bmj.com Rapid Responses for Etminan et al., 0 (2004) 383025040

Rapid Responses to:

PAPERS:
Mayhar Etminan, Bahi Takkouche, Francisco Caamaño Isorna, and Ali Samii
Risk of ischaemic stroke in people with migraine: systematic review and meta-analysis of observational studies
BMJ 2004; 0: bmj.38302.504063.8Fv1 [Abstract]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Oral contraceptive drug type/ dose response data: sarah j mee (15 December 2004)

Hormonal contraceptives cause migraine and ischaemic strokes.: Ellen C G Grant (16 December 2004)

Pooled risk of ischaemic stroke is misleading: E Anne MacGregor (20 December 2004)

Several inconsistencies and errors in the study on migraine and risk of ischemic stroke: Tobias Kurth, and Julie E. Buring. (22 December 2004)

Re: Several inconsistencies and errors in the study on migraine and risk of ischemic stroke: Tobias Kurth, and Julie E. Buring (25 December 2004)

Stroke and migraine: meta-analysis versus basic sciences?: Vinod K Gupta (29 December 2004)

Is it migraine or stroke or both?: Dirk Ulbricht (31 December 2004)

Risk of ischaemic stroke in people with migraine: Authors' response: Mahyar Etminan (4 January 2005)

Most Common Headache?: James Hickman (12 January 2005)

Hormonal contraceptives cause migraine, ischaemic and haemorrhagic strokes.: Ellen C G Grant (13 January 2005)

Migraine; a contraindication to combined oral contraception: Helen K Crispin (4 February 2005)

Oral contraceptive drug type/ dose response data 15 December 2004

sarah j mee,
Pharmacovigilance Manager/ Physician
Alphapharm
Send response to journal:
Re: Oral contraceptive drug type/ dose response data

Could you please inform me if there was any difference found in stroke risk in migraine sufferers taking the Progesterone-only pill compared to the Combined Oral Contraceptive pill compared to a combination product such as ethinyloestradiol/cyproterone acetate. From the metaanalysis one would assume the risk was independent of the type of oral contraceptive involved. Is this correct?
Also, was any dose analysis done? Did low dose contraceptive pills have a lower risk of stroke than high dose preparations?
If the strokes were ischaemic in aetiology would the increased risk of thromboembolism in patients on oral contraception be a possible mechanism for the association? This is not mentioned in the paper. Many Thanks in advance.
Competing interests: None declared

Hormonal contraceptives cause migraine and ischaemic strokes. 16 December 2004

Ellen C G Grant,
physicain and gynaecologist
20 Coombe Ridings, Kingston-upon-Thames, KT2 7JU, UK
Send response to journal:
Re: Hormonal contraceptives cause migraine and ischaemic strokes.

Etminan and colleagues� useful meta-analysis finds that oral contraceptive (OC) use and migraine attacks confer an 8 times increased risk of ischemic stroke compared with headache-free non-users of OCs.1
I investigated the effects of over 60 doses and combinations of seven progestogens and two oestrogens on endometrial glands and blood vessels in the 1960s. When the dose of oestrogen was unchanged, increasing progestogen doses produced peak incidences of firstly irregular bleeding,2 then venous changes including thrombosis,3 then arteriolar changes with headaches, migraine, strokes and heart attacks,4 and finally depression and loss of libido with the highest doses.5 The incidence of headaches and migraine in untreated cycles was 18% (10% and 7%).6 Up to 60% of women deveoped severe headaches during the first year of taking a some combinations. Well developed arterioles were apparent in their endometrial biopsies.4 Venous sinusoids became increasingly dilated before episodes of venous thrombo-phlebitis or thrombosis.3 All oral contraceptives are predominantly progestogenic and have high monoamine oxidase activity for most of the cycle, instead of only for a few days premenstrually in the normal cycle.5
Progestogen-only pills cause less migraine but more irregular bleeding. In a trial of norgestrel 0.05 mg, norethisterone acetate 0.3 mg, megestrol acetate 0.25 mg, and chlormadinone 0.5 mg, the norgestrel dose produced the most arteriolar development and dilatation of venous sinusoids. Irregular bleeding and breast tenderness were common reasons for high drop out rates, while troublesome headaches affected 5 to 11 % of women.7 Also progesterones given alone causes more depression, vaginal dryness and breast cancer than oestrogens.
In the 1970s we found that women taking OCs or other hormone preparations had more migraine attacks than smokers, who more usually have tension headaches. The commonest cause of frequent, severe migraine attacks needing acute treatment was the use of ergot medications (oestrogenic) and OCs or HRT. Stopping taking hormones reduced migraine attacks from 424 to 34, stopping ergotamine from 530 to 52 and stopping smoking from 169 to 16. Reducing migraine attacks 10-fold is a dramatic clinical result.8
Either men or women with headaches are likely to be zinc and/or magnesium deficient, which also commonly causes functional B vitamin deficiencies and blocks in essential fatty acid pathways.9 These impairments to normal homeostasis, amine metabolism and immunity cause increased reactions to common foods and chemicals. The anti-phospholipid syndrome is often due to OC or HRT use and patients can be helped by and repletion of nutrient deficiencies and avoidance of food and chemical allergens and any further hormone use.10,11
HRT is now regarded as too dangerous to use in randomised controlled trials. Why is use of carcinogenic and immunosuppressive progesterones and oestrogens still being used for contraception? It is a poor excuse that the incidence of disease is lower in younger women.
1. Etminan M, Takkouche B, Caama�o Isorna F, Samii A. Risk of ischaemic stroke in people with migraine: systematic review and meta-analysis of observational studies BMJ 2004; 0: bmj.38302.504063.8Fv1
2 Grant ECG. Hormone balance of oral contraceptives. J Obstet Gynaecol Brit Comm 1967;74:908-18.
3 Grant ECG. Relation between headaches from oral contraceptives and development of endometrial arterioles. BMJ 1968;3:402-405.
4 Grant ECG, Pryce Davies J. Effect of oral contraceptives on depressive mood changes and on endometrial monoamine oxidase and phosphatases. BMJ 1968;3:777-80.
5 Grant ECG. Venous effects of oral contraceptives. BMJ 1969;2:73-77.
6 Grant ECG. Relation of arterioles in the endometrium to headaches from oral contraceptives. Lancet 1965;1:1143-44.
7 Mears E, Vessey MP, Andolsek L, Oven A. Preliminary evaluation of four oral contraceptives containing only progestogens. BMJ 1969;2:730-34 (Grant ECG-pathology)
8 Grant ECG. Oral contraceptives, smoking, migraine, and food allergy. Lancet 1978; 2:581-582.
9 Grant ECG. The pill, hormone replacement therapy, vascular and mood over- reactivity, and mineral imbalance. J Nutr Environ Med 1998; 8: 105-116.
10 Grant ECG. Systemic lupus erythematosus. Lancet 2001;358:586.
11 Grant ECG .Food allergy and migraine. Lancet 1979; 2: 358-59.
Competing interests: None declared

