Available online at www.sciencedirect.

com

Comprehensive Psychiatry 52 (2011) 343 – 351

www.elsevier.com/locate/comppsych

From the third month of pregnancy to 1 year postpartum.

Prevalence, incidence, recurrence, and new onset of depression.
Results from the Perinatal Depression–Research & Screening Unit study
Susanna Bantia,⁎, Mauro Mauri a , Annalisa Oppoa , Chiara Borri a , Cristina Rambelli a ,
Daniele Ramacciotti a , Maria S. Montagnani a , Valeria Camilleri a , Sonia Cortopassi a ,
Paola Rucci b , Giovanni B. Cassanoa
a
Department of Psychiatry, Neurobiology, Pharmacology and Biotechnology, School of Medicine, University of Pisa, I-56100 Pisa, Italy
b
Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA

Abstract

Objective: Perinatal depression is a particular challenge to clinicians, and its prevalence estimates are difficult to compare across studies.
Furthermore, to our knowledge, there are no studies that systematically assessed the incidence of perinatal depression. The aim of this study
is to estimate the prevalence, incidence, recurrence, and new onset of Diagnostic and Statistical Manual of Mental Disorders, Fourth
Edition, minor and major depression (mMD) in an unselected population of women recruited at the third month of pregnancy and followed
up until the 12th month postpartum.
Method: One thousand sixty-six pregnant women were recruited. Minor and major depression was assessed in a naturalistic, longitudinal
study. The Edinburgh Postnatal Depression Scale and the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental
Disorders, Fourth Edition, Disorders were administered at different time points during pregnancy and in the postpartum period.
Results: The period prevalence of mMD was 12.4% in pregnancy and 9.6% in the postpartum period. The cumulative incidence of mMD in
pregnancy and in the postpartum period was 2.2% and 6.8%, respectively. Thirty-two (7.3%) women had their first episode in the perinatal
period: 1.6% had a new onset of depression during pregnancy, 5.7% in the postpartum period.
Conclusions: Our postpartum prevalence figures, which are lower than those reported in the literature, may reflect treatment during the
study, suggesting that casting a multiprofessional network around women in need of support may be potentially useful for reducing the
effects of this disorder on the mother and the newborn child. Furthermore, our results indicate that women with a history of depression have a
2-fold risk of developing mMD in the perinatal period.
© 2011 Elsevier Inc. All rights reserved.

1. Introduction pregnancy and depression symptoms. Recent literature

suggests that perinatal mood disorders are not culturally
Perinatal depression (PND) is a particular challenge to bound: they affect women in every society and from every
clinicians. Poorly defined and with uncertain etiology, it socioeconomic background [1].
frequently goes unrecognized because of the overlap with These disorders represent risks for the whole family,
inhibiting the woman's ability to perform daily activities, to
Funding and support: This research was funded with a grant from the
bond with her infant, and to relate to the infant's father [2-5].
Italian Ministry of Health and with liberal grants from the Istituto per la The prevalence rates of clinical depression in the perinatal
ricerca e la prevenzione della Depressione E dell'Ansia (IDEA), the Stella period are comparable to those seen in nonchildbearing
Major Foundations (no-profit advocacy associations), and Pfizer Italia. The groups [6]; however, rates of subclinical symptoms of
authors and the members of the PND-ReScU report no additional financial depression reported at this time are higher than expected [7].
or other relationships relevant to the subject of this article.
⁎ Corresponding author. Tel.: +39 050 835409; fax: +39 050 21581. Literature has been mainly focused on postpartum
E-mail addresses: pndrescu@gmail.com, bantis@libero.it (S. Banti), depression. Studies focused on the perinatal period found
Annalisa.Oppo@gmail.com (A. Oppo). that depression during pregnancy is one of the strongest risk
0010-440X/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.comppsych.2010.08.003
344 S. Banti et al. / Comprehensive Psychiatry 52 (2011) 343–351

