ORIGINAL ARTICLE: CONTRACEPTION

A first-choice combined oral

contraceptive influences general
well-being in healthy women: a
double-blind, randomized,
placebo-controlled trial
Niklas Zethraeus, Ph.D.,a Anna Dreber, Ph.D.,b Eva Ranehill, Ph.D.,c,d Liselott Blomberg, B.Sc.,e
Fernand Labrie, M.D., Ph.D.,f Bo von Schoultz, M.D., Ph.D.,d,e Magnus Johannesson, Ph.D.,b
and Angelica Linde
n Hirschberg, M.D., Ph.D.e,g
a
Department of Learning, Informatics, Management and Ethics, Karolinska Institutet, Stockholm, Sweden; b Department
of Economics, Stockholm School of Economics, Stockholm, Sweden; c Department of Economics, University of Zurich, Zurich,
Switzerland; d Department of Economics, University of Gothenburg, Gothenburg, Sweden;e Department of Obstetrics and
Gynecology, Karolinska University Hospital, Stockholm, Sweden; f Laval University, Quebec City, Montreal, Canada; and
g
Department of Women's and Children's Health, Karolinska Institutet, Stockholm, Sweden

Objective: To determine whether there is a causal effect of oral contraceptive (OC) treatment on general well-being and depressed mood
in healthy women.
Design: Double-blind, randomized, and placebo-controlled trial.
Setting: University hospital.
Patient(s): Three hundred and forty healthy women aged 18–35 years randomized to treatment, of whom 332 completed the data
collection at follow-up evaluation.
Intervention(s): A combined OC (150 mg levonorgestrel and 30 mg ethinylestradiol) or placebo for 3 months of treatment.
Main Outcome Measure(s): Primary outcome measures: global score of Psychological General Well-Being Index (PGWBI) and the
Beck Depression Inventory (BDI); secondary outcome measures: six separate dimensions of the PGWBI.
Result(s): The OC treatment statistically significantly decreased general well-being compared with placebo
4.12 (95% CI,
7.18 to

1.06). Furthermore, OC decreased the following PGWBI dimensions compared with placebo: positive well-being
3.90 (95% CI,
7.78
to
0.01), self-control
6.63 (95% CI,
11.20 to
2.06), and vitality
6.84 (95% CI,
10.80 to
2.88). The effect of OC on depressive
symptoms and on the PGWBI dimension depressed mood were not statistically significant.
Conclusion(s): This study demonstrates a statistically significant reduction in general well-being by a first-choice OC in comparison
with placebo in healthy women. We found no statistically significant effects on depressive symptoms. A reduction in general well-being
should be of clinical importance. (Fertil SterilÒ 2017;-:-–-. Ó2017 by American Society for Reproductive Medicine.)
Key Words: Depression, oral contraceptives, quality of life, randomized clinical trial, well-being

Discuss: You can discuss this article with its authors and with other ASRM members at https://www.fertstertdialog.com/users/
16110-fertility-and-sterility/posts/15214-23476

Received November 29, 2016; revised January 28, 2017; accepted February 24, 2017.
N.Z. has nothing to disclose. A.D. has nothing to disclose. E.R. has nothing to disclose. L.B. has nothing to disclose. F.L. has received personal fees from En-
doCeutics Inc. B.V.S. has nothing to disclose. M.J. has nothing to disclose. A.L.H. has nothing to disclose.
Supported by research grants from Jan Wallander and Tom Hedelius Foundation (P2010–0133:1, P2012–0002:1, and P2013–0156:1), Knut and Alice Wallen-
berg Foundation (Wallenberg Academy Fellows grant to A. Dreber), Swedish Council for Working Life and Social Research (2006–1623), the Swiss Na-
tional Science Foundation (100010–149451), the Swedish Research Council (20324), Karolinska Institutet, and the regional agreement on medical
training and clinical research (ALF) between Stockholm County Council and Karolinska Institutet (20130313). The funders of the study had no role
in study design, data collection, data analysis, data interpretation, or writing of the report. All authors had full access to all the data in the study
and were responsible for the decision to submit for publication.
N.Z., A.D., and E.R. should be considered similar in author order.
Reprint requests: Angelica Linde
n Hirschberg, M.D., Ph.D., Department of Obstetrics and Gynecology, Karolinska University Hospital, Stockholm SE-171 76,
Sweden (E-mail: angelica.linden-hirschberg@sll.se).