Pooled risk of ischaemic stroke is misleading 20 December 2004

E Anne MacGregor,
Director of Clinical Research
The City of London Migraine Clinic EC1M 6DX
Send response to journal:
Re: Pooled risk of ischaemic stroke is misleading

EDITOR,
I was hopeful that the meta-analysis of ischaemic stroke in people with migraine would provide a useful quantification of the risk.(1) This has not been the case and I am concerned that others reading the paper may quote the figures without knowledge of its severe limitations. This is notwithstanding the errors in Table 1 (listed below) and assumes that the correct figures were used in the meta-analysis.
Potential bias and significant differences between the studies of migraine and ischaemic stroke have already been reported.(2) However, a meta-analysis should still be possible provided that like populations are compared. Unfortunately, the overall relative risk reported in this meta-analysis appears to combine data from several quite different populations: one study reported only on men, who were aged 40-84 (3); some reported on the combined risk for men and women, who were aged 18-80,(4) 16-60,(5)15-65,(6) or 15-44 (7, 8); some reported only on women, mostly aged 15-44,(7, 9-11) but also 20-44,(12) and 15-49.(13)
Further concerns relate to the inclusion or exclusion of TIA. Some studies included TIA,(7-9, 11, 14) others specifically excluded TIA.(4, 5, 10, 12, 13, 15) Then there is confusion regarding whether stroke, TIA, or both are used in the analysis. In the Carolei study OR for any migraine, after adjusting for risk factors, included both stroke and TIA (1.9 [1.1 to 3.1]).(7) However, the ORs given for migraine with aura and migraine without aura are unadjusted, relating only to stroke risk and excluding TIA. The CGGS included haemorrhagic stroke.(9) Patients with cardiac abnormalities were sometimes excluded,(6) whereas figures quoted from another study included patients with cardiac and carotid disease.(8)
Diagnostic uncertainty of migraine is another bias. Some studies diagnosed migraine by personal interview using a questionnaire based on IHS criteria,(4, 7, 10, 12) Other studies predated these criteria.(6, 9). Some based their analysis on self-reported migraine.(3, 11) Nightingale identified migraine if the women �had a prescription for a specific antimigraine therapy or diagnosis of migraine and a prescription for a potential antimigraine medication or analgesics in the absence of any other explanatory diagnosis within a year before the index date.�(13) Thus women who had not consulted their doctor for migraine or who had not had migraine previously diagnosed were not identified. This is important since only around half of people with migraine consult a doctor.(16) Then there is concern about accurate diagnosis of migraine aura. Unless carefully questioned, prodromal symptoms that precede both migraine with aura and migraine without aura may be confused with aura. This could dilute the true magnitude of risk associated with aura.
Stroke was also diagnosed in different ways. Some was self- reported,(3, 17) one was based on the National Patient Register,(11) whereas most others were based on hospital diagnosis, although the duration of symptoms for the diagnosis varied from more than 24 hours to more than 48 hours. One study was particularly strict with the diagnosis of stroke made only after a second confirmatory scan.(14) Previous history of stroke was excluded in most but not all studies. Some studies stated that migrainous stroke was excluded,(7, 14) whereas others included it.(5, 6)
Frequency, recent and past type of migraine are important, as Donaghy and Kruit have shown.(18, 19) Chang included an incident history of migraine, whereas most others assessed only prevalent migraine. Lidegaard only included migraine that occurred more often than once a month.(11) These concerns, among others, lead me to consider that the pooled results presented in this meta-analysis are not only unhelpful but are misleading.
