Impact of estrogen type on cardiovascular safety of combined oral contraceptives

Contraception 94 (2016) 328 – 339

Original research article

Impact of estrogen type on cardiovascular safety of combined oral contraceptives☆,☆☆,★ Jürgen Dinger a,⁎, Thai Do Minh b , Klaas Heinemann b b

a Pharmacoepidemiology, Berlin, Germany ZEG—Berlin Center for Epidemiology and Health Research, Berlin, Germany Received 14 March 2016; revised 14 June 2016; accepted 17 June 2016

Abstract Objectives: The International Active Surveillance study “Safety of Contraceptives: Role of Estrogens” (INAS-SCORE) investigated the cardiovascular risks associated with the use of a combined oral contraceptive (COC) containing dienogest and estradiol valerate (DNG/EV) compared to established COCs in a routine clinical setting. Study Design: Transatlantic, prospective, noninterventional cohort study conducted in the United States and seven European countries with two main exposure groups and one exposure subgroup: new users of DNG/EV and other COC (oCOC), particularly levonorgestrelcontaining COCs (LNG). All self-reported clinical outcomes of interest (OoI) were validated via attending physicians and relevant source documents. Main OoI were serious cardiovascular events (SCE), particularly venous thromboembolic (VTEs) events. Comprehensive followup procedures were implemented. Statistical analyses were based on Cox regression models. Results: A total of 50,203 new COC users were followed up for up to 5.5 years (mean value, 2.1 years). Overall 20.3% and 79.7% of these women used DNG/EV and oCOC (including 11.5% LNG users), respectively. A low loss to follow-up of 3.1% was achieved. Based on 47 (VTE) and 233 (SCE) events, the primary analysis (European data set) yielded adjusted hazard ratios for DNG/EV vs. oCOC of 0.4 and 0.5, respectively. The upper bounds of the 95% confidence intervals were 0.98 (VTE) and 0.96 (SCE). The corresponding hazard ratios for DNG/ EV vs. LNG showed similar point estimates but the confidence intervals included unity. Conclusion: DNG/EV is associated with similar or even lower cardiovascular risk compared to oCOC and LNG. Implication Statement: A COC containing DNG and EV is associated with similar or even lower cardiovascular risk compared to COCs containing levonorgestrel or other progestogens. © 2016 Elsevier Inc. All rights reserved. Keywords: VTE; ATE; Combined oral contraceptives; Dienogest; Estradiol valerate; Routine clinical practice

1. Introduction As combined oral contraceptives (COCs) have been further developed over the past decades, their ethinylestradiol (EE) content has been reduced based on the hypothesis ☆

Funding: Unconditional grant from Bayer AG, Germany. Conflict of interest: The study was funded by a manufacturer of hormonal contraceptives. The study was supervised by an independent Safety Monitoring and Advisory Council with full authority over the study (including study protocol, protocol amendments, data analysis and stopping the study). The funder had no access to the source data and did not participate in designing the study or analyzing the data. ★ Registration number at the clinical trials registry of the US National Library of Medicine: NCT01009684. ⁎ Corresponding author. Tel.: +49 171 974 5433; fax: +49 30 945 101 46. E-mail address: [email protected] (J. Dinger). ☆☆

http://dx.doi.org/10.1016/j.contraception.2016.06.010 0010-7824/© 2016 Elsevier Inc. All rights reserved.

that lower EE doses lead to a better safety profile and specifically to a lower incidence of venous thromboembolism (VTE). However, reducing the EE dose led also to a less favorable control of bleeding. Although EE has been used in numerous COCs, efforts have also been made to use estradiol (E2) and estradiol valerate (EV), which have a lower impact on the hepatic system and subsequently on hemostatic parameters [1]. Recently, a new EV-based COC was introduced to the market that appears to combine both reliable contraception and an acceptable bleeding profile [2]. This regimen consists of four phases within a 26-day time frame, with each phase containing different doses of EV, either alone or in combination with different doses of dienogest (DNG): (1) 2 tablets with 3 mg EV, (2) 5 tablets with 2 mg DNG and 2 mg EV, (3) 17 tablets with 3 mg DNG and 2 mg EV, and (4) 2 tablets with 1 mg EV. This

J. Dinger et al. / Contraception 94 (2016) 328–339

329

sequence is followed by 2 days of placebo tablets. The four-phase sequential regimen aims to ensure that sufficient estrogen levels are available during the first half of the cycle in which endometrial proliferation is promoted under the influence of estrogens. Shortening the hormone-free interval from the conventional 7 days to only 2 days and extending the estrogen phase at the end of the progestin phase are expected to be beneficial for cycle control and tolerability. Furthermore, it is conceivable that the short hormone-free interval has a beneficial effect on contraceptive failure rates [3]. This article describes the cardiovascular results from the regular follow-up phase of the International Active Surveillance Study on the Safety of Contraceptives and the Role of Estrogens (INAS-SCORE). Other outcomes will be reported elsewhere. The INAS-SCORE study was conducted as a phase IV commitment to the European regulatory authorities. Two main COC user groups — users of preparations containing DNG/ EV and users of other COCs (oCOC) plus an oCOC subgroup consisting of the users of levonorgestrel-containing COCs (LNG) — were followed throughout the study.

hospitalization, persistent or significant disability/incapacity or require medical/surgical intervention to prevent one of these outcomes) were analyzed, including confirmed angina necessitating hospitalization, acute coronary syndromes and congestive heart failure. Secondary objectives were to ascertain: (1) drug utilization patterns of DNG/EV and established COCs in a study population that is representative for typical use of the individual preparations under routine medical conditions — the study's nonexperimental design was intended to minimize interference of normal drug utilization patterns by study-specific requirements and measures; (2) baseline risk for users of the individual formulations (lifetime history of comorbidity, risk markers, comedication, sociodemographic and lifestyle data); (3) pregnancy-related data on discontinuation of DNG/EV and established COCs, that is, return to fertility and pregnancy outcomes; and (4) risks associated with short- and long-term use of DNG/EV and of established COCs in adolescents below the age of 18 years. The results of the secondary objectives will be reported elsewhere.

