Menopause Live (28 June, 2010) from IMS 

In a recent study, Renoux and colleagues from McGill University in Montreal performed a nested case–control study deriving the data from a cohort of women in the UK General Practice Research Database (GPRD) [1]. Current use of oral and transdermal hormone therapy, based on recorded prescriptions, was compared to no use in 15,710 cases and 59,958 controls. The adjusted rate ratio (RR) for stroke for current use of transdermal estrogens, with or without a progestin, was not increased (RR 0.95; 95% confidence interval (CI) 0.75–1.20) compared with a significant increase associated with oral estrogen, with or without a progestin (RR 1.28; 95% CI 1.15–1.42). This would amount to an attributal risk of 0.8 additional strokes per 1000 women per year. There was an indication of a dose–response relationship; a significant increase in risk was observed with transdermal estrogen doses greater than 50 µg.

The case-control study by Renoux and colleagues is the first major analysis to compare transdermal and oral hormone therapy and conclude that, compared with an increased risk of stroke with oral therapy, there was no increased risk with transdermal treatment at a dose of 50 µg or less. This report is about as strong an observational study as can be achieved. Large numbers of cases (15,710) and controls (59,958) were available for analysis using the well-known UK GPRD. The use of this computerized database precludes selection bias by the investigators and recall bias by the women in the study. The results support the growing conventional wisdom that transdermal therapy at standard doses is free of the cardiovascular risks associated with oral therapy.

Overview
Many clinicians have argued that transdermal administration of estrogen is safer in regard to the risk of venous thromboembolism (VTE) , considering the first-pass effect in the liver to be an important factor in the prothrombotic impact of oral estrogen. A French case–control study concluded that there was a 4-fold increase in VTE with current use of oral estrogen but no increase in risk with transdermal estrogen [2]. In addition, this study reported that oral estrogen treatment adds to the risk of VTE associated with obesity, but transdermal estrogen does not [3]. The French study also reported (although limited by small numbers) that transdermal treatment, in contrast to oral estrogen, does not further increase the risk of VTE associated with Leiden or prothrombin mutations [4]. 

Similar results were reported from the E3N French prospective cohort study, with an increased risk of VTE of 1.7 (95% CI 1.1–2.8) associated with current users of oral therapy (a hazard ratio more in keeping with the usual two-fold increase reported in the literature) and no increase with transdermal estrogen [5]. The French conclusions are supported by a very large case–control study (23,505 cases of VTE) using the UK GPRD (the same investigators and methods as in the report on stroke) [6]. The use of transdermal estrogen alone or combined with a progestin was not associated with an increase in VTE, compared with about a 1.5-fold increase with oral estrogen alone and oral estrogen–progestin.

A greater safety with transdermal administration of estrogen in regards to VTE makes some sense because of the known lesser impact on clotting proteins when the first-pass liver effect is avoided. This is supported by the almost negligible effect of transdermal therapy on activated protein C resistance when compared with oral therapy [7,8]. Therefore, the observational studies support the clinical choice of a transdermal method for women who are at higher risk for VTE. 

But there are some reasons to resist rushing to general promotion of transdermal therapy as ‘safer’. Although the recent report on stroke is a strong study, it is still not a clinical trial, but observational. The characteristics of the women in the study reveal a preponderance among the cases of smoking, hypertension, hyperlipidemia, atrial fibrillation, diabetes, cardiovascular disease, and a history of transient ischemic attack. Of course, these are recognized risk factors for stroke. The crude rate ratios were adjusted for these factors to yield the final conclusions, but a clinician must always be a little wary when statistically significant results require mathematical manipulations.

Observational data over the last 30 years regarding estrogen use and stroke have not been consistent. For example, a prospective cohort study in Denmark recorded an increase in ischemic strokes, but only among hypertensive women, and a large cohort study from Sweden found no link between stroke and hormone therapy [9,10]. The Nurses’ Health Study reported an update of its data on the use of oral hormone therapy and stroke, focusing on the timing of initiation of treatment and the effect of estrogen doses [11]. In the analyses adjusted for age, body mass index, cholesterol levels, diabetes, hypertension, smoking, and family history of early coronary heart disease, the relative risk was increased for ischemic stroke (there was no significant increase in hemorrhagic stroke). Comparing initiation of hormone therapy near menopause with initiation 10 or more years after menopause, there was no major difference. The Nurses’ Health Study reported an increasing risk of stroke with an increasing dose of oral estrogen, similar to the recent stroke study that indicated a dose–response relationship with transdermal treatment. A dose–response relationship is logical, repeatedly demonstrated in studies with oral estrogen–progestin contraceptives.