What can be concluded is what is already known: migraine is an independent risk factor for ischaemic stroke. The magnitude of risk remains unclear, particularly for important sub-populations. Those of us who make clinical recommendations for women with migraine taking combined hormonal contraception are keen to have quantification of these risks.(20) We need accurate data in order not to increase further the risk of ischaemic stroke in women already at risk, such as women with migraine with aura, while ensuring that women at low risk of stroke are not unnecessarily denied combined hormonal contraception.
Errors in table 1:
� Carolei: OR for migraine with aura should be 8.6 [1.0 to 75]; OR for migraine without aura should be 1.0 [0.5 to 2.0].
� Chang: OR for migraine with aura should be 3.81(1.26-11.5)
� Nightingale: OR quoted is from the case-control study, not cohort.
1. Etminan M, Takkouche B, Caama�o Isorna F, Samii A. Risk of ischaemic stroke in people with migraine: systematic review and meta-analysis of observational studies. BMJ 2004:bmj.38302.504063.8Fv1.
2. Tzourio C, Kittner SJ, Bousser M-G, Alp�rovitch A. Migraine and stroke in young women. Cephalalgia 2000;20:190-199.
3. Buring JE, Hebert P, Romero J, Kittross A, Cook N, Manson J, et al. Migraine and subsequent risk of stroke in the physician's health study. Arch Neurol 1995;52:129-134.
4. Tzourio C, Igl�sias S, Hubert J-B, Visy J-M, Apl�rovitch A, al. e. Migraine and risk of ischaemic stroke. BMJ 1993;307:289-92.
5. Haapaniemi H, Hillbom M, Juvela S. Lifestyle-associated risk factors for acute brain infarction among persons of working age. Stroke 1997;28:26-30.
6. Henrich JB, Horwitz LA. A controlled study of ischemic stroke risk in migraine patients. J Clin Epidemiol 1989;42:773-80.
7. Carolei A, Marini C, de Matteis G, and the Italian National Research Council Study Group on Stroke in the Young. History of migraine and risk of cerebral ischaemia in young adults. Lancet 1996;347:1503-1506.
8. Marini C, Carolei A, Roberts RS, Prencipe M, Gandolfo C, Inzitari D, et al. Focal cerebral ischemia in young adults: a collaborative case-control study. Neuroepidemiology 1993;12:70-81.
9. Collaborative Group for the Study of Stroke in Young Women. Oral contraceptives and stroke in young women: associated risk factors. JAMA 1975;231:718-22.
10. Tzourio C, Tehindrazanarivelo A, Igl�sias S, Alp�rovitch A, Chedru F, d'Angle-Chatillon J, et al. Case-control study of migraine and risk of ischaemic stroke in young women. BMJ 1995;310:830-3.
11. Lidegaard �. Oral contraceptives, pregnancy and the risk of cerebral thromboembolism: the influence of diabetes, hypertension, migraine and previous thrombotic disease. Br J Obstet Gynecol 1995;102:153-9.
12. Chang C, Donaghy M, Poulter N, and World Health Organisation Collaboration Study of Cardiovascular Disease and Steroid Hormone Contraception. Migraine and stroke in young women: case-control study. BMJ 1999;318:13-8.
13. Nightingale AL, Farmer RDT. Ischemic stroke in young women. Stroke 2004;35:1574-1578.
14. Schwaag S, Nabavi DG, Frese A, Husstedt I-W, Evers S. The association between migraine and juvenile stroke: a case-control study. Headache 2003; 43:90-95.
15. Schwartz SM, Petitti DB, Siscovick DS, Longstreth Jr WT, Sidney S, Raghunathan TE, et al. Stroke and use of low-dose oral contraceptives in young women. A pooled analysis of two studies. Stroke 1998;29:2277-2284.
16. MacGregor EA, Brandes J, Eikermann A. Migraine prevalence and treatment patterns: the global mgiraine and zolmitriptan survey. Headache 2003;43:19-26.
17. Merikangas KR, Fenton BT, Cheng SH, Stolar MJ, Risch N. Association between migraine and stroke in a large-scale epidemiological study of the United States. Arch Neurol 1997;54:362-368.
18. Donaghy M, Chang CL, Poulter N, on behalf of the European Collaborators of The World Health Organisation Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Duration, frequency, recency, and type of migraine and risk of ischaemic stroke in women of childbearing age. J Neurol Neurosurg Psychiat 2002;73:747-750.
19. Kruit MC, van Buchem MA, Hofman PAM, Bakkers JTN, Terwindt GM, Ferrari MD, et al. Migraine as a risk factor for subclinical brain lesions. JAMA 2004;291(4):427-434.
20. MacGregor EA, Guillebaud J. Recommendations for clinical practice. Combined oral contraceptives, migraine and ischaemic stroke. Br J Fam Planning 1998;24:53-60.
Competing interests: None declared

