Key words: estradiol/meta-chlorophenylpiperazine test/oral contraception/pill-free interval/status migrainosus

Transcription

1 Human Reproduction Vol.20, No.12 pp , 2005 Advance Access publication August 25, doi: /humrep/dei260 Estradiol supplementation modulates neuroendocrine response to M-chlorophenylpiperazine in menstrual status migrainosus triggered by oral contraception-free interval R.E.Nappi 1,2,3, G.Sances 3,4, B.Brundu 2, S.De taddei 2, A.Sommacal 2, N.Ghiotto 3,4, F.Polatti 1 and G.Nappi 3,5,6 1 Research Centre for Reproductive Medicine, University of Pavia, 2 Department of Obstetrics and Gynaecology, IRCCS San Matteo, University of Pavia, 3 University Centre of Adaptive Disorders and Headache (UCADH), University of Pavia, 4 Headache Centre, IRCCS C.Mondino Foundation, Department of Neurology and 5 Chair of Neurology, University La Sapienza, Rome, Italy 6 To whom correspondence should be addressed at: Dept Ob/Gyn, Policlinico S.Matteo, University of Pavia, Piazzale Golgi 2, Pavia, Italy. renappi@tin.it BACKGROUND: Migraine triggered by oral contraception (OC)-free interval is very common and may be extremely severe, long-lasting and poorly responsive to analgesics (status migrainosus). The serotoninergic (5-HT) system is crucially involved in pain threshold and it is sensitive to estradiol (E 2 ). Therefore, we aimed to assess neuroendocrine correlates of OC status migrainosus in response to the direct central 5-HT agonist meta-chlorophenylpiperazine (m-cpp) and to test the effect of transdermal E 2 supplementation of the OC-free interval. METHODS: Clinical investigative protocol, single-blinded placebo-controlled treatment. Oral m-cpp (0.5 mg/kg body weight) challenge test was performed in 10 patients with status migrainosus occurring within 48 h of the discontinuation of a monophasic pill (30 g of ethinyl estradiol and 150 g of desogestrel) and in six healthy women assuming the same OC as controls. In a consecutive menstrual cycle, patients with OC status migrainosus underwent to the same test after they were blindly treated with 2.0 g of percutaneous E 2 gel or placebo daily during the pill-free interval. Plasma prolactin and cortisol levels and clinical characteristics of migraine attacks were evaluated. RESULTS: Women with OC-status migrainosus showed a derangement of prolactin release (F = 4.8; P < 0.01) and a lack of cortisol response (F = 5.8; P < 0.001) after m-cpp in comparison with controls. Transdermal E 2 during the pill-free interval significantly restored prolactin (F = 2.8; P < 0.01) and cortisol responses (F = 18.9; P < 0.001) against placebo and positively affected the duration (P < 0.001), the number of hours in which pain intensity prohibits daily activity (P < 0.001), the episodes of vomiting (P < 0.001) and the consumption of analgesics (P < 0.001). CONCLUSIONS: Status migrainosus triggered by OC-free interval is associated with impaired prolactin and cortisol responses following m-cpp challenge. Transdermal E 2 supplementation is able to restore neuroendocrine response to this specific 5-HT agent, exerting a positive clinical effect on the course of menstrually related migraine. Key words: estradiol/meta-chlorophenylpiperazine test/oral contraception/pill-free interval/status migrainosus Introduction Reproductive hormonal milestones such as menarche, pregnancy and menopause, as well as hormonal interventions, often modulate the course of migraine in women (Silberstein, 1992; MacGregor, 1997). The 2nd edition of the International Classification of Headache Disorders (2004) codified that migraine or headache may increase in frequency or newly develop during the regular use of exogenous hormones, typically for contraception or HRT, or be provoked within 5 days after the cessation of a course of exogenous estrogens. The onset of migraine triggered by oral contraception (OC) during the pill-free interval is, indeed, very common in obstetrics and gynaecology practice (Sulak et al., 2000). Abrupt estrogen withdrawal has been evoked to explain the vulnerability of contraceptive users to migraine occurring during the 7 days free of exogenous estrogens (MacGregor, 2001). That being so, some hormonal techniques such as tricycling the pill, altering the ratio of estrogens to progestins, inducing medical castration with GnRH analogues, using