An overview of the development of combined oral contraceptives containing estradiol: focus on estradiol valerate/dienogest

Natural estrogens such as estradiol (E2) or its valerate ester (E2V) offer an alternative to ethinyl estradiol (EE). E2-containing combined oral contraceptives (COCs) have demonstrated sufficient ovulation inhibition and acceptable contraceptive efficacy. However, earlier formulations were generally associated with unacceptable bleeding profiles. Two E2V-containing preparations have been approved to date for contraceptive use: E2V/cypro-terone acetate (CPA) (Femilar®; only approved in Finland and only in women >40 years or women aged 35–40 years in whom a COC containing EE is not appropriate) and E2V/dienogest (DNG; Qlaira®/Natazia®). The objective of the current review is to provide an overview of the development of COCs containing natural estrogen, highlighting past issues and challenges faced by earlier formulations, as well as the current status and future directions. The majority of information to date pertains to the development of E2V/DNG.


Introduction
Combined oral contraceptives (COCs) consist of estrogen and progestogen components. Although contraceptive effects can largely be achieved with progestogen alone, as is the case for progestogen-only pills, the estrogen component in COCs avoids symptoms of hypoestrogenism [1], enhances contraceptive efficacy and helps regulate bleeding. Although a number of progestogens have been introduced into clinical practice over the last five decades since the introduction of COCs [2][3][4][5], the estrogen component has remained predominantly ethinyl estradiol (EE). This reliance on EE to date is due mainly to its good oral bioavailability (38-48%) [6] relative to other estrogens. Recently, a new COC containing estradiol valerate (E 2 V) combined with dienogest (DNG) has been approved for contraceptive use worldwide.
The objective of the current review is to provide an overview of the development of COCs containing natural estrogen, highlighting past issues and challenges faced by earlier formulations, as well as the present status and future directions.

Reducing the estrogen dose
In 1970, Inman and colleagues demonstrated that COC use was associated with an increased risk of thromboembolic disease [7].
Thrombotic risk was attributed largely to the estrogen component, and increased with increasing doses of estrogen (at that time mestranol or EE) [7].
Efforts were therefore made to reduce the EE dose in COCs. These dose reductions have been highly successful in reducing the risk of venous thromboembolism (VTE); current estimates put the incidence of VTE at between 8 and 10 per 10,000 womenyears in users of COCs containing <50 µg of EE, compared with 4.7 per 10,000 women-years in non-pregnant, non-COC users and around 20 per 10,000 women-years during pregnancy and the post-partum period [8].
Recent years have seen further reductions in the EE dose, to 20 µg and even 15 µg. However, these low doses have been associated with higher rates of discontinuation from clinical trials (mainly due to adverse events including bleeding) and bleeding disturbances (amenorrhea/infrequent bleeding, irregular, prolonged or frequent bleeding or spotting) compared with higher doses of EE [9]. In particular, preparations containing EE 15 μg have a somewhat higher incidence of breakthrough bleeding and/or spotting than COCs containing EE 20 μg [10], and may be associated with premature discontinuation because of bleeding irregularities [11]. These observations, together with the finding that even EE doses as low as 10 µg have been associated with negative effects on hemostatic surrogate parameters [12], mean that it is unlikely that a COC containing lower EE dosages will be well accepted.
At equivalent dosages (i.e. E 2 2 mg has biologic effects that are equivalent to EE 4-20 µg, depending on the target organ) E 2 has been shown to have a lesser impact on metabolic and hepatic parameters than EE in several studies. This is manifested in a more favorable effect of E 2 versus EE on lipids [19] and a reduced effect of E 2 versus EE on the synthesis of hepatic proteins, including sex hormone-binding globulin (SHBG) and angiotensinogen [6,20,21]. In addition, E 2 appears to have a reduced impact on markers of hemostasis than EE [12,22,23].