factors for postpartum depression [8,9]. Moreover, symp- 2. Method

toms of depression are not more common or severe after
childbirth than during pregnancy [10]; and their distribution The Perinatal Depression–Research & Screening Unit
among the different trimesters of pregnancy varies across (PND-ReScU) is based on an ongoing collaboration between
studies. Some authors [11,12] reported the highest rates of the Department of Obstetrics and Gynaecology and the
clinical depression during the first trimester of pregnancy Department of Psychiatry, Neurobiology, Pharmacology,
and a second peak during the third trimester. and Biotechnologies of the Azienda Ospedaliero-Universi-
Bennet et al (2004) [13] indicate that the prevalence of taria Pisana (AOUP) [18]. The primary aim of the PND-
depression during pregnancy is 7.4% (95% confidence ReScU is to assess the prevalence and the incidence of PND
interval [CI], 2.2-12.6), 12.8% (95% CI, 10.7-14.8), and following pregnant women in a naturalistic, longitudinal
12.0% (95% CI, 7.4-16.7) at the first, second, and third study. A further aim is to identify further risk factors for PND
trimester and that overall rates do not differ significantly other than those already known.
across trimesters. Central to our plan was a letter given to each pregnant
Moreover, a recent meta-analysis highlighted that the woman who presented to the local health service to collect
point prevalence of major and minor depression (mMD) the booklet of information prepared by the region of Tuscany
ranges from 8.5% to 11.0% during pregnancy and from 6.5% describing various aspects of pregnancy and maternal health,
to 12.9% during the first year postpartum [14]. the schedule of future obstetrical visits, and the evaluations
The prevalence estimates of postpartum depression are patients will have during their pregnancy. Our letter,
difficult to compare across studies because of variations enclosed within the booklet, provides a very brief description
in definitions (major depression, minor depression, and of PND and informs the woman of the possibility to
elevated depressive symptomatology), rating scales, and participate in a study. Subsequently, all the women who
timing of the assessments [14,15]. The first meta-analysis in presented for their first ultrasound examination at the
this field, based upon dimensional measures as well as Department of Obstetrics and Gynaecology (between the
categorical measures, highlighted the differences in the 12th and 15th gestational week) were asked to participate in
prevalence rates of postpartum depression when self-report the study. Women who agreed and who signed informed
instruments (predominantly the Edinburgh Postnatal De- consent were enrolled and then assessed at 7 time points,
pression Scale [EPDS]) or operationally defined criteria were from the third month of pregnancy until the 12th month after
used [1]. delivery. Recruitment took place between February 2004 and
Most of the studies included in the 2 largest meta-analyses March 2007.
[1,14] were carried out before the advent of the Diagnostic Exclusion criteria for the study were age less than
and Statistical Manual of Mental Disorders, Fourth Edition 18 years, poor knowledge of the Italian language or other
(DSM-IV) [16]. Moreover, because of difficulties in limitations to communication, and no fixed residence.
conducting longitudinal studies in the perinatal period, The Ethics Committee of the AOUP approved the study
evidence in the literature on the incidence of PND is scanty. protocol and the assessment procedures. The Committee also
Recently, Halbreich and Karkun [17] emphasized some required the provision of psychological counseling for
methodological issues that affect the interpretation of the women with mild depressive symptomatology and for all
results of the literature: (1) it is unclear if the point estimates women who requested it, and/or the provision of drug
of postpartum depression reported in many studies are new treatment for women with moderate/severe depression,
occurrences or the continuation of ongoing episodes; (2) according to international guidelines [19,20]. The Ethics
point prevalence is not always distinguished from period Committee allowed us to collect information only after the
prevalence; and (3) analyses do not separate women with a informed consent was signed, as prescribed by the Italian law
history of depression and/or postpartum depression from no. 675 of December 31, 1996, on privacy.
those with a first-ever episode during the postpartum
period. In light of these issues, both Gaynes et al [14] 2.1. Assessments of depression
and Halbreich and Karkun [17] underlined the need of
further research on the prevalence and incidence of PND The diagnostic assessment was conducted at baseline
using a sound methodology. using the Structured Clinical Interview for DSM-IV
Finally, although there are several studies estimating the Disorders (SCID-I) [21] by clinicians trained and
prevalence of depression throughout the perinatal period, certified to the use of the interviews when high levels
investigations on incidence, and in particular the distinction of interrater reliability (N0.90) of their diagnoses with
between new onsets vs. recurrences, are needed. the trainer were achieved. All interviewers had long-
The aim of this article is to provide estimates of the standing experience in the administration of standard-
prevalence, incidence, recurrence, and new onset of perinatal ized interviews.
DSM-IV mMD in a naturalistic, longitudinal study of an The SCID-I is a semistructured interview for making the
unselected population of women followed from the first major DSM-IV diagnoses [16]. The diagnoses of interest for
trimester of pregnancy up to 1 year postpartum. the present study were major depression (including women
 S. Banti et al. / Comprehensive Psychiatry 52 (2011) 343–351 345