estrogen supplementation during the pill-free week, etc., have been proposed to relieve menstrually related migraine (Murray and Muse, 1997; Massiou and MacGregor, 2000; Sulak et al., 2002). Menstrually related attacks may have a typical duration ( 72 h), but in some cases head pain can last several days and may be extremely severe and poorly responsive to analgesics (status migrainosus) (Granella et al., 2004). Migraine at menstruation is different in terms of severity from non-menstrual attacks, even within individuals (MacGregor and Hackshaw, The Author Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved For Permissions, please journals.permissions@oupjournals.org

2 R.E.Nappi et al. 2004). A similar finding may be evident during the pill-free interval in some women presenting unrelenting headache, lasting >72 h, nausea and vomiting, and showing a lack of responsiveness to conventional non-specific and specific therapy, non-steroid antiflammatory drugs and serotonin (5-HT) receptor subtype agonists, respectively. However, given the complexity of the neuroendocrine events driving the menstrual cycle, the sex steroid-dependent mechanisms involved in the vulnerability of the menstrual window to migraine are still unknown. Estrogen variations seem to be highly implicated in modulating the threshold to challenges by altering neuronal excitability, cerebral vasoactivity, pain sensitivity, neuroendocrine axes, etc., during the menstrual cycle (Silberstein and Merriam, 1999). Among the possible targets of estrogenic changes, the 5-HT system is crucially involved in neuroendocrine adaptive phenomena controlling pain threshold (Goadsby et al., 2002) and an impairment of 5-HT function has been described in menstrually related status migrainosus (Nappi et al., 2003). Indeed, we have recently shown that pituitary release of prolactin, and to an even greater extent cortisol, was significantly blunted in ovulatory women with extremely severe migraine attacks triggered by menstruation following a challenge with metachlorophenylpiperazine (m-cpp), a 5-HT agent with a high affinity for several subtypes of 5-HT receptors, especially 1 and 2 (Mueller et al., 1985), in comparison with women with menstrual migraine with typical duration of attacks (4 72 h) and controls. OC is a good model to study the menstrual window vulnerability to severe migraine attacks because the pill-free interval is a well-defined period of abrupt exogenous estrogen withdrawal (MacGregor, 2004) that may eventually represent a major insult to neuroendocrine homeostasis involving 5-HT system. The aim of the present study was to investigate the involvement of 5-HT1 and 5-HT2 receptor subtypes in the occurrence of status migrainosus triggered by OC discontinuation (OC-SM) by using m-cpp challenge test to elicit prolactin and cortisol response during the pill-free interval. In addition, we tested the neuroendocrine and clinical effects of transdermal estradiol (E 2 ) supplementation, against placebo, during the pill-free interval in a consecutive menstrual cycle. Materials and methods Subjects Among the outpatients referred to the Headache Centre of the Department of Neurology, IRCCS C.Mondino, University of Pavia, Italy, there were 10 women who entered the present study, after they had signed informed consent. Each subject filled in the International Headache Society (IHS) criteria (2004) for migraine without aura and had never assumed any prophylactic medications for migraine attacks. Patients were using specific (triptans) and non-specific (non-steroid antiflammatory drugs) symptomatic medications to extinguish migraine attacks.cardiovascular, liver, kidney, psychiatric and gynaecological diseases, as well as metabolic and endocrine dysfunction, were exclusion criteria. Patients [age: 35.4 ± 1.5 years; body mass index (BMI): 22 ± 0.9 kg/m 2 ] had severe, long-lasting migraine (>72 h) without aura attacks occurring within 48 h of the discontinuation of a monophasic OC 3424 containing 30 μg of ethinyl estradiol and 150 μg of desogestrel since at least six menstrual cycles. A retrospective clinical interview revealed that, before assuming OC treatment, each patient was probably already suffering from menstrually related migraine without aura, which has never requested a medical consultation. However, duration