Present
Combinations of estradiol plus different progestogens Estradiol/NOMAC Nomegestrol acetate (NOMAC) is a 17-hydroxy-progesterone derivative [30]. A recent study compared E 2 1.5 mg/NOMAC 2.5 mg with EE 30 μg/drospirenone 3 mg in healthy women (n = 32) (Table I) [30]. In this randomized, six-cycle study, ovulation inhibition was noted in all women in both treatment groups. Ovarian suppression was similar between treatments; progesterone was fully suppressed to levels <2 nmol/L in both groups (Table I) [30]. A more recent publication noted that NOMAC 2.5 mg inhibited both ovulation and follicular maturation, and the antigonadotropic effects of NOMAC 2.5 mg were reinforced when it was combined with E 2 1.5 mg [27]. This dose-finding study underlines the role of the estrogen component in inhibiting ovulation. A third study assessed ovarian activity with two different E 2 (1.5 mg)/NOMAC (2.5 mg) regimens (21/7 [n = 37] and 24/4 [n = 40]) [28]. The 24/4 regimen was associated with greater inhibition of follicular activity. A shorter duration of total and withdrawal bleeding with the 24/4 regimen compared with the 21/7 regimen was described as secondary outcome (p < 0.05) [28]. By cycle 3, the incidence of breakthrough bleeding was similar between regimens, but the duration of breakthrough bleeding was slightly longer with the 21/7 regimen than with the 24/4 regimen [28]. The bleeding profile was also assessed as a secondary outcome in a study comparing the hemostatic effects of E 2 /NOMAC in a 24/4 regimen with those of EE/LNG in a 21/7 regimen. In this short-term study, also based on a limited number of participants and thus not powered to investigate the bleeding profile (n = 45 in each group), the duration of total bleeding, withdrawal bleeding and breakthrough bleeding appeared to be shorter with E 2 /NOMAC than with EE/LNG (Table I) [43]. Meaningful investigations of the bleeding profile of E 2 /NOMAC in a 24/4 regimen and comparisons with other combined oral formulations in a larger more diverse group of women are currently lacking. Contraceptive efficacy was not assessed in these studies.

Estradiol/drospirenone
Two randomized, double-blind, parallel-group Phase II studies have been completed with a COC containing E 2 /drospirenone administered in either a monophasic or triphasic regimen (dosages not defined). Both studies were conducted in healthy women aged 18-35 years. The first study (n = 116) assessed ovulation inhibition (ClinicalTrials.gov identifier: NCT00631124), while the second (n = 575) evaluated cycle control and safety (ClinicalTrials.gov identifier: NCT00653614). Results are anticipated.

Estradiol valerate/CPA
The combination of E 2 V and CPA has been marketed as a COC (Femilar ® ) in Finland since 1993, however it is only indicated in women >40 years or women aged 35-40 years in whom a COC containing EE is not appropriate. This formulation comprises E 2 V 1 mg/CPA 1 mg on days 1-10, E 2 V 2 mg/CPA 2 mg on days 11-21

cycles
The following regimen contained the lowest dose of DNG necessary for suppression of ovulation (defined as ovulation rate <5% with an upper limit of the 95% CI of <10% in cycle 2): E 2 V 3 mg alone for 2 days followed by E 2 V 2 mg/DNG 2 mg for 5 days followed by E 2 V 2 mg/DNG 3 mg for 17 days followed by E 2 V 1 mg alone for 2 days followed by placebo for 2 days.
No safety concerns were raised with any of the regimens studied. Ahrendt  and a pill-free interval on days 22-28. During 12 cycles of treatment with the E 2 V/CPA combination (n = 288), ovulation inhibition was observed in 95% of women. The cumulative pregnancy rate was 0.4% (Table I) [32]. Intermenstrual bleeding/spotting was observed in 35.5% of women in cycle 3 and 24.5% of women in cycle 12 [32]. Bleeding became less frequent over time in the majority of women, and dysmenorrhea subsided [32]. Similar findings were observed in a second study (n = 50) comparing E 2 V/ CPA with E 2 V/norethisterone in a biphasic regimen (E 2 V/CPA: as described above; E 2 V/norethisterone: E 2 V 1 mg/norethisterone 1 mg on days 1-10, E 2 V 2 mg/norethisterone 2 mg on days 11-21 and a pill-free interval on days 22-28) ( Table I) [33]. Ovulation was inhibited in all women (except for one woman who ovulated during the first treatment cycle) in the E 2 V/CPA group. In the E 2 V/ norethisterone group, ovulation occurred in 8 women. One additional woman in this group ovulated during all treatment cycles; treatment was discontinued in this subject. Contraceptive efficacy was not assessed in this study. Menstrual blood loss was reduced in all women in the E 2 V/CPA group. However, in the E 2 V/norethisterone group, menstrual blood loss reduced in 40% and increased in 10% of women. The total number of bleeding days reduced with E 2 V/CPA and increased with E 2 V/norethisterone (Table I) [33].

