Psychotic Illness in First-Time Mothers
with No Previous Psychiatric Hospitalizations:
A Population-Based Study
Unnur Valdimarsdo ´ttir1,2*, Christina M. Hultman1,3, Bernard Harlow4, Sven Cnattingius1, Pa ¨r Spare ´n1
1 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden, 2 Centre of Public Health Sciences, Faculty of Medicine, University of
Iceland, Reykjavı ´k, Iceland, 3 Department of Neuroscience, Psychiatry, Ullera ˚ker, Uppsala University, Uppsala, Sweden, 4 Department of Epidemiology and Community
Health, University of Minnesota School of Public Health, Minneapolis, Minnesota, United States of America
Funding: This research was
supported by grants from the
National Alliance for Research on
Schizophrenia and Depression
(NARSAD) as part of the Toulmin
Research Partners Program, and The
Swedish Council for Working Life
and Social Research (grant # 2013/
2002). The postdoctoral fellowship
of UV is supported by David and
Astrid Hegele ´ns Foundation,
Karolinska Institutet. The funding
sources had no role in study design,
data collection and analysis, decision
to publish, or preparation of the
Competing Interests: The authors
have declared that no competing
Academic Editor: Phillipa Hay,
University of Western Sydney,
Citation: Valdimarsdo ´ttir U, Hultman
CM, Harlow B, Cnattingius S, Spare ´n
P (2009) Psychotic illness in first-time
mothers with no previous
psychiatric hospitalizations: A
population-based study. PLoS Med
6(2): e1000013. doi:10.1371/journal.
Received: October 31, 2006
Accepted: November 25, 2008
Published: February 10, 2009
Copyright: ? 2009 Valdimarsdo ´ttir
et al. This is an open-access article
distributed under the terms of the
Creative Commons Attribution
License, which permits unrestricted
use, distribution, and reproduction
in any medium, provided the
original author and source are
Abbreviations: CI, confidence
interval; ICD, International
Classification of Diseases; LGA, large
for gestational age; SGA, small for
* To whom correspondence should
be addressed. E-mail: firstname.lastname@example.org
A B S T R A C T
Psychotic illness following childbirth is a relatively rare but severe condition with
unexplained etiology. The aim of this study was to investigate the impact of maternal
background characteristics and obstetric factors on the risk of postpartum psychosis,
specifically among mothers with no previous psychiatric hospitalizations.
Methods and Findings
We investigated incidence rates and potential maternal and obstetric risk factors of
psychoses after childbirth in a national cohort of women who were first-time mothers from
1983 through 2000 (n¼745,596). Proportional hazard regression models were used to estimate
relative risks of psychoses during and after the first 90 d postpartum, among mothers without
any previous psychiatric hospitalization and among all mothers. Within 90 d after delivery, 892
women (1.2 per 1,000 births; 4.84 per 1,000 person-years) were hospitalized due to psychoses
and 436 of these (0.6 per 1,000 births; 2.38 per 1,000 person-years) had not previously been
hospitalized for any psychiatric disorder. During follow-up after the 90 d postpartum period,
the corresponding incidence rates per 1,000 person-years were reduced to 0.65 for all women
and 0.49 for women not previously hospitalized. During (but not after) the first 90 d
postpartum the risk of psychoses among women without any previous psychiatric hospital-
ization was independently affected by: maternal age (35 y or older versus 19 y or younger;
hazard ratio 2.4, 95% confidence interval [CI] 1.2 to 4.7); high birth weight (? 4,500 g; hazard
ratio 0.3, 95% CI 0.1 to 1.0); and diabetes (hazard ratio 0).
The incidence of psychotic illness peaks immediately following a first childbirth, and almost
50% of the cases are women without any previous psychiatric hospitalization. High maternal
age increases the risk while diabetes and high birth weight are associated with reduced risk of
first-onset psychoses, distinctly during the postpartum period.
The Editors’ Summary of this article follows the references.
