Beyond “postpartum depressions”: Specific anxiety diagnoses during pregnancy predict different outcomes

Abstract

Objective

Literature underlines that the Edinburgh Postnatal Depression Scale (EPDS) is the most common measure to assess postpartum depression (PPD) worldwide and suggests that the rate of false positives is high. Furthermore, the EPDS does not distinguish between depression and anxiety. This study describes different definitions of PPD and whether pregnancy anxiety disorders are risk factors for different PPDs at both 1 month and 1 year postpartum.

Method

1066 women were recruited during pregnancy and followed until the 12th month postpartum (N = 500). Women were administered the SCID and completed the PDPI-R during pregnancy. During the postpartum women who had an EPDS score of 13 or more were administered the SCID to distinguish minor or major depressive episodes (mMD) from false positives.

Results

41.5% and 44.9% of the PPD assessed with the EPDS were false positives at the 1st month and during the 1st year postpartum respectively. The difference observed in prevalence rates estimated with EPDS and SCID was statistically significant both at the 1st month and during the 1st year postpartum. Overall the effect of anxiety diagnoses in predicting PPD was stronger at the 1st month than during the 1st year postpartum. The role of panic disorder is associated both with probable depression (ES = 0.82) and with mMD (ES = 0.87) at the 1st month postpartum, and predicted mMD during the 1st year postpartum (ES = 0.71). OCD predicted false positives at the 1st month postpartum (ES = 0.89).

Conclusion

An antenatal screening of specific anxiety diagnoses could be extremely useful for the prevention of possible postpartum distress outcomes.

Introduction

In the last decades postpartum depression (PPD) has been widely studied in the literature. Nevertheless, although many studies on prevalence estimates (Gavin et al., 2005, O'Hara and Swain, 1996), prevention (Lumley et al., 2003), screening to identify cases (Austin and Lumley, 2003, Boyd et al., 2005, Buist and Bilszta, 2006, Priest et al., 2008) and both pharmacological and psychological treatments (Austin et al., 2008, Milgrom et al., 2005) have been carried out, there are still significant gaps predominantly regarding methodological issues.

One of the main issues concerns the prevalence estimates that vary widely depending on the type of disorder, diagnostic criteria, location of populations, and measures used to assess mental disorders (O'Hara and Swain, 1996, Gaynes et al., 2005, Austin and Priest, 2005, Halbreich and Karkun, 2006). In their meta-analysis, O'Hara and Swain found that in many of the included studies the prevalence estimates of PPD were based on self-report measures, and that self-report measures yielded significantly higher estimates of PPD than interview-based methods (O'Hara and Swain, 1996). Furthermore, whereas authors found that the participants in PPD studies were almost homogeneous, predominantly women who are aged 25 to 35 years, of middle or high socioeconomic status, and in a marital or equivalent relationship (Ross and McLean, 2006), the differences in the samples and sampling methods across studies and cultures have been contributing to the variation in prevalence rates (Eberhard-Gran et al., 2001).

To date the Edinburgh Postnatal Depression Scale (EPDS) (Cox et al., 1987), a ten-item inventory, is the screening tool of choice worldwide, both for the antenatal (Matthey et al., 2006, Murray, 1988) and the postnatal periods. However, a first review on studies that used EPDS revealed that the likelihood of detecting false positives with the EPDS is about 50% (Eberhard-Gran et al., 2001) and a recent meta-analysis performed on 274 studies found that there was a large variation in estimates of sensitivity and specificity between studies (Gibson et al., 2009). It should be kept in mind that the EPDS was developed to screen women for depression, not as a diagnostic tool, thus clinical confirmation of the diagnosis is essential, because many women without depression might be falsely identified as depressed: these women may exhibit a high EPDS score and have anxiety symptoms rather than depressive ones (Rowe et al., 2008), and as such not satisfy the criteria for a diagnosis of depression.

Women who report high scores on screening questionnaires or who meet the criteria for major depression are heterogeneous: their illnesses include a variety of anxiety, obsessional and post-traumatic stress disorders, together with depression associated with adversity and primary depression linked to bipolar disorder (Brockington, 2004). Furthermore, literature suggests that anxiety is a feature of perinatal depression (Hendrick et al., 2000, Ross et al., 2003) and that many depression screening scales, such as EPDS (Cox et al., 1987), also include items assessing anxiety symptoms.

With regard to the EPDS, although the instrument was conceptualised to detect probable cases of depression, authors reported that the instrument has both a depression factor and an anxiety factor (Brouwers et al., 2001, Phillips et al., 2009). Therefore, even false positives (against criteria for depression) might present clinically relevant symptoms that need treatment. There is growing recognition that in current usage “postpartum depression” is an umbrella term that includes a heterogeneous group of conditions with features of both depression and anxiety, underlying the need for a new definition of the PPD constructs itself.

Thus, both the construct and the prevalence estimates of PPD are strictly related to the instruments used to assess it. In the light of these methodological issues, it could be hypothesized that different, specific risk factors might play a different role in predicting different “postpartum distresses”.

The first aim of this study was to assess the prevalence and the functional impairment of “postpartum distress” conceptualized as: the proportion of women who had an EPDS score of 13 or more indicating probable depressive disorder (probable depression), major or minor depression (MMD) and false positives (EPDS score of 13 or more, without mMD).

The second aim was to assess the EPDS features in these groups in terms of EPDS total score and EPDS anxiety and depression subscale, as defined by Phillips et al. (2009).

The third aim was to assess the specific role of anxiety disorders, diagnosed at the 1st trimester of pregnancy (between the 12th and 15th gestational week), in predicting probable depression (EPDS score of 13 or more), major or minor depression (MMD) and false positives (EPDS score of 13 or more, without mMD).