with bipolar II disorder in a current depressive episode) and The point prevalence of mMD was calculated as the
minor depression; in this last category, both women with percentage of women with depression at the third, sixth, and
DSM-IV minor depression and women with a prior history eighth month of pregnancy and at the first, third, sixth, ninth,
of major depressive disorder having a mild relapse, if they and 12th month postpartum.
met the criteria for minor depression, were included. The The cumulative incidence of mMD during pregnancy was
diagnosis of minor depression proposed in the Appendix to estimated as the cumulative rate of onset of depressive
the DSM-IV requires the presence of 2 to 4 criteria of episodes by the eighth month of pregnancy over the total
depression lasting for at least 2 weeks [16]. number of subjects without depression at the study entry.
At the subsequent visits, the assessment of depression was The cumulative incidence of postnatal depression was
carried out using the EPDS [22] and then administering only estimated as the cumulative rate of onset of mMD between
the Mood section of the SCID to women exceeding the delivery and the 12th month postpartum over the total
EPDS threshold of 12. Furthermore, to reduce the number of number of subjects who completed the eighth-month
false negatives, all women with EPDS scores ranging from assessment and had no depression during pregnancy.
10 to 12 were contacted by phone and asked about their To detect first-ever episodes of depression in the
pregnancy and their well-being; moreover, the Mood section perinatal period, we distinguished new onsets from
of the SCID was administered to women with EPDS scores recurrences. To date, there is no agreement on the time to
less than 12 if an affirmative answer on critical items, such as recovery. In this study, recurrence, as defined by Frank
question 10 (presence of suicidal thoughts), was present. et al [27], referred to the development of a new episode
The EPDS is a 10-item self-report scale specifically following recovery; recovery was ascribed when the period
designed to screen for postnatal depression; items are of remission had been sufficiently sustained and when
scored on a 4-point Likert scale (from 0 to 3), and the total continued well-being could be expected (with or without
score ranges from 0 to 30. The scale rates the severity of continuing treatment).
depressive symptoms experienced over the previous 7 days. Odds ratios (ORs) were used to measure the association
Five of the items explore dysphoric mood, 2 explore between PND and sociodemographic and clinical cha-
anxiety, and 3 assess guilt and suicidal thoughts. The total racteristics. Information on socioeconomic status was
is calculated by summing up the item scores. A score drawn from the Postpartum Depression Predictors Inven-
greater than 12 was used to identify probable cases of tory–Revised [28], which is a self-report instrument
depression with a sensitivity of 86% and a specificity of designed to identify the risk factors for postpartum
78% [22]. The EPDS has been also validated for use during depression. The Postpartum Depression Predictors Inven-
pregnancy [23]. The Italian version of the EPDS had a good tory categorizes socioeconomic status on 3 levels—low,
internal consistency (Cronbach α = .747), and a sensitivity medium, and high—without providing anchor points
of 0.556 and a specificity of 0.989 were associated to a related to the income per year.
score of 12 [24]. All statistical analyses were conducted using STATA,
release 8 (StataCorp, College Station, TX), which includes a
2.2. Procedures family of procedures for analyzing survey data.
Assessments were carried out at the third month of
pregnancy; at the sixth and eighth month of pregnancy; and
3. Results
at the first, third, sixth, ninth, and 12th month after delivery
at the outpatient section of the AOUP Gynaecology Clinic.
Fig. 1 shows the flow of study participants. Of the 2598
2.3. Data analysis women contacted, 2138 were eligible for participation;
1066 (49.9%) signed an informed consent and completed
Data are presented as percentages with 95% CIs. the baseline assessment. A total of 1072 (50.1%) refused to
The weighted period prevalence of mMD was calculated participate for various reasons including lack of time, lack
as the percentage of women with depression at any time of interest in the study protocol, the convictions that they
during pregnancy (third to eighth month of pregnancy) and would never become depressed, or resistance on the part of
in the postpartum period (first to 12th month postpartum), the partner.
adjusted for nonresponse. Mean age was 32.3 years (SD, 3.9), the majority (89.9%)
Weights were obtained using logistic regression on had at least 13 years of education, 92% (n = 981) were
participants at study entry, with response (yes/no) as the married or living with the partner, 82.3% were employed
dependent variable and age group (18-25, 26-30, 31-35, outside the home, 96.2% were living in urban or suburban
36-40, or 41-45) as the independent variable. Individual areas, and 90.8% had a medium socioeconomic status
weight was calculated as the reciprocal of the resulting (Table 1). One third of the women (n = 360) had one or
probability, so that participants with a lower response more children.
probability had a higher weight than participants with high During the study, 116 women received a treatment: the
response probability [25,26]. majority of participants received psychological counseling
346 S. Banti et al. / Comprehensive Psychiatry 52 (2011) 343–351

Fig. 1. Participant flow.

(n = 93), of whom 40 received drug treatment in 3.1. Dropout analysis

combination. Only 23 women received pharmacotherapy
alone (predominantly selective serotonin reuptake inhibi- Overall, 500 women completed the assessment at the 12th
tors). Of the 116 women who received any treatment, 69 month after delivery. During the follow-up, 566 participants
started during pregnancy; and 47 started during the dropped out (53.1%).
postpartum period. Of the 92 women with mMD at the Sociodemographic and clinical characteristics between
study entry, 32 (34.8%) accepted a treatment. the 566 women who dropped out and those who completed
 S. Banti et al. / Comprehensive Psychiatry 52 (2011) 343–351 347

