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
a
Pharmacoepidemiology, Berlin, Germany
b
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 levonorgestrel-
containing 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 follow-
up 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 that lower EE doses lead to a better safety profile and

As combined oral contraceptives (COCs) have been
further developed over the past decades, their ethinylestra-
diol (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.
★ either alone or in combination with different doses of
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: pharmacoepidemiology@t-online.de (J. Dinger).
http://dx.doi.org/10.1016/j.contraception.2016.06.010
0010-7824/© 2016 Elsevier Inc. All rights reserved.
specifically to a lower incidence of venous thromboembo-
lism (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,
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
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.
2. Materials and methods
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 (“Ethik-
Kommission 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 thromboem-
bolism (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
329
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.
2.2. Study population
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, compre-
hensive and reliable documentation of all baseline charac-
teristics 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, cardio-
vascular 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 informa-
tion 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 applica-
ble. The questionnaires were collected in each country by
local study teams. These teams reviewed the question-
naires 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 contact
- friends/relatives
- gynecologist/primary care physician

no contact

3rd level Search in national and international telephone/address contact

directories as well as social networks

no contact

4th level contacaddress inquiry at the National and/or Local

Formal contact
Registry of the appropriate public authority

no contact

Lost to follow-up; Drop out;

exposure and occurrence of (S)AEs after documentation of exposure and reported
the last follow-up unknown (S)AEs till the timepoint the patient with-
drew 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 partic-
ipant 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, (1.5%), (3) did not start OC use after study entry (2.5%), or
estrogen dose of the OC preparation, concomitant medica- (4) declined to sign the informed consent form (1.1%). The
tion, chronic disease, smoking, geographical region, educa- remaining 50,203 quality-controlled computerized data sets
tional level) included in the statistical model as covariates from the women (one per woman) with baseline information
(like in a saturated model). All prognostic factors that had no were analyzed. A total of 30,098 (60.0%) and 20,105 women
relevant impact on the risk estimates were removed from the (40.0%) were recruited in Europe and the United States,
model in a stepwise procedure. The results of the primary respectively (Table 1). In the combined European and US
and secondary models were nearly identical. Therefore, only cohorts, these 50,203 study participants were followed up for
the results of the primary model are reported. 105,761 woman-years (WY) of observation (mean value,
Three exposure groups were compared: users of new 2.1 years per study participant): 73,174 WY (69.2%) in
COCs containing DNG/EV, users of oCOCs and users of Europe and 32,586 WY (30.8%) in the United States
LNG-containing COCs. All VTE and ATE were always (Table 2). The late start of recruitment in the United States
adjudicated for the hormonal contraceptive used by the resulted in a shorter average follow-up in the United States
respective participant at the time. It did not matter whether compared to Europe: 1.6 and 2.4 years, respectively.
the participant was a starter, a switcher or a restarter at the At study entry, 10,191 women received a prescription for
time of the event. If hormonal contraceptive use had been DNG/EV and 40,012 for oCOC. The latter included 5796
stopped during the 3-month period prior to the VTE users of LNG (Table 1). During follow-up, 30%, 35% and
diagnosis, the event was adjudicated to the last hormonal 35% of users of DNG/EV, oCOC and LNG, respectively,
contraceptive used before the event. switched from their baseline prescription to another OC or
The analyses for VTE used two different data sets: all non-oral hormonal contraceptive brand. At the end of the
VTE and what are known as “idiopathic” VTE. The latter INAS-SCORE study, DNG/EV, oCOC and LNG had been
data set excludes cases with acute risk factors (such as used for 12,512, 63,539 and 10,071 WY, respectively. For
pregnancy, delivery, trauma, immobilization, long-haul 4279 WY and 25,431 WY of 105,761 WY, study
travel, surgery and chemotherapy). The analysis of this participants had switched to OHCs (e.g., patches, injections,