and severity of menstrual migraine attacks (disability and analgesic consumption) worsened significantly as soon as OC was started (status migrainosus, OC-SM). In addition, under OC the attack was regularly occurring with the menstruation and no other migraine attacks were present throughout the menstrual cycle. Six matched healthy women (age: 34.6 ± 1.9 years; BMI: 22.3 ± 1.2 kg/m 2 ) who had been using OC for 6 months and never suffering from menstrually related migraine were recruited at the Department of Obstetrics and Gynaecology, IRCCS San Matteo, University of Pavia, Italy, and served as controls, after they had given informed consent. The clinical characteristics of migraine attacks with menstruation and the pattern of bleeding were determined upon a headache diary card that patients filled in prospectively for 3 months prior to the study. Control women also filled in the diary. Figure 1 shows a representative example of a migraine diary filled in by a OC-SM patient. Study protocol The study protocol was approved by the Local University Ethical Committee, and the Department of Obstetrics and Gynaecology provided the clinical setting where m-cpp challenges were performed. Following the 3 month run-in period, each patient with OC-SM and each control woman received oral m-cpp (0.5 mg/kg body weight) challenge test 24 h after the discontinuation of the pill. m-cpp test was performed as previously described (Nappi et al., 2003). Briefly, 45 min after catheter placement into the antecubital vein, an open-label oral dose of m-cpp was administered at 09:00 (0 min). Baseline blood samples were obtained at 30 and 0 min. During the next 210 min, plasma samples were taken every 30 min. Blood samples were centrifuged and serum stored at 20 C until assayed. Side-effects throughout the m-cpp challenge tests were recorded. In a consecutive menstrual cycle, OC-SM women underwent a single-blinded placebo -controlled treatment with transdermal estradiol (E 2 ) during the pill-free interval. Indeed, five patients were randomly supplemented with 2.0 g of percutaneous E 2 gel (Sandrena, NV Organon, The Netherlands), daily for a week, starting the day of discontinuation of OC, while five patients were treated with placebo (dispensed in identical sachets) during the pill-free interval. Women were randomized by a computer-generated number list with odd and even numbers allocating patients to receive blindly transdermal E 2 or placebo. m-cpp test was performed in both groups as described above. A headache diary card was filled by patients under both E 2 or placebo treatment in order to investigate any clinical or side-effect due to the pill-free interval supplementation. Hormone assays Plasma levels of prolactin (ng/ml) and cortisol (nmol/l) were measured by radioimmunoassay using commercially available kits (Immunotech, Marseille, France); the sensitivity of the assay was 0.5 ng/ml and 10 nmol/l respectively, and the intra- and inter-assay coefficients of variation for prolactin were 2.8 and 6.5% respectively, and for cortisol were 3.4 and 6.7% respectively. Statistical analysis Data are expressed as the mean ± SE. Results reported in Table I were analysed by using Student s t-test for independent samples. The days

3 E 2 effect on prolactin and F response to m-cpp in OC status migrainosus Figure 1. Representative example of a migraine diary filled in by a patient with status migrainosus triggered by oral contraception (OC-SM)- free interval. A = analgesic; N = nausea; V = vomiting; P n = phonophobia; P t = photophobia; F = fatigue; S = stress; LS = lack of sleep; OC = oral contraception; MC = first day of expected menstrual cycle; NS = not shown. Table I. Clinical characteristics of migraine attacks triggered by oral contraception (OC) discontinuation under transdermal estradiol (E 2 ) supplementation of the pill-free interval or placebo (PL) E 2 (n = 5) PL (n = 5) P Days with migraine 2.0 ± ± 0.2 < Severity (no. of hours/attack score = 3) 12.4 ± ± 2.1 < Analgesic use 2.4 ± ± 0.4 < Nausea 1.0 ± ± 0.0 NA Vomiting 0.4 ± ± 0.0 < 0.04 Photophobia 1.0 ± ± 0.0 NA Phonophobia 1.0 ± ± 0.0 NA NA = not applicable. with menstrually related migraine in OC-SM, the severity (expressed as no. of hours/attack in which pain intensity was severe and prohibits daily activities; pain