Original Contributions. Risk of Ischemic Stroke Among Users of the Oral Contraceptive Pill. The Melbourne Risk Factor Study (MERFS) Group

Transcription

1 Original Contributions Risk of Ischemic Stroke Among Users of the Oral Contraceptive Pill The Melbourne Risk Factor Study (MERFS) Group Sasitorn Siritho, MD; Amanda G. Thrift, PhD; John J. McNeil, PhD; Roger X. You, PhD; Stephen M. Davis, MD; Geoffrey A. Donnan, MD Background and Purpose Use of oral contraceptives has increased, and there is uncertainty about the stroke risk associated with their use. Our aim was to investigate this issue. Methods Using case-control techniques, we identified consecutive women with ischemic stroke from 4 Melbourne hospitals. All patients were between 15 and 55 years of age and had no prior stroke. Neighborhood-based control subjects were individually age-matched ( 5 years) and geographically matched to subject cases. A questionnaire administered to participants elicited information about prior exposure to various potential risk factors, including the oral contraceptive pill (OCP). Results We included 234 cases and 234 controls (mean age, 42 years). Compared with noncurrent use, current use of the OCP, in doses of 50 g estrogen, was not associated with an increased risk of ischemic stroke (odds ratio [OR] 1.76; 95% CI, 0.86 to 3.61; P 0.124). Factors associated with an increased risk of ischemic stroke were a history of hypertension (OR, 2.18; 95% CI, 1.22 to 3.91), transient ischemic attack (OR, 8.17; 95% CI, 1.69 to 39.6), previous myocardial infarction (OR, 5.64; 95% CI, 1.04 to 30.61), and diabetes mellitus (OR, 5.42; 95% CI, 1.42 to 20.75); family history of stroke (OR, 2.22; 95% CI, 1.12 to 4.43); and smoking 20 cigarettes per day (OR, 3.68; 95% CI, 1.22 to 11.09). Conclusions There was no evidence for an association between ischemic stroke and use of the OCP in low doses ( 50 g estrogen) in young Australian women. Other modifiable risk factors such as hypertension, diabetes mellitus, and smoking are important. (Stroke. 2003;34: ) Key Words: Australia case-control studies risk factors stroke, ischemic The oral contraceptive pill (OCP) was first introduced in the 1960s. The possibility that cerebrovascular accidents (strokes) might be induced by the OCP was first raised by Lorentz 1 in Most studies of the cardiovascular side effects of the OCP carried out during the 1960s and 1970s were based on the use of oral contraceptive formulations that typically contained 80 or 100 g estrogen. There is now widespread use of the OCP in lower doses (30 to 35 g estrogen). Many large studies of OCP use and stroke risk have been performed, but there is disagreement about whether an association remains between OCP use and stroke 2 10 and some concern regarding the potential for serious adverse events among OCP users, 11 including pulmonary embolism, 2,3 myocardial infarction, 12 and recently cerebral venous sinus thrombosis. 13 There are 2 main reasons for reexamining this association. First, the OCP contains lower doses of estrogen and progestogen than previously, and second, current recommendations restrict its use to younger women who do not have other risk factors for cardiovascular disease. 7 Materials and Methods Subjects Cases of ischemic stroke in Melbourne, Australia, occurring between August 1984 and October 1996 were identified by discharge record surveillance of 4 major city hospitals (Austin & Repatriation Medical Centre, Monash Medical Centre, Alfred Hospital, and Royal Melbourne Hospital). Melbourne has a population of 3.5 million. The age-adjusted incidence of stroke among women in part of this region has recently been reported to be 89 per population per year during 1996 to Ischemic stroke was defined as a sudden onset of an acute focal neurological deficit. CT, autopsy, or MRI was used to exclude causes other than ischemia. The duration of cerebral ischemia may be 24 (transient