Contraceptives as possible risk factors for postpartum depression: A retrospective study of the food and drug administration adverse event reporting system, 2004–2015

Abstract Aim Postpartum depression is a mood disorder that commonly affects women during the early postpartum period. The objective of this study was to analyse the association of postpartum depression with drugs (including contraceptive devices and implants) with spontaneously reported adverse events reported in the US Food and Drug Administration Adverse Event Reporting System database. Design Retrospective study. Method Reports of postpartum depression events between 2004–2015 were analysed with a reporting odds ratio (ROR) algorithm. The Medical Dictionary for Regulatory Activities was used to identify postpartum depression. Results The reporting odds ratios (95% confidence intervals, CI) of levonorgestrel (an intrauterine device with progestogen), etonogestrel (a hormonal contraceptive implant), sertraline and drospirenone (an oral contraceptive) were 12.5 (8.7–18.0), 14.0 (8.5–22.8), 12.2 (6.5–23.1) and 5.4 (2.7–10.9) respectively. Among the drugs in the US Food and Drug Administration Adverse Event Reporting System database, the use of contraceptives or an intrauterine device with progestogen might convey risk for postpartum depression.

According to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5; American Psychiatric Association) (American Psychiatric Association, 2013), the criteria for a major depressive episode are as follows: (i) at least five of the following nine symptoms in the same 2-week period: depressed mood, loss of interest or pleasure, change in weight or appetite, insomnia or hypersomnia, psychomotor retardation or agitation, loss of energy or fatigue, feeling worthlessness or guilt, impaired concentration or indecisiveness, recurrent thoughts of death and/or suicidal ideation or attempt; (ii) these symptoms cause significant distress or impairment; (iii) the episode is not attributable to substance abuse or a medical condition; (iv) the episode is not better explained by a psychotic disorder and (v) the patient has never experienced a manic or hypomanic episode. PPD refers to the development of a depressive illness following childbirth and is a usually unipolar illness, but may comprise part of a bipolar condition. In addition to its manifestation as severe, persistent maternity blues that often occurs shortly postpartum, PPD can occur at any time during the first year after delivery and lasts for a few weeks or months (O'Hara & Swain, 1996). PPD is not classified as a separate disease, but rather is diagnosed as an affective or mood disorders according to the DSM-5.
Estimates of the incidence of PPD vary. According to the Cochrane review, the rate of PPD is between 3-25% in the first year following delivery (Dennis & Creedy, 2004) and approximately 10% and 15% of women of childbearing age experience PPD (World Health Organization, 2008). According to a meta-analysis, the average prevalence rate of non-psychotic PPD is 13% (O'Hara & Swain, 1996).
Previous studies have showed risk factors for PPD, including: (i) depression or anxiety during pregnancy; (ii) biological changes in hormone levels and the age of mother; (iii) chronic health problems and antenatal depression; (iv) psychological stress, lack of social support from friends and relatives and a stressful life; (v) obstetric/paediatrics factors, including unwanted pregnancy and history of loss of pregnancy; and (vi) socioeconomic status (Beck, 2001;Mehta & Mehta, 2014).
PPD is a serious health problem that affects not only mothers but also the health of their children as well (Hipwell, Goossens, Melhuish, & Kumar, 2000). An association between hormonal medication and PPD is widely suspected, but most of the publications relating to this issue are case reports or small observational studies (Duke, Sibbritt, & Young, 2007;Luukkainen, Pakarinen, & Toivonen, 2001;O'Connell, Davis, & Kerns, 2007). The lack of systematic studies on the relationship between hormonal medication, including intrauterine device (IUD) and PPD, indicates that the clinical implications of this relationship are yet to be fully elucidated. PPD is a critical public health problem that can impair maternal-infant interactions (Klainin & Arthur, 2009) and even lead to maternal suicide and infanticide (O'Hara & Swain, 1996).
Nurses have an important responsibility to monitor new mothers and give support and continuous care. More information about the connection between drugs and PPD will help nurses give the best possible care to female patients.
Reports of adverse events in clinical practice that are thought to be potentially related to therapies are spontaneously reported by healthcare practitioners and patients to pharmaceutical manufacturers or directly to the US Food and Drug Administration (FDA) (Bates & Evans, 2009 In this study, we performed a retrospective analysis of PPD events recorded in the FAERS database. This is the first report where the possible relationship between drugs, including hormonal contraceptive implants and PPD, was evaluated by using the data available in the FAERS database analysed using the reporting odds ratio (ROR).  To detect PPD incidences, we calculated the ROR, which is established in pharmacovigilance by using a disproportionality analysis (Bates & Evans, 2009). To compare one of the index groups with the reference group, we calculated the crude ROR as (a/c)/(b/d) (Figure 1).