A randomized, double-blind, placebo-controlled secondary prevention trial (the WEST trial) of daily 1 mg estradiol therapy was conducted in postmenopausal women after a recent (within 90 days) ischemic stroke or transient ischemic attack (25% of the women) [12]. After an average of 2.8 years of follow-up (range 16–50 months), there were no significant overall differences comparing the treatment and placebo groups in any of the assessed outcomes, including non-fatal stroke, fatal stroke, coronary death, non-fatal myocardial infarction, or transient ischemic attack. The WEST trial retrospectively analyzed the time course of cerebrovascular events and found a significantly increased risk of stroke only at 6 months, based on 21 strokes in the estradiol group and nine strokes in the placebo group. A major limitation of the WEST study was the reduced compliance with treatment because of the problems associated with unopposed estrogen treatment. It is important to note that a limitation of the recent report on transdermal therapy and stroke, by virtue of the method of data gathering, was an inability to account for patient compliance with treatment.

The Women’s Health Initiative (WHI) reported an overall increase in the estrogen–progestin arm of ischemic stroke, but no increase in fatal strokes [13,14]. The increase in non-fatal ischemic stroke in the estrogen-only arm of the WHI was of similar magnitude [15,16]. In the most recent stroke report from the WHI, when women with prior cardiovascular disease or those older than 60 years were excluded, the risk of stroke in women less than 10 years since their menopause was not significantly increased [17]. The case–control study on transdermal therapy and stroke included women from age 50 to age 79, and the average age of 70.3 years indicated that the study population was composed of a significant number of older postmenopausal women. There was no analysis by age in the report and, hopefully, we will see this in the future.

Considerable variation in estradiol levels has been documented in individuals receiving both oral and transdermal hormone therapy [18,19]. Marketing presentations by the pharmaceutical companies provide mean levels, suggesting stable and smooth maintenance of blood levels; however, the ranges, which are wide, are not revealed. Studies comparing circulating estradiol levels in women receiving oral or transdermal hormone therapy show therapeutic estradiol levels predictive of a good bone response, but they also show large standard deviations, indicating substantial variation among individuals (mean estradiol levels in oral hormone therapy users, 125 ± 71 pg/ml; mean estradiol levels in transdermal hormone therapy users, 61 ± 79 pg/ml) [20]. Furthermore, individual women metabolize estrogen differently, depending on the route of administration, their own liver function, skin absorption, body composition, body size, potential medication interactions, and the presence of binding proteins, all of which contribute to individual variations in serum estradiol levels [19]. 

The only way to accurately compare clinical differences between oral and transdermal estrogen delivery is to establish that the two methods produce similar blood levels and that clinical differences reflect the first-pass effect through the liver. This is difficult to accomplish because the oral first-pass effect raises sex hormone binding globulin (SHBG) levels such that total serum estradiol levels are greatly affected. A study of 18 women showed that oral estrogen increased SHBG by 67% to 171%, whereas transdermal estrogen did not alter SHGB levels [21]. Estrogen-induced changes in SHBG may be clinically significant because estrogen unbound to SHBG determines the estrogen effects of a given regimen. The only study that measured free estradiol levels, compensating for increases in SHBG, showed that, at 12 weeks, serum free estradiol levels in the oral group (3.17 pmol/l) were similar to those in the transdermal group (3.09 pmol/l). However, because these results were derived from only 18 women, the effect of oral and transdermal doses on free estradiol levels has not been reliably established [22]. 

Thus, it is difficult to draw conclusions about clinical differences between oral and transdermal hormone delivery because the studies must adjust for individual variability of dosing to ensure that circulating estrogen levels in the patients being studied are similar. In the French studies of venous thrombosis, oral hormone users used almost exclusively estradiol in doses that averaged 1.5 mg per day. Transdermal users most commonly used an estradiol dose of 50 µg or less daily. To legitimately compare oral and transdermal methods, one would have to be sure the two groups had similar blood levels, to account for the wide variation in metabolism and clearance among individuals.

Leon Speroff
Professor Emeritus of Obstetrics and Gynecology, Oregon Health & Science University, Portland, Oregon, USA