ischemic attack [TIA]) or 24 (cerebral infarction) hours. Received November 6, 2002; final revision received February 13, 2003; accepted February 19, From the National Stroke Research Institute (S.S., A.G.T., R.X.Y., G.A.D.) and Neurology Department (S.S., G.A.D.), Austin & Repatriation Medical Centre, West Heidelberg; Department of Epidemiology and Preventive Medicine, Monash Medical School, Alfred Hospital, Prahran (A.G.T., J.J.M.); Department of Medicine, University of Melbourne, Melbourne (G.A.D.); and Neurology Department, Royal Melbourne Hospital, Parkville (S.M.D.), Australia. Correspondence to Amanda G. Thrift, National Stroke Research Institute, Austin & Repatriation Medical Centre, Neuroscience Bldg, Banksia St, West Heidelberg, 3081, Australia. thrift@austin.unimelb.edu.au 2003 American Heart Association, Inc. Stroke is available at DOI: /01.STR B 1575

2 1576 Stroke July 2003 Cases included patients 15 to 55 years of age who were experiencing their first episode of ischemic stroke. Patients with hemorrhagic stroke were not included, and neither were those residing in nursing homes at stroke onset. Control Subjects Controls were identified with the neighborhood selection method. The nurse-interviewer went to the same street from which the case originated and, starting with neighbors sequentially to the left, attempted to identify the first woman who was the same age ( 5 years) as the matching case. Repeated visits (up to 3 times) were made during evenings and weekends if no one was home during the day. To avoid the possibility of biasing toward the unemployed or those who were immobile for health reasons, it was always established that no potential control existed in a given dwelling before moving on to the next house. Ascertainment of Risk Factors Nurse-interviewers conducted in-person interviews using a structured questionnaire. All questions related to the time period immediately preceding the stroke (cases) or interview (controls). On average, case subjects were interviewed 6 to 7 months after stroke. Controls were interviewed within 2.5 months of cases. Patients who had died or were unable to answer for themselves were included by interview of next of kin. These proxy interviews occurred in 16 cases (6.8%). To avoid information bias, individuals acting as controls by proxy were asked to nominate an equivalent relative to ensure that a similar manner of interview and similar information sources were used as for the corresponding case. Hypertension, previous cardiovascular disease, high cholesterol, and diabetes were considered present when patients reported that their medical practitioner had advised them that they had the condition. A family history of hypertension, stroke, ischemic heart disease, or diabetes mellitus was deemed present when the interviewee reported that either 1 or both parents had suffered the condition. A detailed history of medication use was also recorded. Use of oral contraceptives or hormone replacement therapy was considered present when the subject reported using these medications at any time in her life. All brand names were recorded over the lifetime when available; otherwise, the notation low dose or high dose was used. Each known brand was then categorized as low-dose ( 50 g) or high-dose ( 50 g) estrogen. Self-reported height and weight data were used to calculate body mass index (BMI). We defined a current smoker as a person smoking at least 1 cigarette, cigar, or pipe per day for the preceding 3 months and an ex-smoker as a person who had smoked this amount at some period during her life but had not smoked for the preceding 3 months. Current drinkers were those reporting drinking any amount of alcohol at the time of the stroke (cases) or interview (controls), and previous drinkers were those who drank alcohol in the past. Subjects were considered to have exercised if they had ever undertaken regular physical activity