Table 1 3.3. Incidence of depression

Characteristics of the sample
Age, mean ± SD 32.27 ± 3.95 Of the 974 women without DSM-IV major or minor
Marital status, n (%) depression at baseline (third month of pregnancy), 49 had a
Single 47 (4.4) new episode during the study. The weighted incidence rate
Married/cohabiting 981 (92) was 2.2% (95% CI, 1.1-3.2) during pregnancy and 6.8%
Divorced 30 (2.8)
(95% CI, 4.6-9.2) in the postpartum period (Fig. 3).
Widowed 2 (0.2)
Missing 6 (0.6)
Employment status, n (%)
3.4. Recurrence of depression
Student 22 (2.1)
Unemployed 70 (6.6)
Of the 200 women with a history of depression, 17 had a
Employed 883 (82.8) recurrence during the study. The weighted percentage of
Housewife 60 (5.6) recurrences was 3.7% (95% CI, 0.8-6.7) during pregnancy
Other 17 (1.6) and 7.7% (95% CI, 2.8-12.6) in the postpartum period.
Missing 14 (1.3) The relative risk of depression in the perinatal period was
Educational level, n (%)
Primary school 3 (0.3)
about 2-fold in women with a history of depression
Secondary school 94 (8.8) compared with those without a history (relative risk =
High school (completed) 511 (48.0) 2.27; 95% CI, 1.20-4.29).
University degree 447 (41.9)
Missing 11 (1.0) 3.5. New onset of major or minor depression
Socioeconomic status, n (%)
Low 34 (3.2) Of the 774 women without a previous history of
Medium 968 (90.8) depression, 32 had a new onset during the study. The
High 19 (1.8)
weighted incidence of new onsets was 1.6% (95% CI, 0.6-
Missing 45 (4.2)
Living area, n (%) 2.6) during pregnancy and 5.7% (95% CI, 3.3-8.1) in the
Urban 534 (50.1) postpartum period.
Suburban 491 (46.1)
Rural 26 (2.4) 3.6. Sociodemographic and clinical correlates of PND
Missing 15 (1.4)
First pregnancy, n (%) To determine correlates of PND of the 142 women who
Yes 704 (66.0) had a minor or major depressive episode during the study
No 360 (33.8) course, we have calculated ORs of mMD as a function of
Missing 2 (0.2)
current comorbidity with anxiety disorders (at baseline),
socioeconomic status, educational level, employment, parity,
and marital status.
the 12th-month follow-up assessment (n = 500) were Having an anxiety disorder at study entry (OR = 4.10;
compared: no significant differences were detected on 95% CI, 2.7-6.23), a low socioeconomic status (OR = 2.48;
marital status, socioeconomic status, educational level, 95% CI, 1.02-6.42), and multiparity (OR = 1.95; 95% CI,
living area, and parity. Women who dropped out were 1.32-2.88) were associated with higher odds of having a
significantly younger ([mean ± SD] 31.8 ± 4.1 vs 32.8 ± 3.8, major or a minor depressive episode in the perinatal period.
t[1064] = 3.82, P b .001) and more frequently unemployed After adjusting for age, multiparity was also at higher odds of
(8.6% vs 4.5%, χ2[1] = 7.01, P = .008). having PND (OR = 2.21; 95% CI, 1.47-3.33).
Unemployment, being single, and low educational level
were not associated with PND.
3.2. Prevalence of depression

Overall, 142 women had a diagnosis of mMD

during the study. Ninety-two women (8.6%) met criteria 4. Discussion
for these diagnoses at study entry, of whom 37 (3.5%)
had their first episode of mMD and 55 (5.1%) had To our knowledge, this is the first large naturalistic,
recurrent depression. longitudinal study on incidence, new onset, and recurrence
The point prevalence of mMD decreased from 8.6% at the of PND using self-report and face-to-face structured inter-
third month of pregnancy to 1.7% at the eighth month of views at different time points during pregnancy and in the
pregnancy. In the postpartum period, it ranged between 3.2% postpartum period.
at 1 month postpartum and 1.2% at 9 months postpartum 4.1. Prevalence
(Table 2).
The weighted prevalence was 12.4% (95% CI, 10.2-14.6) Our results indicate that the period prevalence of mMD
during pregnancy and 9.6% (95% CI, 7.0-12.2) in the during pregnancy is higher than in the postpartum period
postpartum period (Fig. 2). (12.4% vs 9.6%), consistent with previous studies [10,14].
348 S. Banti et al. / Comprehensive Psychiatry 52 (2011) 343–351

12th mo (n = 500)

12th mo (n = 457)

12th mo (n = 366)
Our figures are in line with combined estimated reported by

12th mo (n = 91)

3.0% (0.01-.7.1)
Gaynes et al [14] (14.3%-23.3%). However, we found that

1.9% (0.6-3.2)

1.8% (0.5-3.2)

1.6% (0.2-3.0)
7 (1.8%)

2 (2.4%)

5 (1.6%)
8 (1.8%)
the 1-year period prevalence was significantly lower than
that reported by the same authors (39.6%-67.4%). Never-
theless, Gaynes et al stated that the many fewer estimates of
period prevalence allow us to say little about the period
prevalence for mMD [14].
9th mo (n = 534)

9th mo (n = 490)

9th mo (n = 100)

9th mo (n = 390)
2.7% (0.01-6.4)

0.8% (0.01-1.7)
1.2% (0.2-2.2)