PLoS Medicine | www.plosmedicine.orgFebruary 2009 | Volume 6 | Issue 2 | e1000013 0194
P PL Lo oS S MEDICINE
A psychotic illness starting shortly after childbirth is a
relatively rare condition [1,2]. However, the negative impli-
cations of such an illness can be enormous: repeated episodes
of psychoses and hospitalization of the mother , increased
risk of self-harm or suicide , as well as the rare but tragic
occurrences of harm to the newborn infant and infanticide
. An obvious obstacle for effective prevention is that the
etiology of psychotic illness in the postpartum period is
Whether the postpartum period really poses any additional
risk of psychoses has been a matter of speculation: compared
to population- or prepregnancy rates, relative risks from 1.09
 to 12.7  for psychoses during the 3 mo postpartum have
been reported. Apart from the mother’s own or family history
of psychoses [6,7], few background factors have consistently
been associated with increased risks of psychoses in the
postpartum period . In a previous study we found that
almost 14% of women with previous psychiatric hospital-
izations suffered from postpartum psychotic or bipolar
episodes compared to 0.05% of women without previous
psychiatric hospitalizations . Still, some women have their
first and sometimes their only psychotic episode during the
postpartum period [9,10]. It is not known if this is solely due
to a biological vulnerability for psychoses, or if psychosocial
or obstetric factors influence the risk. To date, few studies
have addressed the influence of the mother’s background and
obstetric characteristics on the risk of postpartum psychosis
while controlling for previous psychiatric vulnerability, e.g.,
psychiatric hospitalizations. Moreover, it remains to be
investigated whether episodes of psychosis during the 90 d
postpartum period have risk factors distinct from psychoses
occurring at later times during motherhood.
Studying risk factors for psychoses during the postpartum
period is a methodological challenge, because of low
incidence rates and the confounding effects of previous
psychiatric morbidity. Sweden has excellent conditions for
research in this area, with population-based registers cover-
ing essentially all births and inpatient records. Using these
data sources, we investigated the incidence rate and risk
factors of psychotic illnesses diagnosed during and after the
first 90 d postpartum among all Swedish first-time mothers,
and specifically among mothers without any previous
psychiatric hospitalization. Controlling for previous psychi-
atric hospitalizations, we hypothesized that established risk
factors for nonpuerperal psychosis as well as obstetric
complications would increase the risk of maternal psychoses
during the first 90 d postpartum.
All primiparous women registered in the Swedish Medical
Birth Registry from 1 January 1983 to 31 December 2000 (n¼
745,596) were considered. Diseases in Swedish registers are
coded according to the International Classification of
Diseases (ICD); the 8th version (ICD-8) was used through
1986, the 9th revision (ICD-9) between 1987 and 1996, and the
10th revision (ICD-10) has been used since 1997. Women
without any registered psychiatric diagnosis (ICD-8 codes
290–390; ICD-9 codes 290–316; and ICD-10 codes F00–F99) in
the Hospital Discharge Registry before the date of delivery
were analyzed separately.
Psychoses Criteria during the Postpartum Period
We included information on psychoses registered in the
Hospital Discharge Registry as a primary or secondary
diagnoses (ICD-8 codes 294.4, 295–299; ICD-9 codes 295–
298; and ICD-10 codes F20–31 or F53.1) during the first 90 d
postpartum; or alternatively, later than 90 d until next
pregnancy, emigration from Sweden, death, or end of follow-
up (31 December 2001). A psychosis case could have had more
than one hospitalization for psychoses after childbirth, but we
counted only the first hospitalization. The discharge diag-
noses are made by the attending psychiatrist based on
observations during hospitalization and evaluation of the
patient as well as medical records at discharge. The registered
ICD discharge diagnoses have been reported to have high
agreement with diagnoses based on DSM (Diagnostic and
Statistical Manual of Mental Disorders) criteria  as well as
with diagnoses based on semistructured interviews and
medical records .
Women hospitalized for psychoses were categorized into
the following diagnoses: (1) postpartum psychosis (ICD-8 code
294.4; ICD-10 code F53.1), (2) acute/reactive psychosis (ICD-8
codes 298.00–299.99; ICD-9 codes 298A–X; ICD-10 codes F23
and F28–29), (3) schizophrenia (ICD-8 codes 295.00–295.99;
ICD-9 codes 295A–X; ICD-10 code F20), (4) affective (ICD-8
codes 296.00–296.99; ICD-9 codes 296A–X; ICD-10 codes
F30–3), (5) paranoia (ICD-8 codes 297.00–297.98; ICD-9 codes
297BCDWX; ICD-10 code F22) and, (6) schizoaffective
disorder (ICD-8 code 295.7; ICD-9 code 295H; ICD-10 code
Potential Risk Factors
Independent variables were mainly obtained from the
Medical Birth Registry. The maternal background factors
were: age at birth; years of education (from the Education
Registry); family situation (living with child’s father or not);
country of birth (Sweden; other Nordic country; non-Nordic
European country, North America, Australia and New
Zealand; Asia, Africa, South America); living in one of the
three biggest cities in Sweden (Stockholm, Gothenburg, and
Malmo ¨) or not; hypertensive disease (ICD-8 codes 400–404 or
637; ICD-9 code 642; ICD-10 codes O10–O14); diabetes
(pregestational or gestational; ICD-8 code 250; ICD-9 codes
648A or 648W; ICD-10 code O24); maternal cigarette smoking
at admission to maternity care; and year of birth.