| METHODS
ROR was expressed as point estimates with a 95% confidence interval (CI). For signal detection, general qualitative judgments are viable; whether a signal is detected or not depends on whether the signal indices exceed predefined thresholds: ROR values <1 indicate no potential exposure-event associations and estimates >1 indicate potential exposure-event safety signals. Safety signals were considered significant when the ROR estimates and the lower limits of the corresponding 95% CI were greater than 1. Two or more cases were required to define the signal (Bates & Evans, 2009). Data analyses were performed using the JMP 11.0 (SAS Institute Inc., Cary, NC, USA).

| RESULTS
The FAERS database contains 7,561,254 reports from January 2004 -December 2015. After excluding duplicates according to the FDA's recommendation, 6,157,897 reports were analysed. The number of reports associated with PPD was 253. Cases of PPD were reported from 17 countries, although 72.8% were from the United States (USA).
To evaluate the effect of age on PPD, the reports were stratified into two age groups: less than 30 and over 30 years ( Table 2). The number of reports associated with PPD in patients aged <30 or ≥30 years were 88 and 70 respectively. In those aged <30 years, the five most reported drugs were levonorgestrel, etonogestrel, sertraline, acetaminophen and venlafaxine (serotonin and norepinephrine reuptake inhibitor, SNRI). In F I G U R E 1 Two-by-two contingency

| DISCUSSION
The effects of many drugs, including hormonal contraceptive implants, on mental health are not completely understood. No study has yet focused on the effects of various contraceptive and therapeutic agents on PPD by using an SRS such as the FAERS database. In this study, we suggest that IUDs, hormonal contraceptives and SSRIs might be associated with an increased risk for PPD. The lower limit of the 95% CI for the RORs of levonorgestrel, etonogestrel, sertraline, drospirenone, fluoxetine, topiramate and quetiapine exceeds 1. Our findings point to the importance of drug safety evaluation by using real-world data.
T A B L E 2 Top 10 drugs associated with postpartum depression stratified by age Drug Name