1.1% (0.1-2.1)
As shown in Table 2, 92 women (8.6%) met criteria for
6 (1.2%)

5 (1.1%)

2 (2.1%)

3 (0.8%)
mMD at study entry: 55 (5.1%) women had a diagnosis of
recurrent depression, whereas 37 (3.5%) were at their first
episode of mMD. The diagnostic instrument we used (SCID-
I) does not allow us to ascertain if the mMD episode started
6th mo (n = 600)

6th mo (n = 555)

6th mo (n = 110)

6th mo (n = 444)
2.7% (0.01-5.7) immediately before or after the onset of the pregnancy, so are

0.5% (0.01-1.2)
1.9% (0.8-3.1)

1.0% (0.1-1.8)

unable to establish precisely the mMD incidence rates at

11 (1.9%)

5 (1.0%)

3 (2.8%)

2 (0.5%)
study entry.
We found a significant decrease in the point prevalence
of mMD across trimesters in line with a meta-analysis of
Gavin et al [15] and at variance with a meta-analysis of
3rd mo (n = 663)

3rd mo (n = 606)

3rd mo (n = 121)

3rd mo (n = 485)

Bennet et al [13], suggesting that point prevalence is quite

3.3% (0.06-6.3)
2.9% (1.5-4.2)

1.5% (0.5-2.5)

1.0% (0.1-2.0)
18 (2.7%)

9 (1.5%)

4 (3.3%)

5 (1.0%)

stable across trimesters.

Postpartum

The decline in the point prevalence during the second

and third trimester of pregnancy found in our study can
most probably be attributed to treatment. In fact, as
required by our Ethical Committee, depressed women
1st mo (n = 751)

1st mo (n = 687)

1st mo (n = 132)

1st mo (n = 555)
3.0% (0.06-6.0)

were given psychological counseling and/or drug treatment

3.2% (1.9-4.5)

1.9% (0.9-2.9)

1.6% (0.6-2.7)
24 (3.2%)

13 (1.9%)

4 (3.0%)

9 (1.6%)

if necessary and followed through the entire pregnancy and

the first year postpartum.
At the first month after delivery, our point prevalence
estimate of mMD is significantly lower than that reported by
O'Hara and Swain [1] (95% CI, 3.7-10.7) and by Gonidakis
8th mo (n = 862)

8th mo (n = 788)

8th mo (n = 159)

8th mo (n = 629)
2.5% (0.05-5.0)

0.5% (0.01-1.1)
1.8% (0.9-2.7)

1.3% (0.5-2.1)

et al [29] (12.5%), but within the CI reported by Gaynes et al

15 (1.7%)

10 (1.3%)

4 (2.5%)

3 (0.5%)

[14] (95% CI, 2.2-6.4), who combined results from 4 studies.

It should be noted that comparison with the results of
Gonidakis et al [29] is problematic because these authors
used an EPDS score of 13 or higher rather than a structured
6th mo (n = 935)

6th mo (n = 855)

6th mo (n = 172)

6th mo (n = 683)

clinical interview to determine that depression was present.

1.7% (0.01-3.6)
2.6% (1.5-3.6)

1.3% (0.5-2.0)

1.2% (0.4-2.0)

Furthermore, our point prevalence estimate of mMD in

24 (2.6%)

11 (1.3%)

3 (1.7%)

8 (1.2%)

other postpartum period is significantly lower than that

reported by the Gaynes et al meta-analyses [14].
Prevalence, incidence, recurrence, and new onset of mMD

4.2. Incidence
3rd mo (n = 1066)

3rd mo (n = 974)

3rd mo (n = 200)

3rd mo (n = 774)
8.6% (8.6-10.3)

Of the 974 women without a diagnosis of depression at

92 (8.6%)
Pregnancy

baseline, 2.2% (95% CI, 1.1-3.2) had a new episode of mMD

during pregnancy. The incidence in the postpartum period
was 3-fold as much (6.8%). Therefore, the high prevalence
of depression during pregnancy in our study depends on the
Weighted incidence, % (95% CI)

Weighted incidence, % (95% CI)

Weighted incidence, % (95% CI)

Weighted incidence, % (95% CI)

fact that 92 women (8.6%) were already depressed at

baseline and not on the incidence of new episodes.
Overall, the incidence of postpartum depression in our
Point prevalence, n (%)

study was 6.8% (95% CI, 4.6-9.2), in line with Kitamura et al

Recurrence, n (%)

New onset, n (%)

[30]. Gaynes et al [14] reported postnatal incidence estimates

Incidence, n (%)

from delivery to third month postpartum of 14.5% (95% CI,

10.9-19.2), whereas Chaudron et al [31] reported that 5.8%
Table 2

of women became clinically depressed between delivery and

the fourth month postpartum.
 S. Banti et al. / Comprehensive Psychiatry 52 (2011) 343–351 349

Fig. 2. Prevalence during pregnancy and in the postpartum period (arrows indicate scheduled assessments).