Fertility and Sterility® Vol. -, No. -, - 2017 0015-0282/$36.00

Copyright ©2017 American Society for Reproductive Medicine, Published by Elsevier Inc.
http://dx.doi.org/10.1016/j.fertnstert.2017.02.120

VOL. - NO. - / - 2017 1

ORIGINAL ARTICLE: CONTRACEPTION
T
he introduction of the oral contraceptive pill (OC) more
than 50 years ago has had a tremendous impact on
women's reproductive health and position in society.
Today, it is estimated that over 100 million women worldwide
are current users (1). Although OC pills are recognized as
highly effective and generally safe, there is still remarkably
poor adherence to this contraceptive method. In fact,
numerous surveys show a high discontinuation rate and
irregular use in as many as 60% of women (1–3).
Poor adherence and contraceptive failure have been
related to certain side effects such as negative mood changes
and depressive symptoms (4–6). Some 4% to 10% of OC users
report depressive symptoms (7–9). However, very few
placebo-controlled studies have addressed this issue, and
they have been performed in women with premenstrual
symptoms (10), sterilized women (11), women with dysmen-
orrhea (12), and in those with previous depressive symptoms
using OCs (13). Most of these relatively small studies have not
found any statistically significant effect of OC treatment on
depressive symptoms (10–12). However, one placebo-
controlled study in healthy women demonstrated a statisti-
cally significant impairment of mood symptoms by OC use
(14); another study showed worsening of depressive symp-
toms in women with previous experience of negative mood
symptoms while taking OCs (13). Recently, a large cohort
study also demonstrated an association between hormonal
contraception and subsequent use of antidepressants (15).
The daily intake of OCs and a positive attitude of the user
are a necessary prerequisites for efficacious oral contracep-
tion. Irregular intake might reflect a negative influence on
the individual woman's well-being in a more general sense
than specific depressive symptoms. Apart from depressed
mood, well-being also implies dimensions like vitality, self-
control, and perception of general health. There is very little
information about causal effects of OC use on general well-
being in healthy women. We found only two small placebo-
controlled studies, both with inconclusive results (10, 11).
We have filled a knowledge gap regarding a possible
causal effect of OC treatment on general well-being and
depressive symptoms in healthy women. We performed a
double-blind, randomized, placebo-controlled trial on a com-
bined OC in 340 young, healthy women. Our hypothesis was
that OCs decrease general well-being and increase depressive
symptoms (7–9, 13–15).
MATERIALS AND METHODS
Study Design
The study was a single center, randomized, double-blind, pla-
cebo-controlled, and parallel study conducted at the Karolin-
ska University Hospital, Stockholm, Sweden. The study was
performed in accordance with good clinical practice and the
World Medical Association Declaration of Helsinki's ethical
principles for medical research involving human subjects.
The protocol and the informed consent form were approved
by the local ethics committee in Stockholm (2010/1075-31/
1) and the Swedish Medical Products Agency (151:2010/
44213). The trial is registered at the EudraCT registry, registra-
tion number EudraCT 2010-020824-23.
2 VOL. - NO. - / - 2017
Study Population
The sample comprised women willing to start using OCs. The
inclusion criteria were 18 to 35 years of age, body mass index
(BMI) 19–30, a regular menstrual cycle (25–33 days), using
nonhormonal contraceptive at the study start, and fluency
in the Swedish language. The exclusion criteria were smok-
ing, hypertension, risk factors for thromboembolism and car-
diovascular disease, severe migraine, diabetes mellitus, liver
disease, pancreatitis, history of hormone-dependent cancer,
undiagnosed bleeding, and pregnancy. Intake of any sex ste-
roid hormones during the last 6 weeks before beginning the
study was not allowed. Those fulfilling all the inclusion
criteria and none of the exclusion criteria were included in
the study. All women were instructed to use nonhormonal
contraceptive methods during the study. Free condoms were
distributed.
Randomization
Participants were randomly assigned to treatment with a
combined monophasic pill (Neovletta; Bayer Schering
Pharma), containing 150 mg of levonorgestrel and 30 mg of
ethinylestradiol, or placebo during a period of 3 months.
Randomization was provided by a contract organization
(APL, Pharma Specials). A balanced block randomization
[1:1] was performed with 10 participants in each block of
fixed size. The study investigators, research coordinators,
and the participating women were blinded to the treatment
allocation. Participants took one capsule every day for
21 days, followed by a break during 7 days. Each 28-day
period defined one treatment cycle.
Procedures
During the baseline visit, scheduled for the early follicular
phase of the menstrual cycle (cycle days 1–7), a general health
control was performed that included weight, height, and
blood pressure measurements. Participants also provided
data on education and marital status. At baseline, a blood
sample (20 mL) was collected for analyses of hormones and
binding proteins. After the blood sample collection, partici-
pants filled out two surveys on general well-being and
depressive symptoms: the Psychological General Well-Being
Index (PGWBI) (16) and the Beck Depression Inventory
(BDI) (17). The final data collection was performed at the
end of the third treatment cycle. Participants underwent the
same procedures as at baseline: weight and blood pressure
measurements, blood sampling for hormone determination,
and assessment of general well-being and depressive symp-
toms. All the surveys were computerized and were completed
by the participants at the study center between 08:00 and
18:00 the same day.
Outcomes
Our primary outcome measure of general well-being was the
normalized global score of the PGWBI (16), and our secondary
outcome measures of general well-being were the normalized
scores on each PGWBI dimension. The PGWBI measured