2. Materials and methods

2.2. Study population

A cohort of more than 50,000 new COC users was actively monitored for up to 5 years for the occurrence of rare or unexpected adverse outcomes possibly related to COC exposure. The methodology of the INAS-SCORE study is similar to that of the EURAS, INAS-OC and TASC studies on hormonal contraceptives described elsewhere [4–6], so some methodological details are presented here succinctly. Planning, conduct and evaluation of the study were supervised by an independent Safety Monitoring and Advisory Council, which endorsed all the conclusions presented in this publication. The primary ethical approvals of the study in Europe and the United States were granted by the physicians' association in Berlin, Germany (“EthikKommission der Ärztekammer Berlin”), and the Western Institutional Review Board (WIRB) in Olympia, WA, USA. The study is listed in the public clinical trials registry of the US National Library of Medicine under the number NCT01009684. 2.1. Study objectives The primary objective of the study was to assess the risks of short- and long-term use of DNG/EV, oCOC and LNG in a study population that is representative for the actual users of the individual preparations. The main clinical outcomes of interest for the short- and long-term follow-up were serious cardiovascular events, in particular VTE such as deep venous thrombosis and pulmonary embolism (PE), as well as arterial thromboembolism (ATE) such as acute myocardial infarction (AMI) and cerebrovascular accidents (CVA). In addition to VTE and ATE, all other serious cardiovascular events (i.e., events that result in death, a life-threatening experience, inpatient

The study was conducted in the United States and seven European countries: Austria, France, Germany, Italy, Poland, Sweden and the United Kingdom. A random sample of approximately 10% of all COC prescribers in these countries was contacted for participation in the study. Recruitment of the cohort members was conducted via a network of 880 and 427 COC-prescribing health care professionals in Europe and the United States, respectively. The combined cohort was planned to include 50,000 women, including about 20,000 in the United States and 30,000 in Europe. Recruitment in Europe began in September 2009 and ended in October 2012. Because of the late market introduction of DNG/EV in the United States, recruitment did not commence there until October 2010 and was completed in February 2013. Study participants were followed until February 2015. The individual maximum follow-up therefore varied from 2 to 5.5 years. Study participants were women who received a new prescription for a COC. Participating physicians discussed the study with the subjects only after a COC had been prescribed. This ensured that study participation was not considered a requirement for treatment. Participating women could be starters (first-ever users of COCs), switchers (users who switched from one COC to another — without an intake break or with one of less than 4 weeks) or restarters (users who restarted a COC after an intake break of at least 4 weeks, i.e., at least one complete cycle). More specific inclusion or exclusion criteria were not introduced in keeping with the noninterference approach of the study design. At the centers, all women seeking a prescription for a new COC were asked by their physicians whether they were willing to participate. All new COC users who were willing to sign the informed consent and data privacy forms had to

330

J. Dinger et al. / Contraception 94 (2016) 328–339

be included in the study. The objective was to avoid influencing the prescribing behavior, while at the same time making significant efforts to ensure standardized, comprehensive and reliable documentation of all baseline characteristics and adverse events during the follow-up period. Once enrolled, a subject could switch or discontinue use of the prescribed COC at any time. However, subjects continued to be followed whether or not they continued to use hormonal contraception, provided that they did not withdraw their consent. During the follow-up phase, subjects were asked whether they had switched or discontinued hormonal contraceptive use. Information on dates and reasons for switching/discontinuation during the follow-up phase was also collected. The recruitment procedures for the study centers and participants were the same as those used in the EURAS-OC and INAS-OC studies, where they successfully yielded representative samples of typical COC users with regard to age structure, socioeconomic and lifestyle factors, cardiovascular risk factors, the spectrum of prescribed OCs and percentages of urban and rural COC users [4,5,7–9]. 2.3. Data collection and quality control Baseline data were recorded within the INAS-SCORE study via self-administered questionnaires on participants' state of health, medical history including medication history and history of COC use, and potential prognostic factors for serious diseases, particularly cardiovascular disease. In addition, participants provided their addresses and phone numbers, as well as back-up contacts and contact information for their primary care physicians and/or gynecologists. Baseline questionnaires were completed in the physicians' offices and checked by the physicians or their coworkers. Follow-up assessments for each woman in the INAS-SCORE study were scheduled every 6 months for the first 2 years and annually thereafter. The follow-up questionnaires addressed the occurrence of adverse events — in particular serious adverse events and cardiovascular events. Reasons for discontinuing OC use or for switching to another hormonal contraceptive (OHC) were requested if applicable. The questionnaires were collected in each country by local study teams. These teams reviewed the questionnaires for completeness, plausibility and consistency of the responses. Missing or inconsistent information was clarified directly with the women by phone. In a second quality control step, the central study team in Berlin subjected all data to electronic and manual plausibility checks. A low loss to follow-up rate was essential for the validity of the study. In order to minimize loss to follow-up, a comprehensive four-level process was established [4]. Level 1 activities included mailing the follow-up questionnaire and — in case of no response — two reminder letters. Study participants received a small compensation for each follow-up on returning the questionnaire. If Level 1 activities

did not lead to a response, multiple attempts were made to contact the woman, her friends, relatives and gynecologist/ primary care physician via telephone. In parallel to these Level 2 activities, searches in national and international telephone and address directories as well as social networks were started (Level 3 activities). If this was not successful, an official address search via the respective governmental administration was conducted (in some countries centralized, in others decentralized at community level). The study protocol specified that the total loss to follow-up at the end of the study should be less than 10% of the study population. The four-level follow-up procedures are summarized in Fig. 1. For reported serious adverse events — including ATE and VTE — a group of medical doctors specializing in epidemiology, drug safety and internal medicine (medical reviewer group) contacted the study participants as well as the diagnosing or treating physicians to clarify and validate the information (including diagnosis, diagnostic procedures, exposure and treatment) received from participants [4]. All serious adverse events were classified as confirmed or not confirmed. Events that were confirmed by a diagnostic measure with high specificity (e.g., phlebography for deep venous thrombosis or cerebral magnetic resonance imaging for CVAs) or by a clinical diagnosis supported by a diagnostic test with low specificity (such as d-dimer for VTE) were considered confirmed. Events were considered not confirmed if the diagnosis reported by the participant was excluded by diagnostic measures, if a different medical condition was diagnosed by the attending physician, or if the participant did not contact a health professional to clarify her symptoms and no diagnostic measures were performed [4]. The primary analysis was based on confirmed cases only. Unconfirmed cases were only used for sensitivity analyses. For the analysis, classification of all VTE was verified by independent blinded adjudication. In order to minimize classification bias, all decisions made by the medical reviewer group were reassessed by three independent medical experts specializing in radiology and nuclear medicine, cardiology, as well as internal medicine and vascular diseases. These specialists reviewed all available information on the reported events. Brand names, doses, regimens and compositions of the hormonal contraceptives used by the study participants were rendered anonymous for this process. The adjudicators performed the reviews independently of each other and without knowing the judgment of the other adjudicators or the medical reviewer group. [4]. Events were classified as confirmed if that was the judgment of at least one adjudicator. 2.4. Evaluation The analyses were carried out in accordance with the statistical analysis plan, which was agreed upon with the European regulatory authorities and approved by the Safety Monitoring and Advisory Council prior to the first inferential

J. Dinger et al. / Contraception 94 (2016) 328–339

331

Recruitment Documentation of addresses and phone numbers - patients - relatives and/or friends of the patient - gynecologist and/or primary care physician

Follow-up 1st level

Follow-up questionnaire to all study participants every six months no response: 1st reminder letter no response: 2nd reminder letter no contact

2nd level

Multiple attempts to contact participant per phone using information provided in the baseline questionnaire: - patient - friends/relatives - gynecologist/primary care physician

contact

no contact 3rd level

Search in national and international telephone/address directories as well as social networks

contact

no contact 4th level

contacaddress inquiry at the National and/or Local Formal Registry of the appropriate public authority

contact

no contact

Lost to follow-up; exposure and occurrence of (S)AEs after the last follow-up unknown

Drop out; documentation of exposure and reported (S)AEs till the timepoint the patient withdrew her informed consent.