Several inconsistencies and errors in the study on migraine and risk of ischemic stroke 22 December 2004

Tobias Kurth,
Instructor of Medicine
Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02215, USA,
and Julie E. Buring.
Send response to journal:
Re: Several inconsistencies and errors in the study on migraine and risk of ischemic stroke

To the Editor,
We have read with great interest the systematic review and pooled analysis by Etminan and colleagues of studies evaluating the association between migraine and ischemic stroke (1). The authors pooled data from 14 studies, and classified 11 of these as case-control and 3 as cohort. The pooled relative risk (RR) estimate of ischemic stroke associated with overall migraine was 2.10 (95% confidence interval [CI], 1.61-2.75).
This pooled analysis, however, has several inconsistencies and errors that call the accuracy of their conclusions into question. First, the RR from the study by Merikangas and colleagues (2) is a cross-sectional estimate not the result of a cohort analysis, and following the rules of the pooled analysis by Etminan and colleagues should not have been included. Second, the data in table 1 are confusing and have several mistakes. First, it is unclear to which cases and controls the last two columns are referring. The common convention from case-control studies is that a �case� is a patient with the outcome (here a stroke), and a �control� a patient without the outcome. In the second to last column, �cases� among case-control studies are correctly listed as patients with stroke relative to the number of control persons without stroke. However, for the cohort studies, the �cases� appear to refer to patients with migraine not patients with stroke. For example, in the study by Buring and colleagues, Etminan and colleagues reported 1479 cases and 20,481 cohort size. However, the 1479 is in fact the number of participants reporting migraine and the 20,481 is the number not reporting migraine rather than the total cohort size (3). Furthermore, the number of �cases� in the last column does not always refer to ischemic stroke cases as the title of the table indicates. For example, the study by Chang included 291 women with ischemic, hemorrhagic, or unclassified arterial stroke (4). Further, in the study by Buring and colleagues, a total of 171 participants had ischemic strokes but Etminan and colleagues report instead the number of total strokes from that study. Second, there were errors in the reporting of the individual study findings. The study by Chang and colleagues reported a RR of 3.81 (95% CI, 1.26-11.5) for ischemic stroke associated with migraine with aura, which is not the value reported in the table, i.e., 2.97 (95% CI, 0.66-13.5). Moreover, the RRs presented for migraine with aura and without aura from the study by Carolei and colleagues are interchanged, and moreover, represent the crude risk of ischemic stroke, whereas the RR associated with any migraine represents the adjusted risk of ischemic stroke and transient ischemic attack (5). If the authors had been consistent, the overall RR for Carolei�s data on total migraine should have been 1.3 (95% CI, 0.7-2.4) and not 1.9 (95% CI, 1.1-3.1) as indicated in the table. In the study by Schwaag, the overall RR is 2.11 and not 2.16 as indicated in the table (6). Finally, the study by Donaghy and colleagues (7) provided a RR for migraine with aura, but this was incorrectly indicated in the table as not specified.
These numerous easily identifiable contradictions call into question the results of the pooled estimates since it is unclear which data have been used for the analysis. Because this study is of great importance and the results are likely to be cited frequently, we urge the authors to redo the analysis and confirm their pooled estimates.
References
1. Etminan M, Takkouche B, Caamano I, Samii A. Risk of ischaemic stroke in people with migraine: systematic review and meta-analysis of observational studies. BMJ. 2004.
2. Merikangas KR, Fenton BT, Cheng SH, Stolar MJ, Risch N. Association between migraine and stroke in a large-scale epidemiological study of the United States. Archives of Neurology. 1997;54(4):362-8.
3. Buring JE, Hebert P, Romero J, et al. Migraine and subsequent risk of stroke in the Physicians' Health Study. Archives of Neurology. 1995;52(2):129-34.
4. Chang CL, Donaghy M, Poulter N. Migraine and stroke in young women: case-control study. The World Health Organisation Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. BMJ. 1999;318(7175):13-8.
5. Carolei A, Marini C, De Matteis G. History of migraine and risk of cerebral ischaemia in young adults. The Italian National Research Council Study Group on Stroke in the Young. Lancet. 1996;347(9014):1503-6.
6. Schwaag S, Nabavi DG, Frese A, Husstedt IW, Evers S. The association between migraine and juvenile stroke: a case-control study. Headache. 2003;43(2):90-5.
7. Donaghy M, Chang CL, Poulter N. Duration, frequency, recency, and type of migraine and the risk of ischaemic stroke in women of childbearing age. J Neurol Neurosurg Psychiatry. 2002;73(6):747-50.
Competing interests: None declared

