Sage Journals: Discover world-class research

Abstract

Since 1962, more than 25 studies have been devoted to the relationship between oral contraceptives and stroke. They are all case-control or cohort epidemiological studies and thus contain the difficulties and biases that are inherent in these types of studies. The following conclusions can be drawn from these studies:

High oestrogen content (≥ 50 μg) increases the risk of stroke, all stroke subtypes, and stroke death.

Low oestrogen content (< 50 μg) carries a very low or no risk of stroke.

There are no data on progestogen only oral contraceptives.

Stroke risk is greatly increased if associated risk factors are present, in particular hypertension, cigarette smoking and migraine.

Oral contraceptives, even at low doses, significantly increase the risk of cerebral venous thrombosis, which is further enhanced if congenital thrombophilia is present.

The attributable risk of stroke in young women using oral contraceptives is about 1 per 200 000 woman-years.

The contraceptive and non-contraceptive benefits of low dose oral contraceptives vastly outweigh their risks provided that other risk factors are absent or well controlled.

Introduction

The possibility that stroke might be induced by oral contraceptives (OCs) was first raised by Lorentz (1) in 1962 shortly after OCs became available. Since then, more than 25 studies have been devoted to the OCs, but these studies are fraught with difficulties and biases:

Data are not derived from randomized trials but from case-control and cohort epidemiological studies and provide little or no information about potential biases.

The oestrogen dose has changed from the initial high content of 150 μg to the present 20–35 μg.

Different oestrogens and different routes of administration have been used.

Different progestogens have been combined with oestrogens.

Some studies have compared current use with past use or no use, whereas others have compared ‘ever use’ to ‘never use’.

Some studies were devoted to stroke in general and others to various stroke subtypes, such as cerebral infarction, intracerebral haemorrahge (ICH), or subarachnoid haemorrhage (SAH).

Vascular risk factors, such as hypertension and smoking, which are major confounders, were not always taken into account in the study designs or in the statistical analyses.

Because of these methodological shortcomings, results have changed over time and have been somewhat conflicting. In the 1970s, when high dose OCs were used, the risk of stroke was estimated to be increased fivefold in women using OCs. This increased risk applied to all stroke varieties. It was most marked in women over 35, particularly in current users and those with other vascular risk factors, such as hypertension and cigarette smoking (5–14).

In the following years, most studies reported an increased relative risk of stroke, but magnitude lessened as the oestrogen content of OCs decreased (15–19).

In the last 5 years (1994–1999), the results of six large case-control studies comparing over 3000 young women with stroke with nearly 10 000 controls have helped to clarify the situation. These studies have shown a very low risk and have defined the risk according to stroke subtype and OC generation (20–27).

Overall risk of stroke

Two of these studies give data on the overall risk of stroke. In the Royal College of General Practioners (RCGP) study (20), based on a cohort of 23 000 OC users, 253 women had a first-ever stroke between 1968 and 1990. They were compared with 759 women without stroke. Women who had ever used OC had a significantly increased risk of stroke (odds ratio [OR] 1.5 [1.1–2]) and of a fatal event (OR 2.3 [1.2–4.4]) compared with those who had never used OCs. Current users had an OR of 2.5 (1, 4, 5), whereas former users had no significant increased risk (OR 1.3 [0.9–1.8]).

The World Health Organization (WHO) study (23) is a very large, hospital-based, case-control study that was performed in 21 centres in Europe, Africa, Asia and Latin America. Based on analysis of 2198 cases of stroke and 6086 matched controls, the adjusted OR for all strokes was significantly increased in women currently using high dose OCs (≥ 50 μg), both in Europe (OR 2.71 [1.70–4.32]) and in the developing countries (OR 1.92 [1.48–2.50]). With low oestrogen OCs (< 50 μg), the risk was significantly increased in the developing countries (OR 1.86 [1.49–2.33]) but not in Europe (OR 1.41 [0.90–2.20]).

In the 25-year follow-up of this study (28), the risk of death from stroke was also significantly increased in current and recent users (OR 1.9 [1.2–3.1]), but not in those who had stopped OC use ≥ 10 years previously.

In summary, these studies indicate that:

The overall risk of stroke and stroke death is approximately doubled by current OC use (all doses mixed).