Fig. 1. Cascade of activities to contact patients for follow-up information.

analysis. Statistical analyses were conducted based on the “as treated” (AT) population as well as the “intention to treat” (ITT) population. For the AT analyses, data on outcomes of interest were assigned to the hormonal contraceptive actually used by the respective study participant at the time of the event. For the ITT analyses, all clinical outcomes were assigned to the treatment the study participant had used at study entry, regardless of any switching (or stopping) or of any different (or no) product being used at the time of the event. For a drug safety study, the ITT approach dilutes potential risk differences among treatments [4]. The investigators and the Safety Monitoring and Advisory Council therefore designated the AT approach as the primary analysis for the study data. In this study, however, conclusions based on ITT results did not differ

from those based on AT results. Therefore, only the most important ITT results are reported. Inferential statistics were based on Cox proportional hazard models. Adjustment for potential confounding was based on an a priori defined expert model (primary model). For VTE, this model included age, body mass index (BMI), duration of current hormonal contraceptive use and family history of VTE; for ATE, it included age, BMI, smoking, treated hypertension and a family history of fatal ATE. The prognostic factors were included as time-varying covariates in the statistical model. In addition, a “backward stepwise procedure” was chosen to generate an automated statistical model (secondary model). This procedure started with all available prognostic factors (e.g., age, BMI, duration of current hormonal contraceptive use, family history of

332

J. Dinger et al. / Contraception 94 (2016) 328–339

thromboembolic events, starter, restarter and switcher status, estrogen dose of the OC preparation, concomitant medication, chronic disease, smoking, geographical region, educational level) included in the statistical model as covariates (like in a saturated model). All prognostic factors that had no relevant impact on the risk estimates were removed from the model in a stepwise procedure. The results of the primary and secondary models were nearly identical. Therefore, only the results of the primary model are reported. Three exposure groups were compared: users of new COCs containing DNG/EV, users of oCOCs and users of LNG-containing COCs. All VTE and ATE were always adjudicated for the hormonal contraceptive used by the respective participant at the time. It did not matter whether the participant was a starter, a switcher or a restarter at the time of the event. If hormonal contraceptive use had been stopped during the 3-month period prior to the VTE diagnosis, the event was adjudicated to the last hormonal contraceptive used before the event. The analyses for VTE used two different data sets: all VTE and what are known as “idiopathic” VTE. The latter data set excludes cases with acute risk factors (such as pregnancy, delivery, trauma, immobilization, long-haul travel, surgery and chemotherapy). The analysis of this data set will allow future comparison of the INAS-SCORE data with the results from other scientific groups who often base their analyses on hypothesized “idiopathic” VTE. The study was powered to detect a 2.0-fold and 0.5-fold VTE risk for DNG/EV compared to oCOC, respectively. This calculation was based on the combined US and European study population. However, it turned out that the market introduction of DNG/EV in the United States was less successful than expected by the manufacturer of DNG/ EV. As a result, the proportion of US study participants who used DNG/EV was very small. This fact raised concerns on the part of the European regulatory authorities about basing study conclusions solely on the combined US and European data. They requested that an additional analysis be based on the European arm of the study alone. To compensate for the loss of statistical power, it was decided to extend the INAS-SCORE study in Europe. Despite this, the decision was also made to complete the first study part (transatlantic phase) as planned. This decision was based on the consideration at that time (January 2014) that the first part of the study would have a power of 0.8 to detect a two-fold increased risk of VTE that might potentially be associated with the use of DNG/EV.

3. Results A total of 53,750 women were enrolled by 1327 active study centers. Overall, 3547 of these 53,750 women (6.6%) had to be excluded because they: (1) were enrolled two or more times by one or more study centers (1.5%), (2) continued to use their previous hormonal contraceptive

(1.5%), (3) did not start OC use after study entry (2.5%), or (4) declined to sign the informed consent form (1.1%). The remaining 50,203 quality-controlled computerized data sets from the women (one per woman) with baseline information were analyzed. A total of 30,098 (60.0%) and 20,105 women (40.0%) were recruited in Europe and the United States, respectively (Table 1). In the combined European and US cohorts, these 50,203 study participants were followed up for 105,761 woman-years (WY) of observation (mean value, 2.1 years per study participant): 73,174 WY (69.2%) in Europe and 32,586 WY (30.8%) in the United States (Table 2). The late start of recruitment in the United States resulted in a shorter average follow-up in the United States compared to Europe: 1.6 and 2.4 years, respectively. At study entry, 10,191 women received a prescription for DNG/EV and 40,012 for oCOC. The latter included 5796 users of LNG (Table 1). During follow-up, 30%, 35% and 35% of users of DNG/EV, oCOC and LNG, respectively, switched from their baseline prescription to another OC or non-oral hormonal contraceptive brand. At the end of the INAS-SCORE study, DNG/EV, oCOC and LNG had been used for 12,512, 63,539 and 10,071 WY, respectively. For 4279 WY and 25,431 WY of 105,761 WY, study participants had switched to OHCs (e.g., patches, injections, vaginal rings) or had not used any hormonal contraceptive, respectively. Overall, 12% of the time without exposure to hormonal contraceptives (“no use”) can be attributed to intended and unintended pregnancies. Table 1 Number of women enrolled, excluded and analyzed. Women

Nos.

(%) a

[%] b

A) Who agreed to participate B) excluded because of protocol violations c C) analyzed Cohorts DNG/EV oCOC of which LNG Regions United States Europe European countries Austria France Germany Italy Poland Sweden UK Primary analysis (based on European data) DNG/EV oCOC of which LNG

53,750 3547 50,203

– – (100.0)

[100.0] [6.6] [93.4]

10,191 40,012 5796

(20.3) (79.7) (11.5)

[19.0] [74.4] [10.8]

20,105 30,098

(40.0) (60.0)

[37.4] [56.0]

2208 252 8613 8508 9131 1111 275

(4.4) (0.5) (17.2) (16.9) (18.2) (2.2) (0.5)

[4.1] [0.5] [16.0] [15.8] [17.0] [2.1] [0.5]

9791 20,307 3736

(19.5) (40.4) (7.4)

[18.2] [37.8] [7.0]

a

Percentage of women who agreed to participate. Percentage of women who were in the final analysis. c Women who (1) were enrolled two or more times by one or more study centers, or (2) continued their previous hormonal contraceptive, (3) never started OC use after study entry, or (4) declined to sign the informed consent form. b

J. Dinger et al. / Contraception 94 (2016) 328–339

333

Table 2 User cohorts (USA and Europe combined): number of women, exposure and descriptive statistics on age, weight and BMI at study entry. DNG/EV

n (%) WY (%) Age, mean (SD) Age, minimum Age, percentile 5 Age, percentile 25 Age, median Age, percentile 75 Age, percentile 95 Age, maximum Weight, mean (SD) Weight, minimum Weight, percentile 5 Weight, percentile 25 Weight, median Weight, percentile 75 Weight, percentile 95 Weight, maximum BMI, mean (SD) BMI, minimum BMI, percentile 5 BMI, percentile 25 BMI, median BMI, percentile 75 BMI, percentile 95 BMI, maximum a