Re: Several inconsistencies and errors in the study on migraine and risk of ischemic stroke 25 December 2004

Tobias Kurth,
Instructor of Medicine
Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02215, USA,
and Julie E. Buring
Send response to journal:
Re: Re: Several inconsistencies and errors in the study on migraine and risk of ischemic stroke

To the Editor,
In our previous response, we reported the pooled estimate from the cohort studies (RR=2.10; 95% 1.61-2.75) of the study by Etminan and colleagues. The pooled risk estimate from the case-control studies was 2.18 (95% CI, 1.86-2.56) and the pooled risk estimate from all 14 studies was 2.16 (95% CI, 1.89-2.48).
Competing interests: None declared

Stroke and migraine: meta-analysis versus basic sciences? 29 December 2004

Vinod K Gupta,
Physician
Dubai Police Medical Services, PO Box 12005, Dubai, United Arab Emirates
Send response to journal:
Re: Stroke and migraine: meta-analysis versus basic sciences?

Etminan et al (1) offer a systematic review and meta-analysis of occurrence of stroke in patients with migraine. The only feature that is added by this study is a further clarification of the uncertainty with which stroke might affect patients with migraine, with or without aura; this unpredictability is also extended to female migraine patients using oral contraceptives. Incidence of migraine peaks in the young pre- menopausal female adult and a higher risk of stroke in this cohort � underscored as a significant advance in understanding of migraine mechanisms or phenomenology in this article � simply reflects the effect of occurrence of the illness. In other words, in no female migraine patient can we reasonably predict the occurrence of stroke following use of oral contraceptives. These authors recommend further studies to elucidate these associations but do not specify or clarify what needs to be studied. In their elucidation of the limitations of their analysis, these authors do not acknowledge any of crucial intrinsic limitations of mathematical logic as well as meta-analyses , (2,3) while making further several new methodological assumptions and resting part of their analysis on statistical heterogeneity. Fashions die hard, and the arrival of meta- analysis in medicine manifests humankind�s unabating love for numbers and mathematical analyses but heralds blunting of critical biological analytic ability.
The role of the platelet in migraine is poorly understood. Migraine prophylactic agents such as propranolol promote platelet aggregation but do not predispose to stroke. (4,5). Platelets probably aggregate during every migraine attack but the rarity of ischaemic stroke in migraine patients attenuates their pathophysiological significance in the genesis of this complication. Antiphospholipid antibodies are even more poorly understood in the context of overall predisposition to associated diseases and migraine pathophysiology is no exception. (6) Endless reiteration in the medical literature of these popular concepts has not yielded any insight into disease mechanisms, particularly for migraine. (7)
Quite clearly, there is a genuine physio-anatomical trait � rather than a state-related aberration -- that definitely but only rarely predisposes to development of ischaemic stroke in a particular individual. The predominance of posterior-circulation infarcts in a younger cohort (i.e., migraine patients) suggests the critical role of an anatomic predisposition. The posterior cerebral artery is particularly labile at its origin and is most likely to be affected by any vasoconstrictive influence, rarely reducing perfusion critically to randomly manifest as ischaemic stroke in migraine patients. (8) Simply because such a dynamic perfusion anomaly cannot be measured in routine or under controlled conditions during research, it does not mean that it is unimportant or that it does not ever occur. At the current stage of its understanding, migraine is not a suitable subject to be studied meaningfully through meta -analysis. (3,7,9)
I fervently hope Etminan et al (1) themselves and other responding scientists will vigorously challenge my assertions to produce a lively and illuminating debate rather than the fence-sitting through refusal-to- respond or the enlightened obfuscation that characterizes much of comment in scientific correspondence today. Splitting hairs on the mode or merits of meta-analysis is not the way forward in evolving migraine pathophysiology; it is essential to know and accept the limits of one�s investigative tools. (2) A therapeutically useful quantification of stroke risk of oral contraceptives is virtually impossible because of the multitude of factors that affect stroke incidence; when those variables are combined with the legion of variables that influence migraine predisposition and precipitation, we are saddled with a task that cannot be resolved either by randomized controlled trials or meta-analysis.
References
1. Etminan M, Takkouche B, Isorna FC, Samii Ali. Risk of ischaemic stroke in people with migraine: systematic review and meta-analysis of observational studies. BMJ, doi:10.1136/bmj.38302.504063.8F (published 20 December 2004).
2. Gupta VK. Randomized controlled trials: the hijacking of basic sciences by mathematical logic. BMJ Online. Available at: http://bmj.bmjjournals.com/cgi/eletters/329/7456/2#65969
3. Gupta VK. Meta-analysis of role of metoclopramide for aborting migraine attacks: quo vadis? http://bmj.bmjjournals.com/cgi/eletters/bmj.38281.595718.7Cv1
4. Gupta VK. Platelet-leukocyte adhesion, migraine, and stroke: a bioclinical perspective. J Neurol Neurosrug Psychiatry 2004 [Online 12 July 2004]Available at: http://jnnp.bmjjournals.com/cgi/eletters/75/7/984#197
5. Gupta VK. Platelets and migraine: seeing the trees for the wood? Headache 2005 (In press).
6. Gupta VK. Migrainous stroke: are antiphospholipid antibodies pathogenetic, a biological epiphenomenon, or an incidental laboratory aberration? European Neurology 1996; 36: 110-111.
7. Gupta VK. Classification of primary headaches: pathophysiology versus nosology? BMJ 2004; published online Jan 22. Available at: http://bmj.bmjjournals.com/cgi/eletters/328/7432/119
8. Gupta VK. Regional cerebral blood flow patterns in migraine: what is the contribution to insight into disease mechanisms? European Journal of Neurology. 1995; 2: 586--587 (comments 588-589).
9. Gupta VK. Bureaucratisation of migraine. Lancet Neurology. 2004;3:396.
Competing interests: None declared