Estradiol valerate/DNG
The combination of E 2 V/DNG was approved as a COC in the European Union (EU) in 2008, where it is marketed as Qlaira ® / Klaira ® . FDA approval for Natazia ® was obtained in May 2010. Qlaira ® received regulatory approval in the EU for the treatment of heavy menstrual bleeding (HMB) in October 2010 and in the USA (Natazia ® ) in March 2012. The E 2 V/DNG combination provides early estrogenic dominance to ensure initial endometrial proliferation and endometrial stroma stability during the progestogen-dominated mid-to-late part of the cycle [24]. DNG has potent endometrial activity [44][45][46] and a bioavailability of >90% after oral intake [47]. Early investigations with E 2 V/DNG employed biphasic or triphasic regimens, which provided effective ovulation inhibition but unacceptable bleeding profiles (Table I). The unacceptable bleeding profile with E 2 V/DNG in a biphasic or triphasic regimen prompted the introduction of an E 2 V/DNG combination in an estrogen step-down/progestogen step-up approach. In a pilot study in healthy women (n = 100), a dynamic dosing regimen (E 2 V 3 mg for 3 days, E 2 V 2 mg/DNG 1 mg for 4 days, E 2 V 2 mg/ DNG 2 mg for 16 days, E 2 V 1 mg for 2 days and finally placebo for 3 days) was associated with a far more favorable profile than the biphasic or triphasic regimens (Table I) [34,35].
Four variations of E 2 V/DNG in dynamic phasic regimens were investigated in two sequential Phase II studies designed to determine the optimal daily application and the required dose of DNG for effective inhibition of ovulation (Table I) [31]. In the first study it was shown that a dosing regimen that incorporated 26 rather than 25 days of active treatment was associated with greater ovulation inhibition. In the second study, which examined two 26-day regimens with doses of DNG that were distinctly increased versus those used in the first study, it was shown that a dose of DNG of 2 mg on days 3-7 and 3 mg on days 8-24 (both in combination with E 2 V 2 mg) was the lowest effective dose of DNG for efficient ovulation inhibition. This regimen comprised E 2 V 3 mg alone for 2 days, E 2 V 2 mg/DNG 2 mg for 5 days, E 2 V 2 mg/DNG 3 mg for 17 days, E 2 V 1 mg alone for 2 days then placebo for 2 days (Figure 3 and Table I). There are few data on compliance with COCs, but one would expect that reducing the hormone-free interval to only 2 days (i.e. 26 days of active treatment, 2 days of placebo [26/2 regimen]) would improve tolerability and, in turn, improve compliance.
The efficacy, bleeding profile and safety of E 2 V/DNG in a dynamic dosing regimen has been examined in three Phase III trials ( Table I). The first of these trials enrolled 1377 women aged 18-50 years and was conducted in Europe over twenty 28-day cycles [37]. All women received the regimen outlined above. The E 2 V/DNG combination was associated with an adjusted Pearl Index of 0.34 (upper limit of 95% CI = 0.73), together with good tolerability and a high degree of user satisfaction. Only 2.5% of 1377 women treated for up to 20 cycles prematurely discontinued treatment because of menstrual bleeding irregularities [37]. The second study, conducted in the USA and Canada, was an openlabel, non-comparative study designed to assess the contraceptive efficacy, cycle control, safety and tolerability of E 2 V/DNG. A total of 490 women aged 18-35 years received E 2 V/DNG for up to 28 cycles [48]. The third study, conducted in Europe, compared the E 2 V/DNG regimen with a monophasic COC (EE 20 µg/ LNG 100 µg) over seven cycles (n = 798) [24]. Scheduled withdrawal bleeding occurred in 77.7-83.2% of E 2 V/DNG recipients and 89.5-93.8% of EE/LNG recipients. The maximum intensity of withdrawal bleeding was significantly different in women treated with E 2 V/DNG and EE/LNG; a greater proportion of women who received E 2 V/DNG versus EE/LNG experienced spotting or light bleeding and a smaller portion of women who received E 2 V/DNG versus EE/LNG experienced normal or heavy bleeding. In addition, the mean duration of withdrawal bleeding was reduced with E 2 V/DNG compared with EE/LNG (4.1-4.7 vs. 5.0-5.2 days; p < 0.05 per cycle). Intracyclic bleeding was similar between the two COCs (10.5-18.6% vs. 9.9-17.1% per cycle; p > 0.05) (Table I) [24].
Based on data from the three Phase III trials performed in Europe, the USA and Canada, E 2 V/DNG was associated with a typical-use Pearl Index of 0.79 (upper limit of 95% CI = 1.23) and a perfect-use Pearl Index of 0.42 (upper limit 95% CI = 0.77) in women aged 18-50 years [49]. In women aged 18-35 years, the corresponding Pearl Indices were 1.01 (upper limit 95% CI = 1.59) and 0.51 (upper limit 95% CI = 0.97) [49].