N=10,191 (20.3) 12,512 (11.8) 31.7 (10.0) 11 17 23 31 40 48 59 62.7 (12.2) 30 48 55 60 69 85 172 23.0 (4.2) 12.7 18.0 20.2 22.2 24.8 31.1 60.9

oCOC

Total

All

LNG

N=40,012 (79.7) 63,539 (60.1) 26.0 (7.9) 11 16 20 24 30 42 58 66.6 (16.4) 30 48 55 63 73 99 191 24.5 (5.9) 10.6 18.1 20.4 23.0 26.9 36.2 71.9

N=5796 (11.5) 10,071 (9.5) 26.0 (8.4) 12 16 19 24 31 43 55 66.1 (15.2) 37 48 55 63 73 96 173 24.2 (5.4) 14.2 18.2 20.5 23.0 26.6 35.1 61.6

N=50,203 (100.0) 105,761 (100.0) a 27.1 (8.7) 11 16 20 25 32 44 59 65.8 (15.7) 30 48 55 62 72 96 191 24.2 (5.6) 10.6 18.1 20.4 22.8 26.4 35.3 71.9

Exposure includes 29,710 WY for women who stopped hormonal contraceptive use after study entry or switched to non-COC hormonal contraceptives.

After completing all four levels of the loss to follow-up cascade, the figure for both Europe and the United States was 3.1%. The planned follow-up procedures worked well in all countries. At the end of the transatlantic part, 1557 of the 50,203 women, or 3.1% (3.4% for DNG/EV, 3.0% for oCOC and 3.1% for LNG), were lost to follow-up during that 2- to 5.5-year period. Overall, all loss to follow-up rates were low and balanced across exposure groups. The goal of a loss to follow-up rate of less than 10% was achieved for the total study population, for each of the exposure groups, and for the European and US populations. For each of the main user groups (DNG/EV and oCOC) plus the LNG subgroup, Table 2 shows the number of women with baseline information (N), the exposure, the corresponding proportion of exposure for each of these populations, and descriptive statistics for age, weight, and BMI. At study entry, 20.3% of women were prescribed DNG/EV, 60.0% oCOC and 11.5% LNG. Mean age in the DNG/EV exposure group was 5.7 years higher than that in the COC and LNG (sub)-groups. The age distribution of the oCOC and LNG users — as indicated by the minimum, 5th, 25th, 50th, 75th and 95th percentiles, and the maximum values — corresponds to the typical age profile of oral contraceptive users [4,5]. The DNG/EV users were clearly older as indicated by a median of 31 years and a 75th percentile of 40 years. The latter exceeds the oCOC figure by about 10 years (Table 2). These differences were more

pronounced in the European study population (difference in mean: 6.2 years) and less so in the US study population (difference in mean: 1.5 years). This geographical difference could be explained by regional differences in (1) the use of E2-containing hormone replacement therapy (HRT) preparations and (2) the approved indications for DNG/EV at the time of patient enrollment. European gynecologists traditionally prescribe E2-containing HRT preparations, while HRT preparations with conjugated estrogens are much more prevalent in the United States. The association of E2 use with menopause in Europe probably leads to preferrential prescription of DNG/ EV to premenopausal women in need of contraception. In addition, shortly after its registration in Europe, DNG/EV was indicated for treating heavy menstrual bleeding in women who desired oral contraception, whereas in the United States, it was prescribed throughout nearly the entire study period for pregnancy prevention alone, and only lately was also approved for heavy menstrual bleeding. The higher prevalence of “older” women of fertile age who suffer from heavy menstrual bleeding probably contributes to the higher mean age of European DNG/EV users. Mean weight and mean BMI were similar for all COC (sub)-groups. This is true for both the European and the US study populations. The lower weight and BMI for the DNG/ EV cohort in the total study population (Table 2) is misleading. Given that (1) weight and BMI were substantially

334

J. Dinger et al. / Contraception 94 (2016) 328–339 60

60 DNG/EV

USA

LNG

40 40 38 38

%

23

DNG/EV

EUROPE

oCOC

oCOC

54 52 52

LNG

40

%

25 26

20

20 13 13 13

27 23

23 18

11 11 11

17

11 11 11

15

0

5

4 5

1

2 2

0 <20

[20-25[

[25-30[

[30-35[

35+

<20

[20-25[

[25-30[

[30-35[

35+

BMI

BMI

Fig. 2. Mean BMI by cohort and region.

higher in the United States than in Europe (Fig. 2), and (2) only few DNG/EV users were recruited in the United States, the combined DNG/EV data from the two continents were influenced much more by the relatively low European values than the high US values. Overall, 16,233 women (32.3% of the study population) were starters (first-time users) at study entry, 10,175 women (20.3%) were switchers, and 23,795 (47.4%) were restarters. No substantial differences were observed between the three (sub)-groups of DNG/EV, oCOC and LNG. As to be expected for COCs with progestogens that have been established for decades, the proportion of starters was slightly lower for oCOC (32%) and LNG (30%) compared to the new DNG/EV (34.0%). In addition, more European participants were starters compared to the United States (36% vs. 27%). No notable differences between the COC (sub)-cohorts were seen regarding gynecological history, including mean age at menarche (DNG/EV, 12.9 years; oCOC, 12.8 years; LNG, 12.9) and number of live births (DNG/EV, 1.7; oCOC, 1.7; LNG 1.7). The mean age at first delivery was also similar between cohorts and continents (Europe: 25, 24, and 24 years for DNG/EV, oCOC and LNG; USA: 23, 23 and 23 years for DNG/EV, oCOC and LNG).

Table 3 shows the distribution of prognostic factors for cardiovascular outcomes of interest and also the medical history of selected diseases. As can be seen, major differences between the three cohorts were not found at baseline for most of the risk factors examined. The prevalence of some of these factors seemed to be higher in the DNG/EV cohort. However, these differences (e.g., family history of VTE and family history of fatal ATE) disappeared after adjusting for age. The prevalence of other factors such as overweight and obesity seemed to be lower in the DNG/EV cohort. However, these figures reflect the low usage of DNG/EV and the higher prevalence of overweight and obesity in the United States. Sub-analyses showed that the age difference between the European exposure groups had a substantial impact on the prevalence of age-dependent variables such as hypertension and family history of VTE and fatal ATE (Table 3). The reason for this impact on family history is that the higher age of European DNG/EV users compared to oCOC and LNG users means that they also have older parents/siblings who consequently have a greater probability of VTE or fatal ATE. After age standardization, no substantial differences between the groups were seen, with 3.0%, 3.2% and 3.6% of European DNG/EV, oCOC and LNG users reporting a

Table 3 Prognostic factors for outcomes of interest, past medical history and selected diseases per user (sub)-group: total number and percent of enrolled women. DNG/EV Risk factor Treated high blood pressure High cholesterol Family history of fatal ATE Family history of VTE BMI [25–30 kg/m 2] BMI [30–35 kg/m 2] BMI 35+ kg/m 2 Smoking Heavy smoking a Diabetes Myocardial infarction Stroke PE Deep venous thrombosis Cancer Any surgery a

More than 15 cigarettes per day.