Low dose OCs carry a lower and sometimes non-significant risk.

The risk seems to disappears 10 years after discontinuing OCs.

Risk of ischaemic stroke

The risk of ischemic stroke has been assessed in all recent studies, but the results obtained in the United States of America (USA) have differed from the results obtained in other countries. In the two American studies (21, 25), which have recently been pooled (27), there was no increase in risk in women currently using low dose OCs compared with never users (OR 0.66 [0.29–1.47]) or with past users (OR 1.09 [0.54–2.21]). By contrast, in the four non-USA studies (20, 22, 24, 26), the risk of ischaemic stroke was significantly increased in current OCs users, with OR from 2 to 4. However, this increase in risk was highly variable among different countries and according to the oestrogen-progestogen content of the OCs. Thus, in the European part of the WHO study (22), the risk was increased (OR 5.30 [2.56–11]) with high oestrogen (≥ 50 μg) OCs but not with low dose OCs (OR 1.53 [0.71–3.31]), whereas in the developing countries the risk was significantly increased whatever the oestrogen content of the pill. In the Danish study (26) the risk was increased with OCs containing ≥ 30 μg of ethinyl estradiol but not with those containing 20 μg. In the same study, only users of second generation OCs had an increased risk of ischaemic stroke, whereas in the Transnational study (24) the first generation was associated with a higher risk.

In the Kaiser Permanente and the Transnational studies (24), a significant reduction in the risk of stroke was found in previous users compared with never users (OR 0.49 and 0.6, respectively). This finding that previous OC use could have a protective effect requires careful examination and further exploration.

In summary, the risk of stroke in general and ischaemic stroke in particular increases with the oestrogen content of OCs, but whether low dose OCs carry a risk is still being debated, since the present data indicate no increase in risk in the USA, a slightly increased risk in Europe, and a threefold increased risk in Africa, Asia and Latin America. Whether this is due to the higher prevalence of hypertension and smoking in non-USA countries or to other yet unidentified factors remains unknown.

Risk of intracerebral haemorrhage

Conflicting data regarding the risk of intracebral haemorrhage (ICH) have been reported. In the majority of studies (20, 21, 23, 25, 27), there was no increase in the risk of ICH among OC users. By contrast, the risk was significantly increased in the WHO study that focused on developing countries (23), with an OR of 1.76 [1.34–2.30]. The risk was higher with high-dose OCs and it significantly increased with increased age (≥ 35 years) and associated risk factors such as hypertension and smoking.

Risk of subarachnoid haemorrhage

Increased rates of subarachnoid haemorrhage (SAH) among OC users led to major concerns about the safety of these agents (29–31). Twenty-one studies have addressed this issue, and they have found a relative risk as high as 6.5 (30) and as low as 0.5 (32). Eleven of these studies (33) were included in a recent meta-analysis; nine studies showed an increased risk of SAH in OC users that reached statistical significance in subgroups of three studies, while two studies found non-significant protective effects. The meta-analysis showed a significantly increased risk of SAH among OC users with a relative risk (RR) of 1.42 [1.12–1.80] reaching 1.57 [1.25–1.99] in current users. This increased risk applied to incidence as well as to mortality, and it persisted even after controlling for hypertension and smoking with a slightly greater risk (RR 1.55 [1.26–1.91] (P < 0.0001). High-oestrogen OCs carried a greater risk than low-oestrogen OCs, but the difference was not statistically significant. As in all varieties of stroke, the risk dramatically increased when associated risk factors, such as hypertension and cigarette smoking, were present.

In summary, although some systematic reviews did not find a significant effect of OC use on SAH risk (34), the available data suggest that OC use is associated with a small increase in risk and that this increase is marginal with low-oestrogen preparations but it strongly increases with smoking and hypertension.

Risk of cerebral venous thrombosis

In contrast to arterial stroke, cerebral venous thrombosis (CVT) is a rare condition for which there are no good epidemiological studies. OCs have long been acknow-ledged to be a potential cause of CVT (35–39), but the proportion of OC users in series of CVT is highly variable: it is low in series from poor countries where OC use is limited but postpartum and infections remain major causes of CVT and it is very high in rich countries where non-infectious CVTs are widely preponderant and postpartum CVT has decreased in frequency. Thus the proportion of OC users reached over 90% in two recent series from Italy (40) and the Netherlands (41).