that caused breathlessness at least once a week at any time in their lives. All subjects were asked about their frequency of adding salt to food, whether they removed fat from meat or skin from chicken before eating, and whether they ever followed a vegetarian diet. An ever user of the OCP was a woman who used the OCP at any time in her life before the time of ischemic stroke (cases) or interview (controls). A current user was a person who used the OCP at any time within 5 years before the event (cases) or interview (controls). A past user reported using the OCP in the past but was not a current user. A noncurrent user included never users and past users. The remaining women were classified as having never used these medications. The questionnaire had previously been validated with information from patients medical records. 15 TABLE 1. Demographic Data Ischemic Stroke Cases Controls Variable (n 234) (n 234) Mean (SD) Age, y 41.9 (9.5) 42.5 (10.0) Mean (SD) BMI, kg/m (5.9) 24.9 (4.5) Education, No high school 13 (5.6) 13 (5.6) Some high school 105 (44.9) 119 (50.9) Completed high school 40 (17.1) 49 (20.9) University/other tertiary 75 (32.1) 53 (22.6) Unknown 1 (0.4) 0 (0.0) Statistical Analysis Conditional logistic regression was used to compute odds ratios (ORs) approximating the relative risks of stroke for various exposures. When information on a particular variable under investigation was missing, both the individual and the matched pair were excluded from the analysis. Initially, univariate ORs were calculated for exposure variables of interest and potentially confounding variables. Confounding variables with a value of P 0.10 were retained for use in multivariate analyses. Exercise, hypercholesterolemia, and alcohol consumption were also controlled for in these analyses because of their association with other vascular diseases. Ethics This study was approved by ethics committees at each hospital, and informed consent was obtained from each participant. Results Between August 1984 and October 1996, we identified 244 first-ever cases of ischemic stroke among women 15 to 55 years of age. Five cases refused to participate, and in a further 5 instances, no age-matched control could be identified. This left us with a total of 234 case-control pairs of ischemic stroke (95.9% of ischemic stroke participants). The mean age was 41.9 years for cases and 42.5 years for controls (Table 1). Diabetes, TIA, hypertension, and myocardial infarction were more common among cases than controls (Table 2). Of those currently using an OCP, none reported using a preparation with 50 g estrogen; however, 31% of these cases and 26.9% of controls had taken preparations containing higher doses of estrogen in the past. Table 2 summarizes the crude and adjusted ORs of ischemic stroke for all exposure variables used in the multivariate analysis. Women who reported currently using the OCP (estrogen 50 g) were not at an increased risk of ischemic stroke. The presence of hypertension or a TIA was associated with an increased risk of ischemic stroke (OR, 2.18; 95% CI, 1.22 to 3.91; and OR, 8.17; 95% CI, 1.69 to 39.6, respectively). Women with a history of myocardial infarction and diabetes were also at greater risk of ischemic stroke (OR, 5.64; 95% CI, 1.04 to 30.61; and OR, 5.42; 95% CI, 1.42 to 20.75, respectively). Exercise, alcohol consumption, and hypercholesterolemia were not associated with ischemic stroke in these young women. Interestingly, smoking 20 cigarettes a day was associated with a greater risk of ischemic stroke (OR, 3.68; 95% CI, 1.22 to 11.09). Crude and adjusted ORs of ischemic stroke for age, BMI, dietary factors, and family history are provided in Table 3. Subjects culinary habits such as adding salt to food and