Pharmacokinetics of estradiol valerate/DNG
The pharmacokinetics of E 2 V/DNG were analyzed in 15 healthy women aged 18-50 years who participated in a Phase I, openlabel, single-cycle study [17]. E 2 V/DNG was associated with stable serum levels of E 2 throughout the 28-day period of treatment ( Figure 4) [17,50]. Minimum mean serum E 2 levels during E 2 V administration (days 1-26) were 33.6-64.7 pg/mL, similar to those seen in the mid-follicular phase of normal ovulatory cycles, while minimum mean serum DNG levels were 6.8-15.1 ng/mL during DNG administration (days . Minimum concentrations of DNG showed only minor accumulation within each phase of the regimen during which DNG was administered. On day 24, the geometric mean maximum concentration of DNG was 82.9 ng/mL, while the average concentration and terminal halflife were 33.7 ng/mL and 12.2 h. The median time to maximum observed drug concentration for DNG was 1.5 h. Serum SHBG concentrations increased by 40%, but remained within the normal range. Cortisol-binding-globulin levels remained essentially unchanged.

Metabolic and vascular effects of estradiol valerate/DNG
Data have shown that E 2 V/DNG has an impact on various metabolic and hemostatic parameters that is comparable to or less than that of EE/LNG-containing COCs [51,52]. The effect on prothrombin and D-dimer levels is marginal [52], and the effects on high-and low-density lipoprotein cholesterol [51], insulin [51], and carbohydrate metabolism [51] are, in general, more favorable with E 2 V/DNG than with EE/LNG.
The hemostatic effects of E 2 V/DNG and E 2 /NOMAC have not been compared in a head-to-head trial; however, a comparison of data from different trials suggests that the overall magnitude of changes in hemostatic parameters during treatment with E 2 V/ DNG is comparable to that reported with the E 2 /NOMAC combination [43,51,52]. For example, E 2 V/DNG and E 2 /NOMAC were associated with similar effects on SHBG, prothrombin fragment 1+2, fibrinogen and thrombin generation (either determined by the activated protein C (APC) sensitivity ratio or by APC resistance) as measured by the endogenous thrombin generation method [43,[51][52][53]. This suggests that the hemostatic effects of COCs comprising either E 2 or E 2 V in equimolar dosages are comparable, as one would expect.
Although these results are reassuring, one should bear in mind that the lipid, metabolic and hemostatic parameters are only surrogate markers, and have yet to be fully validated as good predictors of the occurrence of clinical events such as VTE. Furthermore, even oral intake of E 2 is associated with a hepatic first pass effect [6,16] that may impair the biosynthesis and clearance of proteins involved in hemostasis or blood pressure regulation. Further clinical and long-term epidemiological studies of E 2 V/DNG in large populations are needed before any safety conclusions can be made. A large international prospective, controlled, non-interventional cohort active surveillance study (INAS-SCORE) to investigate the occurrence of cardiovascular events over a 3-to 5-year period in COC users (including E 2 V/DNG) is currently underway (ClinicalTrials.gov identifier: NCT01009684).
Additional non-contraceptive benefits of estradiol valerate/DNG E 2 V/DNG has been shown to be effective in women with HMB (defined as menstrual blood loss >80 mL). The registration procedure has been successfully concluded for an indication of HMB without organic pathology in the EU, Switzerland and USA and for heavy and/or prolonged menstrual bleeding in Australia and several Latin American and Asian countries. License applications for the HMB indication have also been submitted to health authorities in other countries. Overall, E 2 V/DNG was associated with an 88% reduction in median menstrual blood loss (from 142 mL to 17 mL/cycle) after 6 months of treatment, compared with a 24% reduction with placebo (from 154 mL to 117 mL/cycle) [54]. Reductions in MBL volume were rapid and sustained and were deemed clinically meaningful [54].
The efficacy of this combination in HMB is thought to be due to its unique dosing regimen, which enables estrogen dominance during the early part of the cycle and progestogen dominance in the late part of the cycle [24]. Women receive 26 days of E 2 V, which supports endometrial stability [17], and 22 days of DNG, a progestogen with high endometrial potency [44][45][46].
In summary, the development of dynamic regimens has yielded acceptable bleeding patterns whilst maintaining a reliable level of contraceptive efficacy. Data have shown this regimen to provide effective ovulation inhibition with an acceptable level of cycle control in healthy women.