intensity was graded hourly on a 3-point scale, where 1 = mild, does not impair daily activities; 2 = moderate, may inhibit, but does not prohibit daily activities; 3 = severe, prohibits daily activities) and the analgesic use (no. of analgesics/attack), as well as the number of days of bleeding were calculated for each subject upon headache diaries over three consecutive menstrual cycles prior to the study and over the menstrual cycle investigated. The presence (no/yes) of nausea, vomiting, photophobia and phonophobia was also recorded and expressed as mean number of episodes for each subject. Figure 2 was obtained by using multiple analysis of variance (MANOVA) with repeated measures, with diagnostic group as the between-subject variable (OC-SM and controls) and time-point (baseline and 30, 60, 90, 120, 150, 180, 210 min after m-cpp) as the withinsubject variables. Figure 2C and D was obtained by using MANOVA with repeated measures, with treatment group as the between-subject variable (E 2 and placebo) and time-point (baseline and 30, 60, 90, 120, 150, 180 and 210 min after m-cpp) as the within-subject variables. Significant differences in plasma hormone levels were analyzed by Duncan s post hoc comparisons. Outcome measures were evaluated by one of us (N.G.) who was blind to the treatment regimen. 3425

4 R.E.Nappi et al. (ng/ml) PROLACTIN (nmol/l) CORTISOL 20 A 800 B C OC-SM C OC-SM 20 C 800 D min Figure 2. Prolactin (A) and cortisol (B) response to meta-chlorophenylpiperazine (m-cpp) (0.5 mg/kg) in women with status migrainosus triggered by oral contraception (OC-SM)-free interval and controls (C). Prolactin (C) and cortisol (D) response to (0.5 mg/kg) in women OC-SM blindly treated with transdermal estradiol (E 2 ) or placebo (PL) during the pill-free interval. See the text for details on MANOVA. Results Neuroendocrine effects of m-cpp in OC-SM and C MANOVA identified significant effects of time (F = 11.9; P < 0.001), diagnosis (F = 8.9; P < 0.01), and diagnosis by time interaction (F = 4.8; P < 0.01) after m-cpp administration on plasma prolactin levels (Figure 2A). Duncan s post hoc comparison revealed that prolactin response to m-cpp test was significantly blunted in OC-SM with respect to controls starting from 60 to 210 min (Figure 2A). The present observations were not altered by the analysis of covariance for age and BMI. No significant difference was found in baseline plasma prolactin levels between OC-SM and control women. Figure 2B shows significant effects of diagnosis by time interaction (F = 5.8; P < 0.001) obtained by MANOVA on plasma cortisol levels after m-cpp challenge test. Indeed, Duncan s post hoc comparison revealed that cortisol response to m-cpp test was absent in OC-SM as opposed to control women who displayed a cortisol increase starting from 90 to 210 min. Even for MANOVA on plasma cortisol levels following m-cpp, the analysis of covariance for age and BMI did not alter the results. No significant difference was found in baseline plasma cortisol levels between OC-SM and C women. Clinical effects of m-cpp in OC-SM patients and controls m-cpp was well tolerated by each patient, without significant side-effects. However, one woman with OC-SM had severe nausea, while one control woman had mild nausea as a consequence of m-cpp administration. No significant difference was found in the duration and severity (disability and analgesic consumption) of migraine attacks (data not shown) occurring after m-cpp test in comparison with the data collected during the 3 months prior to the study. It was interesting to note that patients with OC-SM had prolonged menstrual E 2 PL E 2 PL min attacks (4.3 ± 0.7 days), a high number of hours in which pain intensity prohibited daily activity (33.0 ± 2.3), extremely severe episodes of vomiting (2.7 ± 0.2) and a very high consumption of analgesics (5.9 ± 0.7). The duration of bleeding was similar between OC-SM patients (4.2 ± 0.4 days) and control women (4.1 ± 0.3 days). Figure 1 shows the clinical severity of the migraine attack occurring at menstruation in an OC- SM patient. Neuroendocrine and clinical effects of transdermal E 2 supplementation of the pill-free interval in OC-SM Transdermal E 2 treatment during the pill-free interval significantly augmented prolactin response to m-cpp in OC-SM, whereas