3 Siritho et al Ischemic Stroke Risk in Oral Contraceptive Pill Users 1577 TABLE 2. Crude and Adjusted ORs of Ischemic Stroke for All Variables Included in the Multivariate Regression Model Factor Smoking Cases (n 234), removing fat from meat or skin from chicken before eating showed no association with ischemic stroke. However, having a family history of stroke was associated with an increased risk of ischemic stroke (OR, 2.22; 95% CI, 1.12 to 4.43). Among users of the OCP at doses of 50 g estrogen, there was no association between the number of years using the pill and ischemic stroke (Table 4). Further analyses were undertaken to determine whether there was an interaction between use of the pill and ischemic stroke according to smoking or hypertension status. However, none of these differences reached statistical significance (all interaction values were P 0.05; data not shown). Controls (n 234), Crude* Estimated OR Adjusted (95% CI) Never 94 (39.7) 115 (48.7) 1.00 Current 1 20/d 47 (20.1) 53 (22.6) ( ) Current 20/d 40 (17.1) 18 (7.7) ( ) Past 45 (19.2) 35 (15.0) ( ) Alcohol drinking Never 52 (21.8) 43 (17.9) 1.00 Current 148 (63.2) 158 (67.5) ( ) Past 26 (11.1) 26 (11.1) ( ) Exercise Never 83 (35.0) 75 (31.6) 1.00 Ever 151 (64.5) 159 (67.9) ( ) Cholesterol Normal 189 (80.3) 191 (81.2) 1.00 Hypercholesterolemia 33 (14.1) 28 (12.0) ( ) MI No history 219 (93.2) 232 (98.7) 1.00 History 15 (6.4) 2 (0.9) ( ) Hypertension No 151 (64.1) 188 (79.9) 1.00 Yes 82 (35.0) 46 (19.7) ( ) TIA No history 213 (90.6) 232 (98.7) 1.00 History 21 (9.0) 2 (0.9) ( ) Diabetes mellitus No 207 (88.0) 228 (97.0) 1.00 Yes 27 (11.5) 6 (2.6) ( ) OCP use Noncurrent user 197 (49.6) 202 (86.3) 1.00 Current user 35 (15.0) 28 (12.0) ( ) MI indicates myocardial infarction. *Crude OR indicates univariate conditional logistic regression not adjusted for confounding variables. Adjusted ORs for all factors were obtained from multivariate analyses adjusted for all other factors in the table. The 95% CI is for adjusted OR. Multivariate analysis used a total of 181 matched pairs because of missing values for 1 or both of the cases and controls. Discussion There is still conflicting evidence for a role of the OCP in stroke. 2 7,9,10,16 23 Results have been variable, with some authors finding an increased risk of ischemic stroke in OCP users 2 11,16,17,20,22 27 and a few documenting a protective effect in past users. 21 Results of these case-control and cohort studies may be susceptible to bias because OCP use may be inaccurately measured or other unknown potential confounding factors may be unbalanced between the 2 groups compared. Furthermore, the treatments may vary over time, and particularly in retrospective studies, either women may recall information imperfectly or there may be differential recall between cases and controls. 11 P

4 1578 Stroke July 2003 TABLE 3. Crude and Adjusted ORs of Ischemic Stroke for Various Factors Estimated OR Factor Cases (n 234), Controls (n 234), Matched Pairs, n Crude* Adjusted (95% CI) P Age, y (6.4) 13 (5.6) (18.4) 45 (19.2) ( ) (28.6) 65 (27.8) ( ) (46.6) 111 (47.4) ( ) BMI, kg/m (2.6) 11 (4.7) (51.7) 127 (54.3) ( ) (25.2) 63 (26.9) ( ) (20.1) 33 (14.1) ( ) Vegetarian 167 Not vegetarian 207 (88.5) 203 (86.8) 1.00 Total vegetarian 9 (3.8) 13 (5.6) ( ) Ovolactovegetarian 4 (1.7) 5 (2.1) ( ) Eaten fat on meat/chicken 165 Rarely, never 113 (48.3) 99 (42.3) 1.00 Sometimes 49 (20.9) 59 (25.2) ( ) Usually 57 (24.4) 61 (26.1) ( ) Add salt to food 166 Not at all 96 (41.0) 87 (37.2) 1.00 Sometimes 39 (16.7) 48 (20.5) ( ) Only after tasting 19 (8.1) 23 (9.8) ( ) Often or always 66 (28.2) 62 (26.5) ( ) Hormone replacement therapy 178 No 191 (81.6) 203 (86.8) 1.00 Yes 37 (15.8) 28 (12.0) ( ) Family history of IHD 153 No 110 (47.0) 127 (54.3) 1.00 Yes 104 (44.4) 89 (38.0) ( ) Family history of stroke 153 No 153 (65.4) 167 (71.4) 1.00 Yes 62 (26.5) 45 (19.2) ( ) Family history of DM 152 No 168 (71.8) 184 (78.6) 1.00 Yes 41 (17.5) 31 (13.2) ( ) Family history of HT 139 No 85 (36.3) 97 (41.5) 1.00 Yes 110 (47.0) 108 (46.2) ( ) IHD indicates ischemic heart disease; DM, diabetes mellitus; and HT, hypertension. *Crude OR indicates univariate conditional logistic regression not adjusted for confounding variables. Adjusted ORs for all factors are obtained from multivariate analyses adjusted for all factors in Table 2 (except for hormone replacement therapy, which is adjusted for all factors except OCP use). 