Leon Speroff Professor and Vice Chair for Research, Department of Obstetrics and Gynecology, Oregon Health & Science University, Portland
Dr. Jensen reports that he is a consultant for and receives grant/research support from Bayer, Abbvie, ContraMed, and Merck; receives grant/research support from Medicines 360, Agile, and Teva; and is a consultant for MicroChips and Evofem.
SYNOPSIS: An analysis of a large insurance database showed no consistent effect between postpartum initiation of hormonal contraception and the subsequent diagnosis of depression.
SOURCE: Roberts TA, Hansen S. Association of hormonal contraception with depression in the postpartum period. Contraception 2017;96:446-452.
Prior research has suggested an association between hormonal contraception initiation and depression. Roberts and Hansen sought to determine if starting hormonal contraption in the postpartum interval would influence the risk of depression diagnosis or treatment during the first year following birth. They used the Military Health System Management Analysis and Reporting Tool (M2) to perform a secondary analysis of insurance records from 75,528 postpartum women enrolled in the U.S. military medical system, who delivered between October 2012 and September 2014. The analysis excluded women with a prior history of depression or use of antidepressants in the two years prior to delivery. The M2 contains medical and pharmacy billing records of all individuals who used TRICARE Prime (active duty and retired service members and family members younger than 26 years of age) and includes enrollment status for each month. The database includes information from both military and civilian treatment facilities including electronic medical record and billing statements.
Using these records, the authors identified subjects with a diagnosis of major depression or antidepressant medication prescriptions during the 12 months after delivery (the primary outcome). They assessed the relationship between hormonal contraception use with subsequent antidepressant use or diagnosis with depression in the first 12 months postpartum using Cox proportional hazards regression, with a time-dependent covariate measuring exposure to hormonal contraception. Key demographic variables included age, insurance eligibility status, and the insurance sponsor’s rank (a proxy for socioeconomic status).
The average age at delivery was 28.5 years, most of the women had a senior enlisted sponsor (52.9%) and were family members of active duty service members (74.4%), and the average follow-up period after delivery was 8.9 months. Overall, 7.8% of postpartum women received a prescription for an antidepressant, and 5.0% received a diagnosis of depression. Both antidepressant use and depression diagnosis were associated with younger age, lower socioeconomic status, and a history of military service. Fewer than half of the women (41.7%) initiated hormonal contraception in the 12 months after delivery. The majority (63%) used the norethindrone-only pill. Other contraception choices included the levonorgestrel intrauterine systems (LNG-IUS) (9.8%), etonogestrel implant (8.7%), combined pills (13.9%), and the ethinyl estradiol/etonogestrel vaginal ring (CVR) (4.5%). Women started the norethindrone pill earlier than other methods, while those using long-acting reversible contraception methods contributed more time to follow-up.
Compared to women with no hormonal contraceptive use, use of etonogestrel-containing contraception was associated with a higher risk of antidepressant use (implant: adjusted hazard ratio [HR], 1.22; 95% confidence interval [CI], 1.06-1.41; CVR: HR, 1.45; 95% CI, 1.16-1.80). In contrast, use of norethindrone-only pills was associated with a lower risk of antidepressant use (HR, 0.58; 95% CI, 0.52-0.64) and depression diagnosis (HR, 0.56; 95% CI, 0.49-0.64), and use of the LNG-IUS was associated with a lower risk of depression diagnoses (HR, 0.65; 95% CI, 0.52-0.82). None of the other contraceptive methods showed statistically significant relationships.
The authors concluded that the risk of major depression diagnosis and antidepressant use in the postpartum period varies with the type of hormonal contraception used.
Last year, I highlighted the Danish National database study of Skovlund et al, who reported an association between initiation of hormonal contraception and first diagnosis of depression.1 This new study by Roberts and Hansen uses a U.S. Military Insurance Database to ask a similar question, focusing on contraception initiation in the postpartum interval. Databases offer substantial strengths for epidemiologic study, in particular the linkage of prescription data to clinical outcomes and large subject numbers that yield precise estimates of risk with tight confidence intervals. However, bias must be considered as an explanation of any epidemiologic finding. These include the absence of baseline information concerning important confounders, and the inability to control for prescription bias and the healthy user effect. Readers must keep in mind that database studies are not randomized, controlled trials. Database studies do not have the resolution to exclude chance as the explanation for weak associations, risk estimates < 2.0 (or > 0.5).2
Skovlund et al showed weak but consistent associations for a variety of hormonal methods including combined pills, the patch, and the ring, and also with progestin-only methods such as pills, implants, and the LNG-IUS. Although the Roberts and Hansen paper showed a similar magnitude of increase in the risk of depression associated with use of the etonogestrel implant and ring, a protective effect was seen with both the LNG-IUS and the norethindrone progestin-only pill, and no association with other combined pills. While many women accept a prescription for progestin-only pills and never take them, it is hard to understand why this could result in protection from depression. It requires less imagination to think that women who initiate a more effective method like the implant might be under early pressure to initiate sexual activity. Adjusted risks estimates exceeded 2.0 with teen birth (adjusted HR, 2.03; 95% CI, 1.50-2.76) and low socioeconomic status (e.g., junior enlisted sponsor) (adjusted HR, 2.09; 95% CI, 1.67-2.63). The social context that influences contraception choice also may influence depression risk. In the military, nonuse of contraception may be associated with deployment of a loved one.
Although this study has many flaws, I find the absence of any consistent effect with respect to depression risk and initiation of a progestin-only or combined hormonal method of contraception reassuring. Full evaluation of this question will require well-designed, large, prospective studies that fully control for potential baseline confounders. Perhaps European regulatory authorities will require such a study in the future. Until we have better information, clinicians should continue to openly discuss the risks and benefits of hormonal methods and encourage patients to report symptoms of mood changes. As I have mentioned previously, a frank discussion of mood disorders in a welcoming clinic environment will help ensure that the patient receives appropriate guidance in selecting an appropriate new method of birth control.
- Skovlund CW, Morch LS, Kessing LV, Lidegaard O. Association of hormonal contraception with depression. JAMA Psychiatry 2016;73:1154-1162.
- Grimes DA, Schulz KF. Bias and causal associations in observational research. Lancet 2002;359:248-252.