97), ovulatory (M = 7.95, S. E. = 2.43) and luteal (M = 5.86, S.E. = 1.55) phases of their cycle (all comparisons p < .05), and during the menstrual phase as compared to the luteal phase. Headache activity was also selectively related to hormone levels. Pearson product moment correlations between each hormone and headache activity at each phase of the cycle were calculated. Headache index and progesterone were positively correlated (r = .65, p = .03) during the luteal phase. In order to evaluate the hypothesis that headache activity during the premenstrual phase may be due to the surge in progesterone during the luteal phase, the correlation between luteal phase progesterone and pre-menstrual phase headache activity was calculated. Although failing to reach statistical significance (owing to the small sample size), the correlation was still moderate to high (4 = .41 p = .2). Headache activity was not otherwise significantly correlated with hormone levels, but there were several noteworthy trends. Estradiol and the estradiol/progesterone ratio were related to headache activity in the luteal and premenstrual phases with correlations ranging from .32 to .38, although these correlations were not statistically significant. Menstrual Distress and the Menstrual Cycle. Means for both the somatic and psychological / behavioral factors are shown in Figure 2. Separate repeated measures ANOVAS performed on the two MDQ factors showed that there were significant cycle-related changes for the somatic factor of the MDQ (F (3,30) = 8.43; p = .0005). Post hoc pairwise comparisons using paired t-tests indicated somatic symptoms were highest at the menstrual and premenstrual times of the cycle as compared to the ovulatory and luteal phases of the cycle. Although psychological/behavioral symptoms also varied in the same pattern across the cycle, mean differences were not statistically significant. Study Objective 3: Menstrual Distress, Hormonal Levels and Headache Activity. Pearson product moment correlations between MDQ factors, hormone measures and headache activity across the menstrual cycle are presented in Table 2. Estradiol, the estradiol/progesterone ratio and headache activity were strongly related to the somatic factor scores during the menstrual and premenstrual phases, with higher levels associated with increased somatic symptoms. For the MDQ psychological/behavioral Table 1 Means and Standard Errors for Serum Levels of Estradiol, Progesterone, Estradiol/Progesterone Ratio, FSH and LH in the Menstrual, Ovulatory, Luteal and Premenstrual Phase of One Menstrual Cycle in Women with Migraine. Menstrual Ovulatory Luteal Premenstrual Fvalue Hormone Mean SE Mean SE Mean SE Mean SE Estradiol 43.7(12.8) 144.4(21.0) 94.7(14.3) 99.7(14.3) 6.5** (pmol/I) Progest. 1.0(.05) 1.8(.6) 9.0(1.3) 2.7(.5) 20.6*** (nmol/I) E/P Ratio 43.0(13.0) 130.00(23.7) 11.7(2.6) 39.0(11.6) 17.5'*** FSH 15.4(.93) 18.0(2.8) 9.3(1.0) 12.3(2.4) 2.8* (mU/1) LH 6.9(.60) 39.0(13.5) 6.3(.8) 8.0(1.6) 4.5** (mU/1) *p < 0.05 ** p < 0.01 ***p < 0.001
factor, these patterns were similar. During the menstrual and premenstrual phases, higher estradiol/progesterone ratio and headache activity were associated with more psychological/behavioral symptoms. During the menstrual phase, higher estradiol was related to increased psychological/behavioral symptoms. Headache activity was related to the psychological/behavioral factor during the luteal phase. Table 2 Correlations Between MDQ Factors, Hormone Measures and Headache Across the Menstrual Cycle Phase Menstrual Cycle Phase MDQ Factor Menstrual Ovulatory Luteal Premenstrual Somatic Estradiol .60** .15 .18 .72** Progesterone -.28 .35 .44 .48 E/P ratio1 .61 ** .14 -.06 .72** Headache Index .50* .3 .38 .58* Psychological/Behavioral Estradiol .59** .40 -.23 .23 Progesterone -.36 .02 -.06 -.15 E/P ratio .61** .40 .36 .74** Headache Index .54* .26 .50* .70** *p < .10 **p < .05 1E/P ratio = estradiol/progesterone ratio DISCUSSION
In this study, headache activity was found to be highest during the premenstrual and menstrual phases of the cycle for women with menstrual migraine or menstrually-related migraine. Headache activity during the luteal and premenstrual phases was moderately to strongly related to luteal phase progesterone levels, and there was a trend toward association between premenstrual estradiol levels and headache activity during this period, Contrary to previous studies,4,5 the results of this study suggest that serum progesterone concentration as wall as serum estradiol concentration may be important in understanding the occurrence of menstrual migraine and menstrually-related headache. Menstrual distress also varied across the cycles for this sample. Somatic symptoms were highest at the premenstrual and menstrual phases of the cycle. Psychological/behavioral symptoms, albeit not statistically significantly, followed the same pattern. Hormonal concentrations and headache activity were associated with menstrual distress. Menstrual and premenstrual somatic symptoms were positively related to estradiol, estradiol/progesterone ratio and headache activity. Increased menstrual and premenstrual psychological/behavioral symptoms were related to increased estradiol/progesterone ratio levels and headache activity. Higher estradiol levels were associated with increased psychological/ behavioral symptoms during the menstrual phase of the cycle. The results regarding hormonal levels and menstrual distress are consistent with a study investigating menstrual distress and hormonal levels in women with high levels of menstrual distress, but who did not experience migraine headaches.23 This study provides a preliminary investigation into the covariation of headache, menstrual distress symptoms and ovarian hormones in women with menstrual migraine or menstrually-related headache. However, results need to be evaluated with a larger sample across several menstrual cycles in order to assess the stability of findings. Following a larger group of women across several menstrual cycles would allow for both predictive and pattern analysis of the relationship between headache activity, hormonal levels and menstrual distress. Inclusion of menstrual migraine, menstrually-related migraine and unrelated migraine would allow for comparative analyses. One would predict for example, that migraineurs with headache unrelated to menstrual cycle may report increased menstrual distress, but hormonal levels and headache activity would be unrelated in these women. Comparison to other related disorders, such as late luteal phase dysphoric disorder, could assist in identifying the unique hormonal and symptom complex of menstrual migraine and menstrually-related migraine. The authors would like to thank John C. Barefoot, Thomas L. Haney, Robin M. French, William B. Saun-
ders and John Feaganes for their comments of earlier drafts of this paper. This research was supported part by grant ;1 M01 RR02303-05 awarded to Dr. Penzien by the General Clinical Research Centers Program, Division of Research Resources, National Institutes of Health. REFERENCES
1.