95% CI is for adjusted OR. Perhaps the most useful overall view of the issue comes from a recent large meta-analysis of 16 case-control and cohort studies. 26 Among the case-control studies (published between 1993 and 1999), the relative risk was still elevated among those using low-dose ( 50 g) estrogen (OR, 2.08, 95% CI, 1.55 to 2.80) 5,7,9,18 20,23 but was significantly lower than among those using higher doses ( 50 g) of estrogen (OR, 4.53; 95% CI, 2.17 to 9.50). 2 4,6,16 The overall relative risk, regardless of the OCP dose, was 2.75 (95% CI, 2.24 to 3.38). OCP use appeared to impart a similar ischemic stroke risk among smokers and nonsmokers. Furthermore, there was no difference in the risk imparted by OCP use according to hypertensive status, the presence or absence of migraine history, and different ages. 26

5 Siritho et al Ischemic Stroke Risk in Oral Contraceptive Pill Users 1579 TABLE 4. Adjusted ORs for Ischemic Stroke in Relation to OCP Use and Period of OCP Use Characteristic OCP user Case, Control, Ischemic Stroke Adjusted OR* 95% CI P Never 81 (34.6) 74 (31.6) 1.00 Ever 151 (64.5) 156 (66.7) OCP user Never user 81 (34.6) 74 (31.6) 1.00 Current user 35 (15.0) 28 (12.0) Past user 116 (49.5) 128 (54.7) OCP user Noncurrent user 197 (84.2) 202 (86.3) 1.00 Current user 35 (15.0) 28 (12.0) Period of OCP use, y 5 68 (29.1) 69 (29.5) (17.1) 47 (20.1) (12.0) 26 (11.1) (3.8) 9 (3.8) (0.9) 3 (1.3) *Adjusted ORs are obtained from multivariate analyses adjusted for smoking, alcohol consumption, exercise, cholesterol, myocardial infarction, hypertension, TIA, and diabetes. 95% CI is for adjusted OR. Since this large meta-analysis was undertaken, there has been another study in which the risk of ischemic stroke with several types of OCP has been assessed. 27 In this case-control study of 203 patients with stroke and 925 matched population-based controls, the risk of stroke in women using any type of OCP compared with nonusers was 2.3 (95% CI, 1.6 to 3.3). Also, regardless of progestogen content, use of 50 g estrogen was associated with an increased risk of ischemic stroke, with an OR of 3.1 (95% CI, 1.2 to 7.9), which was slightly higher than that for the use of 50 g estrogen (OR, 2.3; 95% CI, 1.5 to 3.4). In our study, current use of the OCP was not associated with a statistically greater risk of ischemic stroke compared with those not currently using the OCP. Also, there was no evidence that either smoking or hypertension potentiated the risk of stroke in this group. Furthermore, the duration of OCP use did not significantly affect risk estimates of ischemic stroke, even among those who used the OCP for a long period ( 20 years). Similar results have been reported in 2 rigorous, population-based case-control studies of low-dose ( 50 g) estrogen OCPs in the United States. 21 In this population, there was similarly no significant elevation of ischemic stroke risk with current use of low-dose OCPs (OR, 0.66; 95% CI, 0.29 to 1.47). 21 Similar to other investigators, we have reported an association between cigarette smoking and ischemic stroke. 5,10,28 In particular, people smoking 20 cigarettes per day were at greater risk of stroke. Importantly, this increased risk was diminished among past smokers. Furthermore, we have also demonstrated a significant increase in the risk of ischemic stroke among people with a family history of stroke (OR, 2.19; 95% CI, 1.10 to 4.34). This was slightly higher than that reported recently (OR, 1.4; 95% CI, 1.0 to 1.8), although the point estimate of this latter study is incorporated within the 95% CI bounds of the present findings. 