general well-being during the last month and consists of 22
items, where each item has six response choices and the value
of each item ranges from 0 to 5. This scale has had satisfactory
internal construct validity when tested with modern psycho-
metric techniques (18). The PGWBI consists of the following
six dimensions: anxiety (included five items), depressed
mood (three items), positive well-being (4 items), self-control
(3 items), general health (3 items), and vitality (4 items). The
global PGWBI score was computed as the sum of scores on
each item (that is, each of the 22 items was given the same
weight). The global score of PGWBI varies between 0 (poor
quality of life) and 110 (good quality of life). A score of
%72 in PGWBI indicates moderate to severe distress in gen-
eral well-being. The score on each PGWBI dimension (anxi-
ety, depressed mood, positive well-being, self-control,
general health, and vitality) was defined as a secondary
outcome measure. The total score for the different dimensions
was anxiety (0–25), depressed mood (0–15), positive well-
being (0–20), self-control (0–15), general health (0–15), and
vitality (0–20). To facilitate comparison across studies and
for comparisons with other quality-of-life instruments, the
total score on each dimension and the global score of PGWBI
was normalized to 0–100.
The global score of the BDI (17) was our primary outcome
measure of depressive symptoms. The BDI consists of 21 items,
where each item has four response choices, and the value of
each item ranges from 0 to 3. The BDI measured the following
21 depressive symptoms and attitudes during the last week:
sadness, pessimism/discouragement, sense of failure, dissatis-
faction, guilt, expectation of punishment, self-dislike, self-
accusation, suicidal ideation, crying, irritability, social with-
drawal, indecisiveness, body-image distortion, work retarda-
tion, insomnia, fatigability, anorexia, weight loss, somatic
preoccupation, and loss of libido. A global score of BDI is
computed as the sum of scores on each item, and ranges
from 0–63, where a higher value reflects more severe depres-
sive symptoms/attitudes. A score R20 in the BDI indicates
moderate to severe depressive symptoms.
Safety Assessments
Adverse events and severity were assessed at each scheduled
contact and whenever there was an event. The women were
reminded to use nonhormonal contraception. At each visit,
vital signs including blood pressure and weight were re-
corded, and a pregnancy test was performed.
Hormone Measurements
Quantification of testosterone and estradiol in serum was per-
formed by the validated, highly sensitive and specific liquid
chromatography with tandem mass spectrometry (LC-MS/
MS) method performed at Endoceutics (Quebec, Canada) (19).
Serum concentrations of follicle-stimulating hormone (FSH)
and luteinizing hormone (LH) were determined by fluorescence
immunoassays (Auto Delfia; Perkin Elmer), and sex-hormone-
binding globulin (SHBG) by chemiluminescent enzyme immu-
nometric assay (Roche Diagnostics). Concentrations of free
testosterone were calculated from values of total testosterone,
VOL. - NO. - / - 2017
Fertility and Sterility®
SHBG, and a fixed albumin concentration based on an equa-
tion system derived from the law of mass action (20).
Statistical Analysis
Before completing the data collection and before breaking the
randomization codes we posted a predefined analysis protocol
at the Open Science Framework Web site (freely available at
https://osf.io/he8nb/) on August 21, 2015. The main analysis
was performed as an intention-to-treat analysis (ITT) and
included all women who completed the data collection.
To test whether there was a statistically significant differ-
ence in the average value of an outcome measure between the
OC group and the placebo group, we used an independent
samples t-test (not assuming equal variances). The tests
were based on the difference in the value of the outcome mea-
sure between the final data collection and the baseline data
collection for each participant, and all statistical tests were
two-sided (alpha level, 0.05). We specified two subgroup an-
alyses in the protocol for two subgroups of moderate to severe
distress in general well-being and depressive symptoms.
These two subgroups were defined as those who had scores
of %72 in PGWBI (moderate to severe distress) before
normalization of the score and those who had scores R20
in BDI (severity level: moderate to severe). In exploratory an-
alyses, we also estimated the correlation between the pretreat-
ment and posttreatment change in serum levels of total and
free testosterone and the change in general well-being and
depressive symptoms, using the Pearson and the Spearman
correlation coefficients. The sample size of 340 gives us a sta-
tistical power of 80% to find a statistically significant differ-
ence at the 5% level in general well-being (PGWBI) of 4.3
units measured on the PGWBI normalized scale between
0 and 100, and 2.1 units in depressive symptoms measured
on the BDI scale between 0 and 63.
In the Supplemental Appendix we provide a complete
methods section with details about the study population,
study design, treatment, general well-being and depressive
symptom measurements, safety assessments, hormone mea-
surements, and statistical analysis.
RESULTS
About 1,000 women, were initially screened by phone, and
347 women were assessed for eligibility at a screening visit.
Three hundred and forty women were included in the study
between February 3, 2012, and August 24, 2015, and were
randomly assigned to either OC or placebo (169 in the OC
group, 171 in the placebo group). Seven participants were
lost to follow-up evaluation, and one participant did not com-
plete the data collection due to an administrative error (Fig. 1).
All the 332 women who completed the data collection at
follow-up evaluation were included in our main analysis,
including the three women who discontinued treatment but
completed the data collection. There were no statistically sig-
nificant differences in any of the baseline characteristics be-
tween the two treatment groups (P>.05) (Table 1).
We hypothesized that OCs would reduce general well-
being and increase depressive symptoms. Consistent with this
hypothesis, there was a statistically significant difference in
3
ORIGINAL ARTICLE: CONTRACEPTION

FIGURE 1

340 Underwent randomization

169* Were assigned to and received oral

171 Were assigned to and received
contraceptives (150 μg levonorgestrel
placebo
and 30 μg ethinylestradiol)

4 Discontinued drug
and were lost to follow-up 3 Discontinued placebo and
- 1 anxiety were lost to follow-up
- 1 increased appetite - 1 forgot pills when travelling
- 1 depression - 1 depression
- 1 panic disorder - 1 unknown
1 Continued drug but did
not complete data collection
due to an administrative error

2 Discontinued drug
1 Discontinued placebo
but completed data collection†
but completed data collection†
- 1 mood changes
- 1 nausea
- 1 lost the pills

162 Continued drug and completed data 167 Continued placebo and completed
collection† data collection†

Trial profile.* One participant initially randomized to the treatment group discontinued treatment due to an egg donation but was later randomized
into the study again. This explains the uneven number of participants in the two groups. yAll participants that completed the data collection at
follow-up were included in our main analysis (N¼164 in the oral contraceptive group and N¼168 in the placebo group).
Zethraeus. Oral contraceptives and well-being. Fertil Steril 2017.

the change in the global score of PGWBI between the OC and
the placebo groups of
4.12 (95% CI,
7.18 to
1.06;
P¼ .0085) (Fig. 2A). We also found a statistically significant
negative impact of OC on well-being in the hypothesized direc-
tion for the following three PGWBI dimensions (Fig. 2B): pos-
itive well-being
3.90 (95% CI,
7.78 to
0.01; P¼ .0492),
self-control
6.63 (95% CI,
11.20 to
2.06; P¼ .0046), and
vitality
6.84 (95% CI,
10.80 to
2.88; P¼ .0008). The effect
sizes expressed as Cohen's D (21) were 0.29 for the global
PGWBI score, 0.22 for reduced positive well-being, 0.34 for
self-control, and 0.41 for the vitality dimension. These effect
sizes are considered small to medium. For the other three
PGWBI dimensions, anxiety
2.08 (95% CI,
6.15 to 1.99;
P¼ .3155), depressed mood
3.67 (95% CI,
7.51 to 0.17;
P¼ .0613), and general health
2.15 (95% CI,
5.24 to 0.94;
P¼ .1723), we did not find a statistically significant difference
between the OC and the placebo group (Fig. 2B).
For the BDI, our primary outcome measure of depressive
symptoms, the effect was not statistically significant 0.85
(95% CI,
0.66 to 2.36; P¼ .2664) (Fig. 3).
The effect size and statistical significance only changed
marginally in our two robustness tests when we also included
individuals who did not complete the data collection (n ¼
340) or only included participants who did not discontinue
treatment (n ¼ 329) (P¼ .0088 and .0073 for the PGWBI;
.0502 and .0489 for PGWBI dimension ‘‘positive well-being’’;
.0048 and .0047 for the PGWBI dimension ‘‘self-control’’; and
finally .0008 and .0005 for the PGWBI dimension ‘‘vitality’’).
Neither the BDI (P¼ .2684 and .2643), nor any of the other
PGWBI domains (anxiety: P¼ .3168 and .3189; depressed
mood: P¼ .0622 and .0583; and general health: P¼ .1745
and .1110) were statistically significant in any of the two
robustness tests.
The proportion of women with moderate to severe distress
in general well-being (%72 on the PGWBI) and with moder-
ate to severe depressive symptoms (R20 on the BDI) was
comparable between the groups before treatment (Table 1).
We found that 35% the OC and placebo groups had moderate
to severe distress in general well-being at baseline, whereas
7% in the OC and placebo group had moderate to severe