data set will allow future comparison of the INAS-SCORE vaginal rings) or had not used any hormonal contraceptive,
data with the results from other scientific groups who often respectively. Overall, 12% of the time without exposure to
base their analyses on hypothesized “idiopathic” VTE. hormonal contraceptives (“no use”) can be attributed to
The study was powered to detect a 2.0-fold and 0.5-fold intended and unintended pregnancies.
VTE risk for DNG/EV compared to oCOC, respectively.
This calculation was based on the combined US and
Table 1
European study population. However, it turned out that the Number of women enrolled, excluded and analyzed.
market introduction of DNG/EV in the United States was
Women Nos. (%) a [%] b
less successful than expected by the manufacturer of DNG/
EV. As a result, the proportion of US study participants who A) Who agreed to participate 53,750 – [100.0]
B) excluded because of protocol violations c 3547 – [6.6]
used DNG/EV was very small. This fact raised concerns on
C) analyzed 50,203 (100.0) [93.4]
the part of the European regulatory authorities about basing Cohorts
study conclusions solely on the combined US and European DNG/EV 10,191 (20.3) [19.0]
data. They requested that an additional analysis be based on oCOC 40,012 (79.7) [74.4]
the European arm of the study alone. To compensate for the of which LNG 5796 (11.5) [10.8]
Regions
loss of statistical power, it was decided to extend the
United States 20,105 (40.0) [37.4]
INAS-SCORE study in Europe. Despite this, the decision Europe 30,098 (60.0) [56.0]
was also made to complete the first study part (transatlantic European countries
phase) as planned. This decision was based on the Austria 2208 (4.4) [4.1]
consideration at that time (January 2014) that the first part France 252 (0.5) [0.5]
Germany 8613 (17.2) [16.0]
of the study would have a power of 0.8 to detect a two-fold
Italy 8508 (16.9) [15.8]
increased risk of VTE that might potentially be associated Poland 9131 (18.2) [17.0]
with the use of DNG/EV. Sweden 1111 (2.2) [2.1]
UK 275 (0.5) [0.5]
Primary analysis (based on European data)
DNG/EV 9791 (19.5) [18.2]
3. Results
oCOC 20,307 (40.4) [37.8]
of which LNG 3736 (7.4) [7.0]
A total of 53,750 women were enrolled by 1327 active a
Percentage of women who agreed to participate.
study centers. Overall, 3547 of these 53,750 women (6.6%) b
Percentage of women who were in the final analysis.
had to be excluded because they: (1) were enrolled two or c
Women who (1) were enrolled two or more times by one or more study
more times by one or more study centers (1.5%), (2) centers, or (2) continued their previous hormonal contraceptive, (3) never started
continued to use their previous hormonal contraceptive OC use after study entry, or (4) declined to sign the informed consent form.
 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 oCOC Total
All LNG
n (%) N=10,191 (20.3) N=40,012 (79.7) N=5796 (11.5) N=50,203 (100.0)
WY (%) 12,512 (11.8) 63,539 (60.1) 10,071 (9.5) 105,761 (100.0) a
Age, mean (SD) 31.7 (10.0) 26.0 (7.9) 26.0 (8.4) 27.1 (8.7)
Age, minimum 11 11 12 11
Age, percentile 5 17 16 16 16
Age, percentile 25 23 20 19 20
Age, median 31 24 24 25
Age, percentile 75 40 30 31 32
Age, percentile 95 48 42 43 44
Age, maximum 59 58 55 59
Weight, mean (SD) 62.7 (12.2) 66.6 (16.4) 66.1 (15.2) 65.8 (15.7)
Weight, minimum 30 30 37 30
Weight, percentile 5 48 48 48 48
Weight, percentile 25 55 55 55 55
Weight, median 60 63 63 62
Weight, percentile 75 69 73 73 72
Weight, percentile 95 85 99 96 96
Weight, maximum 172 191 173 191
BMI, mean (SD) 23.0 (4.2) 24.5 (5.9) 24.2 (5.4) 24.2 (5.6)
BMI, minimum 12.7 10.6 14.2 10.6
BMI, percentile 5 18.0 18.1 18.2 18.1
BMI, percentile 25 20.2 20.4 20.5 20.4
BMI, median 22.2 23.0 23.0 22.8
BMI, percentile 75 24.8 26.9 26.6 26.4
BMI, percentile 95 31.1 36.2 35.1 35.3
BMI, maximum 60.9 71.9 61.6 71.9
a
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 DNG/EV
USA EUROPE
oCOC oCOC
54 52 52
LNG LNG
40 40
40 38 38
% %
25 26
23
27
20 20 23 23
13 13 13 18
11 11 11
17
11
11 11 15 5 4 5
1 2 2
0 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 oCOC LNG Total
Risk factor
n % n % n % n %
Treated high blood pressure 266 2.6 820 2.0 139 2.4 1086 2.2
High cholesterol 110 1.1 370 0.9 51 0.9 480 1.0
Family history of fatal ATE 278 2.7 764 1.9 105 1.8 1042 2.1
Family history of VTE 400 3.9 1158 2.9 187 3.2 1558 3.1
BMI [25–30 kg/m 2] 1789 17.6 8020 20.0 1549 26.7 1223 21.1
BMI [30–35 kg/m 2] 517 5.1 3476 8.7 202 3.5 441 7.6
BMI 35+ kg/m 2 177 1.7 2492 6.2 4375 7.3 301 5.2
Smoking 2846 27.9 8679 21.7 1549 26.7 11,525 23.0
Heavy smoking a 422 4.1 1096 2.7 202 3.5 1518 3.0
Diabetes 53 0.5 260 0.6 43 0.7 313 0.6
Myocardial infarction 0 0.0 5 0.0 1 0.0 5 0.0
Stroke 5 0.0 15 0.0 4 0.1 20 0.0
PE 1 0.0 5 0.0 0 0.0 6 0.0
Deep venous thrombosis 22 0.2 34 0.1 6 0.1 56 0.1
Cancer 44 0.4 198 0.5 22 0.4 242 0.5
Any surgery 3352 32.9 11,150 27.9 1800 31.1 14,502 28.9
a
More than 15 cigarettes per day.
 J. Dinger et al. / Contraception 94 (2016) 328–339 335