Edelson RN: Menstrual migraine and other hormonal aspects of migraine. Headache 1985; 25:376-379.
2.
Martignoni E, Sances G, Nappi G: Significance of hormonalchanges in migraine and cluster headache Gynecol Endocrinol 1987; 1:295-319.
3.
Nattero G: Menstrual headache. Adv Neurol 1982; 33:215-226.
4.
Silberstein SD, Merrian GR: Estrogens, progestins and headache. Neurology 1991; 41:786-793.
5.
Uknis A, Silberstein SD: Review article: Migraine and pregnancy. Headache 1991; 31:372-374.
6.
Welch KMA, Darnely D, Simkins RT: The role of estrogen in migraine: A review and hypothesis. Cephalalgia 1984; 4:227-236.
7.
Digre K, Damasio H: Menstrual migraine: Differential diagnosis, evaluation, and treatment. Clin Obs Gyn 1987; 30:417-430.
8.
Diamond S, Dalessio DJ: The practicing physician's approach to headache. Baltimore: Williams and Wilkins, 1982.
9.
MacGregor EA, Chia H, Vohrah RC, Wilkinson M: Migraine and menstruation: A pilot study. Cephalalgia 1990; 10:305-310.
10.
Somerville BW: The role of estradiol withdrawal in the etiology of menstrual migraine. Neurology 1972; 22:355-365.
11.
Cockrill CL, Appel MA, Carson RR: The effect of menstrual cycle on physiological and psychological function. Paper presented at the meeting of Society of Behavioral Medicine, Boston, April 1988.
12.
Nattero G, Allais G, De Lorenzo D, Torre E, Ancona M, Benedetto C, Massobrio M: Menstrual migraine: New biochemical and psychological aspects. Headache 1988; 28:103-107.
13.
Brahams D: Pre-menstrual syndrome: A disease of the mind? Lancet 1981; 2:710-718.
14.
Glass GS, Heninger GR, Lansky M, Talan K: Psychiatric emergency related to the menstrual cycle Am J Psychiatry 1971; 128:705.
15.
Dalton K: Cyclical criminal acts in premenstrual syndrome. Lancet 2:1070, 1980.
16.
Collins A, Eneroth P, Langren B: Psychoneuroendocrine stress responses and mood as related to the menstrual cycle. Psychosom Med 1985; 47: 512-527.
17.
Penzien DB, Andrew ME, Knowlton GE, McAnulty RD, Rains JC, Johnson CA, Hursey KG, Jacks SD: Computer-aided system for headache diagnosis with the IHS headache diagnostic criteria: Development and validation. Cephalalgia 1991; 11(Suppl 11):325-326.
18.
Horowitz M, Wilren N, Alvarez W: Impact of event scale: A measure of subjective stress. Psychosom Med 1979; 41:209-218.
19.
Moos RH: The development of a Menstrual Distress Questionnaire. Psychosom Med 1968; 30:853-867.
20.
Moos RH: Typology of menstrual cycle symptoms. Am J Obstet Gynecol 1969; 103:390-402.
21.
Moos RH, Leiderman DB: Toward a menstrual cycle symptom typology. J Psychosom Res 1978; 22:31-40.
22.
Moos RH: Perimenstrual symptoms: A manual and overview of research with the Menstrual Distress Questionnaire, Social ecology Lab. Sanford University School of Medicine and V.A. Medical centers, Palo Alto, CA. 1985.
23.
Dennerstein L, Spencer-Gardner C, Brown JB, Smith MA, Burrows GD: Premenstrual tension-hormonal profiles. J Psychosom Obstet Gynaecol 1984; 3:27-51.