Future directions
In modern contraception, one focus lies in additional health benefits. There is, therefore, a great deal of interest in whether E 2 -containing COCs have known or even new benefits. In order to establish these benefits, extensive efforts have been put into the development program for E 2 V/DNG, with one of the most comprehensive Phase IIIb clinical programs for a COC.
The HARMONY I and II studies are multicenter, randomized, double-blind, active control group studies comparing the efficacy of the E 2 V/DNG combination with that of EE/norgestimate (ClinicalTrials.gov identifier: NCT00754065) or EE/ LNG (ClinicalTrials.gov identifier: NCT00778609) for the treatment of hormone withdrawal-associated symptoms (HWAS), including headache, pelvic pain and bloating. The studies aim to demonstrate the superiority of E 2 V/DNG over the comparators with regards to HWAS improvement after six cycles of treatment. Both studies are underway and results are anticipated in 2011. It is expected that E 2 V/DNG will have a beneficial effect on cycle-related hormone withdrawal symptoms owing to less hormonal fluctuations. Specifically, E 2 V/DNG is associated with levels of E 2 that are stable and are comparable to those during the first week of the follicular phase of a spontaneous menstrual cycle [17]. In addition, the hormone-free interval with E 2 V/DNG (of 2 days) is shorter than that of conventional COCs (7 days). It has been shown previously that hormone-related symptoms are worse during the 7-day hormone-free interval than during active treatment [55], and shortening the hormone-free interval from 7 to 4 days may decrease the number of days of symptoms typically associated with hormone withdrawal [55][56][57].
The STABLE study (ClinicalTrials.gov identifier: NCT 00764881) aims to establish the non-inferiority of E 2 V/DNG over EE/LNG for improving libido in women with OC-associated female sexual dysfunction (FSD). Women presenting with OC-associated FSD are often switched to a COC containing EE/LNG, in the belief that the partial androgenic effects of LNG will alleviate FSD symptoms. STABLE is, to our knowledge, one of the first, if not the first, comprehensive, randomized controlled trial with a comparator arm, using validated questionnaires to investigate the effects of different COCs on OC-associated FSD. The benefit of E 2 V/DNG on libido in women with OC-induced FSD is thought to be due to the low impact of E 2 V/DNG on SHBG and its beneficial effect on vaginal cell maturation, demonstrating the complex and multifactorial nature of OC-associated FSD. Therefore, effects other than just the androgenic nature of the progestogen should also be taken into account in women with OC-associated FSD.
The CALM study (ClinicalTrials.gov identifier: NCT00909857) is investigating the effect of E 2 V/DNG on primary dysmenorrhea. In this multicenter, randomized, double-dummy, parallel-group study, women will receive E 2 V/DNG or EE/LNG for three cycles. Results for both STABLE and CALM are expected in 2011.

Conclusions
The quest to identify an effective, well tolerated COC using natural estrogens, such as E 2 , has encompassed a number of clinical trials over several decades. Earlier attempts to use E 2 in COCs explored various doses, progestogens and regimens, but yielded unfavorable results in terms of bleeding and cycle control. Recent data suggest that the combination of E 2 /NOMAC may yield promising results. Further studies are, however, needed.
Two E 2 V-containing COCs are currently available; E 2 V/CPA has been marketed as Femilar ® in Finland since 1993 (only indicated in women >40 years or women aged 35-40 years for whom a COC containing EE is not appropriate), and E 2 V/DNG has recently been launched in the EU as Qlaira ® /Klaira ® and in the USA as Natazia ® . There are extensive data available regarding the E 2 V/DNG combination, which is administered in a dynamic dosing regimen. This combination has been shown to provide women with reliable contraceptive efficacy whilst maintaining acceptable cycle control. It is also effective for the treatment of HMB, and may offer additional non-contraceptive benefits, such as improvements in OC-associated FSD, dysmenorrhea and hormone withdrawal-associated adverse events.