We cannot exclude that the incidence of depression perinatal period, women with a history of depression had a
reflects the incidence of this disorder in a female population greater risk of having a depressive episode compared with
during the childbearing age. However, although no study has those without a history [32,33].
examined the incidence of mMD in a female population
during the childbearing age with a similar time frame and 4.4. New onset
using a structured clinical interview based on DSM-IV
criteria, differences in incidence rates across these intervals Thirty-two women had their first-ever episode in the
cannot be estimated stably because of the sample size. perinatal period. During pregnancy, of the 774 women
Further research is warranted to replicate these data. without a history of depression, 1.6% had a new onset of
depression. In this period, we found no differences in new
4.3. Recurrence onset at 6 and 8 months of pregnancy; therefore, we did not
identify a specific time when the likelihood of having major
Unique to this study is the identification of both the or minor depression was higher. In the postpartum period,
incidence and the new onset of mMD during pregnancy. The the percentage of new onsets doubled to 5.7% and remained
meta-analysis of Gaynes et al [14] revealed a rate of stable across the assessments (Table 2). Therefore, consistent
incidence of 6.5% for major depression, but did not with Milgrom et al [32], there is no evidence of a peak at 1
distinguish new onsets from recurrences. Furthermore, it month after delivery.
did not consider the diagnosis of minor depression.
Recurrence rates in our study were 2-fold in the 4.5. Sociodemographic and clinical correlates of PND
postpartum period compared with the pregnancy period.
Data from the literature suggest that rates of relapse or Our results indicate that having an anxiety disorder at
recurrence during the perinatal phase are high in women with study entry, low socioeconomic status, and multiparity were
a history of depression and with recurrent mood disorder. In associated with higher odds of having a major or a minor
line with the literature, our results indicated that, in the depressive episode in the perinatal period:

Fig. 3. Cumulative incidence during pregnancy and in the postpartum period (arrows indicate scheduled assessments).
350 S. Banti et al. / Comprehensive Psychiatry 52 (2011) 343–351

It has been speculated that, during the perinatal period, cohabiting (92%), employed women (82.3%) with an
women with low socioeconomic status might be afraid of average socioeconomic status (90.8%). Sociodemographic
being unable to take care of their child [34]. Moreover, characteristics of women who refused to participate in the
having one or more children to look after during the current study are not available because data collection was possible
pregnancy or in the postpartum period might represent an only after the informed consent was signed, as prescribed by
additional life stress that may contribute to the occurrence of the Italian law on privacy; this did not allow us to ascertain if
a depressive episode. nonresponders were mostly of lower socioeconomic class
and had lower education, which may affect the external
4.6. Limitations validity of our results and their generalization. However,
women who dropped out from the study had a similar marital
This naturalistic, observational study has several limitations. status, socioeconomic status, educational level, living area,
First of all, although our response rate is comparable to and parity compared with those who completed the study,
that of similar studies that used self-report and structured although they were significantly younger and more fre-
clinical interviews, and in which the observation period quently unemployed: nonetheless, neither maternal age (in
spanned from pregnancy to postpartum period [30,35], it is samples of women aged N18 years) nor unemployment
moderately low (49.9%). seems to be associated with postpartum depression [33].
Secondly, we used an EPDS cutoff of at least 13 Finally, although postpartum depression is a major health
antenatally and postnatally, which is lower than that of issue for many women from diverse cultures and there are
14/15 suggested during pregnancy and higher than that of well-documented public health consequences associated
9/10 recommended to include minor depression, especially with the disorder, this condition often remains undiagnosed
during the postpartum period [36,37]. However, in this and untreated [40]; however, it must be acknowledged that,
study, EPDS was used as a screening instrument; to confirm even when a correct diagnosis was made, a low percentage
or to exclude a diagnosis of mMD, we administered the (34.8%) of women with mMD at study entry accepted some
Mood section of the SCID at each follow-up visit to women kind of treatment.
exceeding the EPDS threshold of 12 and to women with
EPDS scores less than 12 if an affirmative answer on critical
items, such as question 10 (presence of suicidal thoughts), 5. Conclusion
was present. We are aware that using this cutoff may lead to
an underestimation of some cases because some women with To our knowledge, this is the first study that aims to detect
minor depression may have had EPDS scores of around 10. the incidence of PND and in particular the distinction
Thirdly, we cannot rule out that the study design has between new onsets and recurrences. Our postpartum
contributed to reducing our prevalence and incidence rates of prevalence figures, lower than those reported in the
mMD because of the possible preventive and therapeutic literature, may reflect treatment during the study. Although
effect of conducting serial interviews [32]. Moreover, all the lack of a control group does not allow us to draw
depressed women had the possibility to receive psycholog- conclusions about the impact of providing a psychological/
ical counseling and/or a drug treatment and to be followed psychiatric counseling service in the perinatal period, our
through the entire pregnancy and the first year postpartum. results may suggest that casting a multiprofessional network
This may account for the reduction of prevalence rates of around women in need of support may be potentially useful
mMD both in pregnancy and during the postpartum period: for diminishing the effects of this disorder on the mother and
in fact, 92 women were already depressed at the study entry; the newborn child.
and we can assume that most of them were treated and Further studies that provide information on new onsets of
improved until they no longer satisfied DSM-IV mMD depression at various stages of pregnancy and postpartum
diagnostic criteria at the following evaluation stages. period are highly important to add to the knowledge base of
Fourthly, the percentage of women with at least a high improved clinical prediction of risk in any individual
school diploma (89.9%) is significantly higher than the woman. The present study may contribute to this literature,
percentage in women delivering in Tuscany (66.8%) [38]. even if is difficult to understand the kind of women who are
Our data could be representative of the Italian female generalized in this sample until more detailed research is
population of central-northern Italy if we take into account carried out.
the differences of sociodemographic characteristics among
the 3 geographic areas in which Italy is divided. However, Acknowledgment
our sample does not differ from that of Grant et al [39]
(which consisted of predominantly highly educated The PND-ReScU staff includes Serena Luisi, DPsych;
women, with the large majority [81%] having attained Jascha Bruni, MD; Elisa Cianelli, MD; Rossella Mazzoni,
tertiary level education). MD; Alessia Corradini, DPsych; Caterina Cirri, DPsych;
A further limitation is the homogeneous socioeconomic Sara Di Biase, DPsych; Lucia Casimo, DPsych; Ylenia
status of our sample, including predominantly married/ Giunti, MD; and Benedetta Ciaponi, MD.
 S. Banti et al. / Comprehensive Psychiatry 52 (2011) 343–351 351