10 In addition, similar to recent findings, we found an increased risk of ischemic stroke in people with a history of hypertension and diabetes mellitus. 10 A number of potential biases may have influenced the results of this study. The most important of these may be due to differential recall among cases and controls, because of either intellectual impairment of cases or the fact that cases may have more reasons to ponder the cause of their stroke. This may result in an underestimate or overestimate of the OR, respectively. There are several reasons why our results differ from the findings of the large meta-analysis and the recent study. 10,26 First, data related to the hormone content of the pill used at the time of stroke should be interpreted with care. It has been assumed that any effects of the previous brand disappear as soon as the pill is changed or no longer used. Second, the complex interaction between the type and dose of hormones contained in each pill was not examined. In our study, we were unable to describe which of the drugs that contained 50 g ethinylestradiol also contained varying doses of norethindrone, lynestrenol, and ethynodiol diacetate. It is possible that the risk of ischemic stroke was elevated in women using these combined OCPs, but we were unable to investigate this issue. We may therefore be missing an important association. There were also no data on progestogen-only oral contraceptives in our study. In general, most patients used second-generation progestogens (ie, levonorgestrel). Third, in our study, 45% of the patients were between 45 and 55 years of age, and in this group, only 7.24% were current OCP users. Because so few people were current users of the OCP in this large group of women, it is possible that we have underestimated the association between the OCP and ischemic stroke. To summarize the existing information, OCPs with an estrogen content 50 g seem to be associated with less risk of ischemic stroke than the pill containing a higher dose of estrogen, with ORs of 2 for the former and 4 for the latter. However, because of our findings and the similar nonassociation between the use of pills containing low-dose ( 50 g) estrogen and ischemic stroke in previous well-conducted case-control studies, 21 more data are required. This is particularly the case for the newer-generation OCPs (thirdgeneration progestogen pills) and in patients with migraine. We recommend that the OCP (in low doses) be prescribed with care. Particular care should be taken among women with a history of smoking, hypertension, diabetes mellitus, and myocardial infarction and in those with a family history of stroke. Acknowledgments This work was supported by grants from the Victorian Health Promotion Foundation, Foundation for High Blood Pressure Research, National Health and Medical Research Council, and National Stroke Foundation. We would also like to acknowledge the assistance of Belinda Bardsley, Georgina Neill, and Susan Lang, who found and assessed most of the patients, and Lichun Quang for computer assistance.

6 1580 Stroke July 2003 References 1. Lorentz IT. Parietal lesion and Enavid. BMJ. 1962;2:1191. Letter. 2. Vessey P, Doll R. Investigation of relation between use of oral contraceptives and thromboembolic disease: a further report. BMJ. 1969;2: Sartwell PE, Masi AT, Arthes FG, Greene GR, Smith HE. Thromboembolism and oral contraceptives: an epidemiological case-control study. Am J Epidemiol. 1969;90: Collaborative Group for the Study of Stroke in Young Women. Oral contraception and increased risk of cerebral ischemia or thrombosis. N Engl J Med. 1973;288: Lidegaard O. Oral contraception and risk of thromboembolic attack: results of a case-control study. BMJ. 1993;306: Hannaford PC, Croft PR, Kay CR. Oral contraception and stroke: evidence from the Royal College of General Practitioners Oral Contraception Study. Stroke. 