Effects of transdermal versus oral hormone replacement therapy in postmenopause: a systematic review

Purpose To summarize available evidence comparing the transdermal and the oral administration routes of hormone replacement therapy (HRT) in postmenopausal women. Methods We performed a systematic review of the literature on multiple databases between January 1990 and December 2021. We included randomized controlled trials and observational studies comparing the transdermal and oral administration routes of estrogens for HRT in postmenopausal women regarding at least one of the outcomes of interest: cardiovascular risk, venous thromboembolism (VTE), lipid metabolism, carbohydrate metabolism, bone mineral density (BMD), and risk of pre-malignant and malignant endometrial lesions, or breast cancer. Results The systematic literature search identified a total of 1369 manuscripts, of which 51 were included. Most studies were observational and of good quality, whereas the majority of randomized controlled trials presented a high or medium risk of bias. Oral and transdermal administration routes are similar regarding BMD, glucose metabolism, and lipid profile improvements, as well as do not appear different regarding breast cancer, endometrial disease, and cardiovascular risk. Identified literature provides clear evidence only for the VTE risk, which is higher with the oral administration route. Conclusions Available evidence comparing the transdermal and oral administration routes for HRT is limited and of low quality, recommending further investigations. VTE risk can be considered the clearest and strongest clinical difference between the two administration routes, supporting the transdermal HRT as safer than the oral administration route. Supplementary Information The online version contains supplementary material available at 10.1007/s00404-022-06647-5.


Introduction
Women's life expectancy has increased by almost 10 years over the past half-century and is now approximately 78-86 years in most European countries [1,2].Therefore, a woman can spend almost half of her life in peri-and postmenopause, with a consistent risk of developing a range of estrogen deficiency symptoms and diseases [3].
The transition to menopause is characterized by permanent cessation of ovarian function, leading to bothersome menopausal symptoms and long-term health consequences [4]. In this regard, hormone replacement therapy (HRT) has been proposed as a strategy to relieve menopause symptoms for years and conventionally includes both estrogen and progesterone [5]. Nevertheless, HRT use decreased dramatically after the results of the Women's Health Initiative (WHI) trial in 2002 [6]. The WHI study was stopped 1 3 due to increased myocardial infarction occurrence, thromboembolic events, and breast cancer cases in HRT users [6]. The HRT use dropped to 12% in 2004 and 5% in 2010 [7]. Nevertheless, the WHI trial was criticized for the presence of limitations and biases that should be considered to appropriately interpret study results, such as the inclusion of women aged 60-79 years [6].This crucial consideration pushes toward the foundations of the modern HRT, which consider the importance of starting HRT in the early years after menopause, introducing the concept of "time frame/ window" of opportunity for the benefits of HRT [7,8].
In this scenario, evaluating the best and maybe the safest administration route for HRT is of high relevance. Transdermal HRT is differently metabolized than the oral route, with a lower effective dose [9]. The skin metabolizes estradiol (E 2 ) in a small part, and a reduced amount of hormone is required with lower serum estrone (E 1 ) concentration, similar to premenopausal levels [10]. Based on differences between the oral and transdermal administration routes, the purpose of the present review was to summarize available evidence comparing the transdermal route with the oral administration of the estrogen component of the HRT in postmenopausal women. We focused on cardiovascular risk, venous thromboembolism (VTE), lipid metabolism, carbohydrate metabolism, bone mineral density (BMD), and risk of pre-malignant and malignant endometrial lesions and breast cancer.

Sources and search strategy
A systematic literature search was conducted by two independent reviewers (M.Š.G. and M.M.) according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement for transparent reporting of systematic reviews and meta-analyses [11]. The databases PubMed, clinicaltrials.gov, Scopus, and Web of Science were systematically searched for records from January 1990 to March 2021 using the combination of the medical terms "HRT", "estrogen replacement", "hormonal menopausal therapy", "estrogen replacement therapy", "menopausal therapy", "menopausal hormone therapy", "estrogentherapy", and "estrogen replacement therapy."