The authors thank Giulia Gray for editing the final version [19] Committee on Drugs. Use of psychoactive medication during
of the paper. pregnancy and possible effects on the fetus and newborn. Pediatrics
2000;105:880-7.
The authors thank all the women who participated, [20] Food and Drug Admistration. Labeling and prescription drug
without whom this study would not have been possible. advertising: content and format for labelling for human prescription
drugs. Fed Regist 1979;44:37434-67.
[21] First MB, Spitzer RL, Gibbon M, Williams JBW. Structured clinical
References interview for DSM-IV axis I disorders (SCID). New York: New York
State Psychiatric Institute, Biometrics Research; 1995.
[1] O'Hara M, Swain A. Rates and risk of postpartum depression: a meta- [22] Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression:
analysis. Int Rev Psychiatry 1996;8:37-54. development of the 10 item Edinburgh Postnatal Depression Scale. Br
[2] Beck CT. The effects of postpartum depression on maternal-infant J Psychiatry 1987;150:782-6.
interaction: a metaanalysis. Nurs Res 1995;44:298-304. [23] Murray D. The use of the Edinburgh Postnatal Depression Scale in
[3] Campbell SB, Cohon JF. Prevalence and correlates of postpartum pregnancy. Paper presented at the 4th International Marcé Society
depression in first-time mothers. J Abnorm Psychol 1991;100:594-9. Conference, University of Keele; 1988.
[4] Field T. Infants of depressed mothers. Dev Psychopathol 1992;4: [24] Benvenuti P, Ferrara M, Niccolai C, Valoriani V, Cox JL. The
49-66. Edinburgh Postnatal Depression Scale: validation for an Italian sample.
[5] Philipps LH, O'Hara MW. Prospective study of postpartum depres- J Affect Disord 1999;53(2):137-41.
sion: 4 1/2 year follow-up of women and children. J Abnorm Psychol [25] Bisoffi G, Mazzi MA, Dunn G. Evaluating screening questionnaires
1991;100:151-5. using receiver operating characteristic (ROC) curves from two-phase
[6] O'Hara MW, Schlechte JA, Lewis DA, Varner MW. Controlled (double) samples. Int J Methods Psychiatr Res 2000;9:121-33.
prospective study of postpartum mood disorders: psychological, [26] Dunn G, Pickles A, Tansella M, Vázquez-Barquero JL. Two-phase
environmental and hormonal variables. J Abnorm Psychol 1991;100 epidemiological surveys in psychiatric research. Br J Psychiatry 1999;
(1):63-73. 174:95-100.
[7] Buist A, Ross LE, Steiner M. Anxiety and mood disorders in [27] Frank E, Prien RF, Jarrett JB, Keller MB, Kupfer DJ, Lavori PW, et al.
pregnancy and the postpartum period. In: Castle DJ, Kulkarni J, Abel Conceptualization and rationale for consensus definitions of terms in
M, editors. Mood and anxiety disorders in women; 2006. p. 136-62. major depressive disorder: response, remission, recovery, relapse, and
[8] Gotlib IH, Whiffen VE, Mount JH, Milne K, Cordy NI. Prevalence recurrence. Arch Gen Psychiatry 1991;48:851-5.
rates and demographic characteristics associated with depression in [28] Beck CT. Revision of the postpartum depression predictors inventory.
pregnancy and the postpartum. J Consult Clin Psychiatry 1989;57: J Obstet Gynecol Neonatal Nurs 2002;31(4):394-402.
269-74. [29] Gonidakis F, Rabavilas AD, Varsou E, Kreatsas G, Christodoulou GN.
[9] Yonkers KA, Ramin SM, Rush AJ, Navarrete CA, Carmody T, A 6-month study of postpartum depression and related factors in
March D, et al. Onset and persistence of postpartum depression in an Athens, Greece. Compr Psychiatry 2008;49(3):275-82.
inner-city maternal health clinic system. Am J Psychiatry 2001;158: [30] Kitamura T, Yoshida K, Okano T, Kinoshita K, Hayashi M, Toyoda N,