oCOC

LNG

Total

n

%

n

%

n

%

n

%

266 110 278 400 1789 517 177 2846 422 53 0 5 1 22 44 3352

2.6 1.1 2.7 3.9 17.6 5.1 1.7 27.9 4.1 0.5 0.0 0.0 0.0 0.2 0.4 32.9

820 370 764 1158 8020 3476 2492 8679 1096 260 5 15 5 34 198 11,150

2.0 0.9 1.9 2.9 20.0 8.7 6.2 21.7 2.7 0.6 0.0 0.0 0.0 0.1 0.5 27.9

139 51 105 187 1549 202 4375 1549 202 43 1 4 0 6 22 1800

2.4 0.9 1.8 3.2 26.7 3.5 7.3 26.7 3.5 0.7 0.0 0.1 0.0 0.1 0.4 31.1

1086 480 1042 1558 1223 441 301 11,525 1518 313 5 20 6 56 242 14,502

2.2 1.0 2.1 3.1 21.1 7.6 5.2 23.0 3.0 0.6 0.0 0.0 0.0 0.1 0.5 28.9

J. Dinger et al. / Contraception 94 (2016) 328–339 50

50

USA

EUROPE

40

40

30

30

% 20

335

15.8

12.6

14.5

10

28.4

30.5

33.5

DNG/EV

oCOC

LNG

% 20 10

0

0 DNG/EV

oCOC

LNG

Fig. 3. Current smoking stratified by cohort and region.

family history of VTE. The corresponding proportions of European women with a family history of fatal ATE below the age of 50 years were 1.9%, 1.8% and 1.7% for DNG/EV, oCOC and LNG. The prevalence of smoking was similar across (sub)-cohorts, although substantial continent-specific differences were observed. In the USA, 15.8%, 12.6% and 14.5% of DNG/EV, oCOC and LNG users were current smokers at baseline compared to 28.4%, 30.5% and 33.5% in Europe (Fig. 3). Regular use of concomitant medication was slightly different for the COC (sub)-cohorts: USA 25.5%, 23.7% and 27.9% for DNG/EV, oCOC and LNG; Europe 15.4%, 12.0% and 12.0% for DNG/EV, oCOC and LNG (Fig. 4). The prevalence was substantially higher in the United States compared to Europe (23.7% vs. 13.2%). Psychotropics were the most widely used concomitant medication in the USA, with 10.0% of the US study population taking them regularly compared to 2.3% in Europe. Thyroid drugs (including iodine-containing preparations) were used most frequently in Europe (3.8%). Overall, the oCOC and LNG exposure groups showed typical characteristics of US and European COC user populations regarding age structure, socioeconomic and lifestyle factors and cardiovascular risk factors [4,5,7–9]. The most important difference between the COC (sub)groups was the substantially higher age of DNG/EV users compared to oCOC and LNG. Given this difference, the risk of serious cardiovascular events was a priori higher for DNG/EV users compared to those of oCOCs. It was to be expected that age-adjusted relative risk estimates for

cardiovascular outcomes associated with DNG/EV use would be lower compared to crude risk estimates. 3.1. VTE events A total of 77 VTEs were observed, with a lower incidence rate in the DNG/EV group compared to the oCOC groups: DNG/EV 9 cases for 7.2 VTE per 10,000 WY, oCOC 58 cases for 9.1 VTE per 10,000 WY, and LNG 10 VTE for 9.9 VTE per 10,000 WY. The incidence rate in the “no use” cohort (9 cases for 3.5 VTE per 10,000 WY) was substantially lower compared to the COC (sub)-cohorts. The results for those women who switched after recruitment to OHCs (i.e., injections, implants, LNG-releasing IUDs, or contraceptive patches) are too sparse (1 VTE) for any meaningful analysis. Table 4 shows the number of VTE, point estimates and 95% confidence intervals (95% CIs) for the exposure groups. For 26 of the 77 VTE cases (34%), a PE was observed (DNG/EV cohort: 5 cases; oCOC cohort: 19 cases; LNG sub-cohort: 5 cases; OHC cohort: no case; “no use” cohort: 2 cases). The PE incidence rates for the COC cohorts were similar with a broad overlap of CIs: DNG/EV 5 PE for 4.0 cases per 10,000 WY (95% CI, 1.3–9.3), oCOC 19 PE for 3.0 cases per 10,000 WY (95% CI, 1.8–4.7), and LNG 5 PE for 5.0 cases per 10,000 WY (95% CI, 1.6–11.6). The VTE risk for COC users was approximately 2.5 times higher than that for nonusers. A total of 5 out of 9 VTE in the “no use” cohort were associated with pregnancy and delivery. Exclusion of these cases resulted in an incidence

60

60

Europe

USA 50

50

40

40

% 30

25.5

27.9 23.7

% 30

20

20

10

10

0

15.8 12.0

12.0

oCOC

LNG

0 DNG/EV

oCOC

LNG

DNG/EV

Fig. 4. Regular use of medication by cohort and region.

336

J. Dinger et al. / Contraception 94 (2016) 328–339

Table 4 VTE events: number, incidence and 95% CIs per exposure group. Data set

USA and Europe (confirmed VTE) USA (confirmed VTE) Europe - Confirmed VTE (primary analysis) - Confirmed and potential VTE - “Idiopathic” VTE a

DNG/EV

oCOC a

LNG

n

Incidence (95% CI)

7.2 (3.3–13.7) 0.0 (0.0–81.9)

58 25

7.4 (3.4–14.1) 9.1 (4.5–16.2) 4.9 (1.8–10.7)

33 41 24

n

Incidence (95% CI)

9 0 9 11 6

a

OHC

n

Incidence (95% CI)

9.1 (6.9–11.8) 10.4 (6.8–15.4)

10 4

8.3 (5.7–11.7) 10.4 (7.4–14.0) 6.1 (3.9–9.0)

6 7 6

a

n

Incidence (95% CI)

9.9 (4.8–18.3) 14.3 (3.9–36.7)

1 0

8.2 (3.0–17.9) 9.6 (3.9–19.8) 8.2 (3.0–17.9)

1 1 1

No use a

Total a

n

Incidence (95% CI)

n

2.3 (0.1–13.0) 0.0 (0.0–16.6)

9 5

3.5 (1.6–6.7) 7.7 (2.5–18.0)

77 30

4.0 (0.1–22.5) 4.0 (0.1–22.5) 4.0 (0.1–22.5)

4 6 2

2.1 (0.6–5.4) 3.2 (1.2–6.9) 1.1 (0.1–3.8)

47 59 33

Incidence rates are given in events/10,000 WY.

rate of 1.6 VTE per 10,000 WY. Exclusion of 21 US and European VTE cases associated with acute risk factors for VTE (“idiopathic” VTE) resulted in an overall VTE incidence in COC users of 6.0 per 10,000 WY. The VTE risk for COC users without acute risk factors was about 3.8 times higher than that for non-pregnant nonusers without acute risk factors. Cox regression analysis of the combined US and European data was performed for exploratory reasons only. The following predefined prognostic factors were included in the Cox regression model: age, BMI, duration of current use and family history of VTE. The results are shown in Table 5. The crude hazard ratio (HRcrude) for DNG/EV vs. oCOC was 0.8 (95% CI, 0.4–1.6). The corresponding adjusted hazard ratio (HRadj.) was 0.5 (95% CI, 0.2–1.0). The effect of the adjustment reflects primarily the differences in the age profile of the two cohorts. Alternative analyses using a backwards stepwise procedure for the selection of prognostic factors yielded almost identical results. A comparison of the DNG/EV and LNG groups showed similar results (Table 5): the crude and adjusted VTE HRs were 0.7 (95% CI, 0.3–1.8) and 0.5 (95% CI, 0.2–1.3), respectively.

