Depressive and anxiety disorders in the postpartum period: how prevalent are they and can we improve their detection?

The objectives of this study were: (1) to examine Composite International Diagnostic Interview (CIDI) period prevalence and comorbidity for depression and anxiety disorder in a cohort of women assessed during the first 6-8 months postpartum and (2) to examine the benefits of combining the Edinburgh Postnatal Depression Scale (EPDS) with a simple "interval symptom" question to optimize screening postpartum. Women aged over 18 (N = 1,549) were assessed during late pregnancy and reviewed at approximately 2, 4, and 6-8 months postpartum using the EPDS and an "interval symptom" question. The latter asked about any depressive symptoms in the interval since the last EPDS. Women who scored >12 on the EPDS and/or positive on the "interval symptom" question were then administered the CIDI. A further 65 randomly selected women that screened negative were also administered the CIDI. Loss to postnatal follow-up was very significant, and returns rates were inconsistent across the three postnatal time points. Almost 25% of those who screened positive did not complete a CIDI. For screen-positive status, a total of 314 (24.4%) of those that returned questionnaires (N = 1,289) screened positive at least once across the 6- to 8-month interval. Of these, 79 were lost to follow-up; thus, 235 (74.8%) completed a CIDI. In this group, 34.7% had been positive both on the EPDS and the "interval" question, 15.9% on the EPDS alone, and 49.4% on the "interval" question alone. For the CIDI diagnosis and estimated 6- to 8-month period CIDI prevalence, among those 235 women who screened positive and completed a CIDI, 67.2% met the criteria for a CIDI diagnosis, as did 16.9% of those who screened negative. The breakdown in CIDI diagnoses in the 235 women was 32.8% major depression (± anxiety disorder); 26.4% minor depression alone; and 8.1% with a primary anxiety disorder (approximately half with minor depression). Put another way, 20.4% of these women had an anxiety disorder (approximately two thirds with comorbid depression) and 37.7% of women with a major depressive episode (MDE) had a comorbid anxiety disorder. The estimated 6- to 8-month prevalence rate for a CIDI diagnosis of anxiety or depression (major or minor) was 29.2% (95% CI 26.7%-31.7%). The use of the "interval symptom" question alone was 1.7 times more likely to identify positive CIDI cases than the EPDS alone. Almost 40% of postnatal women with a diagnosis of MDE have a comorbid diagnosis of anxiety disorder. The estimated 6- to 8-month period prevalence for CIDI cases of anxiety and depression was 29.2%. Screening for anxiety and depression using the EPDS alone was associated with a lesser capacity to identify CIDI caseness than a simple "interval symptom" question (for the 2 months prior) which almost doubled the yield. This paper demonstrates that combining the EPDS with the "interval symptom" question improves detection of CIDI caseness.

Maternal trait anxiety, depression and life event stress in pregnancy: relationships with infant temperament.

AIMS: To assess whether links exist between maternal trait anxiety (STAI), perceived life event (LE) stress and depression (Edinburgh scale) and infant temperament. STUDY DESIGN AND SUBJECTS: Women in the third trimester of pregnancy returned psychological self-report questionnaires; infant temperament was evaluated at 4 and 6 months by maternal and paternal report, while depression (concurrent Edinburgh scale) was also assessed at four and six months. As data were returned inconsistently at 4 and 6 months, we combined these two time points for simplicity of reporting and optimisation of numbers. RESULTS: Univariate logistic regressions on 970 subjects indicated that the pregnancy STAI (>40) scores were associated with 2.56- and 1.57-fold increases (maternal and paternal, respectively), in the odds of "difficult" infant temperament at 4 or 6 months. Concurrent Edinburgh scores (OR of 3.06 and 2.64 for maternal reports, respectively) were also predictive of infant temperament. Age, education, income, marital status, obstetric complications, infant gender and prematurity were not predictive of infant temperament. In stepwise multiple logistic regression analyses, the antenatal trait STAI (odds ratio 1.96) significantly predicted maternal reports of "difficult" temperament at 4 or 6 months independent of both antenatal and postnatal depression scores. There were similar trends for paternal reports of "difficult" temperament but these were not significant. Antenatal depression and perceived LE stress were not predictive of temperament. Finally, women (N=14) reporting domestic violence (DV) in pregnancy had highly significant increased Edinburgh and STAI scores. CONCLUSIONS: Maternal trait anxiety was predictive of "difficult" infant temperament, independent of "concurrent" depression and key sociodemographic and obstetric risk factors. These findings, while needing replication using objective measures of infant temperament, suggest that antenatal psychological interventions aimed at minimising anxiety may optimize infant temperament outcomes. There may be some benefit in shaping specific interventions to women reporting specific risk factors such as DV or past abuse.