Obstetric and perinatal exposures were: perinatal death
(stillbirth or infant death within 7 d), congenital malforma-
tions (ICD-8 and ICD-9 codes 740–759, ICD-10 codes Q00–
Q99), gestational age (in completed weeks), and birth weight.
Large for gestational age (LGA) and small for gestational age
(SGA) were defined as a birth weight of more or less than 2
standard deviations from the mean birth weight for gesta-
tional age, respectively, according to the Swedish reference
curve for fetal growth . We also included information on
the infant’s sex, single or multiple births (the index child
being the first-born multiple), mode of delivery (vaginal or
cesarean delivery), dystocia (ICD-8 code 657; ICD-9 code 661;
ICD-10 code O62), and antepartum bleeding (ICD-8 codes
632 and 651; ICD-9 code 641; ICD-10 codes O44–O46).
The study conforms to STROBE guidelines for cohort
PLoS Medicine | www.plosmedicine.orgFebruary 2009 | Volume 6 | Issue 2 | e1000013 0195
studies (Text S1) and was approved by the Ethics Review
Board at Karolinska Institutet, # 03–462.
We calculated the proportion of cases with no previous
psychiatric hospitalizations from the total psychosis cases
during the first 90 d after first births. The incidence rates of
total psychosis cases and first psychosis cases were calculated
within categories of 30 d of the first year postpartum and
then annually until next pregnancy, death, emigration from
Sweden, or towards the end of the observation period (31
December 2001); whichever occurred first. We used Cox’s
proportional hazard regression models to calculate adjusted
hazard ratios for psychoses (95% confidence intervals [CIs])
during and after the first 90 d postpartum. The Wald test for
interactions was performed to test potential interactions
between previous psychiatric hospitalizations and all mater-
nal as well as obstetric characteristics on the risk of psychosis
during the first 90 d postpartum. SAS systems software
version 9.1 was used for the statistical analyses.
Of 745,596 first-time mothers, 892 women (1.2 per 1,000
births) were hospitalized for psychoses during the first 90 d
postpartum. In all, 436 women (49%; 0.6 per 1,000 births) had
not previously been hospitalized for any psychiatric disorder,
and the majority of them were diagnosed with reactive/acute
psychoses or a postpartum psychosis (Table 1).
Incidence rates of maternal psychoses during the first 12
mo postpartum and then annual incidence rates until the end
of the observation period are illustrated in Figure 1.
Incidence rates peaked during the first month following
childbirth; 285 out of the total 892 psychosis cases (32%) were
hospitalized within 7 d after childbirth and 523 (59%) within
14 d. During the first 90 d postpartum, approximately half of
the total number of psychosis cases were accounted for by
women without any previous psychiatric hospitalization; the
30-d incidence rate was 4.99 (95% CI 4.46–5.58) per 1,000
person-years, the 90 d incidence rate was 2.38 (95% CI 2.17–
2.62) per 1,000 person-years, and the incidence rate for the
rest of the observation period was 0.49 (95% CI 0.46–0.52) per
Table 2 shows hazard ratios during and after 90 d
postpartum for all maternal psychoses as well as for women
without any previous psychiatric hospitalization, adjusted for
Maternal age and diabetes specifically affect the risk of
psychoses during the first 90 d postpartum. Higher maternal
age increases the risk of psychoses among mothers without
any previous psychiatric hospitalization during the first 90 d
postpartum. Compared to mothers aged 19 y or younger,
mothers aged 35 y or older had a more than two-fold
increased risk of psychoses during, but not after, the first 90 d
More than 6,000 women were diagnosed with diabetes
(gestational or pregestational); none of these women devel-
oped postpartum psychoses during the first 90 d postpartum,
while the corresponding risk after the 90 d was around 1.0.
In contrast to all women, women without any previous
psychiatric hospitalization had none or a limited increase in
risk of psychoses during the first 90 d postpartum due to low
level of education, not cohabitating with the infant’s father,
and maternal smoking. However, these factors had similar
impact on the risk of psychoses after the first 90 d postpartum
among all women as well as among women without any
previous psychiatric hospitalization. The increased risk of
psychoses among immigrant women (the mother being born
outside Sweden), particularly among mothers born in another
Nordic country, was amplified after the first 90 d postpartum.
Psychosis hospitalizations of all women, before and after the
90 d postpartum period, decreased markedly in Sweden
during the observation period (Table 2).
When we added perinatal death to the model presented in
Table 2, the hazard ratio of psychoses for women 35 y or more
and without any previous psychiatric hospitalizations de-
creased from 2.4 to 1.9 (95% CI 0.9–4.0) during the first 90 d
Pregnancy and Obstetric Characteristics
Women without any previous psychiatric hospitalization
had an increased risk of psychoses when giving birth to a
child with very low birth weight (? 1,500 g). In contrast, high
birth weight (? 4,500 g) and LGA decreased the risk by 70%–
80% during, but not after, the first 90 d postpartum (Table 3).