50 50
USA EUROPE
40 40
28.4 30.5 33.5

30 30
% %
15.8 12.6 14.5
20 20

10 10

0 0
DNG/EV oCOC LNG 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 prepa-
rations) 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
USA Europe
50 50

40 40

% 30 27.9 % 30
25.5 23.7
15.8
20 20
12.0 12.0
10 10

0 0
DNG/EV oCOC LNG DNG/EV oCOC LNG

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 DNG/EV oCOC LNG OHC No use Total
a a a a a
n Incidence n Incidence n Incidence n Incidence n Incidence n
(95% CI) (95% CI) (95% CI) (95% CI) (95% CI)
USA and Europe (confirmed VTE) 9 7.2 (3.3–13.7) 58 9.1 (6.9–11.8) 10 9.9 (4.8–18.3) 1 2.3 (0.1–13.0) 9 3.5 (1.6–6.7) 77
USA (confirmed VTE) 0 0.0 (0.0–81.9) 25 10.4 (6.8–15.4) 4 14.3 (3.9–36.7) 0 0.0 (0.0–16.6) 5 7.7 (2.5–18.0) 30
Europe
- Confirmed VTE (primary analysis) 9 7.4 (3.4–14.1) 33 8.3 (5.7–11.7) 6 8.2 (3.0–17.9) 1 4.0 (0.1–22.5) 4 2.1 (0.6–5.4) 47
- Confirmed and potential VTE 11 9.1 (4.5–16.2) 41 10.4 (7.4–14.0) 7 9.6 (3.9–19.8) 1 4.0 (0.1–22.5) 6 3.2 (1.2–6.9) 59
- “Idiopathic” VTE 6 4.9 (1.8–10.7) 24 6.1 (3.9–9.0) 6 8.2 (3.0–17.9) 1 4.0 (0.1–22.5) 2 1.1 (0.1–3.8) 33
a
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 Exposure Incidence (events/10,000 HR (DNG/EV vs. comparators)
WY)
Point estimate 95% CI Crude estimate 95% CI Adjusted a estimate 95% CI
USA and Europe (confirmed VTE) DNG/EV 7.2 3.3–13.7 – – – –
oCOC 9.1 6.9–11.8 0.8 0.4–1.6 0.5 0.2–1.0
LNG 9.9 4.8–18.3 0.7 0.3–1.8 0.5 0.2–1.3
Europe
Confirmed VTE (primary analysis) DNG/EV 7.4 3.4–14.1 – – – –
oCOC 8.3 5.7–11.7 0.9 0.4–1.8 0.4 0.2–1.0 b
LNG 8.2 3.0–17.9 0.8 0.3–2.4 0.5 0.2–1.5
Confirmed and potential VTE DNG/EV 9.1 4.5–16.2 – – – –
oCOC 10.4 7.4–14.0 0.8 0.4–1.6 0.5 0.3–1.0 b
LNG 9.6 3.9–19.8 0.9 0.3–2.3 0.5 0.2–1.5
“Idiopathic” VTE DNG/EV 4.8 1.8–10.4 – – – –
oCOC 6.3 4.5–8.6 0.8 0.3–1.9 0.4 0.2–1.1
LNG 8.9 4.1–17.0 0.6 0.2–1.9 0.4 0.1–1.4
a
Adjusted for age, BMI, current duration of use and family history of VTE.
b
Statistically significant.
 J. Dinger et al. / Contraception 94 (2016) 328–339 337