Selection criteria
We included only studies reported in the English language. Allowed study designs were randomized and non-randomized controlled trials, observational prospective studies, and retrospective studies. The population of interest was postmenopausal women. Investigated interventions were oral and transdermal estrogen administration for HRT. The definition of transdermal estrogen application included gel, patch, or spray. Studies based on either estrogen monotherapy, combined-cyclic, or combined-continuous HRT were included, as well as studies investigating both natural, synthetic, or conjugated equine estrogens. Studies had to report regarding at least one of the following outcomes of interest: cardiovascular risk (acute coronary disease/myocardial infarction), VTE risk, variation of lipid profile values, alteration of carbohydrate metabolism, risk of pre-malignant and malignant endometrial lesions, risk of breast cancer, and variation in BMD.

Study selection and data extraction
Titles and abstracts of identified studies were screened independently by two review authors (M.Š.G. and M.M.). The full text of the potentially eligible studies was retrieved and independently assessed for eligibility by two other review team members (F.A.F. and S.G.). Any disagreement over the eligibility of studies was resolved through discussion with a fifth author (A.S.L.). The reference lists of all identified studies were systematically revised to identify other eligible publications.

Quality assessment
Two review authors (M.N. and P.T.) independently assessed the risk of bias of included randomized controlled trials according to the Cochrane risk-of-bias tool for evaluating the quality of randomized controlled trials (Rob 2.0) [12]. The following characteristics were considered: adequacy of randomization, allocation concealment, blinding of patients and outcome assessors, reporting of study withdrawals, the performance of an intention-to-treat analysis, and other potential biases.
The same team members assessed the methodological quality of non-randomized studies using the nine-star Newcastle Ottawa Scale (NOS) [13]. Each study was evaluated based on eight items, categorized into three broad perspectives including selection, comparability, and outcome for cohort studies or exposure for case-control studies. We considered studies with a score of 7 or greater as high quality. Disagreements between the review authors over the risk of bias were resolved by discussion with a third author (S.U.).

Literature search
The systematic literature search identified a total of 1369 manuscripts, of which 289 duplicate papers were removed.
After title and abstract screening, 76 potentially relevant articles were identified and underwent full-text assessment for eligibility. Among these, 25 studies were excluded following inclusion and exclusion criteria. A total of 51 studies were finally included in the qualitative synthesis. The PRISMA flowchart summarizing study selection is reported in Fig. 1.
Included manuscripts have been classified into seven groups according to the reported outcome: 6 studies were included in the "cardiovascular risk" group, 10 in the "VTE risk" group, 12 in the "lipid metabolism" group, 7 in the "carbohydrate metabolism" group, 5 in the "endometrial disease risk" group, 7 in the "breast cancer risk" group, and 5 were  Tables 1-8.

Cardiovascular risk
Several studies investigated the effect of HRT (regardless of the route of administration) on the cardiovascular system, particularly affecting the coagulation cascade, inflammatory parameters, lipid composition, intima-media artery thickness, blood pressure, and atherosclerosis progression [14,15]. Nevertheless, only four case-control studies and two cohort studies compared the risk of acute coronary disease (myocardial infarction) in women treated with oral or transdermal HRT [16][17][18][19][20][21]. Characteristics of the selected studies are reported in Table 1. Almost all studies agree with a beneficial effect of HRT, but none of the two routes of administration demonstrated a significant advantage, and heterogeneous results were globally reported. Moreover, none of these studies was designed to compare the two administration routes.

Venous thromboembolism risk
VTE is a rare, but serious risk associated with HRT. Our systematic literature search identified ten studies comparing transdermal versus oral administration and their correlated risk with VTE events. Seven were case-control studies, of which three were multicenter, and three were cohort studies ( Table 2).
Only two studies found no difference between the two routes of administration [17,22]. Conversely, other authors observed that transdermal HRT is a safer choice, especially in women at increased risk for VTE (  thrombotic risk [27]. Among different preparations, other authors suggested that the association of oral estrogen and medroxyprogesterone acetate seemed to correlate with the highest risk [29]. Noteworthy, Straczek et al. investigated the impact of the estrogen administration route on the association between VTE and the most common prothrombotic mutations of factor V of Leiden and prothrombin G20210A (Table 3) [32]. They observed that the oral administration route was associated with a higher increase in the OR for VTE than the transdermal route [32].

