Anxiety disorders are the 24th leading cause of disease burden globally (GBD 2019 Mental Disorders Collaborators, 2022), impacting an estimated 374 million people (Santomauro et al., 2021), and comprising specific phobias, social anxiety disorder, generalized anxiety disorder, panic disorder, and agoraphobia. First-line psychological treatments for anxiety disorders, like exposure therapy, have greater long-term efficacy, and are threefold more preferred by patients, than pharmacotherapy (McHugh et al., 2013). Nevertheless, these treatments, administered in a “one-size-fits-all” approach, are only successful in 50% of people with anxiety disorders (Loerinc et al., 2015). Experts agree that we urgently need mental health treatments that are tailored to individual characteristics (e.g., genes, lifestyle), an approach termed ‘precision medicine’ (Insel and Cuthbert, 2015).

Consideration of sex-specific factors should be a cornerstone of precision medicine for anxiety disorders. The burden of anxiety disorders is greater in women than in men (GBD 2019 Mental Disorders Collaborators, 2022). Anxiety disorders are twice as common, more severe, and more treatment-resistant in women than in men (GBD 2019 Mental Disorders Collaborators, 2022; Li and Graham, 2017, Santomauro et al., 2021). The COVID-19 pandemic has so far led to an additional 51.8 M cases of anxiety disorders in women, more than double that in men (Santomauro et al., 2021). Anxiety disorders increase during periods of hormonal flux, for example, following menarche, during peri-partum and perimenopause, and severity of anxious symptoms fluctuates over the menstrual cycle (Green and Graham, 2022), suggesting a critical role for ovarian hormones in the development and maintenance of anxious symptoms in females. Yet sex is a highly neglected variable in anxiety research, particularly at the preclinical level, where more than 90% of research has focused exclusively on males (Lebron-Milad and Milad, 2012, Zucker and Beery, 2010). Incorporation of sex-based analyses, including systematic examinations of sex-specific factors like differences in endogenous hormone levels, may be critical to identifying the factors that predict for whom, and why, first-line psychological treatments like exposure therapy sometimes fail to yield optimal benefit. This is because exposure therapy modifies neurocircuitry involved in emotional learning and affect regulation, including the prefrontal cortex, hippocampus, and amygdala (Hauner et al., 2012, Mason et al., 2016; Porto et al., 2014; Quidé et al., 2012, Shou et al., 2017, Steiger et al., 2017, Straube et al., 2014), and this neurocircuitry is impacted by the hypothalamic pituitary gonadal (HPG) axis, which regulates the release of ovarian hormones. Indeed, all the brain regions in the key neurocircuitry underlying exposure therapy express sex hormone receptors and are responsive to changes in ovarian hormone levels (Beltz and Moser, 2020, McEwen et al., 2012, van Wingen et al., 2011, van Wingen et al., 2008, Zeidan et al., 2011). Thus, individual variations in HPG functioning, caused by sex and sex-specific factors, may in part account for inconsistencies in the efficacy of exposure therapy.

One sex-specific source of variation in HPG functioning is the use of hormonal contraceptives (HCs), which are used as a form of contraception by >26% of reproductive aged women (amounting to > 248 million women globally; United Nations. Department of Economic and Social Affairs. Population Division, n.d.), and are also used as treatment for a range of conditions, including endometriosis, adenomyosis, and acne (Carey and Allen, 2012, Fraser, 2010). Reproductive aged women who do not take HCs typically experience regular fluctuations in endogenous ovarian hormones, oestradiol and progesterone, over the menstrual cycle (Draper et al., 2018, Schmalenberger et al., 2021). Briefly, at the start of the menstrual cycle (menses; the onset of menstrual bleeding) levels of oestradiol and progesterone are low. Over the course of the first half of the cycle (known as the follicular phase), with the development of the ovarian follicle, oestradiol levels increase and peak, followed by a sharp decrease after ovulation. During the second half of the cycle (the luteal phase), there are sustained increases in oestradiol and progesterone, which decline 3-4 days prior to the next onset of menstruation. In contrast, women who take HCs are exposed to synthetic oestrogen and/or progestins that disrupt these cyclical changes in endogenous ovarian hormones (Hampson, 2020). Combined oral contraceptive pills (hereafter referred to as COCPs) are the most commonly prescribed class of HC, and contain a synthetic oestrogen (typically ethinyl oestradiol) and one of several types of synthetic progestins. The ethinyl oestradiol in COCPs prevents ovulation via negative feedback at the HPG axis (occurring within one day of use), which prevents follicle development (after ∼seven days of use) and therefore suppresses ovarian production of oestradiol and progesterone, leading to levels of endogenous hormones comparable to those experienced during menses in cycling women (Rivera et al., 1999). Less commonly used are a variety of progestin-only containing HCs, including long-acting reversible contraceptives such as intrauterine devices (IUDs) and injectables. Synthetic progestins primarily prevent conception by thickening the cervical mucus to prevent penetration of sperm (Croxatto and Mäkäräinen, 1998, Lähteenmäki et al., 2000). However, progestins also cause negative feedback at the HPG axis, albeit less consistently than ethinyl oestradiol. As such, progestin-only HCs inhibit ovulation in some cycles, which disrupts ovarian synthesis of oestradiol and progesterone in the corresponding cycles (Rivera et al., 1999). The hormonal milieu of women who are cycling, versus those taking HCs, is thus very different – women taking HCs have high and sustained levels of synthetic oestradiol and/or progestins, but often, chronically low endogenous ovarian hormones. In contrast, cycling women experience regular fluctuations from low to high ovarian hormone levels, and are not exposed to synthetic hormones. The half-lives of the hormones in HCs are short, typically clearing from circulation within a few days, and ovulation can occur within two weeks of ceasing HCs (Mishell, 2005). However, there is large individual variability regarding the timeframe within with regular menstrual cycling resumes, and it is recommended that women are not considered regularly “cycling” until they have had at least two consecutive menstrual cycles post-HC cessation (Schmalenberger et al., 2021).

This review will describe evidence for the impact of HCs on exposure therapy, spanning preclinical investigations using rodent models of exposure therapy (fear extinction), and clinical studies of exposure therapy in women with anxiety disorders. Hypotheses about the mechanisms by which HCs impact exposure therapy will be put forward, developed by linking knowledge on the neurobiological effects of HCs with knowledge on the neurobiological mechanisms of exposure therapy, derived from studies of fear extinction. The hope is that this review will provide a foundation and directions for further research to develop a more precise understanding of how HCs impact exposure therapy. Such knowledge is required to enable clients and healthcare providers to make evidence-based choices about which contraceptive methods to use during anxiety treatments, to avoid treatment-interfering effects and maximise treatment benefit.