≧30 70
Levonorgestrel 12 Quetiapine 7 Adalimumab 6 Fluoxetine 5 Citalopram 4 Clonazepam 4 Escitalopram 4 Folic Acid 4 Lamotrigine 4 Sertraline 3 T A B L E 3 Number of reports for top 10 of adverse event associated with levonogestrel, etonogestrel and drospirenone The suggestion that PPD could be induced by an IUD is particularly important. Our results give essential knowledge to improve our understanding of this issue. This information may be particularly beneficial to nurses so that they can provide the best possible counselling to women.
Contraception is of great importance for women's reproductive life from both a social-and a health-related standpoint (Toffol, Heikinheimo, Koponen, Luoto, & Partonen, 2011). Contraceptives such as levonorgestrel, etonogestrel and drospirenone are widely prescribed and are well tolerated. They have a well-described side effect profile overall; however, information concerning the possible side effects of the contraceptives on mental health is relatively scant.
Drospirenone is an oral contraceptive, which is used safely in women with a range of medical conditions, including well-controlled hypertension, uncomplicated diabetes mellitus and depression. In prospective studies, increased anxiety and increased depressive mood were reported in 7% and 10% of women on oral contraceptives respectively (Ernst, Baumgartner, Bauer, & Janssen, 2002).
In our study, the contraceptives levonorgestrel, etonogestrel and drospirenone showed a possible PPD signal (Table 3). Considering these data, we speculate that there may be an association between contraceptive use and PPD. The literature contains inconsistent results about the association between hormonal birth control and mental health. Several studies have shown less severe depressive symptoms and better overall physical function with the use of oral contraceptives . Recent studies have indicated no significant association between contraceptives and depression (Duke et al., 2007;O'Connell et al., 2007). However, relationships between reductions in the concentrations of neuroactive steroids by oral contraceptives and mood or anxiety symptoms in healthy women have not been elucidated (Rapkin, Morgan, Sogliano, Biggio, & Concas, 2006).
The mechanism by which hormonal contraceptives may influence PPD remains unclear. One plausible reason for PPD caused by contraceptives might be adverse hormonal effects. Effects of sex steroids on the central nervous system have been widely characterized (Pluchino et al., 2009). Estrogens are known to regulate many neurotransmitter systems (serotonin, dopamine and noradrenalin) and progestins regulate serotonergic, opioidergic and cholinergic systems. For this reason, it is possible that there are neuropsychological effects associated with the use of contraceptives.
PPD is an area that should be further studied, as it is associated with older maternal age (Matsumoto et al., 2011) and primiparity.
Evidence shows that primiparas have a significantly higher Edinburgh Postnatal Depression Scale (EPDS) score than the multiparas in late PPD. These characteristics of older postpartum primiparas also mean that they are a clinical concern from a nursing perspective. Information about contraceptives should be provided to postpartum women, particularly older postpartum primiparas. A summary of drugs associated with PPD among women of different ages is shown in Table 2. This research is also important for the care of postpartum women suffering autoimmune diseases. As shown in Table 1, adalimumab was mostly used for autoimmune diseases, such as rheumatoid arthritis, Crohn's disease, psoriasis and ulcerative colitis. Multiple sclerosis accounted for 71% of interferon beta-1a cases. Women may need individualized support that would reduce their anticipated difficulty in adapting to hormonal contraceptives. The findings of this study can contribute to providing this individualized care. Kennerley et al. showed that mothers experience maternity blues in the first few days after childbirth (Kennerley & Gath, 1989).
Pitt reported that 50% of mothers suffered a spell of tearfulness and depression (Pitt, 1973). Stein showed that depression showed a peak around days 4-6 (Stein, 1980). However, the usage of contraceptives was not evaluated in these reports. The first postpartum month is considered to be the most difficult and problematic period for many mothers, with poor sleep quality and feelings of despair (Carolan, 2005;Iwata et al., 2015). Therefore, we considered a study of PPD during the first postpartum month to be important. However, women do not usually use contraceptives immediately after delivery.
Because the date of childbirth is not listed in the FAERS database, we could not evaluate the duration from childbirth. Persistent PPD can occur at any time during the first year after delivery and lasts for a few weeks or months (O'Hara & Swain, 1996). The PPD associated with contraceptives and IUD in our results might be different from the maternity blues that often occur shortly postpartum as reported by Kennerley, Pitt or Stein. PPD was observed in approximately 10-19% of Japanese women (Tamaki, 2008), which is similar to that observed in Western countries (Dennis & Creedy, 2004;World Health Organization, 2008). However, no reports were available from Japan in the FAERS database during the time period considered in our study. This might reflect differences in the frequency of drug usage, especially use of contraceptive devices and implants, between the USA and Japan.
Sertraline and fluoxetine are an antidepressant SSRIs. Quetiapine is an atypical antipsychotic for the treatment of schizophrenia and bipolar disorder, which is also concomitantly used with antidepressants to treat depression. Topiramate is an anticonvulsant drug. In our study, RORs with high values were observed for sertraline, fluoxetine, quetiapine and topiramate. We do not have a conclusive explanation for these data. However, associations of these drugs with depression have been reported at low frequencies, as stated on the package insert of the drug. We postulated that these adverse effects, although small, could partially explain the high values of the RORs in our study. Although research on the performance, accuracy and reliability of different data mining algorithms is in progress, there is no recognized gold standard methodology.
Pharmacovigilance indexes, such as the ROR, have been developed to detect drug-associated adverse events in the SRS, but these cannot be applied to evaluate the comparative strength of causalities (Bates & Evans, 2009). The ROR is a clear and easily applicable technique that allows for the control of covariates through logistic regression analysis Suzuki et al., 2015;Ueda et al., 2015;Umetsu et al., 2015). An additional advantage of using the ROR is that non-selective underreporting of a drug or adverse events has no influence on the value of the ROR compared with the population of patients experiencing an adverse event (Van der Heijden, van Puijenbroek, van Buuren, & van der Hofstede, 2002). SRS is one of the primary tools used in pharmacovigilance because they reflect the realities of clinical practice (Poluzzi et al., 2012). With larger numbers of faithful reports, the FAERS database should help for optimizing pharmacotherapy (Sasaoka et al., 2016). Furthermore, Concato et al. reported that the results of well-designed observational studies do not systematically overestimate the magnitude of the effects of treatment as compared with those in randomized, controlled trials on the same topic (Concato, Shah, & Horwitz, 2000). To the best of our knowledge, this is the first study to evaluate the potential risk of contraceptives on PPD in a real-life setting. We hope these data will be the body of information that can be used by nurses to improve the management of women during the early postpartum period.