Lipid metabolism
Our literature search found 12 studies comparing the effect of oral and transdermal HRT on lipid metabolism, as reported in Table 4: 11 randomized controlled trials and 1 cohort study. The 12 studies covered a wide time frame of 20 years and investigated a different combination of estrogens and progestogens. However, HRT reduced LDL values regardless of the administration route in all studies, although results provided by some of them using transdermal estrogens have not demonstrated a statistically significant decrease in LDL concentrations [33][34][35]. Additionally, oral replacement therapy was demonstrated to increase the HDL and triglycerides concentration. Conversely, transdermal therapy had no significant effect on HDL levels and most studies highlighted a significant reduction in triglyceride concentration [36][37][38][39].

Carbohydrate metabolism
Seven studies reporting on carbohydrate metabolism were identified, among which only one randomized controlled trial was found [40][41][42][43][44][45][46] ( Table 5).The only study that directly compared the impact of oral and transdermal HRT on diabetes risk was the French prospective E3N study. The authors reported a higher reduction of diabetes risk in oral HRT users compared to the transdermal arm, although both administration routes reduced the risk as compared to women without HRT and were associated with a reduction in glycated hemoglobin concentration [41]. However, the study did not consider the type of progestogens that was used in combination with the estrogen, and the results could reflect the influence of individual progestogens rather than the estrogen administration route. Notably, Shakir et al. reported a negative impact on glucose tolerance and insulin resistance by medroxyprogesterone acetate and levonorgestrel [40]. Another study was not able to show significant differences between the progestogens used in the HRT [41].
Overall, according to the published results, both oral and transdermal administration routes reduce insulin resistance, with a more evident effect of the oral administration route in non-diabetic women.

Risk of pre-malignant and malignant lesions of the endometrium
Concerning the risk of developing endometrial hyperplasia or endometrial cancer, a small number of studies were designed to compare transdermal and oral estrogenfor HRT (Table 6) [47][48][49][50]. Four randomized controlled trials were identified and included in our results. The authors did not demonstrate an increased risk of malignant and pre-malignant endometrial lesions with transdermal or oral administration in combined therapy or differences between them. Different authors reported a risk of endometrial hyperplasia and cancer with the transdermal route comparable to or even lower than those associated with the oral administration route. Vaginal bleeding was reduced with longer use, and amenorrhea was achieved in an equal percentage of women with oral and transdermal HRT.

Risk of breast cancer
Our literature search identified six cohort studies and one case-control study comparing the oral and transdermal HRT effects on breast cancer risk (Table 7) [16,[51][52][53][54][55][56]. The UK Million Women Study [51] was the first study investigating the effect of the estrogen administration route on breast cancer risk and no differences were found between oral and    [53]. The same authors compared the effect of combined oral and transdermal HRT with progestogens supplementation in 2009 [54]. Once more, no significant differences were found between oral and transdermal estrogen administration. Furthermore, 5 years after cessation of therapy, the risk of breast cancer was the same for non-users [54]. Similarly, the data of the French E3N cohort study reported an increased risk of breast cancer among women receiving HRT without differences between oral and transdermal administration routes, although authors suggested preferring micronized progesterone to synthetic progestins [52,55]. In a study based on the British family physicians database, Opatrny et al. [56] reported an increased risk of breast cancer among HRT users of oral estrogens; however, the same was not observed in users of transdermal products. Nevertheless, intervals were broad, and the overlap with the results from the oral group did not make a statistically significant difference between the two types of HRT administration routes.

Effect on bone mineral density
Four randomized prospective studies and one retrospective case-control study were identified (

Risk of bias assessment
Risk of bias assessment for randomized controlled trials was performed according to the Cochrane risk-of-bias tool and summarized in Supplementary Table 1. Overall, the risk of bias was high for most of the identified randomized controlled trials, followed by unclear risk. Only one randomized controlled trial was classified as having a low risk of bias [44]. Conversely, the risk of bias assessment for observational studies reported a quality score equal to or higher than 6 in all included studies (Supplementary Table 2).

Discussion
Transdermal estrogen preparations are considered as effective for treating menopausal symptoms [63] as the oral administration route [64]. However, the transdermal administration route has different pharmacodynamics as compared to oral administration [65], determining possible different safety profiles and impacts on global women's health [64].
In our systematic review of the literature, we observed relevant differences between the two administration routes that highlight the need to further characterize the similarities and differences between these two administration options.
Nowadays, HRT is not recommended for the primary or secondary prevention of cardiovascular disease [66,67]. This position is primarily based on the WHI trial findings, which raised concerns about an increased risk of acute coronary disease and breast cancer. However, the re-analysis of data from the WHI estrogen-only [68] arm study has demonstrated that the impact of HRT on acute coronary disease risk is related to the woman's age at the beginning of HRT administration. HRT decreased by 44% the acute coronary disease risk in the group of women younger than 60 years. Conversely, the study did not report benefits in the group older than 60 years for the acute coronary disease, but showed an increased risk for stroke. These observations suggested that an early start after menopause is required to achieve benefits on cardiovascular risk. Conversely, when HRT is started regardless of age, no benefits are found for primary or secondary prevention of all-cause mortality, cardiovascular-related death, non-fatal heart infarction, or need for revascularization [69]. Regarding possible differences in cardiovascular risk between the oral and transdermal administration routes, we did not identify studies suggesting differences. Most studies included in our review highlighted a possible beneficial effect of HRT, but none of the two administration routes demonstrated significant advantages over the other [17][18][19]. Consistently, cardiological societies prefer transdermal therapy based on other elements of the safety profile instead of a demonstrated higher efficacy in improving cardiovascular risk. Our systematic review highlights a lack of evidence comparing the incidence of cardiovascular disease between the oral and transdermal administration routes, recommending further investigation [65][66][67].
Regarding VTE, all identified studies are consistent in reporting the transdermal administration route being safer than oral HRT. The avoidance of the first hepatic passage is the main reason explaining the absent increase of procoagulant factors with transdermal products. Three previously  [72], consistently with the conclusion of the meta-analysis by Rovinski et al. [70]. Noteworthy, in women with prothrombotic mutation, oral HRT leads to a 25-fold increased risk for VTE compared to non-users [25], versus a lower fourfold increased risk for VTE determined by transdermal estrogens [32]. Based on identified studies, the transdermal administration route of estrogens appears the preferred choice in terms of VTE risk. After menopause, estrogen reduction causes a pro-atherogenic shift of lipid-lipoprotein profile with an increase in total cholesterol, LDL, lipoprotein-a, triglycerides, and a reduction of HDL levels [73]. In this regard, most of the included studies, the majority randomized controlled trials, did not observe differences between the two administration routes in terms of lipid-lipoprotein profile improvement. However, our systematic review suggests possible differences in lipid-lipoprotein profile changes between the two administration routes; some studies observed a higher improvement of triglyceride levels with the transdermal administration route and a higher impact on cholesterol metabolism with oral HRT.
Menopause is characterized by weight gain, decreased energy expenditure, loss of lean body mass, and an increase in the perivisceral and total amount of fat [74]. These changes have been related to the influence of estrogens on the growth hormone/insulin-like growth factor-I (GH/IGF-I) axis, decreasing IGF-I and increasing GH levels [75,76]. GH and IGF-I play a pivotal role in body composition and in resting energy expenditure and finally influence glucose metabolism [77,78]. Consistently, HRT was reported to significantly reduce the risk of diabetes and insulin resistance, without clear differences between oral and transdermal administration routes. Indeed, only one study directly compared the two administration routes [41], suggesting a higher reduction of diabetes risk in oral HRT users compared to transdermal ones. Nevertheless, further evidence is required to potentially recommend one administration route over the other based on the improvements in glucose metabolisms.
Concerning the risk of invasive breast cancer, literature reported a higher incidence among HRT users than never users. A little risk was demonstrated for a period shorter than 6 months [16], with progressively increasing risk for longer periods, although its magnitude has been found to vary across studies and a safe cutoff for HRT length has not been demonstrated. Moreover, although a higher risk was demonstrated for current users than past users, some studies showed that risk persists after HRT cessation [79].
Concerning the effect on breast cancer risk provided by the administration route, studies included in our systematic review did not show a different risk between the transdermal and oral administration routes, although the E 1 -E 2 ratio is fivefold higher with oral E 2 than in physiological conditions or with the transdermal route [55,56]. Nonetheless, the findings of the included studies are not able to exclude a different risk. Therefore, further investigation is needed to clarify whether the transdermal route carries a lower risk or not [80]. In this regard, other administration routes, such as topical vaginal estrogen preparations, seemed to lead to lower risk, further reducing the systemic exposure [53,81]. Finally, any future study must address differences in progestogens. Breast cancer risks with combined HRT did not appear to differ based on the progestogen; however, some studies suggest a higher risk with norethisterone acetate and a lower risk for dydrogesterone [54].
Endometrial cancer is the most common gynecological cancer in developed countries [82]. More than 90% of cases of endometrial cancer occur in women older than 50 years of age, with a median age at diagnosis of 63 years [83]. Based on the results of our systematic review, the risk of endometrial hyperplasia and endometrial cancer with transdermal HRT is equal to or smaller than with oral HRT [49,50]. Moreover, the proportion of patients reporting amenorrhea did not differ between the two administration routes [50]. Notably, our systematic review confirms that the main risk factor for endometrial cancer, regardless of the administration route, is the use of only estrogen replacement therapy in women with the uterus.
Estrogen deficiency in menopause determines an accelerated bone loss [84]. Osteoporosis affects one-third of women aged over 50 years and is associated with increased social costs, mortality, and worse quality of life [85]. Both the uncombined and combined HRT arms of the WHI study showed a significant increase in BMD and a reduction of hip fractures than controls [85]. Thus, HRT could be considered the first-line therapy for maintenance of BMD in postmenopausal women under the age of 60 years or within 10 years after menopause, as indicated in international guidelines [86][87][88][89]. Meanwhile, after the age of 60 years, a joint Global Consensus Statement in 2016 stated that HRT is a secondline therapy for preventing fractures [90]. In this scenario, our findings showed an osteoprotective effect, with increased BMD, both with oral and transdermal routes [58,61,62]. Therefore, BMD prevention does not appear to guide the choice of the HRT administration route.