Perinatal death, congenital malformations, preterm birth
Table 1. Psychosis Diagnoses of All First-Time Mothers and First-Time Mothers without Any Previous Psychiatric Hospitalizations
during and after the First 90 Days Postpartum
Diagnosis Psychoses within 90 Days PostpartumPsychoses .90 Days Postpartum
All, n (%) No Previous Psychiatric
Illness, n (%)
All, n (%) No Previous Psychiatric
Illness, n (%)
PLoS Medicine | www.plosmedicine.org February 2009 | Volume 6 | Issue 2 | e10000130196
(,32 wk), and cesarean delivery were not statistically
significant risk factors of psychoses among women without
any previous psychiatric hospitalization. However, these
adverse pregnancy outcomes increased the risk of psychoses
among all women during the first 90 d postpartum. After the
first 90 d postpartum, perinatal death was the only obstetric
factor increasing the risk of psychoses among all women, and
multiple births showed a protective effect for psychoses
among mothers without any previous psychiatric hospital-
ization as well as for all mothers.
The Wald test for interactions was performed to test
potential interactions between previous psychiatric hospital-
izations and all maternal as well as obstetric characteristics on
the risk of psychosis during the first 90 d postpartum. No
statistically significant interactions were observed except for
cesarean section (p ¼ 0.0015); maternal age (p ¼ 0.0830) and
gestational age (p ¼ 0.0949) were the only other variables
approaching (but not reaching) statistical significance. We
further tested for interactions between maternal age and
other maternal characteristics on the risk of postpartum
psychosis; none of these proved statistically significant.
Our findings suggest that the immediate time period
following childbirth entails (compared to subsequent periods)
a substantially increased risk of psychotic illness of the first-
time mother. This also holds true for mothers without any
previous psychiatric hospitalization that account for almost
half of the psychosis cases during the first 90 d postpartum.
Among women without any previous psychiatric hospital-
ization, greater maternal age and lower birth weight of the
infant increase the risk of psychoses distinctly during the
postpartum period, while maternal diabetes and high birth
weight of the infant appear to be protective. These findings
have implications for the etiology of psychotic illness
occurring in the immediate postpartum period.
The Incidence of Psychoses during the Postpartum Period
Women without any previous psychiatric hospitalizations
before the date of delivery had a more than ten times higher
incidence rate during the first month postpartum compared
to after the first 90 d postpartum. Thus, our findings provide
strong support for the notion that the period following
childbirth carries an increased risk of psychoses among
women without any previous psychiatric hospitalization.
Whether the rate of first psychosis episodes in first-time
mothers reaches the levels of nulliparous women following
the 90 d postpartum remains to be investigated. Population
rates of first psychoses among women represent a compli-
cated comparison to our study since these typically reflect
rates of nulliparous, primiparous, and multiparous women
during and after the postpartum period.
The fact that one-third of all the psychosis cases during the
first 90 d are concentrated in the first 7 d (one-twelfth of the
total period) indicates that the birth may have a causal role.
As suggested by previous research [1,13] stressful character-
istics of the birth, e.g., perinatal death, might be a triggering
factor. Alternatively, women after childbirth experience an
enormous decrease in levels of estrogens and other hormones
produced by the placenta. The hypothesis that dramatic
reductions in hormone exposure mediate the risk of
psychoses after childbirth has some support from previous
research. Firstly, the higher age of onset of schizophrenia in
women as compared to men, and the increased female
perimenopausal incidence rates of schizophrenia have been
suggested to be related to loss of the protective effects of
estrogens . Secondly, schizophrenic women may be more
Figure 1. Incidence of Psychoses among Swedish First-Time Mothers
Dashed line: all maternal psychoses; solid line: psychoses in mothers without any previous psychiatric diagnoses.
PLoS Medicine | www.plosmedicine.orgFebruary 2009 | Volume 6 | Issue 2 | e1000013 0197
likely to be admitted for their psychoses during the low-
estrogen phase of their menstrual cycle . Thirdly,
improvements of postpartum psychotic symptoms after
estradiol administration were reported in a case series of
about dozen mothers [16,17]. In contrast, no protective
effects were found of estradiol administration on a post-
partum relapse of 29 women previously diagnosed with
affective psychoses . Thus, the evidence is not conclusive
on whether a change in estrogen exposure or other hormones
around childbirth may be a triggering factor for psychoses
during the postpartum period.
Postpartum-Specific Risk Factors
The investigation of risk factors for maternal psychoses
following childbirth is complicated by the risk of confound-
ing by previous psychiatric illness. We therefore specifically
studied potential maternal and obstetric risk factors among
women without any previous psychiatric hospitalization.