1856-63. et al. Multicentre prospective study of perinatal depression in Japan:
[10] Evans J, Heron J, Francomb H, Oke S, Golding J. Cohort study of incidence and correlates of antenatal and postnatal depression. Arch
depressed mood during pregnancy and after childbirth. BMJ 2001;323: Womens Ment Health 2006;9:121-30.
257-60. [31] Chaudron LH, Klein MH, Remington P, Palta M, Allen C, Essex MJ.
[11] Kitamura T, Shima S, Sugawara M, Toda MA. Psychological and Predictors, prodromes and incidence of postpartum depression.
social correlates of the onset of affective disorders among pregnant J Psychosom Obstetr Gynecol 2001;22:103-12.
women. Psychol Med 1993;23:967-75. [32] Milgrom J, Gemmill AW, Bilszta JL, Hayes B, Barnett B, Brooks J,
[12] Kumar R, Robson MK. A prospective study of emotional disorders in et al. Antenatal risk factors for postnatal depression: a large
childbearing women. Br J Psychiatry 1984;144:35-47. prospective study. J Affect Disord 2008;108:147-57.
[13] Bennett HA, Einarson A, Taddio A, Koren G, Einarson TR. [33] Robertson E, Grace S, Wallington T, Stewart DE. Antenatal risk
Depression during pregnancy: overview of clinical factors. Clin factors for postpartum depression: a synthesis of recent literature. Gen
Drug Investig 2004;24:157-79. Hosp Psychiatry 2004;26:289-95.
[14] Gaynes BN, Gavin N, Meltzer-Brody S, Lohr KN, Swinson T, [34] Segre LS, O'Hara W, Arndt S, Stuart S. The prevalence of postpartum
Gartlehner G, et al. Perinatal depression: prevalence, screening depression. The relative significance of three social status indices. Soc
accuracy, and screening outcomes. Evidence report/technology Psychiatr Epidemiol 2007;42:316-21.
assessment no. 119. (prepared by the RTI-University of North Carolina [35] Cox JL, Murray D, Chapman GA. Controlled study of the onset,
Evidence-based Practice Center, under contract no. 290-02-0016.) duration and prevalence of postnatal depression. Br J Psychiatry 1993;
AHRQ publication no. 05- E006-2. Rockville (Md): Agency for 163:27-31.
Healthcare Research and Quality; 2005. [36] Murray D, Cox JL. Screening for depression during pregnancy with the
[15] Gavin NI, Gaynes BN, Lohr KN, Meltzer-Brody S, Gartlehner G, Edinburgh Depression Scale (EPDS). Journal of Reproductive and
Swinson T. Perinatal depression: a systematic review of prevalence and Infant Psychology 1990;8:99-105.
incidence. Obstet Gynecol 2005;106(5):1071-83. [37] Matthey S, Henshaw C, Elliott S, Barnett B. Variability in use of cut-
[16] American Psychiatric Association (APA). Diagnostic and statistical off scores and formats on the Edinburgh Postnatal Depression Scale:
manual of mental disorders. 4th ed. Washington, DC: American implications for clinical and research practice. Arch Womens Ment
Psychiatric Association; 1994. Health 2006;9(6):309-15.
[17] Halbreich U, Karkun S. Cross-cultural and social diversity of [38] Agenzia Regionale Sanità Toscana. Nascere in Toscana Anni
prevalence of postpartum depression and depressive symptoms. 2002-2004. Documenti ARS 21, 2006.
J Affect Disord 2006;91:97-111. [39] Grant KA, Mc Mahon C, Austin MP. Maternal anxiety during the
[18] Borri C, Mauri M, Oppo A, Banti S, Rambelli C, Ramacciotti D, et al. transition to parenthood: a prospective study. J Affect Disord 2008;
Axis-I psychopathology and functional impairment at the 3rd month of 108:101-11.
pregnancy. Results from the Perinatal Depression–Research & [40] Dennis CL, Stewart DE. Treatment of postpartum depression, part 1: a
Screening Unit (PND-ReScU) study. J Clin Psychiatry 2008;69(10): critical review of biological interventions. J Clin Psychiatry 2004;65
1617-24. (9):1242-51.