However, in most series OC use is associated with other risk factors. If one considers cases in which OC use is the only aetiological factor found, the percentage decreases dramatically (10% in our series of 135 patients (42)). The associated factors of congenital thrombophilia (Factor V Leiden and 20210 G – > A prothrombin gene mutation) have recently been shown to play a major role in CVT, particularly in association with OC use (40, 41, 43–48). Thus, the relative risk for CVT in OC users, which goes from 13⁽⁴¹⁾ to 22⁽⁴⁰⁾ with OC use alone, reaches 150 in the presence of the prothrombin gene mutation (40).

Opinions differ regarding the increased risk of venous thrombosis with third generation OCs (49). Conflicting data have been obtained: in the Dutch study, 56% of CVT patients used third generation OCs vs. 38% of controls. This led the authors to conclude that ‘the increased risk for venous thrombo-embolism found in users of third generation products, compared with other OCs was also found for CVT’ (50). By contrast, in the Italian study, the percentage of third generation OC users was similar in CVT patients (81%) and in controls (82%), and it was concluded that the magnitude of the risk was the same as that conferred by other OCs (51).

In summary, OCs, even those with low oestrogen content, are a major risk factor for CVT, particularly in women with hereditary thrombophilia in whom the risk is increased more than a hundred-fold. In developed countries, OCs have replaced infections and postpartum as major risk factors for CVT. Why venous thrombosis develops in the brain in some patients and in the leg in others remains an enigma and suggests the presence of unknown additional risk factors (48).

OC use and vascular risk factors

Vascular risk factors play a major role in stroke in young subjects, since, in most strokes that occur in young people, no immediate indisputable cause is found. However, there is usually an association of different risk factors (52, 53). Among the vascular risk factors identified, three (besides OCs) play a major role in young women: hypertension, smoking and migraine. These factors are frequently associated with OC use and they are thus major confounders that should be taken into account in epidemiological studies of OC use and stroke.

Hypertension is the single most important risk factor for all varieties of arterial stroke at any age. It consistently increases the risk of stroke associated with OC use. Thus, in the RCGP study (20), the odds ratio of first ever stroke in current OC users rises from 2.5 [1.5 to 4] to 4.8 [2.4–9.4] in the presence of hypertension. Similarly in the European part of the WHO study (22, 23), the risk of both cerebral infarction and cerebral haemorrhage rose from non-significant ORs of 1.53 and 1.27, respectively, to highly significant ORs of 10.7 and 10.3. In the non-European part of the study, the ORs reached 14.5 and 14.3 when women with a history of hypertension used OCs. The relationships between hypertension and OC use are complex, since OCs, even those with 30 μg ethinyl-estradiol content, induce a small but statistically significant increase in mean blood pressure (60, 61). These data and the results of the WHO study suggest that blood pressure may have an important role in the mediation of OC-associated stroke risk.

Smoking has long been neglected as a risk factor for stroke, first because its impact on stroke seemed negligible compared with its impact on peripheral arterial disease and on coronary heart disease, and second because the overall relative risk of stroke associated with cigarette smoking was significant but rather low (1.5 [1.4–1.6]) (62). However, in the last 10 years, evidence has accumulated that smoking carries a high attributable risk of stroke, particularly in the young (62, 63) and even more so in women (62)· The risk of all varieties of stroke is increased by smoking, the relative risk in young women reaching 2.5 for ischaemic stroke and 5 for SAH (22, 23). This risk is consistently increased when OC use and smoking are combined. Thus in the WHO study, the risk of stroke, which was negligibly increased by OCs alone, rose to over 3 for cerebral haemorrhage and to 7.2 and 4.8 for cerebral infarction in the European and non-European parts of the study, respectively (22, 23). Many other traditional vascular risk factors such as obesity, diabetes, high alcohol intake and increased cholesterol may be associated with OC use, but the impact of these associations on the risk of stroke has not been properly evaluated.

