Anxiety and trauma disorders are common, particularly in women (McManus et al, 2016) and exposure therapy is effective in the treatment of these (Baldwin et al, 2014). However, in women, oestrogen levels and the use of hormonal contraceptives may affect the therapy’s effectiveness.
Past research shows that ‘fear extinction’ is reduced both in women using hormonal contraceptives, and in those with low oestradiol levels who are not on hormonal contraceptives (Graham and Milad, 2013; Li and Graham, 2016; Wegerer et al, 2014). Fear extinction is a laboratory model of exposure therapy, used to treat experimentally induced fear, whereas exposure therapy is used for fear occurring as part of a disorder.
As there has been no previous research on the effect of oestrogen levels and hormonal contraceptives on exposure therapy, this is an important study.
This is a non-randomised treatment trial of women with spider phobia. The trial examines the effects of exposure therapy in 3 groups:
- Naturally cycling women (NC) with high oestradiol
- NC with low oestradiol
- Women on hormonal contraceptives (HrmC).
The form of exposure therapy used in this trial is ‘One Session Treatment’, which is effective for simple phobias (Ost, 1989). It is an appropriate therapy for this study as treatment occurs in just one hormonal phase in each woman.
The authors predicted that naturally cycling women (NC) with higher oestradiol would have greater symptom improvement than:
- NC with lower oestradiol, and
- Women using hormonal contraceptives (HrmC).
- Aged 18-35
- Principle diagnosis spider phobia
- Regular menstrual cycle or using hormonal contraceptives
- Substance abuse
- Bipolar disorder
- Medication other than SSRI
- Neurological conditions
- Endocrinological conditions
The Behavioral Approach Test (BAT) was the primary outcome measure; increasingly demanding tasks are undertaken approaching a spider; harder tasks result in higher scores (0-9). The Spider Phobia Questionnaire (SPQ), assesses self-reported phobic symptoms; this was the secondary outcome measure, higher scores indicate more fear (0- 31). Depressive symptoms were assessed using the Beck Depressive Inventory – II (BDI-II).
Women were recruited from a public online research participation system in Australia. Following phone screening, eligible patients had an assessment and treatment session of up to 4 hours on another morning. Assessment included a structured clinical interview for specific phobia, BDI-II, SPQ and BAT. Treatment with modelled graded exposure immediately followed assessment. Treatment ended when all steps had been completed or the maximum 3-hour limit had been reached. Blood tests were taken 15 minutes after completion of treatment.
One-week post-treatment, participants completed the BDI- II, SPQ and BAT and were given information about how to maintain improvement. They were encouraged to continue engaging in exposure tasks and keep a record of this. In a follow-up 12 weeks later, participants completed the BDI-II, SPQ and BAT.
After all results were obtained, NC were split in to those with higher and lower oestradiol, based on a median oestradiol split on day of treatment (progesterone levels were also measured).
All 90 eligible women were recruited: 60 NC and 30 HrmC. No-one dropped out between randomisation and post-treatment. 13 could not be contacted to complete the 12-week follow-up.
The groups did not differ in terms of age, number of comorbid phobias, age of phobia onset or oestradiol: progesterone ratio. The NC Low Oestradiol and HrmC groups did not differ in mean oestradiol.
Pre-treatment BAT scores did not differ between groups. All groups improved substantially with treatment. At post-treatment and at follow-up HrmC had lower BAT scores relative to NC High and Low Oestradiol groups (p<0.05). The two NC groups did not differ.
SPQ scores reduced significantly across all 3 groups (p<0.0001) from treatment to post-treatment and post-treatment to follow-up. Groups did not differ on the magnitude of the change. Groups did not differ on the BDI-II or on the magnitude of change in this over time (scores were in the non-depressed range).
During their treatment session, HrmC women completed exposure tasks more slowly than NC women (p<0.05), i.e. “less efficient” treatment. The two NC groups did not differ. The HrmC women reported having completed significantly fewer exposure tasks in the follow up period than the NC women (p<0.05). The two NC groups did not differ.
For all women oestradiol levels (pmol/l) were significantly correlated with BAT and SPQ scores at post-treatment. They were also significantly correlated with treatment efficiency.
All naturally cycling women had greater symptom improvement than women using hormonal contraceptives based on the primary outcome measure (the BAT) though not on the secondary outcome measure (the SPQ).