Is it migraine or stroke or both? 31 December 2004

Dirk Ulbricht,
Neurologist
Service de Neurologie, Centre Hospitalier Emile Mayrisch, L-4240 Esch-sur-Alzette, Luxembourg
Send response to journal:
Re: Is it migraine or stroke or both?

I greatly appreciate the analysis of Etminan et al, but I think the discussion needs to be extended: (1) it appears odd to address medical treatment of migraine and stroke by beta-blockers as confounders. Whether preventive treatment of migraine and successful remission of active migraine leads to fewer strokes, is poorly examined, as stroke in young migraineurs remains rare. (2) A really confounding issue is the presence or absence of persistent foramen ovale (PFO) and related conditions, which might account for recurrent stroke. Interestingly, preliminary data suggests that people suffering from migraine may benefit from surgical or transcatheter closure of a PFO, as well as anticoagulation with a vitamin-K-antagonist, but this issue remains to be validated really. (3) Whether we deal with migraineous stroke or stroke in a migraineur, is a really confounding issue: the current definition of the international headache society defines migraineous stroke as persisting aura symptoms for > 24 hours with confirmation in imaging (or autopsy).
Whether a patient with simple migraine presents an aura for the first time or a first stroke is unclear, and leads, in my experience to some admission on stroke units. (4) It has to be noted that anticontraceptives per se are prothrombotic as they can worsen migraine. (5) Antiphospholipid antibodies per se do not carry an elevated risk for ischemic stroke as demonstrated by a recent study. (6) In a recent article appeared in the JNNP, the authors demonstrated that the risk of extracranial artery dissection remains the same during life, but the accumulation of atherosclerotic risk factors lead to disappearance of dissections as cause of stroke in later life. So the statistical effect of migraine related to contraceptives might be due to the same effect amplified by the fact that anticonception is merely necessary after the age of 45.
Competing interests: None declared

Risk of ischaemic stroke in people with migraine: Authors' response 4 January 2005

Mahyar Etminan,
Fellow
Division of Clinical Epidemiology, Royal Victoria Hospital, Montreal
Send response to journal:
Re: Risk of ischaemic stroke in people with migraine: Authors' response