Migraine has recently emerged as a risk factor for ischaemic stroke in young women, with a remarkably consistent relative risk of approximately 3 (64–68). The risk is doubled in migraine with aura (RR 6) and is markedly increased by smoking (RR 10) and by oral contraceptives (RR 14) (65). These risk factors seem to have more than an additive effect since, in the recent WHO study, the odds ratio for ischaemic stroke in young migrainous patients who take OCs and who smoke reaches 34.4 (3.27–3.61) (68). In a recent pooled analysis of two USA studies devoted to low-dose OC and stroke, the risks of both ischaemic and haemorrhagic strokes in current OC users were significantly increased among women reporting a history of migraine (respectively 2.08 and 2.15) but were not elevated among non-migraineurs (27).

In summary, stroke in young women is a multifactorial condition in which the association of risk factors, particularly hypertension, smoking, migraine and OC use, plays a crucial role, emphasizing the importance of practical preventive measures.

Type of OC

Epidemiological studies devoted to the risk of OC use are made difficult by on-going oestrogen content changes and the type of associated progestogen. Stroke risk (both cerebral infarction and haemorrhage) has decreased consistently with the decreased oestrogen content of the pill, but we are not sure whether this relationship is causal. Many biases might have occurred leading to more alertness and case certifications at the era of high-oestrogen content OCs, or because of the use of different doses in women with different risk profiles. These difficulties are illustrated by the WHO study (22, 23), which is by far the largest case-control study of OC use and stroke, and which has nevertheless led to conflicting results. For all types of stroke, high-dose (≥ 50 μg) oestrogens significantly increased the risk in European and non-European countries (RR 1.86 and 1.92, respectively); and low doses (< 50 μg) significantly increased the risk only in non-European countries (RR 2.71). The risk of haemorrhagic stroke does not differ according to the oestrogen content (< 50 μg and ≥ 50) in the European and non-European parts of the study (23).

For ischaemic stroke, results differ by country: in Europe, the relative risk is not significantly increased with low dose OCs (< 50 μg) (1.53 [0.71–3.31]), but it rises sharply with higher doses (≥ 50 μg) (5.30 [2.56–11]). In Africa, Asia and Latin America, the risk was significantly increased both with the low (3.26 [2.19–4.86]) and the high (2.71 [1.75–4.19]) oestrogen content. These results may be due to different risk profiles: in developing countries, low-dose OCs were given preferentially to higher-risk women and high-dose OCs to lower-risk women, whereas in Europe the opposite pattern prevailed. The European WHO results are consistent with those of the Danish study (16, 26), which found no increased risk of cerebral infarction with < 50 μg ethinyl oestradiol OCs but a significantly increased risk (2, 2, 6, 9) with ≥ 50 μg OCs.

Conflicting results have been obtained for the type of progestins combined with oestrogens. In the Copenhagen study (26), the risk of cerebral infarction was increased only with second-generation progestins, whereas the WHO study (22) showed no significant increase in the risk of ischaemic stroke associated with any of the three generations of progestins. In the pooled USA studies (27), no difference was found in first- and second-generation progestins. The Transnational Euro-pean study (24) showed no difference between second- and third-generation OCs.

The effects of age and the duration of OC use also vary from study to study. No effect of age was found in the Copenhagen study (26) or in the two pooled USA studies (27). However, in Europe and in the developing countries, the risk of both ischaemic and haemorrhagic stroke was found to be significantly higher in older women (22, 23).

Most studies show significant association between duration of OC use and risk of ischaemic or haemorrhagic stroke (5, 8, 11, 22, 23, 26, 27). However, in the WHO study, there is a suggestion that long-term (> 6 years) current OC use may be associated with an increased risk of both types of stroke, at least among older women. This effect could be mediated by a concomitant elevation in blood pressure (22).

In summary, most available data suggest that the risk of stoke has substantially decreased with the decrease in the oestrogen content of OCs, and that it slightly increases with age. By contrast, there is no good evidence for an independent effect of duration of OC use or for a variable risk according to the type of progestins combined with oestrogens (69).

OC use and the risk of stroke in clinical practice

The epidemiological studies that we have reviewed lead to the following conclusions:

Oral contraception does not increase total mortality (28, 70).

Current OC use (all doses mixed) increases the risk of stroke in general, all stroke subtypes, and stroke death.