placebo did not interfere with the pattern of prolactin response. MANOVA identified significant effects of time (F = 6.9; P < 0.001) and treatment (F = 10.2; P < 0.01), and treatment by time interaction (F = 2.8; P < 0.01) after m-cpp administration on plasma prolactin levels (Figure 2C). Cortisol response to m-cpp was completely restored by transdermal E 2 supplementation of the pill-free interval. MANOVA identified significant effects of time (F = 2.6; P < 0.02) and treatment by time interaction (F = 18.9; P < 0.001) after m-cpp administration on plasma cortisol levels (Figure 2D). It was interesting to note that when OC-SM were treated with transdermal E 2, basal plasma cortisol levels (at 30, 0 and 30 min following m-cpp administration) were significantly lower (P < 0.001) in comparison to placebo. Even though the number of OC-SM patients treated with the active drug or placebo was low, it was possible to observe a positive clinical effect of transdermal E 2 supplementation of the pill-free interval. Indeed, the duration (P < 0.002), the number of hours in which pain intensity prohibits daily activity (P < 0.001), the episodes of vomiting (P < 0.04) and the con- 3426

5 E 2 effect on prolactin and F response to m-cpp in OC status migrainosus sumption of analgesics (P < 0.001) were significantly reduced in comparison with placebo (Table I). Any relevant side-effect was evident in OC-SM women treated with transdermal E 2, apart from a modest increase in the duration of bleeding (4.3 ± 0.5 days) in comparison with placebo (3.7 ± 0.3 days; not significant). Discussion The present study demonstrated that OC-SM, a severe form of migraine triggered by pill discontinuation, is associated with an impaired neuroendocrine response following a 5-HT challenge compared with controls, and this may also occur in ovulatory women suffering from status migrainosus during the perimenstrual period. Patients with OC-SM showed a derangement of prolactin release and a lack of cortisol response after m-cpp administered during the pill-free interval. However, the supplementation with transdermal E 2 was able to restore, compared with placebo, prolactin and cortisol responses to this specific 5- HT agent. In addition, it is worth noting that transdermal E 2 supplementation of the pill-free interval was able to induce a significant improvement of OC-SM in terms of both duration and severity. That being so, the involvement of peculiar 5-HT receptor subtypes may be suggested to explain the vulnerability of some women to develop severe migraine attacks triggered by OC. Estrogen withdrawal is likely to be a good candidate in modulating pain control mechanisms mediated by 5-HT2 and 5- HT1 receptors which may be relevant for analgesic responsiveness in hormonally triggered peri-menstrual SM. Headaches are among the most common hormone withdrawal symptoms in OC users, and the course of migraine, in particular, may be affected by the use of exogenous hormones (Sulak et al., 2000, 2002). However, apart from selected circumstances, especially migraine with aura, in which OC should be used with caution to avoid the increased risk of stroke (Curtis et al., 2002), the clinical management of women with migraine under OC is quite controversial (Bousser et al., 2000; MacGregor, 2004). Gynaecologists tried some empirical approaches to manage the recurrence of migraine attacks during the pill-free interval (Murray and Muse, 1997; Massiou and MacGregor, 2000; Sulak et al, 2002), but little is known about the pathophysiological mechanisms involved in the occurrence of migraine triggered by OC free-interval at the time of menstruation. Moreover, data corroborating the E 2 withdrawal theory (Sommerville, 1975) are also lacking in the literature because of a poor bidirectional flow of information between headache centres and reproductive clinics. Martin et al. (2003) have shown that addition of transdermal E 2 to existing GnRH agonist therapy provides a modest preventive benefit in women who are pre-menopausal with migraine, suggesting the importance of minimizing hormonal fluctuations. 5-HT receptors are heterogeneous within the central nervous system and some of them are most relevant to migraine both for prevention and treatment (Goadsby et al., 2002). Sex steroid changes during the menstrual cycle clearly modulate the 5-HT system at various levels, having an impact on pain, mood, cognition and numerous other neuroendocrine and autonomic functions (McEwen, 2002). Women with menstrually related migraine display peculiar neuroendocrine responses during the luteal phase in comparison with controls (Fioroni et al., 1995) reflecting a vulnerability of the menstrual window to external or internal stimuli influencing migraine threshold when sex steroids are falling. The 5-HT system is highly involved, since platelet monoamine oxidase B activity, a good indicator of central 5-HT, is significantly increased in the luteal phase of menstrual migraine sufferers in comparison with asymptomatic controls (Fioroni et al., 1996). Our results, showing a significant imbalance of prolactin and cortisol in response to a 5-HT challenge in OC-SM women which is very similar to that already reported in fertile women with very severe menstrually related attacks (Nappi et al., 2003), are in line with the evidence that OC use and discontinuation eventually represent a trigger factor for severe migraine attacks in those women who display a vulnerable neuroendocrine homeostasis involving the 5-HT system. In this context, it is relevant to note that alterations in 5-HT metabolism are related to depression in susceptible women taking OC (Shaarawy et al., 1982), whereas borderline personality disorders significantly worse during OC are predicted by circulating E 2 levels (DeSoto et al., 2003). In addition, women using OC who experienced high levels of stress displayed an altered uptake of tryptophan in the brain during the pill-free period that may be responsible for mood changes (Tuiten et al., 1995). On the other hand, contraceptive pills may stimulate the production of vasodilatory markers, including serotonin, an effect which can be attributed most likely to the estrogenic component (Seeger et al., 2000). In addition, our data on the ability of transdermal E 2 in restoring neuroendocrine responses by exerting a concomitant positive clinical effect on the severity of OC-SM against placebo further support the notion that E 2 drop disrupts 5-HTmediated adaptive capacities implicated in the inhibition of pain. Such a mechanism may be operating not only in hormonally induced severe migraine attacks but even in naturally occurring peri-menstrual migraine. The potent interaction of the 5-HT system and even short-term E 2 treatment has been recently uncovered in women suffering from depressive symptoms at menopause (Schmidt et al., 2000) and may be a reliable explanation for the improvement of OC-SM in terms of duration and severity (disability and analgesic consumption) during E 2 supplementation. This result was beyond the scope of our study, which has been mainly focused on neuroendocrine features involving the activation of 5-HT2 and 5-HT1 receptors, but it deserves further attention in larger clinical samples. Dennerstein et al. (1988) showed in a double-blind trial that percutaneous E 2 administration for 7 days encompassing menstruation induced a significant reduction in the frequency of migraine and in the amount of pain-killers. In addition, women significantly appreciated the E 2 treatment and expressed the preference for continuing it, without reporting any relevant side-effects similarly to our present experience. In conclusion, abnormalities in the neuroendocrine response to m-cpp test during the pill-free period are related to extremely severe OC-SM attacks. E 2 supplementation by restoring modulatory pain pathways involving 5-HT2 and 5-HT1 3427

6 R.E.Nappi et al. receptors may be a useful approach to a periodically disabling condition induced or aggravated by OC use. In perspective, we believe that E 2 use may have some value in the design of preventive and/or therapeutic strategies for menstrually related migraine, especially with drugs that block or enhance 5-HT release. Acknowledgements This study was supported by a grant of the Italian Ministry of Public Health RC2003. The authors are very grateful to the Reproductive Endocrinology Unit, Dept Ob/Gyn, IRCCS San Matteo (Mrs G.Calabria and Mrs F.Vitale for help in blood sampling), to the Lab Neuroendocrinol, IRCCS C Mondino, University of Pavia, Italy (Dr G.Randine and Dr A.Smeraldi, for their expert technical assistance). The