Editor,
We like to thank the authors of the letters who highlighted the errors in reporting the individual relative risks in Table 1 of our meta- analysis (1). We had mistakenly published the first draft of the table instead of the final draft. We apologize for this mistake that fortunately did not change the results and the core message of our meta-analysis. Except for one figure, the correct numbers had been used in the analysis. The corrected pooled relative risk for migraine with aura is 2.88 (1.89 to 4.39) compared to the previous 2.27, and 1.56 (1.03 to 2.36) for migraine without aura compared to the previous 1.83. The rest of the results have not changed. We have incorporated the new results in both the short and long versions of our manuscript and they will be published in the BMJ shortly.
Kurt and Buring suggest that the study by Merikangas et al (2) should not have been included in our meta-analysis. Although this was initially a survey, the authors point out in the text of the article that �Table 3 presents the analysis of the association of baseline risk factors (including headache) and migraine and stroke reported at follow-up�. Excluding this study from the analysis would not have resulted in any meaningful change in the results of our meta-analysis (pooled relative risk including this study is 2.16 [1.89 to 2.48] and the pooled relative risk without the study is 2.17 [1.87 to 2-53]). The number of cases/controls or cases/cohort size, presented in the last two columns of table1 are not restricted to ischaemic stroke. They simply provide information on the study size as advised by the reviewers; they were not used further in calculation of the pooled relative risk.
As for the study by Carolei et al (3), we followed the rule of always presenting the relative risk estimate that is adjusted by the largest number of variables. If only crude estimates were presented for a sub- group, they were included in the analysis. In Carolei et al., the only data available on migraine with and without aura are those that we presented in table 1. We could have either included data on �any migraine� using the adjusted relative risk the authors presented (partly contaminated by transient ischemic attacks) or the non-contaminated crude relative risk. Since some authors lump together all forms of ischaemic events (4-6), we opted for presenting only the adjusted relative risk of �cerebral ischemia�. Had we opted for using the crude relative risk of 1.3 as suggested by Kurt and Buring instead of the adjusted relative risk of 1.9, the final pooled estimate would have been 2.13 (1.85 , 2.44) instead of 2.16 (1.89 to 2.48) which does not result in a substantial change.
The concerns raised by MacGregor are inherent to meta-analytic procedures as a whole. Pooling only those studies that were carried out on populations with similar base-line characteristics is meaningless. The danger of such an analysis would be pooling similar relative risks across all the studies and potentially defeating the purpose of a meta-analysis. On the contrary, exploring sources of heterogeneity in the different pooled estimates (alternatively using random effects models when heterogeneity is discovered), is probably the most interesting part of conducting a meta-analysis. As for the study by Nightingale et al (7), this is a nested case-control study within a well-defined cohort. It is, therefore, more similar to a cohort study than to a case-control study.
Mee�s questions cannot be answered with the studies that were available in this meta-analysis. Details on the types and dosages of different contraceptives pills were not presented in the studies.
Mahyar Etminan
Bahi Takkouche
Ali Samii
References
1.Etminan M, Takkouche B, Caama�o Isorna F, Samii A. Risk of ischaemic stroke in people with migraine: systematic review and meta- analysis of observational studies BMJ 2004;0: bmj.38302.504063.8Fv1
2.Merikangas KR, Fenton B, Cheng SH, Stolar MJ, Rish N. Association between migraine and stroke in a large-scale epidemiological study of the United States. Arch Neurol 1997;54:362-8.
3.Carolei A, Marini C, De Matteis G. History of migraine and risk of cerebral ischaemia in young adults. Lancet 1996;347:1503-6.
4.Diener HC, Bogousslavsky J, Brass LM, et al. Aspirin and clopidogrel compared with clopidogrel alone after recent ischaemic stroke or transient ischaemic attack in high-risk patients (MATCH): randomised, double-blind, placebo-controlled trial. Lancet 2004; 364:331-7
5.Schwaag S, Nabavi DG, Frese A, Husstedt IW, Evers S. The association between migraine and juvenile stroke: a case-control study. Headache 2003;43:90-5.
6.Marini C, Carolei A, Roberts RS, Prencipe M, Gandolfo C, Inzitari D, et al. Focal cerebral ischemia in young adults: a collaborative case- control study. Neuroepidemiology 1993;12:70-81.
7.Nightingale AL, Farmer RD. Ischemic stroke in young women: a nested case-control study using the UK general practice research database. Stroke 2004;35:1574-8.
Competing interests: None declared

Most Common Headache? 12 January 2005

James Hickman,
GP Principal
North Curry Health Centre, North Curry, Taunton, TA3 6NQ
Send response to journal:
Re: Most Common Headache?

I'm afraid this paper lost me in the first sentence of it's introduction: "Migraine is the most common type of headache in young adults.."
I'm not inclined to pay the $30 required to read the Lancet article referenced. However: Hands up everyone who has had headache in young adulthood. Now hands down everyone whose headache wasn't a migraine. I doubt you'd be left with a majority.
Competing interests: None declared

Hormonal contraceptives cause migraine, ischaemic and haemorrhagic strokes. 13 January 2005

Ellen C G Grant,
physician and medical gynaecologist
Kingston-upon-Thames, KT2 7JU., UK
Send response to journal:
Re: Hormonal contraceptives cause migraine, ischaemic and haemorrhagic strokes.

