Oral contraceptives and stroke

 

Contributors

Sarkis Morales-Vidal MD, author. Dr. Morales of Loyola University of Chicago Stritch School of Medicine has no relevant financial relationships to disclose.

 

Michael J Schneck MD, author. Dr. Schneck of Loyola University of Chicago Stritch School of Medicine has no relevant financial relationships to disclose.

 

Jose Biller MD, author. Dr. Biller of Loyola University of Chicago Stritch School of Medicine has no relevant financial relationships to disclose.

 

Steven R Levine MD, editor. Dr. Levine of the Mount Sinai School of Medicine received honorariums from NCME for speaking engagements.

 

Publication dates

Originally released December 16, 2003; last updated May 14, 2009; expires May 14, 2012

 

Historical note and nomenclature

  Shortly after the introduction of oral contraceptives, strokes associated with oral contraceptive use were first reported (Lorentz 1962), and numerous studies have since estimated the risk of stroke in users of oral contraceptives with varying results (Inman and Vessey 1968; Sartwell et al 1969; Vessey and Doll 1969; Anonymous 1973; Jick et al 1978; Mettinger et al 1984; Chang et al 1986). Several factors have contributed to the significant difficulties in estimating this risk. One major issue is that formulations of oral contraceptives vary considerably and have changed remarkably over the years. When originally introduced, oral contraceptives contained doses of ethinyl estradiol as high as 150 mcg. Presently, no formulations are in use with greater than 50 mcg of ethinyl estradiol, and most contain as little as 20 to 35 mcg. Also, multiple types of progestins are available in different combinations with the various ethinyl estradiol doses. Some studies have evaluated all types of oral contraceptive use, other studies have attempted to address the risk associated specifically with changes in estrogen dose, and other studies evaluated the risks associated with the various progestins. There has been no consistent system of categorization of the different formulations, however, making comparison between studies difficult.

  Additionally, some studies compare current users of oral contraceptives to women who have never used them, whereas other studies compare current users to noncurrent users (women who used oral contraceptives in the past in addition to those who never used them), and still other studies compare women who have ever used oral contraceptives to women who have never used them. Also, in different studies, the definition of current use varies from women using oral contraceptives at the exact time of the qualifying event, to use within several months before the event. One group actually categorized any use within 5 years prior to the event as current use (Siritho et al 2003).

  Furthermore, the type of stroke addressed is not consistent between studies. Some studies evaluated all types of stroke, whereas some selected only certain categories, such as ischemic stroke, intracranial hemorrhage, subarachnoid hemorrhage, cerebral venous thrombosis, or transient ischemic attacks. Furthermore, information about mechanistic ischemic stroke subtypes is limited in most of the studies.

  In addition, other susceptibility factors for stroke are not always considered as comorbidities in the evaluation of stroke risk with oral contraceptive use. Among these factors are arterial hypertension, cigarette smoking, hypercholesterolemia, diabetes mellitus, obesity, increasing age, and migraine. Differences in the frequency as well as control of these risk factors could contribute to diverse results among studies. Newer evidence that inherited prothrombic conditions may be susceptibility factors for oral contraceptive-associated stroke is another potential explanation for differences between studies in various populations (Slooter et al 2005).

  Finally, by necessity, all research on this issue has been observational in design, either as case-control or cohort studies. Both designs allow for the possibility that users differ from nonusers in systematic ways that cannot be readily controlled or adjusted for in the analysis. For example, higher risk individuals may be counseled against oral contraceptive use, and this effect may vary between populations. Case-control studies, which constitute the bulk of the evidence, also have the additional issues of potential recall bias and the appropriateness of the control group.

  There is general acceptance among investigators, however, that there is a small but significant increased risk of stroke among women who use oral contraceptives. The overall degree of risk and the particular risk in special populations is sufficiently uncertain, however, that there are absolute recommendations as to when to avoid oral contraceptives to minimize stroke risk.

  Worldwide, more than 100 million women either ever used or are currently using oral contraceptive medications (Blackburn et al 2000). Furthermore, there are approximately 50 million sexually active (18 to 44 years of age) women in the United States, and of those almost 50% use some type of contraception. At present, there are approximately 40 brands of oral contraceptives in the U.S. market and 21 brands of emergency contraceptive pills (refer to Table 1 for types of hormonal contraception). Different brands are sometimes identical except for packaging; in other instances, brands have different hormones in slightly differing amounts. Thus, when prescribing oral contraceptives, physicians may consider type of estrogen, estrogen dose, progestin type, progestin dose, routes of administration (oral, vaginal, patches), and other relative potencies as important factors. Risk of serious side effects is higher among women who take more than 50 mcg of estrogen. Many physicians now recommend that women take pills containing 35 mcg of estrogen or less. Thus, the ideal drug is the one with the lowest estrogen and progestin doses that will be effective in preventing pregnancy and minimize adverse effects (American College of Obstetricians and Gynecologists).

 

Table 1. Types of Hormonal Contraception

 

Oral contraceptives	Other delivery methods
• Desogestrel	• Medroxyprogesterone acetate SQ
• Drospirenone	• Medroxyprogesterone acetate IM
• Ethynodiol diacetate	• Levonorgestrel intrauterine system
• Levonorgestrel	• Norelgestromin transdermal patch
• Norethindrone	• Etonogestrel implantable device
• Norethindrone acetate	• Etonogestrel vaginal ring
• Norgestimate
• Norgestrel

 

  This clinical summary considers the evidence for stroke risk associated with use of currently available oral contraceptives. Most oral contraceptives are combined formulations of ethinyl estradiol or any of the several different progestins. Doses of estrogen vary from 20 to 50 mcg. Formulations containing progesterone only, in oral and injectable forms, are specifically referred to as such in this summary. Different types of progestins have been included in the various formulations of oral contraceptives and are referred to specifically when possible. Progestins are usually categorized by generation, either first, second, or third. These designations, however, are not uniformly defined, making direct comparisons between studies difficult. Transdermal hormonal contraception is discussed briefly. Implantable forms of hormonal contraception are not discussed, as these have not been individually studied for stroke risk. Estrogen and progesterone preparations used in hormone replacement therapy also are not addressed in this clinical summary.

Clinical manifestations

  The clinical presentation of stroke associated with the use of oral contraceptives is not different from those seen in any young woman. Ischemic stroke, intracranial hemorrhage, and subarachnoid as well as cerebral venous thrombosis have been reported in association with oral contraceptive use. The signs and symptoms of stroke are the same as those described in detail elsewhere.

 

The remainder of this Clinical Summary is available to subscribers only. Current subscribers, please log in at the MedLink Home page. If you are not a subscriber, please choose one of the following options: