The Pill Project - Oral Contraceptive and Serotonergic Brain Signaling

Background:

More than 100 million women worldwide use combined oral contraceptives (OCs), i.e. the most common types that combine estrogen and gestagen. Register-based epidemiological work has recently shown that starting on hormonal contraceptives is associated with an increased risk of developing a depressive episode, and apparently this includes the most severe cases that involve suicide. In addition, the use of OCs has also been linked to reduced sexual desire and interest, one of the common symptoms in depressive disorder. It is unknown why OCs can trigger a depressive episode, at least in a subgroup of women, or how this subgroup at high-risk may be identified in advance.

OCs act by suppressing the hypothalamic-pituitary-gonadal hormonal axis, which drives the reproductive system, resulting in downregulated sex hormones. Such changes in sex-steroid milieu shape human brain biology and function including serotonergic neurotransmission, which is key in maintaining mental health, including regulation of sexual desire. Preliminary and independent data from the Center for Integrated Molecular Brain Imaging (Cimbi) database, strongly suggest that healthy women who use OCs differ in terms of serotonergic brain architecture. In particular when using a potential marker of brain serotonergic function, the investigators have in a cross-sectional study shown, that OC users display lower serotonin 4 receptor (5-HT4R) brain binding relative to non-users. This difference is remarkably large and comparable to what is seen in depressed individuals compared to in healthy controls and to the effects provoked by pharmacological tools as traditional antidepressant drugs (SSRI). The 5-HT4R is relevant in the context of depression, since rodent work points to a fast-acting antidepressant- and anxiolytic-like effect of 5-HT4R stimulation and also, notably, 5-HT4R activation is implicated in behavioral and neurogenic effects of SSRIs. Therefore, a lowered 5HT4R agonism capacity in OC users would offer a plausible explanation for why OC use may provoke depressive symptoms.

The initial cross-sectional observation of a difference in 5-HT4R binding between OC users and non-users warrants replication and validation in an independent dataset and in a study design where causality can be inferred. With this study, the investigators apply longitudinal study design to determine if OC use versus placebo cause changes in 5-HT4R brain binding and if such changes map onto relevant signatures of brain functions and mental states, including sexual desire. The investigators anticipate that this work will substantially advance the understanding of how changes in sex-hormone milieu increase susceptibility for manifest depressive episodes and ideally provide novel preventive and therapeutic opportunities.

Study design:

The investigators will conduct a single-blind randomized placebo-controlled trial with a 3-month intervention paradigm of either OC or placebo. The investigators will include up to 50 healthy women (until 20 women have completed in each arm) at 18-22 years of age. The participants will be allocated to receive a 2nd generation OC (Femicept) or placebo for three pill cycles of 28 days. The participants will undergo an investigational program at baseline (before treatment allocation) and at follow-up (at a specific time in the third pill cycle dependent on treatment arm). The programs consist of 11C-SB207145 PET and MRI/fMRI scan of the brain, neuropsychological testing, collection of blood and saliva samples, completion of questionnaires regarding various trait- and state-related measures including a month at baseline and follow-up of daily questionnaires measuring psychometrics of mood/affect, sexual desire and sleep quality. The participants will be instructed to contact us by e-mail every time they have their first cycle day, i.e. first day of bleeding, during the study, which will be used to estimate menstrual cycle phases.

Randomization:

Participants will be randomized in random block sizes to either Femicept or placebo. The randomization will be carried out by an administrative staff member at Neurobiology Research Unit (Rigshospitalet, Denmark), who will not otherwise be involved in the study or enrollment. The sequence generation will be created via the online service, Sealed EnvelopeTM (www.sealedenvelope.com). Region Hovedstadens Apotek will carry out allocation concealment by sequentially numbered containers. Enrollment and intervention assignment will be carried out by the clinical investigators. The assignment to active or placebo treatment will happen after the baseline assessment. The investigators will be unblinded after a certain time after allocation so they will be able to plan the follow-up investigational program. To evaluate the efficacy of blinding, participants will be asked which group they believed they belonged to prior to unblinding.

Intervention:

Femicept (CampusPharma) is a monophasic 2nd generation combined OC, the recommended type of combined OC due to the lower thromboembolic risk and it is the most used combined OC in Denmark (www.medstat.dk). One packet consists of 21 active pills, containing 150 micrograms levonorgestrel and 30 microgram ethinylestradiol. One pill is taken every day approximately at the same time of the day (with no more than 12 hours variation) for 21 days. Normally, this will be followed by 7 days off the pill before starting on a new pill cycle, however, to decrease complexity and to increase compliance and the timing of pill cycles, a 7-day regime with placebo pills will be added, during which participants will experience withdrawal bleeding. The placebo pill cycle treatment constitutes 28 days with a daily placebo pill.

Evaluation of compliance:

Pill count will be carried out at every pill cycle and plasma estradiol and progesterone will be measured during pill cycle day 18-28 to capture non-compliance based on non-suppressed endogenous ovarian hormone levels indicating if ovulation had occurred. If the participants have missed more than two pills during a pill cycle or if plasma sex steroid levels indicate non-compliance during a pill cycle, it is allowed to extend the treatment with one or two more pill cycles before follow-up, hence, for some participants the intervention may last 4-5 months.

Statistical analysis plan:

The investigators will use linear latent variable models on log-transformed 5-HT4R binding potentials to estimate the multiplicative group effect of OC use (i.e., the percentage change from baseline in 5-HT4R binding potentials) in the regions of interest (neocortex, neostriatum, and the hippocampus) and compare it between the groups. The log-transformed binding potentials will be adjusted for the injected tracer mass per kg body weight. The investigators will compute model diagnostics by using score tests to detect model misspecifications and add parameters until no misspecification can be detected.

Correlation analyses will be used to investigate the association between changes in 5-HT4R binding potentials and secondary outcome measures; affect and sexual desire scores, reward fMRI BOLD signals and verbal memory performances. Mediation analysis will be conducted to test the relation between OC-induced changes in 5-HT4R binding potential and the cortisol awakening response (CAR).

Further, in an explorative approach, the investigators will conduct pooled analyses of baseline data (both OC and placebo group) and potential coupling between baseline imaging outcomes and behavioral outcomes will be performed. Also, analyses of psychometric variations potentially mapping onto menstrual cycle phases will be performed.

Sample size calculation:

Calculations based on intra-subject variability of the 5-HT4R binding on 16 subjects with scan-rescan show that a group size of 20 is required to detect 8-11% difference in 5-HT4R BPND in the region of interest (neostriatum, neocortex, and hippocampus) between OC- and placebo group from baseline to follow-up with a power of 0.8 and a significance level of 0.05 (two-sided). This is equivalent to the effect size seen in the cross-sectional finding for these regions. For the primary analyses, only participants with 5-HT4R follow-up data will be included. Missing data as well as dropouts will be reported in the relevant publications. Dropouts with baseline data can be included in the explorative pooled analyses of the baseline data.

Hypotheses

Primary hypothesis:

1. OC use reduces neostriatal, neocortical, and hippocampal 5-HT4R brain binding from baseline relative to placebo. Secondary hypotheses:
2. OC use induces a reduction in positive affect and day to day positive affect variability possibly in a manner dependent on the magnitude of 5-HT4R binding decrease in OC users.
3. OC use induces mood disturbance and lower day to day mood disturbance variability possibly in a manner dependent on the magnitude of 5-HT4R binding decrease in OC users.
4. OC use induces a reduction in sexual desire and day to day sexual desire variability in a manner dependent on the magnitude of 5-HT4R binding decrease in OC users.
5. OC use reduces brain circuit responses to reward stimuli compared to placebo in ventral striatum in a manner dependent on the magnitude of 5-HT4R binding changes.
6. OC use changes resting state functional connectivity brain networks compared to placebo.
7. OC use results in reduced gray matter volumes of the hippocampus.
8. OC use changes white matter integrity.
9. OC use induces a blunted cortisol awakening response (CAR) possibly in a manner dependent on the magnitude of change in 5-HT4R binding and/or changes in hippocampal volume from baseline.
10. The association between 5-HT4R binding and verbal memory performance is dependent on OC use status, such that lower 5-HT4R binding is associated with worse verbal memory performance in the OC use state in contrast to better performance in non-OC use state.
11. OC use results in a bias in processing of emotional information compared to placebo.
12. OC use reduces sleep quality.
13. Within the OC users, the estradiol and/or allopregnanolone levels are associated with the 5-HT4R level, working memory performance, cortisol dynamics and hippocampal volume.
14. OC use induces changes in gene transcript and/or DNA methylation profiles compared to placebo.
15. OC use increases anxiety in a manner dependent on the magnitude of change in 5-HT4R binding.

Ethical considerations:

• Side effects to the intervention: The side effect of most concern is thromboembolic events. However, this risk is only increased from about 1 to 3 per 10.000 women when using 2nd generation OCs, or even less due to the exclusion of those with thromboembolic risk factors, hence it is considered a relatively small risk and acceptable. Further, the blinding of the treatment also raises an ethical concern regarding the risk of pregnancy; even though participants are unaware if they are on OCs or not, they may behave less carefully in regards to preventing pregnancy. To address this, participants will be thoroughly informed to use protection during intercourse and will be offered condoms for free during the study period.
• Exposure to radioactivity during the PET scans: The total exposure will not exceed 10 mSv equal to 3 years of natural background radiation in Denmark, which should be seen in the light of the new knowledge this study can generate.

The study will be conducted in agreement with the Declaration of Helsinki.