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, chemo-
therapy) 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
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).
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 LNG OHC No use Total
a a a a a
n Incidence (95% CI) n Incidence (95% CI) n Incidence (95% CI) n Incidence (95% CI) n Incidence (95% CI) n
Complete data set
All ATE 1 0.8 (0.0–4.5) 15 2.4 (1.3–3.9) 1 1.0 (0.0–5.5) 0 0.0 (0.0–7.0) 2 0.8 (0.1–2.8) 18
of which AMI 0 0.0 (0.0–2.4) 3 0.5 (0.1–1.4) 1 1.0 (0.0–5.5) 0 0.0 (0.0–7.0) 1 0.4 (0.0–2.2) 4
Ischemic stroke 1 0.8 (0.0–4.5) 9 1.4 (0.6–2.7) 0 0.0 (0.0–3.0) 0 0.0 (0.0–7.0) 0 0.0 (0.0–1.2) 10
TIA 0 0.0 (0.0–2.4) 1 0.2 (0.0–0.9) 0 0.0 (0.0–3.0) 0 0.0 (0.0–7.0) 1 0.4 (0.0–2.2) 2
Peripheral ATE 0 0.0 (0.0–2.4) 2 0.3 (0.0–1.1) 0 0.0 (0.0–3.0) 0 0.0 (0.0–7.0) 0 0.0 (0.0–1.2) 2
European data
All ATE 1 0.8 (0.0–4.6) 12 3.0 (1.6–5.3) 1 1.4 (0.0–7.6) 0 0.0 (0.0–12.1) 2 1.1 (0.1–3.8) 15
of which AMI 0 0.0 (0.0–2.4) 2 0.5 (0.6–1.8) 1 1.4 (0.0–7.6) 0 0.0 (0.0–12.1) 1 0.5 (0.0–2.9) 3
Ischemic stroke 1 0.8 (0.0–4.6) 8 2.0 (0.9–4.0) 0 0.0 (0.0–4.1) 0 0.0 (0.0–12.1) 0 0.0 (0.0–1.6) 9
TIA 0 0.0 (0.0–2.4) 0 0.0 (0.0–0.8) 0 0.0 (0.0–4.1) 0 0.0 (0.0–12.1) 1 0.5 (0.0–2.9) 1
Peripheral ATE 0 0.0 (0.0–2.4) 2 0.5 (0.6–1.8) 0 0.0 (0.0–4.1) 0 0.0 (0.0–12.1) 0 0.0 (0.0–1.6) 2
US data
All ATE 0 0.0 (0.0–82.0) 3 1.3 (0.3–3.7) 0 0.0 (0.0–10.7) 0 0.0 (0.0–16.6) 0 0.0 (0.0–4.6) 3
of which AMI 0 0.0 (0.0–82.0) 1 0.4 (0.0–2.3) 0 0.0 (0.0–10.7) 0 0.0 (0.0–16.6) 0 0.0 (0.0–4.6) 1
Ischemic stroke 0 0.0 (0.0–82.0) 1 0.4 (0.0–2.3) 0 0.0 (0.0–10.7) 0 0.0 (0.0–16.6) 0 0.0 (0.0–4.6) 1
TIA 0 0.0 (0.0–82.0) 1 0.4 (0.0–2.3) 0 0.0 (0.0–10.7) 0 0.0 (0.0–16.6) 0 0.0 (0.0–4.6) 1
Peripheral ATE 0 0.0 (0.0–82.0) 0 0.0 (0.0–1.3) 0 0.0 (0.0–10.7) 0 0.0 (0.0–16.6) 0 0.0 (0.0–4.6) 0
a
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 cardio-
vascular 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, con-
trolled, 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 com-
bines 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.
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
339
European Active Surveillance study on Oral Contraceptives based on
142,475 women-years of observation. Contraception 2007;75:344–54.
[5] 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.
[6] Dinger J, Möhner S, Heinemann K. Cardiovascular risk associated
with the use of an etonogestrel-containing vaginal ring. Obstet Gynecol
2013;122:800–8.
[7] 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].
[8] Güther B. Access panel – quality aspects and experience. In: Streich
W, Braun B, & Helmert U, editors. Surveys in public health—deve-
lopments and perspectives in research and political counseling. Asgard
Verlag: Sankt Augustin; 2005, pp. 91–02.
[9] Potthoff P, Güther B. Access panel in the context of health research.
Forum Public Health 2005;13:16–8.
[10] 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.
[11] Susser M. What is a cause and how do we know one? A grammar for
pragmatic epidemiology. Epidemiol 1991;133:635–48.
[12] Rothman KJ, Poole C. A strengthening programme for weak
associations. Epidemiol 1988;17:955–9.
[13] 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.
[14] Shapiro S. Bias in the evaluation of low-magnitude associations: an
empirical perspective. Epidemiol 2000;151:939–45.
[15] Shapiro S. Causation, bias and confounding: a hitchhiker's guide to the
epidemiological galaxy. Part 2. Principles of causality in epidemio-
logical research: confounding, effect modification and strength of
association. J Fam Plann Reprod Health Care 2008;34:185–90.
[16] Shapiro S. Causation, bias and confounding: a hitchhiker's guide to the
epidemiological galaxy. Part 3: principles of causality in epidemio-
logical 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.
[17] Taubes G. Epidemiology faces its limits. Science 1995;269:164–9.
[18] Hill AB. The environment and disease: association or causation? Soc
Med 1965;58:295–300.
[19] 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.
[20] Belcaro C, Geroulakos G, Nicolaides AN, Myers KA, Winford M.
Venous thromboembolism from air travel: the LONFLIT study.
Angiology 2001;52:369–74.
[21] 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.
[22] 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.