4 VOL. - NO. - / - 2017

 Fertility and Sterility®

In exploratory analyses, we tested whether the size of the

TABLE 1
pretreatment and posttreatment changes in serum levels of
Baseline characteristics of the intention-to-treat population.
total and free testosterone were correlated with the change
in general well-being and depressive symptoms. None of
Placebo OC
Characteristic (n [ 171) (n [ 169) these correlations were statistically significant in the OC
group (Pearson and Spearman correlations, P>.05, two-
Age (y) 24.1  4.0 23.3  3.6
Educationa
sided), and only two correlations were statistically significant
Primary school 5 (2.9) 6 (3.6) in the placebo group (Supplemental Tables 2 and 3).
Secondary school 53 (31.2) 58 (34.7) We detected adverse events in 21.0% of the women in the
University studies 64 (37.6) 56 (33.5) OC group and 12.0% in the placebo group (Supplemental
(ongoing)
University degree 48 (28.2) 47 (28.1) Table 4). There were four serious adverse events in total,
In a stable relationshipa including two pregnancies and one ovarian cyst in the pla-
Yes 100 (58.8) 97 (57.7%) cebo group and one panic disorder in the OC group. The
BMI (kg/m2) 22.6  2.8 22.5  3.1
Participants with moderate to severe distress in general well-being
most common reported adverse events in the OC group were
%72 in the PGWBI 60 (35.1) 59 (34.9) bleeding disturbances (n ¼ 14), anxiety and mood changes
Participants with moderate to severe depressive symptoms (n ¼ 12), acne (n ¼ 5), and appetite changes (n ¼ 3). Anxiety
R20 in the BDI 12 (7.0) 12 (7.1) and mood changes (n ¼ 4) were the most common reported
Score on PGWBIb 68.5  14.6 69.0  13.5
Scores on PGWBI dimensions b adverse events in the placebo group.
Anxiety 63.5  18.2 62.9  18.5
Depressed Mood 79.3  16.3 78.8  15.2
Positive Well-Being 60.7  18.0 61.6  17.3 DISCUSSION
Self-Control 71.1  19.6 71.7  18.9
General Health 84.5  13.6 84.7  12.9 Here we report results from the first large randomized,
Vitality 60.2  17.1 63.0  15.8 placebo-controlled trial of a combined OC on general well-
Score on BDI c
7.3  7.3 7.8  6.9 being and depressive symptoms in healthy women. We found
Serum levels of blood sample testsd
FSH (IE/L) 5.4  1.7 5.4  1.7
a statistically significant reduction in general well-being in
LH (IE/L) 4.9  2.4 4.8  2.2 women who used a levonorgestrel-containing OC for
Total T (pg/mL) 246.4  93.7 244.3  88.7 3 months compared with placebo, as measured with the estab-
SGBG (nmol/L) 78.1  46.7 72.6  27.3 lished PGWBI instrument. There were also negative effects on
Free T (pg/mL) 3.2  1.3 3.2  1.3
Estradiol (pg/mL) 33.5  19.1 34.4  25.1 the specific PGWBI dimensions of positive well-being, self-
Note: Data presented as mean  standard deviation or n (%). The body mass index (BMI) is control, and vitality. Thus, the results were consistent with
the weight in kilograms divided by the square of the height in meters. There were no statis- our hypothesis that in some women OC use could cause a
tically significant differences between the two treatment groups (P>.05), as assessed with
the use of the chi-square test for categorical data and the independent samples t-test (not decrease in general well-being.
assuming equal variances) for continuous data. BDI ¼ Beck Depression Inventory; BMI ¼
body mass index; FSH ¼ follicle-stimulating hormone; LH ¼ luteinizing hormone; OC ¼
The difference in the primary outcome measure PGWBI
oral contraceptive; PGWBI ¼ Psychological General Well-Being Index; SHBG ¼ sex between the OC and the placebo group corresponds to a
hormone-binding globulin; T ¼testosterone.
a
There were three missing values on the education variable—two in the OC group and one 6.0% reduction of the baseline value. The corresponding re-
in the placebo group—and two missing values on the marital status variable (one in the OC
group and one in the placebo group).
ductions in the three significant PGWBI dimensions (positive
b
The PGWBI (and each PGWBI dimension) ranges from 0 to 100, with higher values indi- well-being, self-control, and vitality) were 6.4%, 9.3%, and
cating greater general well-being. The PGWBI is defined as the average of the normalized
score on each PGWBI item where each item is given the same weight. 11.1%, respectively. These small to medium reductions, in
c
The Beck Depression Inventory (BDI) scores range from 0 to 63, with higher scores indi- terms of Cohen's D terminology (21), in well-being are likely
cating more severe depressive symptoms.
d
There were two missing values in the OC group for FSH, LH, SHBG, and free T and one to be of clinical importance for individual women and may be
missing value for total T and estradiol in the OC group.
one explanation for the high discontinuation rate.
Zethraeus. Oral contraceptives and well-being. Fertil Steril 2017.
For the other primary outcome measure, depressive
symptoms as assessed by the BDI, we found no statistically
significant negative effect. Thus, our hypothesis of increased
depressive symptoms. After treatment, 44% in the OC group depressive symptoms could not be confirmed. Moreover, the
and 38% in the placebo group had moderate to severe distress PGWBI dimension of depressed mood was not statistically
in general well-being, whereas 7% in the OC group and 7% in significantly different from the controls. The non-
the placebo group had moderate to severe depressive symp- statistically-significant effect on depressive symptoms is in
toms. There was no statistically significant treatment effect line with most previous placebo-controlled studies (10–12).
for the primary outcome measures or the secondary outcome However, higher scores of depressed mood with OC
measures in these subgroups, but it should be noted that the treatment in comparison with placebo have been reported in
study was not powered for comparisons within subgroups. two previous placebo-controlled studies (13, 14).
As expected, the serum levels of FSH, LH, and total and Furthermore, an association between hormonal
free testosterone decreased after 3 months of active treatment, contraception and a first diagnosis of depression, especially
and the change in hormone levels was statistically signifi- among adolescents, was recently suggested in a large cohort
cantly different between the OC and the placebo groups, sup- study (15).
porting that the participants were compliant with the study The mechanism behind impaired well-being by OC use in
treatment (independent samples t-test [not assuming equal healthy women is not known. We found an increase in SHBG
variances], P< .001) (Supplemental Table 1). and consequently a decrease in total and free testosterone