Dr Hickman is annoyed that Etminan et al�s first sentence is, �Migraine is the most common type of headaches in young adults�, because headaches are obviously more common than migraine in the general population.
Migraine becomes the most common type of headache in young adults attending clinics. This is because doctors prescribe ergot medications, or, like Anne Macgregor, want to �ensure that women at low risk of stroke are not unnecessarily denied combined hormonal contraception�. Do neurologists not appreciate that progesterones are immunosuppressive, vasoactive and carcinogenic and headaches or migraines warn of intolerance?
Etminan et al�s estimated prevalence of migraine is 4% before puberty and as high as 25% in women by their mid to late 30s�. However before the Pill was tested in fertile women in the 1960s in London the migraine incidence was only 7%.2
The synergistic effects of ergot medications, exogenous hormones, smoking and were clearly demonstrated in the 1970s at Charing Cross Hospital Migraine Clinic. These were the main precipitants of frequent severe migraines in groups of 30 patients over a 3 month period and were the main reasons for patients needing emergency hospital treatment for acute attacks.3,4
On questioning many patients told me that their menstrual headaches had changed into frequent severe migraine attacks after attending Migraine Clinics. Ergot medications causes such severe nausea and vomiting in about half of the patients, that further is avoided. Other patients had continued to use ergot for more frequent and severe attacks until, after a few years of �treatment�, many of them were taking ergot several times each week or even daily. This is why ergot users totalled the most frequent migraine attacks over a 3 month period if the need to take ergot was regarded as due to prodromal signs of an imminent attack. We found that ergot uses tended to have abnormally high or low red blood cell levels of the enzyme catechol-o-methyl transferase. Progesterone users have elevated monoamine oxidase activities. Both of these changes in enzyme activities impair amine pathways.
Smoking alone causes more tension headaches than migraine. Patients susceptible to headaches had usually learnt to avoid alcoholic drinks. Pettiti et al found a synergistic effect of �low-estrogen� oral contraceptive preparations and smoking, with an increased risk of haemorrhagic stroke of 3.64 (0.95-13.87). 5 All available hormonal contraceptives are progesterone dominant.6
In the 1960s the third woman I gave the Pill to developed migraine and benign intracranial hypertension and the twelfth developed hypertension and a stroke which was due to cerebral artery thrombosis with a secondary haemorrhage. Her brain arterial and venous changes were similar to Pill-induced endometrial blood vessel changes and are pictured in the 1998 June issue of the Journal of Nutritional and Enviromental Medicine, which is available at www.tandf.co.uk/journals.7
Progestogen-only emergency contraceptives can also cause severe long- lasting migraine attacks and immunosuppression, in my experience.
1 Etminan M, Takkouche B, Isorna FC, and Samii A. Risk of ischaemic stroke in people with migraine: systematic review and meta-analysis of observational studies. BMJ 2005; 330: 63.
2 Grant ECG. Clinical review of headache. http://bmj.com/cgi/eletters/325/7369/881#26539, 27 Oct 2002.
3 Grant ECG, Albuquerque M, Steiner TJ, Clifford Rose F. Oral contraceptives, smoking and ergotamine in migraine. In: Greene R, ed. Current Concepts in Migraine Research. New York: Raven Press, 1978: 97.
4 Grant ECG, Clifford Rose F. Smoking and migraine. In: Greenhalgh RM, ed. Smoking and Arterial Disease Bath: Pitman Medical, 1981:29-34.
5 Petitti DB, Sidney S, Bernstein A, Wolf S, Quesenberry C, Ziel HK. Stroke in users of low dose contraceptives. N Engl J Med 1996; 335: 8-15
6 Grant ECG. Hormonal contraceptives cause migraine and ischaemic strokes. http://bmj.com/cgi/eletters/330/7482/63#89277, 15 Dec 2004
7 Grant ECG. The pill, hormone replacement therapy, vascular and mood over-reactivity, and mineral imbalance .J Nutr Environ Med 1998; 8: 105- 116.
Competing interests: None declared

Migraine; a contraindication to combined oral contraception 4 February 2005

Helen K Crispin,
House Officer
John Radcliffe Hospital, Headley Way, Oxford, OX3 3DZ
Send response to journal:
Re: Migraine; a contraindication to combined oral contraception

This systematic review and meta-analysis by Etminan et al. provides further evidence to support the association between migraine and ischaemic stroke[1]. Importantly this study also finds an increased relative risk of stroke in female migraineurs taking the combined oral contraceptive.
In the UK there is cautious use of the combined oral contraceptive (COC) in migraineurs. The NHS guidelines recommend that the COC should not be prescribed to patients who experience migraine with aura but may be given to those who experience migraine without aura and up to one other risk factor for ischaemic stroke (www.prodigy.nhs.uk).
It can be inferred from the guidelines that the risk of ischaemic stroke is greater in patients who experience aura. Two case-control studies in young women (18-45 years) have specifically examined the risks of ischaemic stroke in migraineurs according to migraine subtype[2,3]. The odds ratios of ischaemic stroke are greater for these patients, 6.2[2] and 3.81[3] compared to 3.5[2] and 2.97 [3] for migraine without aura. However, this data also indicates that the risk of ischaemic stroke is not statistically different between the two migraine subtypes (the 95% confidence intervals of migraine with and without aura overlap)[2,3]. Similarly, the pooled analysis by Etminan et al. (in male and female subjects) found that the odds ratio for migraine with aura was 2.28 (1.89 - 4.39) and for migraine without aura 1.56 (1.03 - 2.36)[1].
The current evidence is limited, but supports the hypothesis that both simple and classical migraine are risk factors for ischaemic stroke. However, there is no conclusive evidence that migraine with aura confers a greater risk of stroke than migraine without aura and arguably they should be subject to the same prescribing guidelines.
References
1. Etminan M, Takkouche B, Caama�o Isorna F, Samii A. Risk of ischaemic stroke in people with migraine: systematic review and meta- analysis of observational studies. BMJ 2005;330:63-65.
2. Tzourio C, Tehindrazanarivelo A, Iglesias S, Alperovitch A, Chedru F, d'Anglejan-Chatillon J, Bousser M-G. Case-control study of migraine and risk of ischaemic stroke in young women. BMJ 1995;310:830-833.
3. Chang CL, Donaghy M, Poulter N, and World Health Organisation Collaborative Study of Cardiovascular disease and Steroid Hormone Contraception. Migraine and stroke in young women: case-control study. BMJ 1999;318:13-18.
Competing interests: None declared