A total of 47 VTE were observed in the European study population (primary analysis) with a lower incidence rate in the DNG/EV cohort compared to the oCOC (sub)-cohorts: DNG/EV 9 cases for 7.4 VTE per 10,000 WY, oCOC 33 cases for 8.3 VTE per 10,000 WY, and LNG 6 VTE for 8.2 VTE per 10,000 WY. The incidence rate in the “no use” group (4 cases for 2.1 VTE per 10,000 WY) was substantially lower compared to the COC exposure groups. Number of VTE, point estimates and 95% CIs for the (sub)-cohorts are presented in Table 4. The HRcrude for DNG/EV vs. oCOC was 0.9 with a 95% CI of 0.4 to 1.8. The HRadj. was 0.4 (Table 5) with an upper 95% confidence limit of 0.98. The effect of the adjustment reflects primarily the differences in the age profile of the two cohorts. Alternative analyses using a backward stepwise procedure for the selection of prognostic factors yielded almost identical results. A comparison of the DNG/EV cohort with the LNG sub-cohort showed similar point estimates with wider CIs: the crude and adjusted VTE hazard ratios were 0.8 (95% CI, 0.3–2.4) and 0.5 (95% CI, 0.2–1.5), respectively.

Table 5 VTE incidence rates, HRcrude and HRadj, and 95% CIs. Data set

USA and Europe (confirmed VTE)

Europe Confirmed VTE (primary analysis)

Confirmed and potential VTE

“Idiopathic” VTE

a b

Exposure

Incidence (events/10,000 WY)

HR (DNG/EV vs. comparators)

Point estimate

95% CI

Crude estimate

95% CI

Adjusted a estimate

95% CI

DNG/EV oCOC LNG

7.2 9.1 9.9

3.3–13.7 6.9–11.8 4.8–18.3

– 0.8 0.7

– 0.4–1.6 0.3–1.8

– 0.5 0.5

– 0.2–1.0 0.2–1.3

DNG/EV oCOC LNG DNG/EV oCOC LNG DNG/EV oCOC LNG

7.4 8.3 8.2 9.1 10.4 9.6 4.8 6.3 8.9

3.4–14.1 5.7–11.7 3.0–17.9 4.5–16.2 7.4–14.0 3.9–19.8 1.8–10.4 4.5–8.6 4.1–17.0

– 0.9 0.8 – 0.8 0.9 – 0.8 0.6

– 0.4–1.8 0.3–2.4 – 0.4–1.6 0.3–2.3 – 0.3–1.9 0.2–1.9

– 0.4 0.5 – 0.5 0.5 – 0.4 0.4

– 0.2–1.0 b 0.2–1.5 – 0.3–1.0 b 0.2–1.5 – 0.2–1.1 0.1–1.4

Adjusted for age, BMI, current duration of use and family history of VTE. Statistically significant.

J. Dinger et al. / Contraception 94 (2016) 328–339

337

for the DNG/EV group compared to the oCOC and LNG (sub)-groups, but the DNG/EV point estimate is based on one case only and the 95% CIs overlap widely. The 4 AMIs break down among the (sub)-groups as follows: DNG/EV no case, oCOC 3 cases, LNG 1 case, OHC no case, and “no use” 1 case. This corresponds to AMI incidence rates of 0.0 ATE/10,000 WY for the DNG/EV cohort, and of 0.5, 1.0, 0.0 and 0.4 for the oCOC, LNG, OHC and “no use” cohorts, respectively (Table 6). Overall, 10 cases of ischemic strokes and 2 cases of TIA occurred: DNG/ EV 1 and 0 case, oCOC 9 and 1 case(s), LNG 0 and 0 case, OHC 0 and 0 case, and “no use” 0 and 1 case. This corresponds to the following incidence rates: DNG/EV, 0.8 and 0.0 events/10,000 WY; oCOC, 1.4 and 0.2 events/ 10,000 WY; LNG, 0.0 and 0.0 events/10,000 WY; OHC, 0.0 and 0.0 events/10,000 WY; and “no use” 0.0 and 0.4 events/ 10,000 WY (Table 6). The results for Europe alone are similar to the overall ATE results (Table 6). The US data were too sparse for any meaningful comparison between the (sub)-groups (Table 6). Cox regression analysis was not done, in accordance with the analysis plan which stipulated that hazard ratios were only to be calculated if a minimum of 5 confirmed events were available in each of the comparison groups. This requirement was not fulfilled for two of the three COC (sub)-groups (DNG/EV and LNG). For the European data set (primary analysis), the ATE incidence rate ratios for DNG/ EV vs. oCOC and LNG were 0.3 (95% CI, 0.0–1.8) and 0.6 (95% CI, 0.0–47.0).

No VTEs were observed in US users of DNG/EV. Therefore, no meaningful analysis of the US VTE data is possible. Number of VTE, point estimates and 95% CIs for the (sub)-cohorts are presented in Table 4. The validation process for reported VTE identified 19 events that did not represent VTE according to the definition above but could still not be completely ruled out. The blinded adjudicators unanimously classified them as non-VTE. As such, these cases were probably not VTEs. In order to assess possible error, an additional evaluation was performed in which these potential VTEs were combined with confirmed VTEs. The results are shown in Tables 4 and 5. Overall, these results are similar to those for confirmed VTE. For the analysis of “idiopathic” VTE, 26 and 14 VTE cases with acute risk factors (such as pregnancy, delivery, trauma, immobilization, long-haul travel, surgery, chemotherapy) were excluded from the combined US and European and the primary (Europe only) data set, respectively. Tables 4 and 5 show the results compared to those of confirmed VTE in Europe. Overall, Cox regression analyses yielded only slight differences between the results of confirmed and “idiopathic” VTE. 3.2. Arterial thromboembolic events A total of 18 ATE were observed in the study (Table 6): 4 AMIs, 10 ischemic strokes, 2 transient ischemic attacks (TIAs) and 2 complete thromboses of a peripheral artery. The ATEs break down among the (sub)-groups as follows: DNG/ EV 1 case, oCOC 15 cases, LNG 1 case, OHC 0 cases and “no use” 2 cases. This corresponds to ATE incidence rates of 0.8 ATE/10,000 WY for the DNG/EV cohort, and of 2.4, 1.0, 0.0 and 0.8 for the oCOC, LNG, OHC and “no use” (sub)-cohorts, respectively. The incidence rates were lower

3.3. Serious cardiovascular events A total of 233 serious cardiovascular events were observed in the European study population: DNG/EV 33

Table 6 Arterial thromboembolic events: number, incidence and 95% CIs per exposure group. Category