We found that maternal diabetes and high infant birth
weight seemed protective of first-onset psychoses during, but
not after, the first 90 d postpartum. We are not aware of any
previous study reporting such findings. Regarding diabetes,
the possibility of a chance finding cannot be excluded;
however, similar findings were reported in a German study: in
313 people newly diagnosed with type 1 diabetes, none had a
probable nonpuerperal psychotic disorder as compared to
1.5% in representative comparison sample of 2,046 individ-
uals . Intense surveillance of diabetic mothers during
pregnancy might reduce the risk of overt psychotic illness: a
recent report indicates that such a program for diabetic
mothers reduces the risk of adverse pregnancy outcomes as
well as risk of depression 3 mo postpartum . Alternatively,
diabetic mothers tend to have larger babies , and high
birth weight was also associated with protection from
psychoses during the first 90 d postpartum. Pregnancies
involving large babies  as well as multiple births  have
both been associated with high levels of estrogen. However,
pregnancy hormones are produced by the placenta and
hormone levels decrease rapidly after the expulsion of the
placenta. Thus, although possible, it still remains uncertain
whether the possible protective mechanisms of diabetes or
high birth weight include hormonal effects.
The majority of previous studies addressing mean maternal
age [1,9,24,25] or the proportion of cases within different age
groups  have not found age differences between women
with and without psychoses during the postpartum period.
Similar to our findings on all first-time mothers, a recent
Swedish study observed that mothers aged 40–44 y faced a
more than 6-fold increased risk of psychoses during the
Table 2. Swedish First-Time Mothers’ Characteristics and Health as Potential Risk Factors for Psychoses during and after the First 90
Mothers’ CharacteristicsBirths, n Psychoses within the First
90 Days Postpartum, Adjusted
Hazard Ratio (95% CI)a
Psychoses .90 Days Postpartum,
Adjusted Hazard Ratio (95% CI)a
All (n ¼ 892) No Previous Psychiatric
Illness (n ¼ 436)
All (n ¼ 1,834) No Previous Psychiatric
Illness (n ¼ 1,395)
Age (years) 19 or younger
35 or older
15 or more
9 or less
Country of birth
Cohabit. with father Yes
Yes, 0–9 cigarettes/d
Yes, 10þ cigarettes/d
Living in big city
Year of birth
aHazard ratios adjusted for: maternal age, education, country of birth, cohabiting with father, living in big city, diabetes, hypertension, smoking, and year of birth.
PLoS Medicine | www.plosmedicine.orgFebruary 2009 | Volume 6 | Issue 2 | e10000130198
postpartum period compared to mothers aged 20–24 y .
These age-related risks may be exaggerated by confounding
effects of uncontrolled previous psychiatric morbidity; when
we excluded women with any kind of previous psychiatric
hospitalizations, we found that older women (? 35 y) still had
a more than 2-fold increased risk of psychoses during the first
90 d postpartum.
The fact that the relationship between maternal age and
psychoses is restricted to the first 90 d indicates that our
findings are reasonably not explained by a ‘‘natural’’ age of
onset of psychotic disorders. Further, previous findings
suggest that the onset of nonpuerperal psychotic disorders
peaks at a somewhat different age level than was observed in
the present study among first-time mothers [14,27]. Moreover,
the hypothesis that the association is explained by the notion
that undetected ‘‘premorbid women’’ delay their childbearing
is also undermined by the specific ‘‘postpartum nature’’ of the
relationship. The notion that the effect of maternal age is
mediated by hormonal differences remains uncertain: some
have found estrogen levels during pregnancy to vary with
maternal age  while others have not . Alternatively, the
risk of adverse pregnancy outcomes, e.g., perinatal death, has
been reported to increase with maternal age [29,30]. Probably
because of small numbers, perinatal death was not statistically
significantly associated with psychoses among women without
any previous psychiatric hospitalizations, and the age-related
risk was reduced in an additional analysis when including
perinatal death as a covariate. Thus, it is possible that such
severe obstetric stress partly mediates the relationship
between maternal age and the risk of psychoses.
Our findings provide no evidence that the relationship
between severe obstetric hazards (perinatal death or con-
genital malformation) and postpartum psychosis is different
for all mothers versus mothers without previous psychiatric
hospitalizations. While a previous investigation reported an
increased risk of postpartum psychosis after cesarean delivery
, our interaction tests show that cesarean delivery is
associated with increased risk for postpartum psychoses only
among women with previous psychiatric hospitalizations,
whereas there is a tendency towards protection among
women without previous psychiatric hospitalization.