VOL. - NO. - / - 2017 5

ORIGINAL ARTICLE: CONTRACEPTION

FIGURE 2

Mean changes from baseline to end of follow-up in scores for the Psychological General Well-Being Index (PGWBI). (A) Results for the global score
of PGWBI (the average of the score on each item). (B) Results for each PGWBI dimension. The global score of PGWBI and the score on each PGWBI
dimension range from 0 to 100, with higher scores indicating increased well-being. Error bars denote 95% confidence intervals.
Zethraeus. Oral contraceptives and well-being. Fertil Steril 2017.

accompanying OC treatment. Some studies have demon-
strated positive effects of testosterone treatment on general
well-being in postmenopausal women (22, 23).
Consequently, a reduction in free testosterone by OC use
might cause reduced general well-being. However, we found
no statistically significant correlations in the OC group be-
tween the change in serum levels of total and free testosterone
and changes in general well-being or depressive symptoms.
Another possible mechanism whereby OCs might affect
general well-being is a direct negative effect of the progestin
component on the brain. Sex hormones can be classified as
neurosteroids because they can be synthesized in the central
and peripheral nervous system (24). They bind to their specific
nuclear receptors and act via genomic effects but could also
influence brain function via more rapid nongenomic mecha-
nisms and indirectly through neurotransmitter systems,
including the serotonin and gamma-aminobutyric acid
(GABA) systems (25). Estrogen has been shown to enhance
brain excitability, whereas progesterone seems to have an
opposite and inhibitory influence (26). Progesterone and
several progesterone metabolites exert a sedative effect on
the brain. Some metabolites (e.g., allopregnanolone) could
even be regarded as anesthetic agents (26). Clearly, such influ-
ence could affect the sense of vitality and reduce general well-
being. Cyclic negative mood symptoms such as irritability,
loss of impulse control, depression, and fatigue have been
related to an increase of progestogenic neurosteroids during
the luteal phase of the menstrual cycle (27, 28). Different
progestins and combined OCs have been reported to
provoke similar symptoms (29). Although we did not find
statistically significant effects on depressive symptoms, it is
possible that a direct progestin-induced central nervous sys-
tem effect could be the underlying mechanism behind reduced
well-being, self-control, and vitality in the OC group.