DNG/EV

oCOC a

Complete data set All ATE of which AMI Ischemic stroke TIA Peripheral ATE European data All ATE of which AMI Ischemic stroke TIA Peripheral ATE US data All ATE of which AMI Ischemic stroke TIA Peripheral ATE a

LNG a

OHC a

No use a

Total a

n Incidence (95% CI) n

Incidence (95% CI) n Incidence (95% CI) n Incidence (95% CI) n Incidence (95% CI) n

1 0 1 0 0

0.8 0.0 0.8 0.0 0.0

(0.0–4.5) (0.0–2.4) (0.0–4.5) (0.0–2.4) (0.0–2.4)

15 3 9 1 2

2.4 0.5 1.4 0.2 0.3

(1.3–3.9) (0.1–1.4) (0.6–2.7) (0.0–0.9) (0.0–1.1)

1 1 0 0 0

1.0 (0.0–5.5) 1.0 (0.0–5.5) 0.0 (0.0–3.0) 0.0 (0.0–3.0) 0.0 (0.0–3.0)

0 0 0 0 0

0.0 (0.0–7.0) 0.0 (0.0–7.0) 0.0 (0.0–7.0) 0.0 (0.0–7.0) 0.0 (0.0–7.0)

2 1 0 1 0

0.8 0.4 0.0 0.4 0.0

(0.1–2.8) (0.0–2.2) (0.0–1.2) (0.0–2.2) (0.0–1.2)

18 4 10 2 2

1 0 1 0 0

0.8 0.0 0.8 0.0 0.0

(0.0–4.6) (0.0–2.4) (0.0–4.6) (0.0–2.4) (0.0–2.4)

12 2 8 0 2

3.0 0.5 2.0 0.0 0.5

(1.6–5.3) (0.6–1.8) (0.9–4.0) (0.0–0.8) (0.6–1.8)

1 1 0 0 0

1.4 (0.0–7.6) 1.4 (0.0–7.6) 0.0 (0.0–4.1) 0.0 (0.0–4.1) 0.0 (0.0–4.1)

0 0 0 0 0

0.0 (0.0–12.1) 0.0 (0.0–12.1) 0.0 (0.0–12.1) 0.0 (0.0–12.1) 0.0 (0.0–12.1)

2 1 0 1 0

1.1 0.5 0.0 0.5 0.0

(0.1–3.8) (0.0–2.9) (0.0–1.6) (0.0–2.9) (0.0–1.6)

15 3 9 1 2

0 0 0 0 0

0.0 0.0 0.0 0.0 0.0

(0.0–82.0) (0.0–82.0) (0.0–82.0) (0.0–82.0) (0.0–82.0)

3 1 1 1 0

1.3 0.4 0.4 0.4 0.0

(0.3–3.7) (0.0–2.3) (0.0–2.3) (0.0–2.3) (0.0–1.3)

0 0 0 0 0

0.0 (0.0–10.7) 0.0 (0.0–10.7) 0.0 (0.0–10.7) 0.0 (0.0–10.7) 0.0 (0.0–10.7)

0 0 0 0 0

0.0 (0.0–16.6) 0.0 (0.0–16.6) 0.0 (0.0–16.6) 0.0 (0.0–16.6) 0.0 (0.0–16.6)

0 0 0 0 0

0.0 0.0 0.0 0.0 0.0

(0.0–4.6) (0.0–4.6) (0.0–4.6) (0.0–4.6) (0.0–4.6)

3 1 1 1 0

Incidence rates are given in events/10,000 WY.

338

J. Dinger et al. / Contraception 94 (2016) 328–339

cases, oCOC 150 cases, LNG 32 cases, OHC 7 cases, and “no use” 43 cases. This corresponds to incidence rates of 27.2 cases/10,000 WY (95% CI, 18.7–38.1) for DNG/EV, and of 37.9 (95% CI, 32.1–44.4), 43.9 (95% CI, 30.1–62.0), 28.3 (95% CI, 11.4–58.2) and 22.7 (95% CI, 16.4–30.5) for oCOC, LNG, OHC and “no use”, respectively. The HRadj for the comparison of DNG/EV and oCOC was 0.6 with an upper confidence limit of 0.96. The point estimate of the HRadj. for the comparison of DNG/EV vs. LNG was also 0.6 but the 95% CI included unity (0.4–1.1). The results for the combined US and European data set were similar to those for the European data set (primary analysis).

4. Discussion The incidence rates for VTE, ATE and serious cardiovascular events were lower for DNG/EV compared to oCOC and LNG. The primary statistical analysis (European data set) of these outcomes yielded HRadj of 0.4 for VTE and 0.6 for serious cardiovascular events for the comparison of DNG/EV vs. oCOC. The corresponding 95% CIs did not include unity and suggested superiority of DNG/EV. The data on ATE were too sparse for meaningful statistical testing. Comparisons vs. LNG yielded point estimates of the HRadj of 0.5 and 0.6 for VTE and serious cardiovascular events (i.e., similar to the comparison of DNG/EV vs. oCOC) but the 95% CIs included unity. The fact that the point estimates of the HRs were similar for DNG/EV vs. oCOC and DNG/EV vs. LNG suggests that potential differences between the cohorts were not “diluted” by the inclusion of COCs that are potentially associated with an increased risk of VTE. Furthermore, analyses that included potential VTE or were restricted to “idiopathic” VTE showed similar results to the analysis of confirmed VTE. In addition, sensitivity analyses that included all other available prognostic factors for VTE confirmed the results of the primary analyses. The methodological limitations of prospective, controlled, noninterventional, active surveillance cohort studies such as INAS-SCORE have been discussed elsewhere [4–6,10]. Like in other nonexperimental studies, it is not possible to entirely eliminate potential effects of bias and residual confounding [11–16]. Therefore, the ability to infer causation is limited [11] and relative risk estimates that are close to unity may not allow differentiation between causation, bias and confounding [15,16]. In general, it is difficult to interpret a relative risk of 1.5 to 2.0 and 0.5 to 0.67 in observational research; a valid interpretation of relative risks of 1.0 to 1.5 and 0.67 to 1.0 is almost impossible [17,18]. Within these limitations, the INAS-SCORE study combines several methodological strengths that support the validity of its results: (1) a prospective, comparative cohort design; (2) availability of important confounder information (e.g., BMI and family history of cardiovascular outcomes);