Low oestrogen content OCs (< 50 μg) confer a lower risk than high-dose OCs (≥ 50 μg), but the risk varies from country to country with no or marginal increase in risk in the USA and Europe and a threefold increase in Asia, Africa and Latin America.

The risk of stroke in OC users is greatly increased in the presence of other risk factors, particularly hypertension, smoking and migraine.

The risk in current users increases slightly with age, independent of the duration of OC use.

There is no good evidence that different progestins (first, second and third generation) carry different risks.

The absolute risk of stroke in OC users is low: 6.7 per 100 000 woman-years in users of low-dose OCs and 12.9 in users of high-dose OCs (23). Women younger than 35 years of age who use low-dose OCs have an estimated attributable risk of stroke of about 1 per 200 000 woman-years.

There are no data concerning the risk of stroke with progestogen-only contraception.

OC use (even OCs with a low oestrogen content) is a major risk factor for CVT, with a more than tenfold increase in risk. This risk is increased to over a hundred-fold in subjects with congenital thrombophilia.

How can the clinician use these epidemiological data to establish useful guidelines for his day-to-day practice? A recent paper has reviewed the epidemiology literature to try and answer the question ‘What is the risk of cardiovascular disease among users of currently available low dose oral contraceptives who are aged less than 35 years, do not smoke and do not have a medical condition known to increase the risk of vascular disease?’ (2).

Seventy-four papers about the relationship between current low-dose OC use and cardiovascular disease were identified and only five were found to provide information that directly addressed the clinical question. All five papers were related to deep vein thrombosis and pulmonary embolism. The authors conclude that much of the epidemiological data about the vascular risk of OC use is not useful to clinicians who must extrapolate from study populations to their practice populations, which may greatly vary.

The clinician is therefore left with a case-by-case estimate of the benefit/risk balance of OC use. In young women who have none of the other vascular risk factors, the contraceptive and non-contraceptive benefits of low-dose OCs outweigh the potential slight increase in stroke risk. The major problem arises when other risk factors are associated: not so much hypertension, since it is uncommon in this population and since blood pressure should be checked before and during OC use, but smoking, since more and more women smoke and at an ever-younger age, particularly in developing countries. Tobacco smoking carries a considerable attributable risk of stroke in young women and it should be discouraged in any young woman who wants to use OCs. Smoking is the single most important modifiable risk factor for stroke in the young and strategies must be developed world-wide to respond to this major health threat.

Some specialists believe that migraine is a contra-indication for OC use. We favour a less rigid position, particularly in migraine without aura, provided that other vascular risk factors are absent. Young migraine sufferers who want to use OCs should definitely stop smoking. The situation is more complex in migraine with aura, which carries a higher risk of stroke. Again the decision will be made on a case-by-case basis, taking into account the expected benefits and risks of this and other types of contraception. If the decision to use OCs is made, a progestogen-only OC might be recommended, although there are no data on the risk of stroke with this variety of OC. There is no indication for aspirin use, since aspirin has never been found to be effective in the primary prevention of stroke (in contrast to its efficacy in the primary prevention of myocardial infarction and the secondary prevention of stroke).

It would be unrealistic and not cost-effective to do a systematic screening for thrombophilia before prescribing OC use, but a screening should be performed if there is a family history of venous thrombosis; if congenital thrombophilia is found, oestro-progestogen OC is contraindicated.

In conclusion, the vascular risk of OC use has been a ‘hot’ topic for over 30 years. Despite numerous epidemiological studies, it is still difficult to draw practical guidelines from the results of these studies. Using the low oestrogen-content pill, controlling associated vascular risk factors, in particular hypertension and smoking, and encouraging women to develop potential protective factors via diet and regular physical activity are the most obvious practical steps to take.

References

1 Lorentz

. Parietal lesion and Enavid. BMJ 1962; ii: 1991.

2 Hannaford

Owen-Smith

. Using epidemiological data to guide clinical practice: review of studies on cardiovascular disease and use of combined oral contraceptives. BMJ 1998; 316: 984–7.

3 Realini

Goldzieher

. Oral contraceptives and cardiovascular disease: a critique of the epidemiologic studies. Am J Obstet Gynecol 1985; 152: 729–98.

4 Katerndahl

Realini

Cohen

. Oral contraceptive use and cardiovascular disease: is the relationship real or due to bias? J Fam Pact 1992; 35: 147–57.