6 VOL. - NO. - / - 2017


FIGURE 3
Mean changes from baseline to end of follow-up in scores for the Beck Depression Inventory (BDI). The BDI ranges from 0 to 63, with higher scores
indicating more depressive symptoms—that is, a positive change reflects an increase in depressive symptoms. Error bars denote 95% confidence
intervals.
Zethraeus. Oral contraceptives and well-being. Fertil Steril 2017.
Combined OCs are more or less equally effective as con-
traceptive agents, but they may differ in other aspects such
as risk profile, side effects, and compliance. Most combined
OCs contain the synthetic ethinylestradiol as an estrogen
component, whereas a few recently developed contain estra-
diol. In contrast, there are several different progestins with
varying degrees of androgenic activity combined with the es-
trogen component. Androgenic progestins in combination
with ethinylestradiol (second-generation OCs) seem to be
associated with a lower risk of venous thromboembolism
and are therefore recommended as the first choice for contra-
ception in some countries (30). In the present study, we used a
levonorgestrel-containing OC, which is the drug most
commonly prescribed in Sweden. On the other hand, negative
side effects such as mood changes, acne, increased appetite,
weight gain, and sexuality may be more frequent with andro-
genic OCs compared with antiandrogenic OCs (third- and
fourth-generation OCs) (31-33).
Antiandrogenic OCs may even have positive effects on
mood. In placebo-controlled trials, combined OCs containing
drospirenone (antiandrogenic fourth-generation OC) have
demonstrated beneficial effects in women with premenstrual
syndrome (34). Also, a shorter hormone-free interval by the
24/4 regimen may contribute to the beneficial effect because
adverse mood symptoms are most pronounced during the pill-
free interval (9). Thus, our findings may not be valid for all
OCs. Additional studies on antiandrogenic OCs are needed.
If the results from alternative OCs show less or no adverse ef-
fect on general well-being, this may be reason enough to
reconsider recommendations of first-choice combined
contraception.
The strengths of our present study are the large number of
participants and the double-blind placebo-controlled design.
VOL. - NO. - / - 2017
Fertility and Sterility®
However, one limitation was the short study period because it
is well-known that several adverse effects vanish over time.
Still, mood changes usually do not disappear, and a further
deterioration is also possible. Furthermore, positive and nega-
tive expectations by the women under treatment may have
influenced the results. The use of OCs will change the amount
of and cyclicity of menstrual bleeding and thus patient blind-
ing in reality may not have been complete. However, it is also
known that women who usually have regular cycles can
become anovulatory during a test situation. Even if questions
on menstrual cyclicity were avoided, women may have sus-
pected that they were given active or inactive treatment.
Such bias, however, could work in both directions. Additional
strengths of the study are the use of well-established and vali-
dated instruments for assessment of general well-being
(PGWBI) and depressed mood (BDI), and the highly specific,
validated, and accurate method for testosterone analysis.
CONCLUSION
Our present study demonstrates a statistically significant
reduction in general well-being with the use of a popular
levonorgestrel-containing OC for 3 months in comparison
with placebo in healthy women. Furthermore, the PGWBI di-
mensions of positive well-being, self-control, and vitality
were reduced. However, we found no statistically significant
effects on depressive symptoms in this group of healthy
women. Even a modest reduction in general well-being dur-
ing OC intake should be of clinical importance and may
explain the high discontinuation rate and irregular use of
this contraceptive method. This possible side-effect thus has
to be recognized and discussed during contraceptive coun-
seling. It is vitally important to evaluate alternative OC
7
ORIGINAL ARTICLE: CONTRACEPTION
combinations in large placebo-controlled studies to identify
those with the most beneficial profile.
Acknowledgments: The authors thank Elisabeth Berg and
Mesfin Tessma for valuable comments on the analysis proto-
col; Berit Legerstam, Siv R€
odin Andersson, and Agneta Berge
for research assistance; and Johan Almenberg for proof-
reading the manuscript.
REFERENCES
1. Brynhildsen J. Combined hormonal contraceptives: prescribing patterns,
compliance, and benefits versus risks. Ther Adv Drug Saf 2014;5:201–13.
2. Potter L, Oakley D, de Leon-Wong E, Can ~amar R. Measuring compliance
among oral contraceptive users. Fam Plann Perspect 1996;28:154–8.
3. Josefsson A, Wir
ehn A-B, Lindberg M, Foldemo A, Brynhildsen J. Continua-
tion rates of oral hormonal contraceptives in a cohort of first-time users: a
population-based registry study, Sweden 2005–2010. BMJ Open 2013;3:
e003401.
4. Sanders SA, Graham CA, Bass JL, Bancroft J. A prospective study of the ef-
fects of oral contraceptives on sexuality and well-being and their relationship
to discontinuation. Contraception 2001;64:51–8.
5. Ott MA, Shew ML, Ofner S, Tu W, Fortenberry JD. The influence of hormonal
contraception on mood and sexual interest among adolescents. Arch Sex
Behav 2008;37:605–13.
6. Hall KS, O'Connell White K, Rickert VI, Reame N, Westhoff C. Influence of