In contrast to the postpartum-specific risk factors, our
findings indicate that not cohabitating with the child’s father,
maternal smoking, and lower maternal education are general
risk factors for maternal psychoses; it is unlikely that these
factors have any causal role for psychoses that are limited to
the immediate postpartum period. Further, we observed an
increased risk of psychoses among mothers born outside of
Sweden. Being born in another Nordic country was a risk
factor for maternal psychoses during and after the first 90 d
postpartum. Additional analyses among the women born in
other Nordic countries revealed that a Finnish nationality
carried the highest risks (unpublished data). A previous
Swedish study indicates that after controlling for sociodemo-
Table 3. Exposures Related to the Child or Obstetric Complications and the Risk of Psychoses during First 90 Days Versus 91 Days or
Later Postpartum among Swedish First-Time Mothers
Exposure Category Perinatal/
Births, nPsychoses within the First
90 Days Postpartum, Adjusted
Hazard Ratio (95% CI)a
Psychoses .90 Days Postpartum,
Adjusted Hazard Ratio (95% CI)a
All (n ¼ 892) No Previous Psychiatric
Illness (n ¼ 436)
All (n ¼ 1,834)No Previous Psychiatric
Illness (n ¼ 1,395)
Perinatal death No
Gestational age (weeks)
Birth weight (grams)
Birth weight for gestational age
Mode of delivery
aHazard ratios adjusted for: maternal age, education, cohabitation, diabetes, smoking, and calendar period of birth.
PLoS Medicine | www.plosmedicine.orgFebruary 2009 | Volume 6 | Issue 2 | e10000130199
graphic factors, elevated relative risks of psychotic conditions
remain high among first- and second-generation Finnish
Based on a nationwide cohort of 745,596 first-time mothers,
to our knowledge our study is the largest to date to address
risk factors of maternal psychosis during the postpartum
period. Strengths of the study include the use of complete
population-based registers of births  and hospital dis-
charges  with high validity, exploration of the influence of
potential risk factors of psychoses by time, and the separate
analyses of mothers without any previous psychiatric hospital-
izations. Information on pre- and postpartum psychiatric
disorders was limited to inpatients, and no information is
available about women with clinical or subclinical syndromes
who are treated as outpatients. The 50% decrease in psychosis
hospitalizations during the last two decades probably reflects
reforms of psychiatric care in Sweden rather than an actual
decrease in incidence . Consequently, the incidence rates
of psychoses after childbirth are probably underestimated.
Concerning the postpartum-specific risk factors, we judge the
risk of potential confounding due to undetected psychotic
outpatients as minimal, since the hazard ratios stay intact
when controlling for calendar period of birth. However, we
cannot exclude the possibility that some unmeasured con-
founders, e.g., social characteristics or sources of support,
may explain our findings.
Conclusion and Implications
The immediate period following childbirth carries, com-
pared to subsequent periods, high incidence rates for
psychoses in first-time mothers, even among those without
any previous psychiatric hospitalization. We found the risk of
such first-onset psychotic illness during the 90 d postpartum
period to be increased with maternal age and reduced among
mothers with diabetes and those giving birth to infants with
high birth weight. These findings may have implications for the
etiology of psychotic disorders during the postpartum period.
Text S1. STROBE Guidelines for Reporting Cohort Studies
Found at doi:10.1371/journal.pmed.1000013.sd001 (89 KB DOC).
Author contributions. All authors (UV, CMH, BH, SC, PS)
participated in the planning, conception of research questions, and
design of the study. PS was responsible retrieving the data/localizing
the dataset, and PS and UV were responsible for analyzing the data.
UV drafted the article and all authors participated in interpreting
data and critically revising the manuscript for important intellectual
content. The final version to be published has been approved by all
1. Kendell RE, Chalmers JC, Platz C (1987) Epidemiology of puerperal
psychoses. Br J Psychiatry 150: 662–673.
2. Terp IM, Mortensen PB (1998) Post-partum psychoses. Clinical diagnoses
and relative risk of admission after parturition. Br J Psychiatry 172: 521–526.
3. Terp IM, Engholm G, Moller H, Mortensen PB (1999) A follow-up study of
postpartum psychoses: prognosis and risk factors for readmission. Acta
Psychiatr Scand 100: 40–46.
4. Appleby L, Mortensen PB, Faragher EB (1998) Suicide and other causes of
mortality after post-partum psychiatric admission. Br J Psychiatry 173:
Spinelli MG (2001) A systematic investigation of 16 cases of neonaticide.
Am J Psychiatry 158: 811–813.
Jones I, Craddock N (2001) Familiality of the puerperal trigger in bipolar
disorder: results of a family study. Am J Psychiatry 158: 913–917.
Harlow BL, Vitonis AF, Sparen P, Cnattingius S, Joffe H, et al. (2007)
Incidence of hospitalizations for postpartum psychotic and biopolar
episodes in women with and without prior pre-pregnancy or prenatal
psychiatric hospitalizations. Arch Gen Psychiatry 64: 42–28.