5 Collaborative Group for the Study of Stroke in Young Women. Oral contraceptives and stroke in young women. JAMA 1975; 231: 718–22.

6 Fogelhom

Aho

. Ischemic cerebrovascular disease in young adults. Acta Neurol Scandinav 1973; 49: 415–27.

7 Fuertes de la Haba

Cuevas

JOC

Pelegrina

Bangdiwala

. Deaths among users of oral contraceptives. Obstet Gynecol 1970; 36: 597–602.

8 Inman

WHW

Vessey

. Investigation of deaths from pulmonary, coronary and cerebral thrombosis and embolism in women of child-bearing age. BMJ 1968; 2: 193–9.

9 Jick

Porter

Rothman

. Oral contraceptives and nonfatal stroke in healthy young women. Ann Intern Med 1978; 89: 58–60.

10.

10 Mettinger

Söderström

Neiman

. Stroke before 55 years of age at Karolinska Hospital 1973–77. A study of 399 well-defined cases. Acta Neurol Scand 1984 Year?; 70: 415–22.

11.

11 Sartwell

Masi

Arthes

Greene

Smith

. Thromboembolism and oral contraceptives: an epidemiologic case-control study. Am J Epidemiol 1969; 90: 365–80.

12.

12 Vessey

Doll

. Investigation of relation between use of oral contraceptives and thromboembolic disease. A further report. BMJ 1969; 2: 651–7.

13.

13 Vessey

Lawless

Yeates

. Oral contraceptives and stroke: findings in a large prospective study. BMJ 1984; 289: 530–1.

14.

14 Stadel

. Oral contraceptives and cardiovascular disease (second of two parts). NEJM 1981; 305: 672–7.

15.

15 Thorogood

Mann

Murphy

Vessey

. Fatal stroke and use of oral contraceptives: findings from a case-control study. Am J Epidemiol 1992; 136: 35–45.

16.

16 Lidegaard

. Oral contraception and risk of a cerebral thromboembolic attack: results of a case control study. BMJ 1993; 306: 956–63.

17.

17 Lindenstrøm

Boysen

Nyboe

. Lifestyle factors and risk of cerebrovascular disease in women – the Copenhagen City Heart Study. Stroke 1993; 24: 1468–72.

18.

18 Marini

Carolei

Roberts

et al.Focal cerebral ischemia in young adults: a collaborative case-control study. Neuroepidemiology 1993; 12: 70–81.

19.

19 Lidegaard

. Decline in cerebral thromboembolism among young women after introduction of low-dose oral contraceptives: an incidence study for period 1980–93. Contraception 1995; 52: 85–92.

20.

20 Hannaford

Croft

Kay

. Oral contraception and stroke- evidence from the Royal College of General Practitioners' oral contraception study. Stroke 1994; 25: 935–42.

21.

21 Petitti

Sidney

Bernstein

et al.Stroke in users of low-dose oral contraceptives. NEJM 1996; 335: 8–15.

22.

22 WHO, Collaborative Study. Cardiovascular Disease, Steroid Hormone Contraception. Ischemic stroke and combined oral contraceptives: results of an international, multicentre, case-control study. Lancet 1996b;348: 498–505.

23.

23 WHO, Collaborative Study. Cardiovascular disease, steroid hormone contraception. Hemorrhagic stroke, overall stroke risk, and combined oral contraceptives: results of an international, multicentre, case-control study Lancet 1996a; 348: 505–10.

24.

24 Heineman

LAJ

Lewis

Thorogood

et al.The Transnational Research Group on Oral Contraceptives and the Health of Young Women. BMJ 1997; 315: 1502–4.

25.

25 Schwartz

Siscovick

Longstreth

et al.Use of low-dose oral contraceptives and stroke in young women. Ann Intern Med 1997; 127: 596–603.

26.

26 Lidegaard

Kreiner

. Cerebral thrombosis and oral contraceptives: a case-control study. Contraception 1998; 57: 303–14.

27.

27 Schwartz

Petitti

Sincovick

et al.Stroke and use of low-dose oral contraceptives in young women. A pooled analysis of two US studies. Stroke 1998; 29: 2277–84.

28.

