In September, a group of Danish investigators published in the JAMA Psychiatry an article entitled “Association of hormonal contraception with depression.”1 The authors concluded that “use of hormonal contraception, especially among adolescents, was associated with subsequent use of antidepressants and a first diagnosis of depression, suggesting depression as a potential adverse effect of hormonal contraceptive use.” The story was immediately picked up by most news outlets and spread virally. This enthusiasm occurred in spite of the fact that many earlier studies had found no association or even a reduced risk of depression symptoms.2 Some patients have started to ask again about this issue, and many more may be troubled by this new risk to their health.
So, what should we tell the patients?
The study included records of 1,061,997 women, with over 24 million person-years of use of a wide variety of hormonal methods with different progestogens, so it sounds very impressive. However, these numbers were generated from national registries (National Prescription Register and the Psychiatric Central Register) in Denmark. From these registries and “statistics Denmark,” the investigators were able to control for age, education level, diagnosis of PCOS and endometriosis. In addition, they got information about BMI and smoking for women (from the National Birth registry) at the time of their last pregnancy (please note that few of the teens would have had information available for BMI or smoking). Also, education levels of teens may be less relevant than for older women. A 16-year old woman is most likely to have no more than a high school education.
The authors tried to exclude women who had previously been diagnosed with any psychiatric condition by excluding women who had used antidepressant drugs or had depression diagnosis at any time for at least five years before study entry.
What were the results?
The incidence rate of first use of antidepressants among…
Hormonal contraceptive users: 2.2/100 person-years
Non-users* of hormonal contraceptives: 1.71/100 person-years
* Non-users = never or former hormonal users (> 6 months). These women may not be sexually active.
Detailed analyses were done by type of contraceptive method, type of progestin, amount of estrogen, and time between initiation of hormonal methods and treatment/diagnosis of depression.
How can you interpret these results?
Even though two things seem to be related, they must meet well-tested criteria before any claim can be made for causation. These are called the Bradford Hill criteria:
- Strength: The larger the association, the more likely it is causal. (The association between tobacco and lung cancer was 8.0; in this study the odds ratio was 2.3).
- Consistency: The authors themselves cited the conflicts seen in previous studies ranges from increased risk to reduced risk.
- Specificity: The authors were not able to control for many confounding factors.
- Temporality: The effect did occur after the start of pills, but longer exposure to pills actually resulted in lower risk of depression.
- Biologic gradient: (aka dose response). This was violated. The method with the lowest hormone levels (Mirena) had one of the highest relative risks.
- Plausibility: Several mechanisms were suggested, but these were weak.
- Coherence: Experiment, analogy. Not applicable.
Given that the strength of the effect was modest, except perhaps in teens, that there is no consistency among studies, that there was no dose effect, and that the risk decreased with prolonged exposure, it is hard to see why this paper was accepted by such a prestigious journal. Probably because it carried bad news for the pill. I wish you all the best trying to explain all this to your patient. Maybe saying the risk increased in only one woman in 200 could help.1
- Skovlund CW, Mørch LS, Kessing LV, Lidegaard Ø. Association of hormonal contraception with depression. JAMA Psychiatry. 2016;73(11):1154-1162.
- Keyes KM, Cheslack-Postava K, Westhoff C, Heim CM, Haloossim M, Walsh K, Koenen K. Association of hormonal contraceptive use with reduced levels of depressive symptoms: a national study of sexually active women in the United States. Am J Epidemiol. 2013;178(9):1378-88.
Submitted by: Anita Nelson, MD, professor emeritus of obstetrics and gynecology at the David Geffen School of Medicine at the University of California-Los Angeles and clinical professor at the University of Southern California