Seyfried LS, Marcus SM (2003) Postpartum mood disorders. Int Rev
Psychiatry 15: 231–242.
Videbech P, Gouliaev G (1995) First admission with puerperal psychosis: 7–
14 years of follow-up. Acta Psychiatr Scand 91: 167–173.
10. Robertson E, Jones I, Haque S, Holder R, Craddock N (2005) Risk of
puerperal and non-puerperal recurrence of illness following bipolar
affective puerperal (post-partum) psychosis. Br J Psychiatry 186: 258–259.
11. Kristjansson E, Allebeck P, Widstedt B (1987) Validity of the diagnosis of
schizophrenia in a psychiatric inpatient register. Nord J Psychiatry 41: 229–
12. Ekholm B, Ekholm A, Adolfsson R, Vares M, Osby U, et al. (2005) Evaluation
of diagnostic procedures in Swedish patients with schizophrenia and
related psychoses. Nord J Psychiatry 59: 457–464.
13. Marsal K, Persson PH, Larsen T, Lilja H, Selbing A, et al. (1996) Intrauterine
growth curves based on ultrasonically estimated foetal weights. Acta
Paediatry 85: 843–848.
14. Kendell RE, Rennie D, Clarke JA, Dean C (1981) The social and obstetric
correlates of psychiatric admission in the puerperium. Psychol Med 11:
15. Hafner H Gender differences in schizophrenia. (2003) Psychoneuroendoc-
rinology 28: 17–54.
16. Huber TJ, Borsutzky M, Schneider U, Emrich HM (2004) Psychotic
disorders and gonadal function: evidence supporting the oestrogen
hypothesis. Acta Psychiatr Scand 109: 269–274.
17. Ahokas A, Aito M (1999) Role of estradiol in puerperal psychosis.
Psychopharmacology (Berl) 147: 108–110.
18. Ahokas A, Aito M, Rimon R (2000) Positive treatment effect of estradiol in
postpartum psychosis: a pilot study. J Clin Psychiatry 61: 166–169.
19. Kumar C, McIvor RJ, Davies T, Brown N, Papadopoulos A, et al. (2003)
Estrogen administration does not reduce the rate of recurrence of affective
psychosis after childbirth. J Clin Psychiatry 64: 112–118.
20. Petrak F, Hardt J, Wittchen HU, Kulzer B, Hirsch A, et al. (2003) Prevalence
of psychiatric disorders in an onset cohort of adults with type 1 diabetes.
Diabetes Metab Res Rev 19: 216–222.
21. Crowther CA, Hiller JE, Moss JR, McPhee AJ, Jeffries WS, et al. (2005) Effect
of treatment of gestational diabetes mellitus on pregnancy outcomes. N
Engl J Med 352: 2477–2486.
22. Ehrenberg HM, Mercer BM, Catalano PM (2004) The influence of obesity
and diabetes on the prevalence of macrosomia. Am J Obstet Gynecol 191:
23. Kaijser M, Granath F, Jacobsen G, Cnattingius S, Ekbom A (2000) Maternal
pregnancy estriol levels in relation to anamnestic and fetal anthropometric
data. Epidemiology 11: 315–319.
24. Thomas HV, Murphy MF, Key TJ, Fentiman IS, Allen DS, et al. (1998)
Pregnancy and menstrual hormone levels in mothers of twins compared to
mothers of singletons. Ann Hum Biol 25: 69–75.
25. Kumar R, Marks M, Platz C, Yoshida K (1995) Clinical survey of a
psychiatric mother and baby unit: characteristics of 100 consecutive
admissions. J Affect Disord 33: 11–22.
26. Sharma V, Smith A, Khan M (2004) The relationship between duration of
labour, time of delivery, and puerperal psychosis. J Affect Disord 83: 215–20.
year of delivery associated with hospital admission for postpartum psychosis?
A study of 500,000 first-time mothers. Acta Psychiatr Scand 112: 47–53.
28. Welham JL, Thomis R, McGrath JJ (2003) Age-at-first-registration and
heterogeneity in affective psychoses. Aust N Z J Psychiatry 37: 66–69.
29. Panagiotopoulou K, Katsouyanni K, Petridou E, Garas Y, Tzonou A, et al.
(1990) Maternal age, parity, and pregnancy estrogens. Cancer Causes
Control 1: 119–124.
30. Cnattingius S, Forman MR, Berendes HW, Isotalo L (1992) Delayed
childbearing and risk of adverse perinatal outcome. A population-based
study. JAMA 268: 886–890.
31. Cnattingius S, Stephansson O. (2002) The epidemiology of stillbirth. Semin
Perinatol 26: 25–30.