28 Beral

Hermon

Kay

et al.Mortality associated with oral contraceptive use: 25 year follow-up of cohort of 46000 women from the Royal College of General Practitioners' oral contraception study. BMJ 1999; 318: 96–100.

29.

29 Royal College of General Practitioners' oral contraception study principal investigators. Mortality among oral contraceptive users. Lancet 1977; 2: 727–31.

30.

30 Petitti

Wingerd

. Use of oral contraceptives, cigarette smoking and risk of subarachnoid haemorrhage. Lancet 1978; 2: 234–5.

31.

31 Thorogood

Adam

Mann

. Fatal subarachnoid hemorrhage in young women: role of oral contraceptives. BMJ 1981; 283: 762.

32.

32 Bonita

. Cigarette smoking, hypertension and the risk of subarachnoid hemorrhage: a population-based case-control study. Stroke 1986; 17: 831–5.

33.

33 Claiborn Johnston

Colford

Gress

. Oral contraceptives and the risk of subarachnoid hemorrhage: a meta-analysis. Neurology 1998; 51: 411–8.

34.

34 Teunissen

Rinkel

GJE

Algra

van Gijn

. Risk factors for subarachnoid hemorrhage: a systematic review. Stroke 1996; 27: 544–9.

35.

35 Atkinson

Fairburn

Heathfield

KWG

. Intracranial venous thrombosis as a complication of oral contraception. Lancet 1970; 1: 914–8.

36.

36 Buchanan

Brazinsky

. Dural sinus and cerebral venous thrombosis. Incidence in young women receiving oral contraception. Arch Neurol 1970; 22: 440–4.

37.

37 Fairburn

. Intracranial venous thrombosis complicating oral contraception: treatment by anticoagulant drugs. BMJ 1973; 2: 647.

38.

38 Estanol

Rodriguez

Conte

et al.Intracranial venous thrombosis in young women. Stroke 1979; 6: 680–4.

39.

39 Bousser

Chiras

Sauron

Castaigne

. Cerebral venous thrombosis. A review of 35 cases. Stroke 1985; 16: 199–213.

40.

40 Martinelli

Sacchi

Landi

et al.High risk of cerebral vein thrombosis in carriers of a prothrombin-gene mutation and in users of oral contraceptives. NEJM 1998; 338: 1793–7.

41.

41 De Bruijn

SFTM

Stam

Koopman

Vandenbroucke

. Case-control study of risk of cerebral sinus thrombosis in oral contraceptive users and in carriers of hereditary prothrombotic condition. The Cerebral Venous Sinus Thrombosis Study Group. BMJ 1998; 316: 589–92.

42.

42 Bousser

Barnett

HJM

. Cerebral venous thrombosis. In: Barnett

HJM

Mohr

Stein

Yatsu

eds. Stroke. Pathophysiology, diagnosis and management , volume 1, 3rd edn. Philadelphia: Churchill Livingstone, 1998: 623–47.

43.

43 Deschiens

Conard

Horellou

et al.Coagulation studies, factor V Leiden and anticardiolipin antibodies in 40 cases with cerebral venous thrombosis. Stroke 1996; 27: 1721–3.

44.

44 Zuber

Toulon

Marnet

Mas

. Factor V Leiden mutation in cerebral venous thrombosis. Stroke 1996; 27: 1721–3.

45.

45 Biousse

Conard

Brouzes

et al.Frequency of the 20210 G→A mutation in the 3-untranslated region of the prothrombin gene in 35 cases of cerebral venous thrombosis. Stroke 1998; 29: 1398–400.

46.

46 Reuner

Ruf

Grau

et al.Prothrombin gene G 20210→A transition is a risk factor for cerebral venous thrombosis. Stroke 1998; 29: 1765–9.

47.

47 Bertina

Rosendaal

. Venous thrombosis: the interaction of genes and environment. NEJM 1998; 338: 1840–1.

48.

48 Vandenbroucke

. Cerebral sinus thrombosis and oral contraceptives. BMJ 1998; 317: 483–4.

49.

49 WHO Scientific Group Meeting on Cardiovascular Disease and Steroid Hormone Contraceptives. Summary conclusions. WHO Wkly Epidemiol Rec 1997; 48: 361–3.

50.