The authors concluded that hormonal contraceptives, and potentially, endogenous oestradiol levels, are related to the efficiency and effectiveness of exposure therapy for specific phobia.
Strengths and limitations
The population studied, and the intervention, were well defined. The women were objectively split into 3 groups. BAT ratings are clearly specified and objective. SPQ has test-retest stability and internal consistency (Muris P & Merckelbach H, 1996). Those seeing participants for assessment and treatment had not screened them and blood test results were not accessed until all data had been collected, so assessors were blind to participants’ hormonal status. Depressive symptoms were measured at each assessment and found not to vary between groups. The groups did not differ in non-hormonal characteristics measured.
This study was registered with ClinicalTrials.gov (NCT02622087). The entry does not suggest there were originally plans to look for a correlation between actual oestradiol levels and outcome measures, or that treatment efficiency would be calculated. These may have been post hoc analyses. Another secondary outcome measure was planned: change from baseline in diagnostic criteria meet for specific phobia. This was removed in an update in September 2017, after the last data was collected. It is not stated why this secondary outcome measure data was not published.
Supplementary data includes a CONSORT flow diagram. The study had a high retention rate (86% to 12-weeks follow up). Completers were compared to those lost to follow up, finding one significant difference: those lost to follow up were more likely to have a higher score on SPQ at post-treatment. There were no significant differences in retention rates between the 3 groups. It is not stated how missing data was handled.
As this is not a randomised controlled trial, the differences between HrmC and NC women may be due to factors associated with them taking hormonal contraceptives, rather than due to hormonal levels or the contraceptives themselves. A study of longer duration would provide information about whether apparent reduced efficacy associated with hormonal contraceptives was persistent.
Implications for practice
Patients very often report varying in how able they feel to face up to anxiety-provoking situations. This study suggests that varying levels of oestrogen, at different times of the menstrual cycle, may be one factor underlying this. It provides evidence that hormonal contraceptives may make women less responsive to exposure therapy and likely to complete fewer exposure tasks.
Possibly, timing exposure tasks so they are tried more often when oestrogen levels are higher (mid and late cycle) would be helpful. Switching from a hormonal contraceptive to a non-hormonal contraceptive may increase the effect of exposure therapy; but the evidence does not appear sufficient to suggest this at present.
This is just one non-randomised clinical trial. The results of this trial need replicating, ideally in randomised controlled trials. At this stage they cannot be considered reliable. These results apply only to women with specific phobia. Testing in clinical trials is necessary to explore if they are true in other anxiety disorders.
Conflicts of interest
The author acknowledges support from the NIHR Oxford cognitive health Clinical Research Facility. The views expressed are those of the author and not necessarily those of the NHS, the NIHR or the Department of Health.
Graham B, Li S, Black M, Ost L-G. (2018) The association between estradiol levels, hormonal contraceptive use, and responsiveness to one-session-treatment for spider phobia in women. Psychoneuroendocrinology 90 (2018) 134- 140 [PubMed abstract]
Baldwin D, Anderson I, Nutt D et al (2014) Evidence-based pharmacological treatment of anxiety disorders, post-traumatic stress disorder and obsessive-compulsive disorder: A revision of the 2005 guidelines from the British Association for Psychopharmacology (PDF). Journal of Psychopharmacology 1– 37
Graham B and Milad M (2013) Blockade of estrogen by hormonal contraceptives impairs fear extinction in female rats and women. Biological Psychiatry 73, 371 – 378 [Abstract]
Li S and Graham B (2016) Estradiol is associated with altered cognitive and physiological responses during fear conditioning and extinction in healthy and spider phobic women. Behavioral Neuroscience, 130(6), 614-623 [PubMed abstract]
McManus S, Bebbington P, Jenkins R, Brugha T. (2016) Mental health and wellbeing in England: Adult Psychiatric Morbidity Survey 2014. Leeds: NHS Digital.
Muris P & Merckelbach H (1996) A comparison of two spider fear questionnaires. Journal of Behavior Therapy and Experimental Psychiatry, 27(3), 241-244 [Abstract]
Ost L-G (1989) One Session Treatment for Specific Phobias. Behav. Res.Ther. Vol 27,1- 7 [PubMed abstract]
Wegerer M et al 2014 Low levels of estradiol are associated with elevated conditioned responding during fear extinction and with intrusive memories in daily life. [Abstract]