Strengths and limitations
The strength of this study lies in the methodological approach of the comprehensive literature search, the following of The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA), and the quality of the studies assessment. Conversely, our conclusions are limited by the disadvantage of most studies, which are based on small numbers of enrolled subjects, and by the limited number of studies comparing the two administration routes. Notably, most reports are observational, and most randomized controlled trials reported a high or unclear risk of bias.

Conclusion
According to our systematic review, available evidence comparing the transdermal and oral administration routes for HRT is limited. Most studies are observational, and the majority of randomized controlled trials present a high or medium risk of bias. In this scenario, available literature comparing the oral and transdermal administration routes for HRT provides clear evidence only for the VTE risk, which is higher with the oral administration route. Conversely, oral and transdermal administration routes do not appear different regarding an improvement of BMD, glucose metabolism, and lipid profile changes, as well as they do not appear different regarding the risk of breast cancer, endometrial disease, or cardiovascular risk. Considering that the effect on VTE can be considered the clearest and strongest clinical difference between the two administration routes, our systematic review supports the transdermal HRT as safer than the oral administration route. Nevertheless, the final choice of the type of therapy must be tailored and discussed with the patient according to baseline risks and her preferences. Further larger and well-designed studies are mandatory to provide evidence able to guide the personalized choice of HRT.