(3) validation of outcomes of interest and exposure for the relevant cases; (4) comprehensive long-term follow-up and very low loss to follow-up to minimize underreporting; (5) independent, blinded adjudication of critical outcomes; (6) relevant statistical analyses (e.g., stratified analyses by user status and exposure period; comparison of isochronous, new user cohorts; sensitivity analyses on the impact of prognostic factors); (7) study population with baseline characteristics similar to OC users under routine clinical conditions; (8) reproducibility of the typical time pattern of VTE risk; and (9) supervision by an independent Safety Monitoring and Advisory Council as well as scientific independence from the study funder. In our judgment, selection bias was probably not a major issue in the INAS-SCORE study because adverse events for both inpatients and outpatients were included in the analyses and because participants' demographic characteristics were similar to those in the EURAS- and INAS-OC studies, which were representative for typical COC users [4,5]. Also, misclassification bias probably had no substantial impact on the results because precise information on the exposure and the outcomes of interest were available. In addition, reliable information on duration of current use was available. Furthermore, the low loss to follow-up rate of 3.1% is noteworthy. In theory, a disproportionately high percentage of serious adverse events (including VTE, ATE and serious cardiovascular events) could have occurred in those patients who were lost to follow-up, because these events could be the reason for the break in contact with the investigators. An advantage of the INAS-SCORE study design, however, is that the investigator team had direct contact with the participants; contact was not lost if the women changed their gynecologists, for example (e.g., due to change of residence or dissatisfaction with treatment). In contrast, it was impossible to exclude diagnostic bias. Clinical symptoms of VTE cover the spectrum from a complete absence or unspecific, slight symptoms to dramatic, acute, life-threatening symptoms [19–21]. A high awareness of potential cardiovascular risks of COC use might have led to more diagnostic procedures and, therefore, to more detected VTE, ATE or other serious cardiovascular outcomes. It is conceivable that this potential bias leads to an overestimate of the relative risk for new COCs like DNG/EV. Therefore, diagnostic bias should not result in underestimating the cardiovascular risk carried by DNG/EV. Unlike many other observational studies on hormonal contraceptives, a strength of this study was the availability of information on many important prognostic factors for the outcomes of interest. Nevertheless, we acknowledge that due to the noninterventional character of the study, information on specific gene mutations was only available for VTE cases but not for the vast majority of study participants. This limitation was mitigated by information on family history of VTE which has a higher predictive value for VTE compared to gene mutations [22].

J. Dinger et al. / Contraception 94 (2016) 328–339

Overall, the methodological quality of the study probably allows for a cautious interpretation of relative risks of 1.5–2.0 (or higher) and 0.67–0.5 (or lower), respectively. We conclude that the study is methodologically valid to exclude a twofold VTE and serious cardiovascular event risk for DNG/EV users compared to users of oCOCs or LNG. Data on ATE are sparse but do not suggest an increased ATE risk for DNG/EV. In addition, the hazard ratios for the risk of VTE and serious cardiovascular events are sufficiently different from unity (approximately 0.5) to show superiority of DNG/EV compared to oCOC and LNG given that the statistical power is sufficient. Therefore, it will be interesting to reanalyze the study data at the end of the extension phase when the statistical power of the analyses will be substantially higher compared to that of the analyses presented here.

[5]

[6]

[7]

[8]

[9] [10]

Acknowledgments The authors would like express their appreciation to the members of the independent Safety Monitoring and Advisory Council for their constructive criticism and unfailingly fair scientific discussion. The authors would also like to highlight the contributions of numerous colleagues who were responsible for the field work in the individual countries. They clarified data inconsistencies and missing data, validated patient-reported adverse events with patience, care and tenacity, and their untiring commitment enabled a remarkably low loss to follow-up rate. The authors' special thanks are due to Dr. Sabine Möhner for data management and Ms. Marlene Schoofs for editorial support in preparing the manuscript. References

[1] Wiegratz I, Lee JH, Kutschera E, Winkler UH, Kuhl H. Effect of four oral contraceptives on hemostatic parameters. Contraception 2004;70:97–06. [2] Nelson A, Parke S, Mellinger U, Zampaglione E, Schmidt A. Efficacy and safety of a combined oral contraceptive containing estradiol valerate/dienogest: results from a clinical study conducted in North America. J Women's Health 2014;23:204–10. [3] Graziottin A. The shorter, the better: a review of the evidence for a shorter contraceptive hormone-free interval. Contracept Reprod Health Care 2015;20:1–3. [4] Dinger JC, Heinemann LAJ, Kühl-Habich D. The safety of a drospirenone-containing oral contraceptive: final results from the

[11] [12] [13]

[14] [15]

[16]

[17] [18] [19]

[20]

[21]

[22]

339

European Active Surveillance study on Oral Contraceptives based on 142,475 women-years of observation. Contraception 2007;75:344–54. Dinger J, Bardenheuer K, Heinemann K. Cardiovascular and general safety of a 24-day regimen of drospirenone-containing combined oral contraceptives: final results from the International Active Surveillance Study of Women Taking Oral Contraceptives. Contraception 2014;89:253–63. Dinger J, Möhner S, Heinemann K. Cardiovascular risk associated with the use of an etonogestrel-containing vaginal ring. Obstet Gynecol 2013;122:800–8. Potthoff P, Heinemann LAJ, Güther B. A household panel as a tool for cost-effective health-related population surveys: validity of the Healthcare Access Panel. Ger Med Sci 2004;2 [Doc05/20040716]. Güther B. Access panel – quality aspects and experience. In: Streich W, Braun B, & Helmert U, editors. Surveys in public health—developments and perspectives in research and political counseling. Asgard Verlag: Sankt Augustin; 2005, pp. 91–02. Potthoff P, Güther B. Access panel in the context of health research. Forum Public Health 2005;13:16–8. Dinger J, Möhner S, Heinemann K. Cardiovascular risks associated with the use of drospirenone-containing combined oral contraceptives. Contraception 2016;93:378–85, http://dx.doi.org/10.1016/ j.contraception.2016.01.012. Susser M. What is a cause and how do we know one? A grammar for pragmatic epidemiology. Epidemiol 1991;133:635–48. Rothman KJ, Poole C. A strengthening programme for weak associations. Epidemiol 1988;17:955–9. Khoury MJ, James LM, Flanders WD, Erickson JD. Interpretation of recurring weak associations obtained from epidemiologic studies of suspected human teratogens. Teratology 1992;46:69–77. Shapiro S. Bias in the evaluation of low-magnitude associations: an empirical perspective. Epidemiol 2000;151:939–45. Shapiro S. Causation, bias and confounding: a hitchhiker's guide to the epidemiological galaxy. Part 2. Principles of causality in epidemiological research: confounding, effect modification and strength of association. J Fam Plann Reprod Health Care 2008;34:185–90. Shapiro S. Causation, bias and confounding: a hitchhiker's guide to the epidemiological galaxy. Part 3: principles of causality in epidemiological research: statistical stability, dose- and duration response effects, internal and external consistency, analogy and biological plausibility. J Fam Plann Reprod Health Care 2008;34:261–4. Taubes G. Epidemiology faces its limits. Science 1995;269:164–9. Hill AB. The environment and disease: association or causation? Soc Med 1965;58:295–300. Scurr JH, Machin SJ, Bailey-King S, Mackie IJ, McDonald S, Smith PDC. Frequency and prevention of symptomless deep-vein thrombosis in long-haul flights: a randomised trial. Lancet 2001;357:1485–9. Belcaro C, Geroulakos G, Nicolaides AN, Myers KA, Winford M. Venous thromboembolism from air travel: the LONFLIT study. Angiology 2001;52:369–74. Schwarz T, Siegert G, Oettler W, Halbritter K, Beyer J, Frommhold R, et al. Venous thrombosis following long-haul flights. Arch Intern Med 2003;163:2759–64. Bezemer ID, van der Meer FJ, Eikenboom JC, Rosendaal FR, Doggen CJ. The value of family history as a risk indicator for venous thrombosis. Arch Intern Med 2009;169:610–5.