| LIMITATIONS
We should note several limitations of our study. The definition of PPD is a subject of discussion. According to the Introductory Guide MedDRA Version 19.0, the PT is a distinct descriptor (single medical concept) for a symptom, sign, disease, diagnosis, therapeutic indication, investigation, surgical, or medical procedure and medical, social, or family history characteristic (International Council for Harmonisation, 2016). PTs should be unambiguous and as specific and self-descriptive as possible in the context of international requirements. Therefore, eponymous terms are only used when they are recognized internationally. The granularity/specificity of the PT level is such that clinical pathologic or etiologic qualifiers of the descriptors are represented at the PT level (http://www.meddra. org/sites/default/files/guidance/file/intguide_19_0_english.pdf).
It is difficult to confirm the criteria used to define PPD events by volunteers at the time of reporting. They only reported adverse events were according to ICH E2B, the international safety reporting guidelines and relied on the definitions provided by MedDRA.
The second limitation is a robustness of disproportionality analysis. The ROR is an index for the detection of an adverse event signal; however, it should be considered as exploratory in a context of signal detection. The ROR does not give sufficient evidence on causality and only offers a rough indication of signal strength. We should note that our study lacked detailed patient information related to possible confounding factors, such as psychological stress and obstetric/paediatrics factors, unwanted pregnancy and primiparity. As these factors are important in mood change and depression, robust epidemiological studies are recommended.
A third limitation is the timing of contraceptive administration, e.g., whether it was at the point of PPD or before pregnancy, is an important factor for the interpretation of our results. If it was before pregnancy it might not affect biological mechanisms at play after delivery and would therefore be irrelevant. In the FAERS database, administration data and the onset date of adverse events for each drug are not coded (U. S. Food and Drug Administration, "ASC_NTS.DOC"). There is no one-to-one correspondence between a drug and an adverse event.
Therefore, we could not evaluate the precise timing of the administration of contraceptives. On the other hand, reports in the FAERS database have been reported by healthcare professionals. In our analysis, only reports with the drug code of "primary suspect" drug (PS) were included. We consider that the association of PPD and contraceptive drugs is therefore suggested by our data.
The FAERS database is a computerized spontaneous reporting system to which healthcare professionals and consumers send adverse event reports voluntarily through the MedWatch program (http:// www.fda.gov/Safety/MedWatch/, cited 12 Jan 2017). To date, the FDA has only accepted electronic submissions of Individual Case Safety Reports in the XML format, prepared in accordance with ICH E2B to transmit information directly from database to database using standardized data elements. According to the "Instructions for Completing Form FDA 3500," FDA staff knows the name, mailing address, phone number and E-mail address of the person who can be contacted to give information on the event if follow-up is necessary (http:// www.fda.gov/Safety/MedWatch/HowToReport/DownloadForms/ ucm149236.htm, cited 17 October 2016). Despite their best efforts, the FAERS database does not always contain enough patient background information to properly evaluate an event. In this study, we have not adjusted data for different regulatory activities. It could be possible by the logistic regression analysis, if we included the terms for various reporting nations, such as the USA and Japan among others, in the logistic model. The covariates should be evaluated with respect to a variety of patients' backgrounds by using well-organized epidemiologic studies in the future.
According to the DSM-5, onset of PPD can begin at 4 weeks after delivery (American Psychiatric Association, 2013). Contraception is important for women who are postpartum, including those who are breastfeeding (Tepper, Phillips, Kapp, Gaffield, & Curtis, 2015;World Health Organization, 2010). As combined hormonal contraceptive may adversely affect milk production, these methods are generally not recommended for use during breastfeeding until 6 months postpartum. However, combined hormonal methods are an important contraceptive option during the postpartum period for non-breastfeeding women (World Health Organization, 2010). We need to carefully consider the timing of contraceptive administration.
Initiation of contraception during the postpartum period is important to prevent unintended pregnancy and short birth intervals (World Health Organization, 2006). Hormonal contraceptives are prominent among contraceptive options. As the benefit and tolerability of contraceptives have been accepted worldwide, our results do not provide any justification for the restriction of contraceptive usage.

| CONCLUSION
Among the drugs in the FAERS database, the use of contraceptives or an IUD with progestogen might pose a risk for PPD. We showed the potential risk of contraceptives on PPD in a real-life setting. These data will enhance the information available to nurses and clinicians in advising patients on contraception and/or treating PPD and may be useful in the management of women's health during the early postpartum period. Considering the causality restraints of the current analysis, further epidemiological studies are recommended.

ETHICAL CONSIDERATIONS
Data from the FAERS database were obtained from the FDA website (www.fda.gov). Research ethics committee approval was not required for this study.