authors are also grateful to Dr G.Branca (Pavia, Italy) for preparing m-cpp. References Bousser MG, Conard J, Kittner S, de Lignieres B, MacGregor EA, Massiou H, Silberstein SD and Tzourio C (2000) Recommendations on the risk of ischaemic stroke associated with use of combined oral contraceptives and hormone replacement therapy in women with migraine. The International Headache Society Task Force on Combined Oral Contraceptives & Hormone Replacement Therapy. Cephalalgia 20, Curtis KM, Chrisman CE and Peterson HB (2002) Contraception for women in selected circumstances. Obstet Gynecol 99, Dennerstein L, Morse C, Burrows G, Oats J, Brown J and Smith M (1988) Menstrual migraine: a double-blind trial of percutaneous estradiol. Gynecol Endocrinol 2, DeSoto MC, Geary DC, Hoard MK, Sheldon MS and Cooper L (2003) Estrogen fluctuations, oral contraceptives and borderline personality. Psychoneuroendocrinology 28, Fioroni L, Martignoni E and Facchinetti F (1995) Changes of neuroendocrine axes in patients with menstrual migraine. Cephalalgia 15, Fioroni L, D Andrea G, Alecci M, Cananzi A and Facchinetti F (1996) Platelet serotonin pathway in menstrual migraine. Cephalalgia 16, Goadsby PJ, Lipton RB and Ferrari MD (2002) Migraine current understanding and treatment. N Engl J Med 346, Granella F, Sances G, Allais G, Nappi RE, Tirelli A, Benedetto C, Brundu B, Facchinetti F and Nappi G (2004) Characteristics of menstrual and nonmenstrual attacks in women with menstrually-related migraine. Cephalalgia 24, International Classification of Headache Disorders (2004) 2nd edn, Cephalalgia 2004(Suppl 1). MacGregor EA (1997) Menstruation, sex hormones, and migraine. Neurol Clin 15, MacGregor EA (2001) Hormonal contraception and migraine J Fam Plann Reprod Health Care 27, MacGregor EA (2004) Oestrogen and attacks of migraine with and without aura. Lancet Neurol 3, MacGregor EA and Hackshaw A (2004) Prevalence of migraine on each day of the natural menstrual cycle. Neurology 63, Martin V, Wernke S, Mandell K, Zoma W, Bean J, Pinney S, Liu J, Ramadan N and Rebar R (2003) Medical oophorectomy with and without estrogen addback therapy in the prevention of migraine headache. Headache 43, Massiou H and MacGregor EA (2000) Evolution and treatment of migraine with oral contraceptives. Cephalalgia 20, McEwen B (2002) Estrogen actions throughout the brain. Recent Prog Horm Res 57, Mueller EA, Murphy DL and Sunderland T (1985) Neuroendocrine effects of m-chlorophenylpiperazine, a serotonin agonist, in humans. J Clin Endocrinol Metab 61, Murray SC and Muse KN (1997) Effective treatment of severe menstrual migraine headaches with gonadotropin-releasing hormone agonist and addback therapy. Fertil Steril 67, Nappi RE, Sances G, Brundu B, Ghiotto N, Detaddei S, Biancardi C, Polatti F and Nappi G (2003) Neuroendocrine response to the serotonin agonist M-chlorophenylpiperazine in women with menstrual status migrainosus. Neuroendocrinology 78, Schmidt PJ, Nieman L, Danaceau MA, Tobin MB, Roca CA, Murphy JH and Rubinow DR (2000) Estrogen replacement in perimenopause-related depression: a preliminary report. Am J Obstet Gynecol 183, Seeger H, Ludtke R, Graser T, Wallwiener D and Mueck AO (2000) Effect of oral contraceptives on the urinary excretion of biochemical markers indicating vasoactive action. J Clin Pharm Ther 25, Shaarawy M, Fayad M, Nagui AR and Abdel-Azim S (1982) Serotonin metabolism and depression in oral contraceptive users. Contraception 26, Silberstein SD (1992) The role of sex hormones in headache. Neurology 42, Silberstein S and Merriam G (1999) Sex hormones and headache 1999 (menstrual migraine). Neurology 53, S3 S13. Sommerville BW (1975) Estrogen-withdrawal migraine. Neurology 25, Sulak PJ, Scow RD, Preece C, Riggs MW and Kuehl TJ (2000) Hormone withdrawal symptoms in oral contraceptive users. Obstet Gynecol 95, Sulak PJ, Kuehl TJ, Ortiz M and Shull BL (2002) Acceptance of altering the standard 21-day/7-day oral contraceptive regimen to delay menses and reduce hormone withdrawal symptoms. Am J Obstet Gynecol 186, Tuiten A, Panhuysen G, Koppeschaar H, Fekkes D, Pijl H, Frolich M, Krabbe P and Everaerd V (1995) Stress, serotoninergic function, and mood in users of oral contraceptives. Psychoneuroendocrinology 20, Submitted on April 28, 2005; resubmitted on July 4, 2005; accepted on July 18,