32. Hjern A, Wicks S, Dalman C (2004) Social adversity contributes to high
morbidity in psychoses in immigrants—a national cohort study in two
generations of Swedish residents. Psychol Med 34: 1025–1033.
33. Cnattingius S, Ericson A, Gunnarskog J, Kallen B (1990) A quality study of a
medical birth registry. Scand J Soc Med 18: 143–148.
34. Osby U, Hammar N, Brandt L, Wicks S, Thinsz Z, et al. (2001) Time trends
in first admissions for schizophrenia and paranoid psychosis in Stockholm
County, Sweden. Schizophr Res 47: 247–254.
PLoS Medicine | www.plosmedicine.org February 2009 | Volume 6 | Issue 2 | e10000130200
Background. The first cries of a new life echo around the delivery suite:
this is a time of great joy for most women. Yet, in the following days and
weeks (the postpartum period), up to 80% of new mothers experience
some sort of mental disturbance. Usually, this is the ‘‘baby blues,’’ a
normal reaction to childbirth that is characterized by short-lived mood
swings or postnatal depression. However, about one in 1,000 women
develop postpartum psychosis, a serious mental disorder that needs
immediate medical attention. Postpartum psychosis usually develops
suddenly in the first 2–3 weeks after delivery and, like other forms of
psychosis, is characterized by a loss of contact with reality. Women with
postpartum psychosis may have false ideas about current events and
about themselves (delusions) and see and hear things that are not there
(hallucinations). They sometimes stop eating or sleeping and may
become anxious and agitated. In the worst cases, they can have suicidal
thoughts or even threaten their baby’s life. Treatment for postpartum
psychosis includes antipsychotic drugs, counseling, and hospital
admission if the woman is a danger to herself or others.
Why Was This Study Done? Women with a personal or family history of
psychosis have an increased risk of developing postpartum psychosis,
but what causes this disorder is unknown. The rapid changes in hormone
levels that occur after delivery are likely to be involved—but might social
circumstances, stress, other illnesses, or the birth itself also affect
whether a woman develops postpartum psychosis? If additional risk
factors for postpartum psychosis could be identified, it might be possible
to prevent some cases of this serious mental disorder. In this study, the
researchers investigate the incidence rate (the rate at which new cases
occur in a population) and risk factors for psychotic illnesses diagnosed
among first-time mothers registered in the Swedish Medical Birth
Registry between 1983 and 2000.
What Did the Researchers Do and Find? The researchers identified
three-quarters of a million first-time mothers and, from the Swedish
Hospital Discharge Registry, found that 892 of these women (1.2 per
1,000 births) had been admitted to hospital because of psychosis within
90 days of giving birth. Put another way, the incidence rate of psychosis
over the first 90 days postpartum in this population was 4.84 per 1,000
person-years. Almost half of the women who developed postpartum
psychosis had not been previously admitted to hospital for any
psychiatric disorder. Among this subset of women, the incidence rate
of postpartum psychosis was highest during the first month after
delivery but dropped to less than a tenth of this initial rate after 90 days
postpartum. Furthermore, the risk of developing psychosis during the
first 90 days postpartum (but not after) increased with age—women
older than 35 years were more than twice as likely to develop psychosis
than those aged 19 years or less—but was reduced in women who had
large babies or who had diabetes. Many other factors (including smoking
and not living with the infant’s father) did not affect the risk of psychosis
during the first 90 days postpartum in these women.
What Do These Findings Mean? These findings indicate that the
occurrence of psychotic illness severe enough to require hospitalization
peaks shortly after giving birth for the first time, even in women with no
previous psychiatric illness. Indeed, women with no history of mental
disorders account for almost half the women admitted to hospital for
postpartum psychosis, at least in Sweden. The timing of the peak of
postpartum psychosis supports the idea that either giving birth or the
hormonal changes that occur shortly after may trigger the development
of psychosis, and the findings that maternal diabetes and high infant
birth weight reduce the risk of postpartum psychosis whereas increasing
maternal age increases the risk provide new clues about the causes of
postpartum psychosis. Most importantly, however, these findings
highlight the importance of carefully monitoring women for psychosis
during the first month after delivery.
Additional Information. Please access these Web sites via the online
version of this summary at http://dx.doi.org/10.1371/journal.pmed.
? This paper is further discussed in a PLoS Medicine Perspective by
? The MedlinePlus Encyclopedia contains a page on MedlinePlus
encyclopedia psychosis (in English and Spanish); MedlinePlus also
provides links to information on psychotic disorders
? The UK National Health Service Direct Health encyclopedia has
information on psychosis and on postnatal depression
? Mental Health America has a fact sheet on postpartum disorders
PLoS Medicine | www.plosmedicine.org February 2009 | Volume 6 | Issue 2 | e10000130201