depressed mood and psychological stress symptoms on perceived oral con-
traceptive side effects and discontinuation in young minority women.
Contraception 2012;86:518–25.
7. Ernst U, Baumgartner L, Bauer U, Janssen G. Improvement of quality of life in
women using a low-dose desogestrel-containing contraceptive: results of an
observational clinical evaluation. Eur J Contracept Reprod Health Care 2002;
7:238–43.
8. Kelly S, Davies E, Fearns S, McKinnon C, Carter R, Gerlinger C, et al. Effects
of oral contraceptives containing ethinylestradiol with either drospirenone
or levonorgestrel on various parameters associated with well-being in
healthy women. Clin Drug Investig 2010;30:325–36.
9. Poromaa IS, Segebladh B. Adverse mood symptoms with oral contracep-
tives. Acta Obstet Gynecol Scand 2012;91:420–7.
10. Graham CA, Sherwin BB. The relationship between mood and sexuality in
women using oral contraceptive as a treatment for premenstrual symptoms.
Psychoneuroendocrinology 1993;18:273–81.
11. Graham CA, Ramos R, Bancroft J, Maglaya C, Farley TM. The effects of
steroidal contraceptives on the we1l-being and sexuality of women: a
double-blind, placebo-controlled, two-centre study of combined and
progestogen-only methods. Contraception 1995;52:363–9.
12. O’Connell K, Davis AR, Kerns J. Oral contraceptives: side effects and depres-
sion in adolescent girls. Contraception 2007;75:299–304.
13. Gingnell M, Engman J, Frick A, Moby L, Wikstro €m J, Fredrikson M, et al. Oral
contraceptive use changes brain activity and mood in women with previous
negative affect on the pill–—a double-blinded, placebo-controlled random-
ized trial of a levonorgestrel-containing combined oral contraceptive. Psy-
choneuroendocrinology 2013;38:1133–44.
14. Cullberg J. Mood changes and menstrual symptoms with different gesta-
gen/estrogen combinations: a double blind comparison with a placebo.
Acta Psychiatr Scand Suppl 1972;236:1–86.
8 VOL. - NO. - / - 2017
15. Skovlund CW, Mørch LS, Kessing LV, Lidegaard Ø. Association of hormonal
contraception with depression. JAMA Psychiatry 2016;73:1154–62.
16. Chassany O, Dimen€as E, Dubois D, Wu A, Dupuy H. The Psychological Gen-
eral Well-Being Index (PGWBI) user manual. Lyon, France: MAPI Research
Institute; 2004.
17. Beck AT, Steer RA, Ball R, Ranieri WF. Comparison of Beck Depres-
sion Inventories -IA and -II in psychiatric outpatients. J Pers Assess
1996;67:588–97.
18. Lundgren-Nilsson  A, Jonsdottir IH, Ahlborg G Jr, Tennant A. Construct val-
idity of the psychological general well-being index (PGWBI) in a sample of
patients undergoing treatment for stress-related exhaustion: a Rasch anal-
ysis. Health Qual Life Outcomes 2013;11:2.
19. Ke Y, Bertin J, Gonthier R, Simard JN, Labrie F. A sensitive, simple and robust
LC–MS/MS method for the simultaneous quantification of seven androgen-
and estrogen-related steroids in postmenopausal serum. J Steroid Biochem
Mol Biol 2014;144:523–34.
20. So€dergard R, B€ackstro
€m T, Shanbag V, Carstensen H. Calculation of free and
bound fractions of testosterone and estradiol-17 beta to human plasma pro-
teins at body temperature. J Steroid Biochem 1982;16:801–10.
21. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Mah-
wah: NJ: Lawrence Erlbaum Associates; 1988.
22. Shifren JL, Braunstein GD, Simon JA, Casson PR, Buster JE, Redmond GP,
et al. Transdermal testosterone treatment in women with impaired sexual
function after oophorectomy. N Engl J Med 2000;343:682–8.
23. Somboonporn W, Davis S, Seif MW, Bell R. Testosterone for peri and post-
menopausal women. Cochrane Database Syst Rev 2005:CD004509.
24. Compagnone NA, Mellon SH. Neurosteroids: biosynthesis and function of
these novel neuromodulators. Front Neuroendocrinol 2000;21:1–56.
25. Simoncini T, Genazzani AR. Non-genomic actions of sex steroid hormones.
Eur J Endocrinol 2003;148:281–92.
26. Turkmen S, B€ackstrom T, Wahlstrom G, Andreen L, Johansson IM. Tolerance
to allopregnanolone with focus on the GABA-A receptor. Br J Pharmacol
2011;162:311–27.
27. Rapkin AJ, Akopians AL. Pathophysiology of premenstrual syndrome and
premenstrual dysphoric disorder. Menopause Int 2012;18:52–9.
28. B€ackstro
€m T, Bixo M, Johansson M, Nyberg S, Ossewaarde L, Ragagnin G, et al.
Allopregnanolone and mood disorders. Prog Neurobiol 2014;113:88–94.
29. Joffee H, Cohen LS, Harlow BL. Impact of oral contraceptive pill use on pre-
menstrual mood: predictors of improvement and deterioration. Am J Obstet
Gynecol 2003;189:1523–30.
30. Lidegaard O, Hougaard Nielsen L, Wessel Skovlund C, Lokkegaard E.
Venous thromobosis in users of non-oral hormonal contraception: follow-
up study, Denmark 2001–10. BMJ 2012;344:e2990.
31. Kurshan N, Epperson CN. Oral contraceptives and mood in women with and
without premenstrual dysphoria: a theoretical model. Arch Womens Ment
Health 2006;9:1–14.
32. Caruso S, Sareri MI, Agnello C, Romano M, Lo Presti L, Malandrino C, et al.
Conventional vs. extended-cycle oral contraceptives on the quality of sexual
life: comparison between two regimens containing 3 mg drospirenone and
20 mg ethinyl estradiol. J Sex Med 2011;8:1478–85.
33. Zethraeus N, Dreber A, Ranehill E, Blomberg L, Labrie F, von Schoultz B, et al.
Combined oral contraceptives and sexual function in women–a double-
blind, randomized, placebo-controlled Trial. J Clin Endocrinol Metab
2016;101:4046–53.
34. Lopez LM, Kaptein AA, Helmerhorst FM. Oral contraceptives containing dro-
spirenone for premenstrual syndrome. Cochrane Database Syst Rev 2012:
CD006586.