De Bruijn

SFTM

Stam

Vandenbroucke

for the cerebral venous sinus thrombosis study group. Increased risk of cerebral venous sinus thrombosis with third generation oral contraceptives. Lancet 1998; 351: 1404.

51.

51 Martinelli

Taioli

Palli

Manucci

. Risk of cerebral vein thrombosis and oral contraceptives. Lancet 1998; 352: 326.

52.

52 Rohr

Kittner

FeeSeries

et al.Traditional risk factors and ischemic stroke in young adults. The Baltimore-Washington cooperative young stroke study. Arch Neurol 1996; 53: 603–7.

53.

53 Thompson

Grenberg

Meade

. Risk factors for stroke and myocardial infarction in women in the United Kingdom as assessed in general practice: a case-control study. Br Heart J 1989; 61: 403–9.

54.

54 Al-Roomi

Heller

Holland

Floate

Wlodzarczyk

. The importance of hypertension in the etiology of infarctive and hemorrhagic stroke. The Lower Hunter Stroke Study. Med J Australia 1992; 157: 452–5.

55.

55 Chasan-Taber Willet

Manson

Spiegelman

et al.Prospective study of oral contraceptives and hypertension among women in the United States. Circulation 1996; 94: 483–9.

56.

56 Sharpe

. Smoking among oral contraceptive users in Quebec in 1987. J Clin Epidemiol 1994; 47: 313–23.

57.

57 Jeune

Wielandt

. Prevalence of smoking and oral contraception in a sample of Danish young women. Scan J Soc Med 1991; 19: 44–6.

58.

58 Prospective Studies Collaboration. Cholesterol, diastolic blood pressure, and stroke: 13 000 strokes in 450 0 00 people in 45 prospective cohorts. Lancet 1995; 346: 1647–53.

59.

59 Bousser

. Stroke in women. The 1997 Paul Dudley White International lecture. Circulation 1999; 99: 463–7.

60.

60 Khaw

Pearl

. Blood pressure and contraceptive use. BMJ 1982; 285: 403–7.

61.

61 Fraser

Weir

. Cardiovascular system: blood pressure. In: Goldzieher

Fotherby

eds. Pharmacology of the contraceptive steroids. New York: Raven Press, 1994: 299–308.

62.

62 Shinton

Beevers

. Meta-analysis of relation between cigarette smoking and stroke. BMJ 1989; 298: 789–94.

63.

63 Love

Biller

Jones

Adams

Bruno

. Cigarette smoking. A risk factor for cerebral infarction in young adults. Arch Neurol 1990; 47: 693–8.

64.

64 Tzourio

Iglesias

Hubert

et al.Migraine and risk of ischemic stroke. BMJ 1993; 308: 289–92.

65.

65 Tzourio

Tehindrazanarivelo

Iglesias

et al.Case-control study of migraine and risk of ischaemic stroke in young women. Br Med J 1995; 310: 830–3.

66.

66 Lidegaard

. Oral contraceptives, pregnancy and the risk of cerebral thromboembolism: the influence of diabetes, hypertension, migraine and previous thrombotic disease. Br J Obst Gynec 1995; 102: 153–9.

67.

67 Carolei

Marini

De Matteis

& the Italian National Research Council Study Group in Stroke in the Young. History of migraine and risk of cerebral ischaemia in young adults. Lancet 1996; 347: 1503–6.

68.

68 Chang

Donaghy

Poulter

. WHO collaborative of cardiovascular disease and steroid hormone contraception. BMJ 1999; 318: 13–8.

69.

69 Lewis

. Myocardial infarction and stroke in young women: what is the impact of oral contraceptives? Am J Obste Gynecol 1998; 179: S68–77.

70.

70 Skegg

DCG

. Oral contraception and health. BMJ 1999; 318: 69–70.

71.

71 Steering Committee of the Physicians' Health Study Research Group. Final report on the aspirin component of the ongoing Physicians' Health Study. NEJM 1989; 321: 129–35.

72.

72 Hansson

Zanchetti

Carruthers

et al.for the Hot Study Group. Effects of intensive blood-pressure lowering and low-dose aspirin in patients with hypertension: principal results of the hypertension optimal treatment (HOT) randomised trial. Lancet 1998; 351 (1755):17–62.