1994;25: WHO Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Ischaemic stroke and combined oral contraceptives: results of an international, multicentre, case-control study. Lancet. 1996;348: WHO Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Haemorrhagic stroke, overall stroke risk, and combined oral contraceptives: results of an international, multicentre, case-control study. Lancet. 1996;348: Haapaniemi H, Hillbom M, Juvela S. Lifestyle-associated risk factors for acute brain infarction among persons of working age. Stroke. 1997;28: Lidegaard O, Kreiner S. Contraceptives and cerebral thrombosis: a five-year national case-control study. Contraception. 2002;65: Lewis MA. The epidemiology of oral contraceptive use: a critical review of the studies on oral contraceptives and the health of young women. Am J Obstet Gynecol. 1998;179: Lewis MA, Heinemann LA, Spitzer WO, MacRae KD, Bruppacher R. The use of oral contraceptives and the occurrence of acute myocardial infarction in young women: results from the Transnational Study on Oral Contraceptives and the Health of Young Women. Contraception. 1997; 56: de Bruijn SF, Stam J, Vandenbroucke JP. Increased risk of cerebral venous sinus thrombosis with third-generation oral contraceptives: Cerebral Venous Sinus Thrombosis Study Group. Lancet. 1998;351: Thrift AG, Dewey HM, Macdonell RAL, McNeil JJ, Donnan GA. Stroke incidence on the East Coast of Australia: the North East Melbourne Stroke Incidence Study (NEMESIS). Stroke. 2000;31: You R, McNeil JJ, O Malley HM, Davis SM, Donnan GA. Risk factors for lacunar infarction syndromes. Neurology. 1995;45: Collaborative Group for the Study of Stroke in Young Women. Oral contraceptives and stroke in young women: associated risk factors. JAMA. 1975;231: Royal College of General Practitioners Oral Contraception Study. Incidence of arterial disease among oral contraceptive users. J R Coll Gen Practitioners. 1983;33: Petitti DB, Sidney S, Bernstein A, Wolf S, Quesenberry C, Ziel HK. Stroke in users of low-dose oral contraceptives. N Engl J Med. 1996; 335: Schwartz SM, Siscovick DS, Longstreth WT Jr, Psaty BM, Beverly RK, Raghunathan TE, Lin D, Koepsell TD. Use of low-dose oral contraceptives and stroke in young women. Ann Intern Med. 1997;127: Heinemann LA, Lewis MA, Spitzer WO, Thorogood M, Guggenmoos- Holzmann I, Bruppacher R. Thromboembolic stroke in young women: a European case-control study on oral contraceptives: Transnational Research Group on Oral Contraceptives and the Health of Young Women. Contraception. 1998;57: Schwartz SM, Petitti DB, Siscovick DS, Longstreth WT Jr, Sidney S, Raghunathan TE, Quesenberry CP Jr, Kelaghan J. Stroke and use of low-dose oral contraceptives in young women: a pooled analysis of two US studies. Stroke. 1998;29: Mant J, Painter R, Vessey M. Risk of myocardial infarction, angina and stroke in users of oral contraceptives: an updated analysis of a cohort study. Br J Obstet Gynaecol. 1998;105: Lidegaard O. Smoking and use of oral contraceptives: impact on thrombotic diseases. Am J Obstet Gynecol. 1999;180:S357 S Mann JI, Thorogood M, Waters WE, Powell C. Oral contraceptives and myocardial infarction in young women: a further report. BMJ. 1975;3: Lidegaard O, Bygdeman M, Milsom I, Nesheim BI, Skjeldestad FE, Toivonen J. Oral contraceptives and thrombosis: from risk estimates to health impact. Acta Obstet Gynecol Scand. 1999;78: Gillum LA, Mamidipudi SK, Johnston SC. Ischemic stroke risk with oral contraceptives: a meta-analysis. JAMA. 2000;284: Kemmeren JM, Tanis BC, van den Bosch MA, Bollen EL, Helmerhorst FM, van der Graaf Y, Rosendaal FR, Algra A. Risk of Arterial Thrombosis in Relation to Oral Contraceptives (RATIO) study: oral contraceptives and the risk of ischemic stroke. Stroke. 2002;33: Vessey MP, Lawless M, Yeates D. Oral contraceptive and stroke: findings in a large prospective study. BMJ [Clin Res Ed]. 1984;289: