Abstract

Postpartum depression is a depressive episode that begins around childbirth. DSM-5-TR does not list it as a separate disorder but as the specifier with peripartum onset, covering onset in pregnancy or within four weeks of delivery. It is usually detected with the Edinburgh Postnatal Depression Scale and national bodies disagree on whether universal screening improves outcomes. Related presentations include the postpartum blues, generally regarded as normal though a risk factor for depression, rare postpartum psychosis and depression in fathers, while cognitive studies report biased processing of infant cues, rumination and negative self-beliefs whose role as cause or consequence remains unresolved. Hormone sensitivity is supported in a subgroup by small experiments, child outcomes show small effects on language and IQ and variably supported behavioural effects, psychological therapy helps, antidepressant evidence is low certainty and the durability of neuroactive steroids is unknown.

Keywords: postpartum depression, postnatal depression, peripartum onset, perinatal mental health, mother-infant interaction

Postpartum depression is a depressive episode that begins around childbirth. The current American manual does not define it as a disorder of its own: it is a depressive episode carrying the specifier with peripartum onset, which applies to onset during pregnancy as well as after delivery (American Psychiatric Association, 2022), a merging of the two onsets that a DSM-5 commentary criticised, together with the four-week window, arguing for separating them and for a window of six months (Sharma & Mazmanian, 2014). The merging reflects how often the episode starts before the birth: among women who screened positive after delivery in a large case series, 40.1% reported onset after the birth, 33.4% during pregnancy and 26.5% before pregnancy (Wisner et al., 2013). Detection usually relies on the Edinburgh Postnatal Depression Scale, whose best balance of sensitivity and specificity fell at a cut-off of 11 or higher in an individual participant data meta-analysis (Levis et al., 2020). Depressed and anxious mothers processed infant emotion with a negative bias in a systematic review (Webb & Ayers, 2015), and psychotherapy outperformed control conditions with moderate to large effects in a meta-analysis of 43 randomised trials (Cuijpers et al., 2023). This page covers depression that begins around childbirth; depressive mood in general is covered on the depression page, and the diagnosis to which the peripartum-onset specifier attaches is covered on the major depressive disorder page.

The article first sets out what postpartum depression is and whether it is a distinct disorder, then traces the history of the concept, the symptoms, diagnosis and measurement of the condition, screening as a programme, and the experience of postpartum depression as mothers and fathers describe it. A section on the types of postpartum depression covers the specifier and its ICD-11 counterpart, first-onset and recurrent cases, anxious and psychotic presentations, data-derived subtypes, the postpartum blues, postpartum psychosis, bipolar postpartum depression, perinatal anxiety and obsessive-compulsive presentations, depression in fathers and in adoptive and other non-birthing parents, and depression after perinatal loss. Prevalence, course and recovery and the risk factors follow. Four sections then take up cognition: the cognitive model in the perinatal period; stress, interpersonal and evolutionary accounts; attention to infant cues, memory and executive function; and testing whether cognition is cause or consequence, followed by the validity of laboratory and observational measures. The biology of postpartum depression, sleep and circadian rhythms, and brain networks and imaging come next, then mother-infant interaction, child outcomes, breastfeeding and infant feeding, culture and social context, comorbidity, suicide and self-harm, intrusive thoughts of harm to the infant, and related conditions and normal adjustment. The treatment sections cover prevention, treatment evidence, how treatments work, neuroactive steroid drugs, hormones and brain stimulation, emerging treatments, parent-infant interventions, and access to care and service models. A worked example shows what a positive screen means for one woman, and the article closes with the current research directions, a discussion of what is settled and what is open, and the reader aids.

Key Takeaways

  • DSM-5-TR treats postpartum depression as a depressive episode with the specifier with peripartum onset, which covers onset in pregnancy as well as after delivery; whether it is a distinct disorder remains open.
  • Many episodes found after birth began earlier: in a screening series of ten thousand mothers, about 40% of screen-positive women had onset after delivery, a third during pregnancy and a quarter before pregnancy.
  • A high Edinburgh scale score is a positive screen, not a diagnosis; at a cut-off of 11 roughly half of positive screens are not major depression at an assumed prevalence of 12%, and national bodies disagree on universal screening, with the Canadian task force recommending conditionally against routine questionnaire screening on very uncertain evidence.
  • The postpartum blues are a brief state generally regarded as normal, although they are a risk factor for postpartum depression, postpartum psychosis is a rare psychiatric emergency tied mainly to the bipolar spectrum, and depression in fathers affected about one in ten in a meta-analysis, although fatherhood brought no rise in psychiatric admission in a register study.
  • A negative self-schema in pregnancy predicted later onset and brooding predicted symptom change, and depressed mothers showed negatively biased processing of infant emotion, but the one cognitive training trial was null on its primary outcome, so causation is not established.
  • Sensitivity to normal hormone change is supported in a subgroup by small experiments; an independent review found that zuranolone improves depression against placebo with moderate certainty, and found no clear difference for intravenous brexanolone with low certainty, in short, industry-sponsored trials.
  • Maternal depression is linked to child problems, though one critical review judged the effects on language and IQ small and the behavioural effects variably supported, and in a sibling comparison only concurrent, mother-reported maternal depression remained associated.
  • Psychological treatment helps, with gains lasting six to twelve months in a meta-analysis; a Cochrane review judged with moderate certainty that zuranolone improves depression compared with placebo over at most 45 days; and treating the mother's depression alone did not reliably improve the relationship with the infant.
  • Counselling interventions lowered the onset of perinatal depression, mostly in women at raised risk, in a review for the US Preventive Services Task Force, which recommends them with moderate certainty, whereas a Cochrane review of two small trials found antidepressant prevention unproven.
  • Intrusive thoughts of harming the baby were normative after birth in a prospective cohort and were not associated with physical aggression towards the infant in a province-wide cohort, whereas the rare filicides in a review of childbearing psychoses clustered in psychoses with depressive features.

What Postpartum Depression Is

MeSH defines postpartum depression as depression in women after childbirth that usually begins within four weeks of delivery and ranges in degree from a mild, passing depression to depressive disorders of neurotic or psychotic severity; its scope note credits the definition to DSM-IV (MeSH, D019052). MeSH files the descriptor under both depressive disorder and puerperal disorders, and lists eight entry terms that are spelling variants of one another: post-natal depression, post-partum depression, postnatal depression and postpartum depression, and post-natal dysphoria, post-partum dysphoria, postnatal dysphoria and postpartum dysphoria.

A Specifier, Not a Separate Diagnosis

The current American manual does not list postpartum depression as a disorder in its own right. It treats the condition as a depressive episode that carries the specifier with peripartum onset (American Psychiatric Association, 2022). The episode itself is the one described under major depressive disorder and, more broadly, depression. The specifier also merges onset during pregnancy with onset after delivery, a choice a DSM-5 commentary criticised, arguing that the two onsets may have different aetiologies and should be separated, and that the four-week window excludes many postpartum episodes and should be lengthened to six months (Sharma & Mazmanian, 2014).

The window has been contested for longer than that. The DSM-IV specifier allowed four weeks, whereas ICD-10, the older international manual, allowed six weeks; a commentary written before DSM-5 called the wording of the definition a problem needing careful reconsideration, though it also judged disease-specific time frames premature on the limited data then available (Wisner et al., 2010). A later review argued that the onset specifier should be extended to the first year following childbirth (Radoš et al., 2024).

The international classification organises the condition differently. ICD-11 has its own grouping for mental or behavioural disorders associated with pregnancy, childbirth or the puerperium (World Health Organization, 2024). A comparative review found the two manuals' overall organisation substantially similar in ICD-11 and DSM-5, with nineteen ICD-11 categories absent from DSM-5 and seven DSM-5 categories absent from ICD-11 (First et al., 2021). ICD-11 was restructured relative to ICD-10 (Reed et al., 2019).

Vocabulary

The literature uses several time words that are not interchangeable. Postpartum and postnatal both mean after birth, and in this literature they are used interchangeably for the mother; puerperal is the older medical term for the period after delivery. Antenatal means during pregnancy, peripartum spans late pregnancy and the weeks after birth, and perinatal spans pregnancy and the period after birth, which is how it is used for depression in both periods.

Prevalence in this field also depends on how a case is defined. Scoring above a cut-off on a screen like the Edinburgh Postnatal Depression Scale flags a woman for attention, but a diagnosis comes only from an interview, for example the Structured Clinical Interview for DSM-IV. The two do not give the same figure, and the estimates vary with the definition and the postpartum period chosen (O'Hara & Wisner, 2014). A systematic review of perinatal depression drew on studies using a structured clinical interview, and still reported wide confidence intervals and significant uncertainty in the true rates, restricted to English-language studies from developed countries (Gavin et al., 2005).

Is It a Distinct Disorder?

Whether postpartum depression is a distinct condition, or depression that happens to follow birth, has been argued since the manuals first gave it a name. A narrative review reports that studies with non-childbearing control groups found no difference in depression rates, while recent large-scale studies gave some evidence of raised postpartum risk, and it rates the data as equivocal (O'Hara & Wisner, 2014); the figures are set out under Prevalence, Course and Recovery.

Other designs bear on the question differently. In a five-year follow-up of 55 depressed first-time mothers and 40 controls, women whose first-ever episode followed childbirth were at raised risk only of later postnatal episodes, and not of episodes outside the postpartum period; the authors suggested that postnatal depression may be a specific nosological concept (Cooper & Murray, 1995). A cross-sectional comparison of 143 women with postnatal depression, 131 with depression outside the postpartum period and 173 controls found that both depressed groups differed from controls in neuroticism, dysfunctional attitudes and self-esteem, while the two depressed groups did not differ on any measure; the study found no trait specific to postnatal depression, although its design could not show cause and it had a high refusal rate (Jones et al., 2010). In a family study of 120 sibling pairs with recurrent unipolar depression, postpartum depression clustered in families when onset was defined as within four weeks, and showed no significant clustering when the definition was six months; familiality peaked at six to eight weeks (Forty et al., 2006).

The reviews do not agree. One found conflicting evidence on the disorder's validity and concluded that postpartum depression is probably a complex, heterogeneous phenotype (Di Florio & Meltzer-Brody, 2015). Another found mixed evidence, with early postpartum depression (the first eight weeks) possibly distinct from major depression and later cases possibly resembling it, so that the verdict depends on how the postpartum period is defined (Batt et al., 2020). A third reported that postpartum depression may differ from major depression, showing more anxiety, psychomotor symptoms, obsessive thoughts, fatigue and impaired concentration but less sad mood and less suicidal ideation, with a similar prevalence; it recommended modifying the diagnostic criteria while noting that more comparative research is needed (Radoš et al., 2024).

The question remains open. Early-onset cases may be distinct, according to one review (Batt et al., 2020), and the manuals currently treat the condition as a specifier of a depressive episode, not as a separate disorder (American Psychiatric Association, 2022).

Scale and Seriousness

Whatever its status, the condition is a significant clinical problem. One narrative review described it as common and serious, harming mothers and offspring, with prevalence estimates of 13% to 19% and the first six months after delivery a possible high-risk time (O'Hara & McCabe, 2013). A clinical review described it as common, with effects that reach the whole family (Stewart & Vigod, 2016). A Lancet Series review found that perinatal and other disorders share similar risk factors, while treatment considerations differ in pregnancy and breastfeeding (Howard et al., 2014). A disease primer reported that postpartum depression and anxiety are common and psychosis rare (Meltzer-Brody et al., 2018b). A review of puerperal psychiatric disorders noted a wide variety of them, including puerperal psychosis at about one in 1000 deliveries, and concluded that specialist multidisciplinary teams have an important place (Brockington, 2004).

History of the Concept

MeSH's own indexing history records how recently the condition gained a heading of its own: Depression, Postpartum dates from 1996, and before that the literature was indexed under depressive disorder (1981 to 1995), depression (1966 to 1995) and puerperal disorders (1966 to 1995) (MeSH, D019052).

The clinical interest is much older than the heading. Marcé published a treatise in 1858 on the mental disorders of pregnant, recently delivered and nursing women (Marcé, 1858). A later historical commentary judged it a landmark early contribution to perinatal psychiatry that anticipated modern findings, and noted that the work was largely ignored for 150 years (Trede et al., 2009). Puerperal insanity was a recognised diagnosis in the nineteenth century, and a historical account of it exists (Loudon, 1988).

Twentieth-century evidence first concerned the severe end. A record-linkage study followed a population of 470 000 people over 12 years and 54 087 births, and found that the risk of admission for psychosis was highest in the first 30 days after delivery. The authors concluded that the puerperal psychoses were probably manic-depressive illness and unrelated to schizophrenia (Kendell et al., 1987).

Non-psychotic depression after childbirth was far less well described. A 1968 study of what it called atypical depression following childbirth noted that postpartum depression was poorly characterised and had received little evaluation, with the early-puerperium tearfulness known as the blues better known (Pitt, 1968). A practical tool arrived later: the Edinburgh Postnatal Depression Scale, a 10-item self-report scale that takes about 5 minutes, showed satisfactory accuracy in screening postnatal depression in a validation study of 84 mothers (Cox et al., 1987).

A controlled prospective study complicated the picture. Childbearing women and matched controls did not differ in rates of minor or major depression, although childbearing women showed more depressive symptoms and poorer social adjustment in late pregnancy and the early puerperium (O'Hara et al., 1990). That result bears on whether childbirth itself causes depression, and not only on who becomes depressed.

Classification followed a contested course. A historical case study, drawing on archival and interview data, linked the exclusion of postpartum depression as its own category during the drafting of DSM-IV to questions about suitable hierarchies and definitions of evidence (Godderis, 2013). The postpartum onset specifier that resulted was criticised as problematic, in part because DSM-IV and ICD-10 set different windows (4 weeks against 6 weeks), and the authors judged that the DSM-V wording needed careful reconsideration while calling disease-specific time frames premature (Wisner et al., 2010). DSM-5 then introduced a with peripartum onset specifier (American Psychiatric Association, 2022). A critique argued that this merges prepartum and postpartum onset, which may have distinct aetiologies, and that its 4-week window excludes many postpartum episodes. The authors argued for splitting the onset types and lengthening the window (Sharma & Mazmanian, 2014). DSM-5 attaches its peripartum specifier to episodes that begin during pregnancy or within four weeks of delivery, while ICD-10 allows up to six weeks (Perry et al., 2021).

The concept therefore stands today as a specifier within the depressive disorders, framed as perinatal and no longer strictly postpartum (American Psychiatric Association, 2022; World Health Organization, 2024).

Symptoms, Diagnosis and Measurement

Postpartum depression is recognised by the symptoms of a depressive episode, and the two classification systems name the category and its peripartum specifier (American Psychiatric Association, 2022; World Health Organization, 2024). Measurement is a separate matter from diagnosis, and the two are easily run together. This section separates the diagnostic standard from the questionnaires built to approximate it, then takes the Edinburgh scale, the other instruments and the problem of somatic items in turn. How screening is organised in services is covered under Screening for Postpartum Depression.

Criteria and the Reference Standard

The presentation is that of a depressive episode: a sustained low or empty mood or loss of interest, accompanied by changes in functioning that a clinician weighs against the person's usual state. The manual sets out the symptom list and the specifier for onset around childbirth (American Psychiatric Association, 2022). This citation supports only the existence of that description; the criteria themselves are not reproduced here. A diagnosis rests on a clinical or structured interview. That interview is the reference standard against which questionnaires are validated, and the large Edinburgh scale analyses draw their major depression classification from structured or semi-structured interviews such as the Structured Clinical Interview for DSM Disorders, the Mini International Neuropsychiatric Interview and the Composite International Diagnostic Interview (Levis et al., 2020). Earlier validation work used the same logic with other interview schedules (Cox et al., 1987). A review of Edinburgh validation studies likewise treated interview-defined depression as the criterion for judging the scale (Gibson et al., 2009).

A questionnaire score is therefore a screen or a severity measure. A score above a cut-off flags a woman for assessment; it does not establish that she meets criteria for an episode.

The Edinburgh Postnatal Depression Scale

The Edinburgh Postnatal Depression Scale (EPDS) is a ten-item self-report screen that takes about five minutes to complete. It was developed and validated in a study of 84 mothers, where it showed satisfactory accuracy for screening postnatal depression against interview-based criteria (Cox et al., 1987). Questions of reliability and validity in measurement of this kind belong to psychometrics.

The cut-off dispute

The score at which a woman counts as screen-positive has been contested. A systematic review of 37 validation studies published between 1987 and 2008 found sensitivity ranging from 34 to 100 percent and specificity from 44 to 100 percent, with positive likelihood ratios from 1.61 to 78. It reported cut-offs of 9/10 for possible and 12/13 for probable postpartum depression, and 14/15 for antepartum depression. The authors noted marked heterogeneity and that the studies might not be directly comparable, and concluded that the scale may not suit all settings (Gibson et al., 2009).

An individual participant data meta-analysis later pooled 58 of 83 eligible studies, covering 15,557 women, 2,069 of them with major depression. A cut-off of 11 or higher balanced sensitivity and specificity best (sensitivity 0.81, specificity 0.88, Youden's J 0.73), and 13 or higher was more specific. Against 10 or higher (sensitivity 0.85, specificity 0.84, Youden's J 0.70), the 11 cut-off gave up some sensitivity for specificity; at 13 sensitivity fell to 0.66 and specificity rose to 0.95 (Youden's J 0.66). At a cut-off of 11 or higher and a major depression prevalence of 5% to 25%, positive predictive values ranged from 26 to 69 percent and negative predictive values from 93 to 99 percent (Levis et al., 2020). Accuracy did not differ significantly between pregnant and postpartum women, by country human development index or by risk-of-bias domain, and older age was associated with higher specificity across cut-offs 9 to 15. Compared with an earlier meta-analysis, the 11 cut-off here sat below that analysis's optimum of 12, with sensitivity 6 to 13 percent lower and specificity 4 to 7 percent higher. Limits were that 25 of 83 eligible datasets were missing, heterogeneity was moderate, few studies came from low and middle income countries, and subgroups by culture or trimester were not analysed (Levis et al., 2020). On the current evidence, then, 11 or higher is the best-supported balanced cut-off and 13 or higher the choice when specificity matters more.

How the scale is used in practice

Departures from validated cut-offs are not rare in published work. A methodological review of studies, drawn from a convenience sample and from one author's data, found unvalidated cut-offs and altered formats becoming increasingly common, and showed that an unvalidated cut-off could shift the proportion of high scorers substantially. Its recommendation was to use validated cut-offs and to justify any deviation; the effect of different formatting was not known (Matthey et al., 2006).

The scale does not measure depression alone. Items 3, 4 and 5 form an anxiety subscale, the EPDS-3A, which in a diagnostic accuracy study of 238 women and 218 men at six weeks postpartum was usable for screening probable anxiety, with a cut-off of 6 or higher in women and 4 or higher in men. The factor structure in men was unclear (Matthey, 2008). A high total score can therefore reflect anxiety as well as depression.

Use with fathers

The scale has also been validated in fathers, at a lower cut-off than for mothers. In a couples cohort recruited antenatally (208 fathers, 230 mothers), fathers endorsed seven of the ten items less often than mothers, the crying item most markedly, and the optimum cut-off was 5/6 in fathers against 7/8 in mothers at six weeks (Matthey et al., 2001). A later screening study of 192 UK fathers at about seven weeks postpartum, using a structured interview as the standard, reported a cut-off above 10 (sensitivity 89.5 percent, specificity 78.2 percent, AUC 0.916), and above 8 when generalised anxiety disorder was counted as a case. Its caveats were a modest participation rate, selection of fathers by high scores, an interview gap of 4.8 weeks and limited generalisability (Edmondson et al., 2010). The two paternal studies therefore disagree by several points on the best cut-off, and the question remains open; paternal presentations are taken up under Types of Postpartum Depression.

Other Instruments

Other tools differ in kind. The Postpartum Depression Screening Scale (PDSS) is a 35-item self-report screen developed for this population. In a sample of 525 new mothers its scores showed reliable and valid interpretation across seven dimensions, including sleeping and eating disturbances and loss of self, with alpha values of 0.83 to 0.94. Its sensitivity, specificity and predictive values were still pending in that report (Beck & Gable, 2000).

The Patient Health Questionnaire-9 (PHQ-9) is a general self-report depression measure, not one built for the perinatal period. In a meta-analysis of 35 studies of 19,760 women, its screening accuracy was nearly identical to the Edinburgh scale: at a cut point of 10 or higher, seven pooled criterion studies gave sensitivity 0.84 and specificity 0.81. Categorical agreement between the two instruments was only moderate, and only seven studies were pooled (Wang et al., 2021a).

The clinician-rated 17-item Hamilton Rating Scale for Depression is a different kind of instrument again. It appears in this article as the primary endpoint of the brexanolone trials, not as a screen (Meltzer-Brody et al., 2018a).

Somatic Symptoms After Birth

Sleep, fatigue and appetite are standard depression items, but they also change with pregnancy and the care of a newborn, which makes them harder to read as signs of illness. The evidence comes from pregnancy, not the postpartum period. A cross-sectional multigroup confirmatory factor analysis compared 255 pregnant women with 820 community women on eight scales of the Inventory of Depression and Anxiety Symptoms. Appetite aside, the somatic symptoms worked as valid signs of depression during pregnancy; both appetite loss and appetite gain performed poorly as indicators among pregnant women. Loadings differed by group: lassitude and appetite gain loaded lower in pregnancy than in the community samples, while insomnia and well-being loaded higher. A one-factor model fit adequately in both groups. The authors concluded that somatic symptoms other than appetite change are usable prenatally (Nylen et al., 2013). This result concerns prenatal depression and does not show how the same items behave after delivery.

Screening for Postpartum Depression

Screening is a programme, not a questionnaire. It consists of deciding who is asked, when, with what instrument, and what happens to the answer. The instrument's accuracy (covered under Symptoms, Diagnosis and Measurement) is one component; whether asking women improves what happens to them is a separate question, and the evidence for it is much thinner than the evidence for the instruments.

Where and When Women Are Asked

Two sets of guidance place screening in different clinics. The American College of Obstetricians and Gynecologists, in a committee opinion, advised that perinatal patients be screened for depression and anxiety with a standardized, validated tool, and observed that while screening by itself may bring some clinical benefit, the greatest benefit comes when it leads to starting treatment or making a referral (American College of Obstetricians and Gynecologists, 2018). In 2023 the college followed this with an evidence-based clinical practice guideline on screening for and diagnosing perinatal mood and anxiety disorders (American College of Obstetricians and Gynecologists' Committee on Clinical Practice Guidelines–Obstetrics, 2023). A clinical report on paediatric practice extended the task to the 1-, 2-, 4- and 6-month well-child visits, advising that practices screen, coordinate and refer. The same report noted that positive screens often lack follow-up and that diagnosed mothers often miss evidence-based treatment (Earls et al., 2019).

In practice, being asked varies widely by place. A cross-sectional surveillance survey, the 2018 Pregnancy Risk Assessment Monitoring System, covered 32,659 women with a recent birth across 31 sites. Overall, 79.1% reported being asked about depression prenatally and 87.4% postpartum, but the site figures ranged from 51.3% (Puerto Rico) to 90.7% (Alaska) prenatally, and from 50.7% (Puerto Rico) to 96.2% (Vermont) postpartum. Inquiry rose over 2016 to 2018, from 76.2% to 79.3% prenatally and from 84.1% to 88.0% postpartum. The survey relied on self-report, was subject to recall and social-desirability bias, and could not show whether screening or referral was actually performed (Bauman et al., 2020).

What a Positive Screen Means

A positive screen is a prompt for assessment, not a diagnosis. In an individual participant data meta-analysis of 58 studies (15,557 women), the Edinburgh Postnatal Depression Scale was compared against structured diagnostic interviews. A cut-off of 11 gave sensitivity 0.81 and specificity 0.88, and was judged to balance the two best; a cut-off of 13 lost sensitivity (0.66) but was more specific (0.95). Accuracy did not differ significantly between pregnant and postpartum women. At a cut-off of 11 or higher and a prevalence of 5% to 25%, the positive predictive values ranged from 26% to 69%, and the negative predictive values from 93% to 99% (Levis et al., 2020). The pooled data had limits: 25 of 83 eligible datasets were missing, heterogeneity was moderate, and few studies came from low- and middle-income countries. A positive result therefore leaves a substantial share of screen-positive women without major depression, and the interactive demonstration below shows how that share moves with the cut-off. The Worked Example applies the figures to a single case.

Cut-off, prevalence and what a positive screen means

Choose an Edinburgh scale cut-off and set how common major depression is among the women screened.

010,000 women screened
Sensitivity and specificity at this cut-off0.81 and 0.88
Women who screen positive2,028
Of those, women with major depression (positive predictive value)972 (47.9%)
Women with major depression who screen negative228
Negative predictive value97.1%
Sensitivity and specificity are the values reported by Levis et al. (2020) for the Edinburgh Postnatal Depression Scale against major depression. The prevalence setting and the 10,000 women are illustrative arithmetic, not data from any study.

Does Screening Improve Outcomes?

Detecting depression and improving outcomes are different things, and the reviews disagree about the second. A systematic review for a US task force, covering pregnant and postpartum women, found that screening programmes were associated with lower depression risk than usual care. The review pooled 6 trials (n = 11,869), with relative reductions of 18% to 59% and absolute reductions of 2.1% to 9.1%; the authors concluded that screening may reduce symptoms and prevalence, and noted that evidence was sparser for pregnant women (O'Connor et al., 2016).

A systematic review of randomised trials, limited to the screening question itself, reached a more cautious reading. It found one postpartum trial (462 mothers), none in pregnancy, and called the evidence unsound: the risk of bias was high, outcome measures were changed after the fact, and the implied effect per woman treated was implausibly large. Symptoms at 6 months favoured the screened group (SMD = 0.34), yet the authors argued that guidelines recommending screening should be reconsidered (Thombs et al., 2014).

A UK evidence synthesis with a decision model examined cost as well as validity. It judged the Edinburgh scale reasonably valid, found clinical-effectiveness evidence only suggestive, and found no existing cost-effectiveness studies. The model concluded that screening was not value for money, a result driven by false positives, and that the National Screening Committee criteria were not currently met. The authors called for randomised trial evidence (Hewitt et al., 2009).

The Care Cascade After a Positive Screen

Identification is only the first step. A systematic review of weighted mean rates, searching English-language articles through March 2015, described a perinatal depression treatment cascade with losses at each stage. For postpartum depression, 30.8% of cases were identified, 15.8% were treated, 6.3% received adequate treatment and 3.2% reached remission. Identification, adequate treatment and remission were lower than for antenatal depression, whereas treatment was higher (15.8% versus 13.6%) (Cox et al., 2016). Services after screening are described under Access to Care and Service Delivery.

Part of the loss is at disclosure. A meta-synthesis of 24 UK qualitative studies, all concerned with postnatal depression, found that recognising the problem and continuity of care went with help-seeking, while stigma, an unclear role for the health professional and negative experiences of childbirth care went against it. Women were more likely to seek help from a familiar professional than from an unfamiliar one (Button et al., 2017).

National Recommendations Compared

National bodies have read the same evidence differently. The US Preventive Services Task Force recommended (grade B) screening of adults, including pregnant and postpartum women, on condition that systems for diagnosis, treatment and follow-up are in place (Siu, 2016). Its 2023 update kept a B recommendation for adults aged 19 and older, again including pregnant and postpartum women, with moderate certainty of moderate net benefit; for suicide-risk screening it judged the evidence insufficient (an I statement) (US Preventive Services Task Force, 2023). The American obstetric college advises screening all perinatal patients (American College of Obstetricians and Gynecologists, 2018).

The Canadian task force went the other way. Its guideline made a conditional recommendation against routine questionnaire screening of postpartum women, on very low certainty evidence. The evidence rested on one small trial (n = 462) with very serious risk-of-bias concerns; no trials were found in pregnancy, and the evidence on harms was limited (Lang et al., 2022).

The bodies are reading the same thin trial evidence differently, and no study cited on this page reports a later trial that decides between them. The question remains open: the US task force and the American obstetric college currently advise screening, and the Canadian task force currently recommends conditionally against routine questionnaire screening.

The Experience of Postpartum Depression

Qualitative research records what the diagnostic criteria cannot: how the condition is lived, named and concealed. Its evidence comes from interviews and focus groups, so it describes meanings and patterns of experience and does not estimate how common any of them are.

Mothers' Descriptions

A metasynthesis of 18 qualitative studies of postpartum depression, using the Noblit and Hare approach, produced four overarching themes. One is the gap between what a mother expected motherhood to be and what she found. Another is a downward spiral. The third is pervasive loss, and the fourth, making gains, records that the experience was not only one of deterioration. The authors drew implications for practice and theory from the synthesis (Beck, 2002).

Accounts of unhappiness after childbirth are not confined to one health system. Focus groups and key-informant interviews in 15 centres across 11 countries, with four to six focus groups per centre, found morbid unhappiness after birth in every centre, among mothers of babies aged 5 to 7 months. Fathers and grandmothers described themes that closely matched the mothers' themes. Lack of social support, family conflict, sleeplessness and tiredness, and problems with the baby were each named as sources of unhappiness, and the mother-in-law was named in every centre but Sweden (Oates et al., 2004). The authors concluded that morbid unhappiness is universal while doubting that it amounts to an equivalent illness everywhere. They also noted that informants per centre were few, that no representative sample was possible, and that the findings cannot be generalised (Oates et al., 2004).

The accounts also differed by setting. The term postnatal depression was used less in Portugal, Switzerland, Uganda and among UK Asians than elsewhere, and antidepressants were mentioned only in the USA. Lay informants located remedies in psychosocial support, whereas health professionals asked for more health services (Oates et al., 2004).

Why Mothers Do Not Seek Help

A qualitative systematic review of 40 articles published from 1966 to 2005 found that women did not proactively seek help. The barriers it identified were nondisclosure, myths about the condition and gaps in services. Where treatment preferences were reported, talking therapies were preferred to drug treatment, though few studies had examined the acceptability of treatment (Dennis & Chung-Lee, 2006).

A later meta-ethnography of 24 UK qualitative studies, all concerned with postnatal depression, explained help-seeking behaviour through three influences. Recognising that something was wrong made a mother more likely to seek help, and continuity of care did the same. Contact with a familiar health professional favoured help-seeking over contact with an unfamiliar one, while an unclear professional role and negative experiences of care in childbirth reduced it. Social stigma worked in the same direction, and the review linked it to self-silence and to the fear of being judged a bad mother (Button et al., 2017). The authors called for continuity of care and staff training. They also said the impact for milder distress was unclear and that more research was needed on how distress is identified (Button et al., 2017).

Sexual and Gender Minority Parents

A systematic review used qualitative meta-aggregation across 26 studies with 1199 LGBTQ2S+ childbearing participants. Perinatal distress was described alongside stigma, exclusion and isolation. Heteronormativity and cisnormativity were associated with greater distress, and gaps in provider knowledge with poorer care. The authors concluded that systems-level inequities and exclusion were largely responsible (Kirubarajan et al., 2022).

Fathers' Descriptions

In a UK interview study using thematic analysis, 19 fathers described distress and doubted that they were entitled to support. The authors concluded that men may hesitate to seek help (Darwin et al., 2017). The cohort was not diverse in ethnicity or socioeconomic position, all fathers lived with a partner, interviews took place 5 to 10 months after the birth, and data saturation was not achieved (Darwin et al., 2017). The accounts therefore show that such experiences exist and not how widespread they are.

Recovery and Course

Longitudinal studies published from 1985 to 2012 indicate that most women recover and that a subgroup follows a chronic course. The review's authors judged that conceptual and methodological limitations preclude strong conclusions (Vliegen et al., 2014). Course and recovery are treated in Prevalence, Course and Recovery.

Types of Postpartum Depression

A type in this section is a recognised presentation or classification, not necessarily a distinct disease. Two diagnostic manuals supply the formal vocabulary: the DSM-5-TR names a specifier, with peripartum onset (American Psychiatric Association, 2022), and the ICD-11 names a grouping, mental or behavioural disorders associated with pregnancy, childbirth or the puerperium (World Health Organization, 2024). The research literature adds subtypes derived from data. Around these sit the postpartum blues, a brief state generally regarded as normal, boundary presentations that belong to other diagnoses, and depression in parents who did not give birth.

Cognitive findings are uneven across the list. Some types have laboratory or imaging studies of their own; for others the studies cited on this page report none, and each entry below says which is the case. Table 1 lists the types in the order in which they are discussed.

Table 1

Types of Postpartum Depression and Neighbouring Presentations

TypeStandingKey instrument
Peripartum onset: pregnancy versus postpartum onsetDSM-5-TR specifierEdinburgh Postnatal Depression Scale
ICD-11 mental or behavioural disorders associated with pregnancy, childbirth or the puerperiumICD-11 grouping–
First-onset versus recurrent postpartum depressionCourse distinction–
With anxious distressDSM-5-TR specifier–
Postpartum depression with psychotic featuresDSM-5-TR specifier–
Data-derived subtypes and trajectoriesResearch constructEdinburgh Postnatal Depression Scale
Postpartum bluesGenerally regarded as normal; a risk factor for depression–
Postpartum psychosisBoundary presentation–
Bipolar postpartum depressionBoundary presentation–
Perinatal anxiety and obsessive-compulsive presentationsBoundary presentationEPDS-3A (items 3, 4 and 5 of the Edinburgh scale)
Paternal postpartum depressionPresentation outside the specifierEdinburgh Postnatal Depression Scale, with a cut-off set for fathers
Depression in adoptive and other non-birthing parentsPresentation outside the specifierEdinburgh Postnatal Depression Scale
Depression after perinatal lossBoundary presentationPatient Health Questionnaire-2 (adapted), in one surveillance survey

Note. Standing gives each entry's place in the manuals or in the research literature. An en dash marks a type for which the studies cited in this section name no instrument of its own; for the postpartum blues, no established criteria exist.

Peripartum Onset: Pregnancy Versus Postpartum Onset

The DSM-5-TR specifier is named with peripartum onset (American Psychiatric Association, 2022). One label therefore spans two onsets. A commentary on the DSM-5 wording observed that the specifier merges episodes that begin before delivery with episodes that begin after it, and that its 4-week postpartum window leaves out many episodes that start later (Sharma & Mazmanian, 2014).

Much of the depression detected after birth did not begin there. In a sequential case series, ten thousand postpartum mothers were screened with the Edinburgh Postnatal Depression Scale, and 1396 (14.0%) scored 10 or higher. Among the screen-positive women the episode had begun after delivery in 40.1%, during pregnancy in 33.4% and before pregnancy in 26.5%, so onset often predated the birth (Wisner et al., 2013). These proportions rest on incomplete assessment: only 59.2% of the screen-positive women (826) completed the home visit, and only 10.5% (147) the telephone interview (Wisner et al., 2013).

When the episode began, and which definition counts it

Each definition counts a different share of the depressive episodes found by screening after delivery.

26.5%Before pregnancy33.4%During pregnancy40.1%After delivery

Counted under this definition: 40.1% of the episodes. Left out: 59.9%.

The three proportions are the onset timings reported by Wisner et al. (2013) among women who screened positive after delivery and completed assessment. The three definitions are simplified for illustration and are not the wording of any diagnostic manual.

Symptom levels in pregnancy are at least as high as those after delivery. In a longitudinal cohort of pregnant women in Avon, Edinburgh scale scores were higher at 32 weeks of pregnancy than at 8 weeks postpartum (difference in means 0.88, 95% confidence interval 0.79 to 0.97): 1222 women (13.5%) scored above threshold at 32 weeks, 821 (9.1%) at 8 weeks after birth and 147 (1.6%) throughout, which led the authors to conclude that depression was no more prevalent after birth than before it (Evans et al., 2001). Antenatal symptoms also predict what follows. In a prospective community study of 8323 women in England, anxiety in pregnancy predicted postnatal depression with antenatal depression controlled (OR = 3.22, p < 0.001), on self-report measures and with selective attrition (Heron et al., 2004).

Timing of onset is linked to what the depression looks like. A cluster analysis of 663 women aged 19 to 40, drawn from seven of an international consortium's 19 sites, found five subtypes that differed in the quality of their symptoms and in when the symptoms began (Putnam et al., 2017). Two narrative reviews ask whether timing marks a separate disorder. One judged the evidence that postpartum depression is distinct from major depressive disorder to be mixed and dependent on how the postpartum period is defined; it concluded that depression arising in the early postpartum period may be distinct, while later-onset cases may resemble major depression (Batt et al., 2020). The other concluded that peripartum depression may be distinct, describing more anxiety and impaired concentration and less sad mood and suicidal ideation than in major depressive disorder at a similar prevalence; it argued that the diagnostic criteria should be modified and called for more comparative research (Radoš et al., 2024).

Critics of the single label have asked for it to be changed. The DSM-5 commentary argued for splitting prepartum from postpartum onset and for a longer postpartum window, on the ground that the two onsets possibly have distinct causes (Sharma & Mazmanian, 2014). Setting the 4-week window of DSM-IV beside the 6-week window of ICD-10, an earlier commentary argued that DSM-V must reconsider the time bound for postpartum onset, but cautioned that, with so little data available, it may be too early to set time frames specific to each disease (Wisner et al., 2010). The current position is that one specifier covers both onsets. Whether depression that begins in pregnancy and depression that begins after delivery differ in cause remains open.

Cognitive Implications

Depression in the perinatal period has to be separated from the cognitive change that accompanies pregnancy itself. A random-effects meta-analysis of 20 studies (709 pregnant women and 521 non-pregnant women) found poorer overall cognition in the pregnant group (SMD 0.52, 95% CI 0.07-0.97, P = 0.025). Cross-sectional comparisons showed significantly lower general cognition, memory and executive functioning in the third trimester, with no such reduction in the first two; when the same women were tracked over time, scores fell from the first trimester to the second, which led the authors to conclude that the differences arise mainly during the first trimester (Davies et al., 2018).

Against that background, a small case-control pilot gave tests of working memory to 28 women at 34-36 weeks of gestation, 10 of them depressed, and to 26 non-pregnant controls. The depressed pregnant women made more errors than the non-depressed pregnant women, with large effects (Spatial Working Memory task: d = 0.99, p = .014; Self-Ordered Pointing: d = 1.24, p = .006). The pregnant group as a whole did not differ from controls, paragraph recall did not separate the two pregnant groups, and the depressed women's errors fell after delivery (Hampson et al., 2015). The study had no postpartum-depressed comparison group, and its authors describe the data as pilot-level and in need of replication in a larger cohort (Hampson et al., 2015).

Attention to emotional material gave a different picture. A cross-sectional emotional Stroop study of pregnant and postpartum Swedish women, 40 with antenatal and 33 with postpartum depressive disorder, found no attentional bias in untreated peripartum depression. Interference in antenatal depression did not differ significantly from that of controls, and women with postpartum depression responded faster to emotional than to neutral stimuli, a pattern the authors read as possible emotional numbing (Edvinsson et al., 2017).

Subjective complaints track symptoms more consistently than test performance does. A systematic review of 42 studies of perinatal women, with exploratory meta-analyses, found that depressive symptoms went with more self-reported cognitive complaints (six studies), that associations with verbal memory and recognition memory were not significant, and that results for working memory, attention and cognitive flexibility were inconsistent. It called the evidence for objective impairment inconclusive and noted that most studies were cross-sectional (Sobral et al., 2026). No study cited on this page compares cognition in pregnancy-onset and postpartum-onset depression.

ICD-11 Mental or Behavioural Disorders Associated With Pregnancy, Childbirth or the Puerperium

The ICD-11 counterpart to the peripartum specifier is a grouping, mental or behavioural disorders associated with pregnancy, childbirth or the puerperium, whose categories are named for the absence or the presence of psychotic symptoms (World Health Organization, 2024). Its predecessor differed from the American manual on timing: ICD-10 used a 6-week postpartum window where DSM-IV used 4 weeks (Wisner et al., 2010).

ICD-11 restructured the classification of mental disorders relative to ICD-10 (Reed et al., 2019). A comparative review of 103 diagnostic entities shared by ICD-11 and DSM-5 found the two manuals closer than before: 31 entities (30.1%) were essentially identical, 42 (40.8%) had minor definitional differences and 20 (19.4%) had major ones, and nineteen ICD-11 categories had no DSM-5 counterpart against seven DSM-5 categories absent from ICD-11 (First et al., 2021). Both reviews are cited here for the classifications as a whole, not for the perinatal grouping.

Cognitive Implications

The grouping is a classification device, and no study cited on this page reports cognitive findings by ICD-11 category. The findings that apply are those for the depressive syndrome itself, described in the section Attention to Infant Cues, Memory and Executive Function.

First-Onset Versus Recurrent Postpartum Depression

Whether a postnatal episode is a woman's first or continues an earlier history changes what follows it. A prospective cohort followed 55 primiparous women who had been depressed after delivery, and 40 controls, for five years. Women whose postnatal episode was their first had more later postnatal episodes than controls but no more non-postpartum episodes; women whose postnatal episode was a recurrence showed the reverse, with more non-postpartum episodes than controls and no excess of postnatal ones (Cooper & Murray, 1995). The authors read the pattern as possible support for postnatal depression as a specific nosological concept (Cooper & Murray, 1995).

Register data agree on recurrence. Among 457,317 primiparous Danish women with no prior psychiatric history, with 789,068 births, postpartum affective disorder was rare, at 4,550 episodes (0.6%), yet a woman who had had one was 26.9 to 46.4 times as likely as a woman without that history to have another after a later birth, with recurrence between 15% and 21% (Rasmussen et al., 2017). Recurrence was higher when the first episode had involved hospital contact than when it had been identified through medication (RR = 1.7, 95% CI 1.1-2.7), and an earlier postpartum episode also raised the rate of affective disorder outside the postpartum period, 6.2 to 6.6 times. The registers identify treated cases, so milder episodes were possibly missed and the cases did not necessarily meet DSM-IV criteria (Rasmussen et al., 2017).

A family study adds a third strand. Among 120 sibling pairs with recurrent unipolar depression, postpartum depression clustered in families when it was defined narrowly, by onset within 4 weeks of delivery, but not significantly when the definition was widened to 6 months; familiality peaked at a definition of 6-8 weeks (Forty et al., 2006).

These findings are consistent with a birth-specific vulnerability in some women: recurrence that follows births, and familial clustering confined to early onset. They do not establish a distinct disease. The follow-up study rests on 55 depressed women, the register study counts treated cases and found episodes outside the postpartum period raised as well, and the familial effect was not significant under the broad definition. Whether first-onset postpartum depression is a separate condition remains open.

Cognitive Implications

The one relevant comparison on this page found nothing specific to the postnatal form. A cross-sectional study compared 143 women with postnatal depression, 131 with non-postnatal depression and 173 controls on the Eysenck Personality Questionnaire, the Dysfunctional Attitudes Scale and a self-esteem questionnaire. The postnatal group scored higher than controls on neuroticism and dysfunctional attitudes and lower on self-esteem, but did not differ significantly from the non-postnatal depressed group on any measure, and the authors found no evidence that these traits confer risk specific to the postnatal period (Jones et al., 2010). The design cannot separate cause from consequence, the personality and cognitive measures were limited, and onset in pregnancy was not examined (Jones et al., 2010). No study cited on this page compares cognition in first-onset and recurrent cases.

With Anxious Distress

The DSM-5-TR also names a specifier with anxious distress (American Psychiatric Association, 2022). Anxiety is prominent in perinatal depression whichever label is applied. A meta-analysis of 66 studies covering 162 120 women in 30 countries found co-morbid anxiety and depression in the perinatal period to be common, with prevalence that did not differ significantly by year of publication or by country income (Falah-Hassani et al., 2017).

Data-derived classifications separate anxious presentations as well. Greater anxiety severity went with membership of the most severe class in a latent class analysis of consortium data (Postpartum Depression: Action Towards Causes and Treatment (PACT) Consortium, 2015), and anxiety was one of three symptom dimensions in a cluster analysis that found five subtypes (Putnam et al., 2017). Both analyses are described under Data-Derived Subtypes and Trajectories, and co-occurring anxiety disorders in the Comorbidity section.

Cognitive Implications

No study cited on this page compares the cognition of anxious and non-anxious postpartum depression.

Postpartum Depression With Psychotic Features

The DSM-5-TR names a further specifier, with psychotic features (American Psychiatric Association, 2022). It marks a depressive episode in which delusions or hallucinations occur.

The specifier describes symptoms within a depressive episode. Postpartum psychosis is a different usage: a name for an illness that one narrative review links to the bipolar spectrum and treats as an emergency (Bergink et al., 2016), and that another relates in most cases to bipolar disorder (Perry et al., 2021). The line between the two is not sharp, because postpartum psychosis has no universally accepted definition and its nosological status remains contentious (Perry et al., 2021). The Postpartum Psychosis entry below covers that condition.

Cognitive Implications

No study cited on this page reports cognitive findings specific to postpartum depression with psychotic features. The one imaging study of women at risk of postpartum psychosis is described in the Postpartum Psychosis entry.

Data-Derived Subtypes and Trajectories

Statistical classification of symptom data has produced subtypes that neither manual contains. A latent class analysis of pooled consortium data (17 912 unique records from 19 institutions in seven countries) identified three classes with distinct phenotypes. The most severe class had more suicidal ideation and more obstetric complications than the other two, and its depression had more often begun in pregnancy (67% against 34% in the second class) (Postpartum Depression: Action Towards Causes and Treatment (PACT) Consortium, 2015). The consortium presents the analysis as hypothesis-generating only: the data came from sites with differing protocols, most of the women were white, bipolarity was not assessed, and latent class analysis may overestimate the number of classes (Postpartum Depression: Action Towards Causes and Treatment (PACT) Consortium, 2015). A second analysis, of 663 women from seven of the consortium's sites, extracted three symptom dimensions (depressed mood, anxiety and anhedonia) and clustered the women into five subtypes that differed in symptom quality and in onset (Putnam et al., 2017).

Course has been classified in the same way. In a population-based birth cohort of 4866 mothers assessed at 4, 12, 24 and 36 months postpartum, about three-quarters (74.7%) kept low symptoms throughout, 12.6% began at a medium level and declined, 8.2% began low and rose, and 4.5% stayed high. A history of mood disorder and gestational diabetes both went with the high-persistent path, and the authors suggest that later screening is possibly warranted (Putnick et al., 2020). A systematic qualitative review of longitudinal studies published from 1985 to 2012 likewise found that most women recover while a subgroup follows a chronic course, although it judged the conceptual and methodological limitations of the research to preclude strong conclusions (Vliegen et al., 2014).

Severity and persistence together matter for children. In a prospective cohort of 9848 mothers and 8287 children in the UK, postnatal depression that was both persistent and severe carried the worst outcomes: child behaviour problems, low mathematics grades and depression at 18 years were all more likely than when the mother had not been depressed. The persistent severe group was small (n = 83), and the scale thresholds were not clinical diagnoses (Netsi et al., 2018). The Child Outcomes section gives the detail.

One proposed subtype is defined by mechanism instead of symptoms. In a double-blind hormone manipulation, 16 euthymic women, eight with a history of postpartum depression and eight without, had gonadal steroids (estradiol and progesterone) withdrawn. Mood symptoms followed in five of the eight women with a history (62.5%) and in none of the eight without, which the authors took to show differential sensitivity in a subgroup of women (Bloch et al., 2000). A narrative review reads this and related work as enough evidence for a hormone-sensitive phenotype, in which hormonal shifts trigger depression in a subgroup only. The same review notes that absolute hormone levels do not distinguish women with postpartum depression from those without, that the treatment studies are small and lack controls, and that the causes remain poorly understood (Schiller et al., 2015). The hormone-sensitive phenotype is therefore a proposed subtype that rests on small experiments.

None of these subtypings is part of the DSM-5-TR or the ICD-11.

Cognitive Implications

These studies classified women by symptoms, severity and timing. No study cited on this page reports cognitive measures by subtype or trajectory.

Postpartum Blues

The postpartum blues, also called the maternity blues or the baby blues, are a brief disturbance of mood in the days after delivery. A 1968 report described the Blues as transitory tearfulness in the early puerperium, commonplace on maternity wards and generally regarded as normal, and named the 3rd, 4th or 10th day (Pitt, 1968). A later narrative review gives the time course: the symptoms begin in the first week postpartum, peak between day 3 and day 5 and have largely diminished by days 10 to 12 (O'Hara & Wisner, 2014). The same review notes that no established criteria for the blues exist and that prevalence estimates run from 26% to 84% as the criteria differ (O'Hara & Wisner, 2014). A meta-analysis of 26 studies (5,667 postpartum women) pooled the prevalence at 39.0% (95% CI [32.3, 45.6]), across study estimates from 13.7% to 76.0% and with great heterogeneity (I² = 96.6%); the figure for Africa, 49.6%, was higher than for other regions (Rezaie-Keikhaie et al., 2020).

The blues are not evenly distributed among women. The blues were predicted, in a prospective cohort that tracked 182 women from pregnancy to the ninth postpartum week, by a personal or family history of depression, by stressful life events and by free and total estriol levels, and the authors concluded that the state belongs on the affective-disorder spectrum (O'Hara et al., 1991). A narrative review linked the blues to dysphoria in pregnancy, past depression and neuroticism, treated the blues and early postpartum elation as an apparent index of affective vulnerability, and found the biological research inconclusive (Henshaw, 2003).

Two statements about the blues and later depression have to be read together. In a prospective community cohort of 853 German women, the maternity blues predicted postpartum depression and anxiety disorders, with odds ratios of 3.8 and 3.9 for the two outcomes; the blues were assessed retrospectively, which the authors flag as a source of bias, and the sample was middle class and low-risk (Reck et al., 2009). In a meta-analysis of 84 studies from the 1990s, the maternity blues were one of 13 significant predictors of postpartum depression (mean effect sizes .25 to .31), ten of which had moderate effect sizes (Beck, 2001). Yet the review that sets out the time course also judges postpartum depression to be mostly independent of the blues (O'Hara & Wisner, 2014). The statements are compatible: a predictor can raise a woman's odds without accounting for most cases.

What separates the blues from depression is time. The blues begin in the first week, peak within days and have largely diminished by days 10 to 12 (O'Hara & Wisner, 2014); a depressive episode persists. The section Related Conditions and Normal Adjustment sets the blues beside other states that are generally regarded as normal.

Cognitive Implications

A 1968 report on emotional and cognitive disturbance in the early puerperium treated pregnancy and the months after delivery as what is generally regarded as a maturational crisis, identified the first month after birth as the point at which overt mental illness peaks, and expected that the cause would turn out to involve psychosocial factors acting together with endocrine ones (Kane et al., 1968).

Postpartum Psychosis

Postpartum psychosis is a rare psychiatric emergency and a boundary presentation, not a severe form of postpartum depression. A systematic review of 6 population-based studies reported incidence estimates of 0.89 to 2.6 in 1000 women, from data that were not pooled and that used differing definitions and assessments (VanderKruik et al., 2017). A narrative review cites 0.25 to 0.6 per 1,000 births for first-lifetime onset in register studies of psychiatric admissions, and describes the condition as an emergency (Bergink et al., 2016).

Onset is early. A record-linkage study covering 54 087 births found the risk of admission for psychosis highest in the first 30 days, raised after a first baby, and greater with a history of manic-depressive illness than with a history of schizophrenia; its authors concluded that puerperal psychoses are probably manic-depressive illness (Kendell et al., 1987). In a Danish register cohort, first-time mothers at 10 to 19 days postpartum had 7.31 times the risk of psychiatric admission of women 11 to 12 months after delivery (95% CI, 5.44-9.81), a figure that counts admission for any mental disorder (Munk-Olsen et al., 2006).

Most of the evidence places the condition on the bipolar disorder spectrum. One narrative review concludes that most cases are related to bipolar disorder, with childbirth as the trigger, and that bipolar disorder, primiparity and a family history of mood disorder predict it whereas absolute steroid hormone levels do not (Perry et al., 2021). Among 120,378 women in a Danish register cohort, those whose psychiatric onset fell within 0 to 14 days of delivery had a relative risk of 4.26 (95% confidence interval 3.11 to 5.85) of later converting to bipolar affective disorder, and the authors regard such early onset as a possible marker of bipolarity (Munk-Olsen et al., 2012). Perinatal episodes are also common across the mood disorders generally: among 1785 parous women the perinatal risk was approximately 50% in bipolar I disorder and approximately 40% in bipolar II disorder and in recurrent major depression (Di Florio et al., 2013).

Some women have psychosis only after childbirth. In a clinic cohort of 70 high-risk pregnant women, none of those with that history who took prophylaxis from delivery relapsed, against 44.4% of those who declined it, although the authors acknowledge that the timing of prophylaxis remains controversial (Bergink et al., 2012). Across 37 articles and 5,700 deliveries, a meta-analysis put overall postpartum relapse at 35% (95% CI 29 to 41), about a third, with severe episodes more frequent after a history of postpartum psychosis (29%) than in bipolar disorder (17%) (Wesseloo et al., 2016).

Treatment studies are sparse; one review counted fewer than 30 publications (Bergink et al., 2016). In 64 women consecutively admitted with first-onset postpartum psychosis and treated by a four-step algorithm, 98.4% reached remission and no patient needed electroconvulsive therapy, the last step; the authors call the algorithm possibly effective and prefer lithium for maintenance (Bergink et al., 2015). Sleep loss is a suspected trigger: one literature review argues that it is possibly the final common pathway (Sharma & Mazmanian, 2003), and another notes that sleep loss is a frequently cited precipitant of mania but that studies of it as a trigger of postpartum psychosis are scarce (Perry et al., 2021).

The emergency status reflects risk to the mother and, less often, the infant. A review of over 4000 published cases and a personal series of 321 found suicide rare in the acute episode and more frequent later, and filicide more frequent in psychoses with depression (4.5%) than in those without (less than 1%); the review notes that its suicide figures report lifetime risks rather than outcomes of the acute episode (Brockington, 2017).

Classification is unsettled. A narrative review states that no universally accepted definition exists and that the nosological status of the condition remains contentious: under DSM-5 these episodes are usually classified as severe mood episodes carrying the peripartum specifier when onset falls in pregnancy or within 4 weeks of delivery, and the older ICD-10 allowed a puerperal category only for onset within the first 6 weeks and only when the episode could not be classified elsewhere (Perry et al., 2021). In the current manuals the relevant labels are the peripartum specifier (American Psychiatric Association, 2022) and the ICD-11 grouping for disorders associated with pregnancy, childbirth or the puerperium (World Health Organization, 2024). The current view is that postpartum psychosis is mostly a bipolar-spectrum illness triggered by childbirth; the status of the isolated form, confined to the postpartum period, remains open.

Cognitive Implications

One imaging study bears on cognition. In a functional magnetic resonance imaging case-control study, twenty-four women at risk of postpartum psychosis (11 of whom developed an episode and 13 of whom remained well) and 20 healthy postpartum women performed an n-back working memory task and a fearful faces task. The at-risk women showed altered activation and connectivity on both tasks, including differences in the connectivity of the right dorsolateral prefrontal cortex and of the left amygdala, while task performance did not differ significantly between the groups; the authors call the evidence preliminary (Kowalczyk et al., 2021). It is the only cognitive study of the condition on this page.

Bipolar Postpartum Depression

Bipolar postpartum depression is a depressive episode after childbirth in a woman with bipolar disorder. It is a boundary presentation because the episode can resemble unipolar postpartum depression. A narrative review drawing on 3 databases found a bipolar diagnosis in 21.4-54% of women with postpartum depression, a range that varied with the population studied, and counted younger age at onset, a first depression arising after childbirth, onset immediately after delivery and a family history of bipolar disorder among the features pointing to bipolarity (Sharma et al., 2017). In the case series of screen-positive mothers described under Peripartum Onset, 22.6% of primary diagnoses were bipolar disorders and 68.5% unipolar, subject to that study's incomplete assessment (Wisner et al., 2013).

Timing is itself a signal. Over a 15-year follow-up of the Danish register cohort, a later bipolar diagnosis came to 14% of women whose psychiatric contact came soon after childbirth, compared with 4% of women whose contact had no link to giving birth (Munk-Olsen et al., 2012). Women with established bipolar disorder relapse often after delivery. In a meta-analysis, postpartum relapse occurred in 66% of those who were medication-free against 23% of those taking prophylaxis, and prophylactic medication in pregnancy appeared highly protective (Wesseloo et al., 2016). In the study of parous women with mood disorders, episodes in bipolar I disorder were more frequent after delivery than in pregnancy (Di Florio et al., 2013).

Management is weighed case by case. Because the perinatal period raises the risk of severe mental disorder, one review concludes that care has to weigh risks against benefits woman by woman, although few studies exist to guide that judgement (Jones et al., 2014); for bipolar postpartum depression, a separate review found no controlled treatment data (Sharma et al., 2017). The practical consequence drawn by the case series and by that review is that bipolar disorder needs to be told apart from unipolar depression, and identified promptly, when a woman presents with depression after childbirth (Wisner et al., 2013; Sharma et al., 2017).

Cognitive Implications

No study cited on this page compares cognition in bipolar and unipolar postpartum depression.

Perinatal Anxiety and Obsessive-Compulsive Presentations

Anxiety and obsessive-compulsive presentations share the perinatal period with depression and often accompany it. Perinatal anxiety is common. A systematic review and meta-analysis of 102 studies (221 974 women in 34 countries) estimated that 15.2% (95% CI 9.0–21.4) of women had any anxiety disorder antenatally and 9.9% (95% CI 6.1–13.8) postnatally, with higher rates in low- and middle-income countries than in high-income ones; the studies used diverse measures, cut-offs and time periods (Dennis et al., 2017).

Prevalence of obsessive-compulsive disorder rose in steps in a meta-analysis of 29 studies that used structured diagnostic interviews: a mean of 1.08% among women in the general population, 2.07% in pregnancy and 2.43% postpartum. The aggregate risk ratio of 1.79 came from an analysis the authors call exploratory (Russell et al., 2013).

Anxiety and depression co-occur often in this period, as the meta-analysis described under With Anxious Distress found (Falah-Hassani et al., 2017), and the standard depression screen registers both. Items 3, 4 and 5 of the Edinburgh Postnatal Depression Scale form an anxiety subscale, the EPDS-3A (range 0-9), which a diagnostic accuracy study of 238 women and 218 men at 6 weeks postpartum judged usable for screening for probable anxiety disorder, at a cut-off of 6 or more in women (Matthey, 2008). Whether the two conditions are biologically separable is an early question: a review argues that anxiety and depression arising after childbirth carry neural signatures specific to that period, distinct from those seen when either condition appears at another time, and that the brain systems the two disorders involve overlap, while conceding that the neural correlates remain poorly understood (Pawluski et al., 2017). The Comorbidity section covers co-occurring disorders in full.

Cognitive Implications

The characteristic cognitive phenomenon here is the unwanted intrusive thought. In a prospective cohort of one hundred women recruited in pregnancy and assessed three times, including at 4 and 12 weeks postpartum, intrusive thoughts of harm coming to the infant were normative. Thoughts of accidental harm were more frequent and took up more time than thoughts of intentional harm, which were the more distressing; parenting stress went with more thoughts of intentional harm and social support with fewer, and the study found little evidence that intrusive thoughts were linked to aggressive parenting (Fairbrother & Woody, 2008). The section Intrusive Thoughts of Harm to the Infant takes the topic further.

Paternal Postpartum Depression

Fathers undergo none of the physiology of the puerperium, so in men the term means depression arising during a partner's pregnancy or in the months after the birth. On symptom measures it is not rare. A meta-analysis of 43 studies with 28 004 participants estimated the rate of paternal prenatal and postpartum depression at 10.4% (95% CI, 8.5%-12.7%), rising to 25.6% (95% CI, 17.3%-36.1%) at 3 to 6 months postpartum, and found a positive correlation with maternal depression (r = 0.308; 95% CI, 0.228-0.384); the rates were substantially heterogeneous (Paulson & Bazemore, 2010). A later meta-analysis of 47 studies (20,728 fathers) described paternal depression as relatively common both before and after the birth, at 9.76% prenatally and 8.75% postpartum (Rao et al., 2020).

Severe illness shows a different picture. In the Danish register cohort of 2,357,942 persons, fatherhood brought no increase in psychiatric admission or outpatient contact, unlike motherhood; severe disorders in the first 3 months after childbirth ran at 1.03 per 1000 births in mothers and 0.37 per 1000 births in fathers, and the authors concluded that, for severe disorders serious enough to need hospital admission or outpatient contact, first-time mothers carried a raised risk for several months after the birth, whereas fathers showed no such excess (Munk-Olsen et al., 2006). A commentary on the diagnostic definition draws the same contrast, an elevated risk in mothers and no increase in fathers (Wisner et al., 2010).

Measurement in fathers relies on the Edinburgh Postnatal Depression Scale, with a threshold of its own. In a cohort of couples recruited antenatally (208 fathers and 230 mothers) and assessed six weeks postpartum, the scale was valid for fathers, who endorsed seven of the ten items less than mothers did, most markedly the crying item, and whose optimum cut-off for distress caseness was 5/6 against 7/8 for mothers (Matthey et al., 2001). A validation against a structured clinical interview in 192 UK fathers, 7 weeks after the birth, found a higher cut-off optimal for depression, a score of more than 10 (sensitivity 89.5%, specificity 78.2%, AUC = 0.916), falling to more than 8 when generalised anxiety disorder was included; participation was modest and the interviewed fathers had been selected for high scores (Edmondson et al., 2010). The two studies agree that the scale works in fathers and differ on where the threshold lies.

Paternal depression matters for children. In a prospective population cohort that included 8431 fathers, depression in fathers 8 weeks after the birth predicted emotional and behavioural problems in their children at age 3.5 years (adjusted OR 2.09, 95% CI 1.42-3.08), and conduct problems in boys in particular (OR 2.66, 95% CI 1.67-4.25). The associations held after maternal depression and later paternal depression were controlled, though the child outcomes were reported by mothers (Ramchandani et al., 2005). A systematic review of twenty-one prospective studies likewise linked paternal depression to emotional and behavioural problems in offspring, with heterogeneity that hindered strong conclusions (Sweeney & MacBeth, 2016). A meta-analysis of 28 studies found small effects on parenting, with less positive and more negative parenting among depressed fathers and effect sizes comparable to those in mothers (Wilson & Durbin, 2010).

What fathers say about the experience is known from qualitative work. Interviews with 19 UK fathers, 5-10 months postpartum, found men who were distressed yet questioned the legitimacy of their own stress and their entitlement to support, and who gave priority to protecting the partnership. The authors conclude that men may hesitate to seek help, from a sample that lacked ethnic and socioeconomic diversity and did not reach data saturation (Darwin et al., 2017).

An expert narrative review argues that fathers' risk of perinatal depression and anxiety is elevated and that tailored screening of fathers early in the perinatal period is essential (Fisher et al., 2021). The two bodies of evidence need not conflict: the prevalence figures come from symptom scales, and the register finding concerns illness severe enough for admission or outpatient psychiatric contact. Whether paternal postpartum depression is a perinatal condition in its own right or depression that happens to occur in new fathers remains open.

Cognitive Implications

The studies cited here report parenting behaviour and child outcomes for depressed fathers. None is a laboratory study of fathers' attention, memory or processing of infant cues.

Depression in Adoptive and Other Non-Birthing Parents

Parents who take on an infant without giving birth offer a natural test of hormonal against psychosocial accounts. A cross-sectional survey matched 147 adoptive mothers with 147 postpartum women. Depressive symptoms did not differ significantly between the groups, and the proportions scoring above 13 on the Edinburgh Postnatal Depression Scale, 7.5% and 8.8%, did not differ significantly either; the adoptive mothers reported greater well-being and less social anxiety, panic and traumatic intrusion (Mott et al., 2011). Sleep deprivation, a past psychological disorder and distress about infertility predicted more depressive symptoms, and better marital adjustment predicted fewer (Mott et al., 2011).

The comparison shows that symptom levels like those of postpartum women can occur without pregnancy or childbirth. It cannot show that hormonal change is unimportant. The survey was a single self-report assessment of women from one private agency, with a 19.7% response; the groups differed in age, income and education; and agency screening possibly led to under-reporting (Mott et al., 2011). Nor does a group average bear on a hormone-sensitive subgroup of the kind described under Data-Derived Subtypes and Trajectories.

Two further sources concern sexual and gender minority parents, some of whom gave birth. A cross-sectional survey of 49 lesbian and bisexual perinatal women found that the biological mothers scored higher on the Edinburgh scale than a previously published sample of heterosexual women, which the authors took to mean that perinatal depression is not rarer, and is possibly commoner, in this group (Ross et al., 2007). A qualitative meta-aggregation of 26 studies with 1199 LGBTQ2S+ childbearing participants tied perinatal distress to stigma, exclusion and isolation, noted gaps in providers' knowledge, and held systems-level inequities largely responsible (Kirubarajan et al., 2022).

Cognitive Implications

No study cited on this page reports cognitive findings for adoptive or other non-birthing parents with depression.

Depression After Perinatal Loss

Depression after the death of a baby sits on the boundary with grief. A systematic review and meta-analysis of 29 studies from 17 countries compared 31,072 participants who had experienced a perinatal loss with 1,261,517 controls. Loss was followed by a higher risk of depressive disorder (RR = 2.14, 95% CI = 1.73–2.66, k = 22) and of anxiety disorder (RR = 1.75, 95% CI = 1.27–2.42, k = 9), and by higher depression and anxiety scores (SMD = 0.34 and SMD = 0.35); post-traumatic stress did not differ significantly, on only three studies (Herbert et al., 2022).

Survey data point the same way. In a United States surveillance survey of 32,659 women with a recent birth across 31 sites, postpartum depressive symptoms were reported by 48.7% of women whose infant had died against 13.0% of those whose infant was alive; the measure was self-reported symptoms on an adapted Patient Health Questionnaire-2, not a diagnosis (Bauman et al., 2020).

Neither study separates a depressive episode from grief. Telling the two apart after a loss is a clinical judgement that the studies cited here do not settle.

Cognitive Implications

No study cited on this page reports cognitive findings for depression after perinatal loss.

Prevalence, Course and Recovery

Estimates of how many mothers become depressed after birth, how long the depression lasts and whether it returns depend heavily on how the question is asked. This section sets out the pooled prevalence figures and the reasons they differ, the long-running dispute over whether childbirth itself raises the risk of depression, and the evidence on course, recovery and recurrence.

How Common It Is

An early meta-analysis put the prevalence of non-psychotic postpartum depression at about 13%, and it found higher figures for self-report measures than for other measures and for longer postpartum windows than for shorter ones (O'Hara & Swain, 1996). A later systematic review judged the true levels of perinatal depression uncertain: 28 of 109 reviewed articles met its inclusion criteria, the confidence intervals were wide, and the review was limited to English-language studies from developed countries; the authors called for studies with larger and more representative samples (Gavin et al., 2005).

Meta-analyses since then have pooled far more data and have landed on different numbers. Across 96 studies and 140 usable estimates, perinatal depression had a pooled prevalence of 11.9% (95% CI 11.4-12.5) (Woody et al., 2017). A review of 58 articles covering 37,294 women with no prior depression reported a prevalence of 17% (95% CI 0.15-0.20) (Shorey et al., 2018). A meta-analysis of 291 studies in 56 countries gave 17.7% (95% CI 16.6-18.8%), above the 13% of the earlier estimate (Hahn-Holbrook et al., 2018). The largest synthesis, 565 studies of 1,236,365 women in 80 countries, found 17.22% (95% CI 16.00-18.51), with heterogeneity of I² = 99.70% (Wang et al., 2021b).

Why the Figures Differ

Four sources of spread recur across the studies cited here.

- Measurement. Symptom scales gave higher prevalence than diagnostic instruments (OR 1.6) (Woody et al., 2017). Self-report scales probably overestimate relative to interviews, and a lower Edinburgh Postnatal Depression Scale cut-off raised prevalence (Hahn-Holbrook et al., 2018). Not every synthesis agrees: the review of healthy mothers found no significant effect of diagnostic tool or screening time point (Shorey et al., 2018), and the 565-study analysis found diagnostic scale did not explain its heterogeneity (Wang et al., 2021b). - Timing. Longer postpartum windows gave higher prevalence (O'Hara & Swain, 1996), while later assessment gave slightly lower prevalence in the 291-study analysis (Hahn-Holbrook et al., 2018). Timing differences in the 565-study analysis were not significant (p = 0.08) (Wang et al., 2021b). - Country mix. Low- and middle-income countries had higher perinatal prevalence than high-income ones (OR 1.8) (Woody et al., 2017). Developing countries averaged 19.99% against 14.85% for developed ones, with Denmark at 6.48% and Southern Africa at 39.96% (Wang et al., 2021b). National figures ran from 3% in Singapore to 38% in Chile, and income inequality (GINI index), maternal mortality, infant mortality and long working hours for women were each linked to the national variation (Hahn-Holbrook et al., 2018). A review of 143 studies in 40 countries found prevalence from almost 0% to almost 60% and judged the familiar 10-15% figure probably unrepresentative of the world (Halbreich & Karkun, 2006). Regional differences also appear among healthy mothers, with 26% in the Middle East against 8% in Europe (Shorey et al., 2018). - Study size and heterogeneity. Larger studies gave lower prevalence than small ones in two syntheses (19.44% for small against 12.97% for large studies) (Wang et al., 2021b; Hahn-Holbrook et al., 2018), and the heterogeneity was very high in both.

The figures therefore agree on a broad order of magnitude, roughly one mother in six to one in eight, with large and probably real differences between countries. The exact global number depends on instrument, cut-off and the countries sampled.

Antenatal and Postnatal Levels

Depression is not confined to the months after birth. In a Avon cohort, 13.5% of women were above the Edinburgh Postnatal Depression Scale threshold at 32 weeks of pregnancy and 9.1% at 8 weeks postpartum, and scores were higher in late pregnancy than after birth. The authors concluded that postnatal prevalence did not exceed antenatal prevalence (Evans et al., 2001).

Does Childbirth Raise the Risk of Depression?

Controlled studies gave the first answer, and it was surprising. A prospective study of childbearing women and matched controls found no excess of minor or major depression diagnoses among the childbearing women, although their depressive symptoms and social adjustment problems were greater in late pregnancy and the early puerperium (O'Hara et al., 1990). A matched two-stage screening study of 232 postnatal women found six-month point prevalence of 9.1% against 8.2% in controls, and period prevalence of 13.8% against 13.4%, neither difference significant. Onset within five weeks of childbirth was threefold more frequent, and the authors offered long duration and childbirth stress only tentatively as explanations (Cox et al., 1993).

Population-level data point the other way for some outcomes. Past-year major depression, in a US national cross-sectional survey, affected 9.3% of postpartum women and 8.1% of non-pregnant women, an odds ratio of 1.59 (95% confidence interval 1.15 to 2.20); pregnant women showed no such elevation, and the authors called the postpartum increase possible (Vesga-López et al., 2008). Danish registers showed sharply raised psychiatric admission among first-time mothers in the early postpartum weeks (RR 7.31, 95% CI 5.44-9.81, at 10 to 19 days), whereas fathers showed no increase (Munk-Olsen et al., 2006). An older record-linkage study of 54,087 births found the risk of admission with puerperal psychosis highest in the first 30 days, and it probably reflected manic-depressive illness rather than schizophrenia (Kendell et al., 1987).

These findings can be read together. The clearest excess after birth is for severe illness needing admission, and for early onset. For depressive diagnoses overall the excess is smaller or absent in the controlled studies. The question is not fully settled, because the studies differ in design (matched cohorts, a cross-sectional survey, national registers), in outcome and in country.

Course, Recovery and Recurrence

A systematic review of longitudinal studies from 1985 to 2012 found that most women recover, with a chronic course in a subgroup, though conceptual and methodological limitations prevented strong conclusions (Vliegen et al., 2014). In a cohort of 4,866 mothers assessed at 4, 12, 24 and 36 months, four symptom trajectories emerged: low-stable (74.7%), low-increasing (8.2%), medium-decreasing (12.6%) and high-persistent (4.5%). A history of mood disorder and gestational diabetes were linked to the high-persistent group, and the authors suggested that later screening may be warranted (Putnick et al., 2020).

Persistence matters for children. In the ALSPAC cohort, persistent severe postnatal depression was linked to child behaviour problems, lower mathematics grades and depression at 18 years. The authors treated persistent severe depression as a priority for treatment, although the severe persistent group was small and attrition was high (Netsi et al., 2018).

Recurrence follows two patterns. In a five-year follow-up of 55 depressed first-time mothers, women whose first-ever episode was postnatal were at risk of later postnatal episodes but not of non-postpartum ones, which suggested postnatal depression may be a specific nosological concept (Cooper & Murray, 1995). Danish register data on 457,317 first-time mothers showed that women with an earlier postpartum affective disorder had a far higher rate of recurrence (RR 26.9 to 46.4), while 27.9% were still being treated at one year and 5.4% at four years. The data identify treated cases and probably miss milder ones (Rasmussen et al., 2017).

Bipolar disorder changes the picture. A meta-analysis of 37 articles found that about a third of women with bipolar disorder or earlier postpartum psychosis relapsed after delivery (35%, 95% CI 29-41), and relapse in bipolar disorder was 66% without medication against 23% with prophylaxis in pregnancy (Wesseloo et al., 2016). In Danish registers, a psychiatric contact in the first 14 days postpartum carried a relative risk of 4.26 for a later bipolar diagnosis, which the authors read as a possible marker of bipolarity (Munk-Olsen et al., 2012).

Risk Factors

Risk factors for postpartum depression are characteristics measured before or soon after birth that go with a later depressive episode or higher symptom scores. The evidence comes mostly from meta-analyses of observational studies, and each subsection below gives the design and size behind its predictors. Prediction is not explanation, a limit taken up in the last subsection.

Psychiatric History and Antenatal Symptoms

Mood and anxiety during pregnancy, together with earlier depression, head the lists in the largest syntheses. A meta-analysis of 84 studies from the 1990s found 13 predictors of postpartum depression, most of moderate size (Beck, 2001). Prenatal depression (.44 to .46), prenatal anxiety (.41 to .45) and a history of previous depression (.38 to .39) were among them, while self-esteem (.45 to .47) and childcare stress (.45 to .46) had ranges at least as high as prenatal depression. An earlier meta-analysis likewise found past psychopathology and prenatal psychological disturbance to be strong predictors (O'Hara & Swain, 1996). Its findings generally mirrored those of earlier qualitative reviews.

A later systematic review drew on those two major meta-analyses and on nearly 10,000 additional subjects, on top of over 14,000 in the meta-analyses. According to the review, five antenatal factors carried the strongest predictive weight: antenatal depression, antenatal anxiety, a prior history of depression, stressful life events and poor social support (Robertson et al., 2004). The review noted that lower socioeconomic status and teenage mothers were understudied. An umbrella review of 21 reviews published between 1996 and 2016 identified 25 statistically significant risk factors and ranked prenatal depression and current abuse among the strongest (Hutchens & Kearney, 2020). Two further factors were inconclusive, and the reviews differed too much in method for the results to be aggregated in a standard way.

Anxiety in pregnancy has a prospective test of its own. A community cohort of 8,323 women in England assessed anxiety and depression at 18 and 32 weeks of gestation and again at 8 weeks and 8 months after birth. Antenatal anxiety predicted postnatal depression (odds ratio 3.22, p < 0.001) with antenatal depression controlled (Heron et al., 2004). Symptoms were self-reported and attrition was selective, so the estimate comes from those who stayed in the study.

The same predictors have been studied one step earlier, as risks for antenatal depression and anxiety. A systematic review of 97 papers from 2003 to 2015, without a meta-analysis, linked a personal history of mental illness, a lack of partner or social support, abuse or domestic violence, adverse life events and unplanned pregnancy to antenatal depression and anxiety (Biaggi et al., 2016). Age and parity were inconsistent predictors, and the review excluded high-risk populations and papers not in English.

Stress, Support and Adversity

Stress and relationships form the second cluster. Across its 84 studies, the meta-analysis found moderate predictive strength for life stress (.38 to .40), the marital relationship (.38 to .39) and social support (.36 to .41), but only a small one for socioeconomic status (.19 to .22) (Beck, 2001). The systematic review of antenatal predictors likewise placed stressful life events and low social support among its five strongest factors (Robertson et al., 2004). In the umbrella review, high life stress, lack of social support, past abuse and marital dissatisfaction were also significant predictors (Hutchens & Kearney, 2020).

Domestic violence has its own meta-analysis, covering 67 papers and perinatal depression, anxiety and post-traumatic stress. Partner violence during pregnancy carried an odds ratio of 3.1 for postnatal depression, and lifetime partner violence an odds ratio of 2.9 (Howard et al., 2013). The authors advised stronger evidence, since the studies were heterogeneous, of variable quality and unclear about the direction of causation.

Population surveillance data show the same gradients. A cross-sectional analysis of the 2018 Pregnancy Risk Assessment Monitoring System covered 32,659 recent-birth women at 31 sites. Postpartum depressive symptoms, measured with an adapted Patient Health Questionnaire-2, were reported by 13.2% (Bauman et al., 2020). They were more common with intimate partner violence (33.1% against 12.5%), at age 19 or under (22.2% against 10.8% at 35 or over) and under Medicaid coverage (17.2% against 10.1% under private insurance). Parity made no difference: first and later births each had 13.2%. The symptoms were self-reported rather than diagnosed, and recall and social desirability biases may apply.

Obstetric and Infant Factors

Several pregnancy and neonatal conditions have been examined, with uneven evidence. Across 26 articles in a systematic review, depression in the early postpartum period was more common among women whose babies were born preterm than among those who delivered at term, and mothers of very-low-birth-weight infants showed the excess across the whole first year (Vigod et al., 2010). Rates in some samples were as high as 40%. Most of the studies ignored confounding by depression in pregnancy. The surveillance survey found a higher rate after preterm than term birth (17.1% against 12.8%) (Bauman et al., 2020).

Gestational diabetes was linked to postpartum depressive symptoms in a meta-analysis of 10 cohort studies with 2,000,002 women: the pooled relative risk was 1.32 (95% CI 1.09 to 1.60), with high heterogeneity (I² = 69.1%) (Arafa & Dong, 2019). Design mattered. Prospective cohorts gave a relative risk of 1.95 and retrospective cohorts 1.16 (P for interaction = 0.003), and the authors called for well-designed prospective studies to confirm the raised risk. A systematic review and meta-analysis of preeclampsia, with 13 studies and 8 in the meta-analysis, found that women with preeclampsia had higher postpartum depressive severity than women without it. Eight of the 13 studies reported an association (Caropreso et al., 2020), and the studies were highly heterogeneous, so the authors advised caution.

How the birth was experienced also relates to mood. In a systematic review of 15 studies from 2000 to 2015, eleven found that a negative birth experience went with postnatal depression (Bell & Andersson, 2016). The studies used heterogeneous instruments for birth experience and carried a potential for bias.

Infant characteristics belong here too. In a longitudinal path analysis of 55 married women at 3 months postpartum, a difficult infant temperament predicted depression directly and also through lower parenting self-efficacy (Cutrona & Troutman, 1986). Infant temperament was a moderate predictor in the later meta-analysis (.33 to .34) (Beck, 2001). At the extreme, the surveillance survey found symptoms in 48.7% of women whose infant had died, against 13.0% where the infant was alive (Bauman et al., 2020).

What Risk Factors Cannot Show

A predictor is an association, not a cause. Nearly all of the findings above are correlations or odds ratios between a factor and a symptom-scale score, and the domestic violence meta-analysis itself noted that causal direction was unclear (Howard et al., 2013). Gestational diabetes shows the point on a small scale: the effect was larger in prospective than in retrospective cohorts, and the authors still asked for prospective confirmation (Arafa & Dong, 2019).

The factors are also not specific to the postpartum period. According to a review in a Lancet series, what puts women at risk of perinatal mental disorders largely mirrors what puts people at risk of the same disorders outside the perinatal period (Howard et al., 2014). That is consistent with the prominence of depression history and prenatal symptoms above, and it leaves open whether the period adds risk of its own. The methods that separate cause from consequence, and what they have found, are covered under Testing Whether Cognition Is Cause or Consequence.

The Cognitive Model in the Perinatal Period

Beck's General Formulation

Beck's cognitive model holds that early adverse experience, such as trauma, can lay down dysfunctional beliefs, that later stressors congruent with those beliefs can trigger depression, and that the model has since incorporated information-processing biases, cognitive distortions and automatic thoughts (Beck, 2008). The transition to parenthood offers a natural test of this sequence, and the perinatal studies discussed below measure the beliefs during pregnancy, before the stressor of the birth and early infancy arrives. The same review sketched possible genetic and neurochemical pathways that interact with cognition, and associated a hyperactive amygdala and hypoactive prefrontal regions with depression (Beck, 2008). A later unified version recasts depression as an energy-conserving adaptation to the perceived loss of a vital resource, in which negative processing and stress reactivity feed depressogenic beliefs and those beliefs feed appraisals of loss (Beck & Bredemeier, 2016). The authors suggest that severe depression may be an evolutionary program that has become anachronistic (Beck & Bredemeier, 2016). A neural reading of the model attributes the negative biases in attention, memory and rumination to heightened limbic drive together with weakened top-down cognitive control, although the review noted that the mechanisms differ across elements of the model and that earlier neural findings had not been integrated (Disner et al., 2011). The full account of the model belongs to the Major Depressive Disorder page. For this page, the model counts as supported: it accounts for cognitive vulnerability that operates after childbirth much as it does at other times, though the prospective evidence for that rests mostly on self-report questionnaires (Evans et al., 2005), and its most recent statement is the unified energy-conservation version (Beck & Bredemeier, 2016).

Predictions at the Transition to Parenthood

Pregnancy offers a natural test, because vulnerability can be measured before any stressor of birth or infancy arrives. In ALSPAC, a prospective cohort of 12,003 pregnant women in Avon who completed postal questionnaires, a six-item negative self-schema score predicted later onset of depression among the 8,540 women who were not depressed at baseline (Evans et al., 2005). Onset was 13.4% in the highest tertile of schema score and 4.8% in the lowest, and the association held after adjustment for other measured risks (adjusted OR = 1.6, 95% CI 1.27 to 2.02), including baseline Edinburgh Postnatal Depression Scale score (Evans et al., 2005). It remained present from eight weeks to 33 months postpartum, and the authors read it as support for a key element of Beck's theory (Evans et al., 2005). Its limits are stated: depression was assessed by a brief self-report screen without diagnostic interviews, the schema measure was drawn from a personality instrument and taken once, personality itself was unmeasured, and the finding may not generalise beyond pregnancy (Evans et al., 2005).

A smaller prospective study followed 94 healthy first-time mothers from the third trimester. Self-devaluation and a lack of specificity in autobiographical memory retrieval predicted depressive symptoms eight weeks after birth, but not low mood at two weeks, which the authors linked to a larger role for biological factors in the earliest weeks (Hipwell et al., 2004). The findings held after controlling for neuroticism, education and dysphoria (Hipwell et al., 2004).

Attributional style gave a more partial result. In 170 women recruited in the second trimester and followed to three months postpartum, attributional style predicted postpartum depression alongside prepartum depression, delivery stress and stressful life events (O'Hara et al., 1982). In first-time mothers followed from pregnancy, attributional style measured with the Attributional Style Questionnaire predicted postpartum depression among women not depressed in pregnancy, but not among those who already were, and the author concluded that modifications to the model may be indicated (Cutrona, 1983). Possible problems in measuring attributions were noted as a caveat (Cutrona, 1983).

Content specific to motherhood has also been examined. In two cross-sectional online studies of 136 and 288 first-time mothers in the United States, scores on the Attitudes Toward Motherhood Scale, which taps beliefs about others' judgments, maternal responsibility and idealisation of the maternal role, were positively related to depressive and anxiety symptoms (β = 0.43 for each) (Sockol et al., 2014). The association with depression remained beyond general dysfunctional attitudes (β = 0.24), and beyond social support and marital satisfaction as well (β = 0.15) (Sockol et al., 2014). The authors suggested a possible perinatal-specific risk factor, while noting that the design cannot establish cause, that the samples were small, homogeneous and prone to dropout, and that longitudinal replication is needed (Sockol et al., 2014).

A further cognitive route runs through parenting self-efficacy. In 55 married women assessed to three months postpartum, path analysis showed infant difficulty acting on depression directly and through self-efficacy, while social support acted on depression mainly indirectly, again through self-efficacy (Cutrona & Troutman, 1986).

Is the Vulnerability Specific to Postpartum Episodes?

A cross-sectional trait comparison addressed this directly. It compared 143 women with postnatal depression, 131 parous women with depression unrelated to childbirth and 173 controls on neuroticism, dysfunctional attitudes and self-esteem, among other measures (Jones et al., 2010). Women with postnatal depression scored higher than controls on neuroticism and dysfunctional attitudes and lower on self-esteem, yet did not differ from the non-postnatal group on any measure, and the authors found no evidence that these traits confer postnatal-specific risk (Jones et al., 2010). The study was cross-sectional, its cognitive and personality measures were limited, refusal was high, and antenatal onset was not examined (Jones et al., 2010).

Standing

The model is best regarded as supported for vulnerability that applies after childbirth as at other times, with some motherhood-specific content on top. The prospective evidence (Evans et al., 2005; Hipwell et al., 2004; O'Hara et al., 1982; Cutrona, 1983) is consistent with antenatal cognitive vulnerability preceding postpartum symptoms, but it relies largely on self-report, and the attributional findings are only partial. The motherhood-specific attitudes finding is cross-sectional, and whether a vulnerability unique to the postpartum period exists remains open, since the one direct comparison found none. Experimental tests of cause are few; the section on Testing Whether Cognition Is Cause or Consequence takes up that question.

Stress, Interpersonal and Evolutionary Explanations

Five families of explanation recur in the literature on depression after childbirth: stress acting on prior vulnerability, interpersonal disruption, depressive thinking styles, sensitivity to reproductive hormones, and evolutionary signalling or mismatch. They differ sharply in standing. The first rests on prospective prediction, the second on treatment trials, the third on mixed perinatal evidence, and the evolutionary accounts remain proposed and contested.

Stress, Vulnerability and Social Support

The stress account holds that depression after birth arises when a woman's earlier vulnerability meets the demands of childbearing and too little support. Its evidence is prospective prediction, not experiment. In a longitudinal study of 170 women first assessed in the second trimester, depression during pregnancy, attributional style, stress around delivery and stressful life events each predicted depression at three months postpartum. Menstrual-problem history, parity, education and income did not (O'Hara et al., 1982).

Drawing on two large meta-analyses of over 14,000 participants and later studies of almost 10,000 more, one systematic review ranked five antenatal factors highest: antenatal depression, antenatal anxiety, a previous depressive history, stressful life events and poor social support. The authors flagged gaps, including understudied teenage mothers and women of lower socioeconomic status (Robertson et al., 2004).

Support appears to act partly through the mother's confidence in her own parenting. In a path analysis of 55 married women followed to three months postpartum, an infant with a difficult temperament predicted depression directly and also through lower parenting self-efficacy. Social support influenced depression mainly indirectly, through self-efficacy (Cutrona & Troutman, 1986). The account is therefore the study of how psychological stress accumulates on a vulnerable mother. None of the studies cited here manipulated stress or support experimentally, so the causal reading rests on prospective association alone. Risk factors are treated in full under Risk Factors.

The Interpersonal View

The interpersonal account locates the problem in role transition and in disputes with those closest to the mother. Its practical expression is interpersonal psychotherapy (IPT). In a randomised trial of 120 postpartum women with major depression, twelve weeks of IPT lowered Hamilton ratings from 19.4 to 8.3, against 19.8 to 16.8 on a waiting list. Recovery by Beck Depression Inventory criteria was 43.8% with IPT and 13.7% in the waiting-list group, and social adjustment also favoured IPT. Some women did not complete the protocol (O'Hara et al., 2000). A meta-analysis of 28 perinatal studies later found IPT superior to comparison conditions for depressive symptoms and for the prevalence of depressive episodes in prevention, and for depression, anxiety, relationship quality and social support in treatment. Moderator results were inconsistent across outcomes and few were significant (Sockol, 2018).

Efficacy does not show that the theory is right. A theory paper set out four hypothesised interpersonal change mechanisms, among them social support, emotional processing and interpersonal skills, and noted that little is known about how IPT works. These mechanisms are hypothesised, little research has tested whether they mediate change, and the authors called for trials to test candidate mediators (Lipsitz & Markowitz, 2013). The standing is thus asymmetric: the treatment is supported, the proposed mechanisms remain largely untested. Treatment evidence is reviewed under Treatment Evidence and How Treatments Work.

Attachment adds a related risk. A systematic review of 20 studies of 2306 mothers (14 longitudinal and 6 cross-sectional) found insecure adult attachment to be an additional risk for postnatal depression, with anxious attachment more strongly linked to symptoms than avoidant or dismissing styles. One study reported no association. The authors rated each study on whether it reported direct cause-effect relations, and called for large international longitudinal studies (Warfa et al., 2014). Causality between insecure object attachment styles and postnatal depression is therefore unestablished.

Hopelessness and Rumination

Two cognitive theories have been carried into the perinatal period. Hopelessness theory proposes that a negative cognitive style, meeting stress, produces a hopelessness form of depression; its authors judged it promising but unproven, with symptom-specific tests lacking and the stability of the cognitive diathesis unresolved (Abramson et al., 1989). It descends from the learned helplessness tradition. Response styles theory holds that rumination prolongs depression whereas distraction shortens it. Its original review found support in the studies it examined (Nolen-Hoeksema, 1991). A later review found that rumination predicts the onset of depression consistently and its duration less consistently, and noted findings that contradicted original predictions (Nolen-Hoeksema et al., 2008). Both theories were developed outside the perinatal period.

The perinatal test most relevant to both is a three-wave prospective study of 101 women recruited in the third trimester. Brooding predicted symptom change to two months, and negative inferential style predicted it only among women with higher initial symptoms. Reflection did not predict change, and none of the cognitive variables predicted change from pregnancy to one month. Every measure was self-report, and childbirth stress was not assessed individually (Barnum et al., 2013). A systematic review of ten perinatal studies, which also proposed a preliminary cognitive model of perinatal rumination, found the evidence mixed. The studies were few, the experimental findings were largely trend-level, and the authors judged that the longitudinal studies did not allow definite causal conclusions, while the experimental studies suggested a causal role for rumination (DeJong et al., 2016).

Standing, then, differs by theory. Response styles theory is well supported for depression generally and only partly confirmed perinatally. Hopelessness theory is promising but unproven generally, and its perinatal test covers a single negative-inferential-style finding limited to women with higher initial symptoms. The cognitive model is treated in The Cognitive Model in the Perinatal Period, and the question of cause versus consequence in Testing Whether Cognition Is Cause or Consequence.

Reproductive-Hormone Sensitivity

The hormone-sensitivity account is covered in The Biology of Postpartum Depression and is stated only briefly here. In a double-blind manipulation, withdrawal of gonadal steroids produced mood symptoms in five of eight euthymic women with a history of postpartum depression and in none of eight without one (Bloch et al., 2000). A narrative review judged the evidence enough to support a hormone-sensitive phenotype affecting a subgroup of women, with causes still poorly understood (Schiller et al., 2015).

Evolutionary Explanations

Evolutionary accounts ask why a depressive response to birth should exist at all. One proposal treats postpartum depression as a possible adaptation that tracks low social support and infant health problems: it informs the mother that she has suffered a fitness cost, motivates reduced investment in offspring under certain circumstances, and may help her negotiate more investment from others (Hagen, 1999). This sits within a more general argument that low mood may be adaptive in unpropitious situations, with hypotheses the author considered testable but needing specific tests (Nesse, 2000). A second proposal sees a mismatch between modern conditions and those in which mothers evolved. A narrative review linked postpartum depression risk to features of modern life such as low seafood intake, early weaning and weak social support. The data were largely correlational, and the authors said that establishing causal relationships will require further research (Hahn-Holbrook & Haselton, 2014).

The adaptive reading meets a counterweight. In a retrospective online survey of 306 post-reproductive women, depression after the first birth was associated with lower completed fertility, and the authors took this cost to make an adaptationist view doubtful. Recall bias was a stated limitation (Myers et al., 2016). No study cited on this page puts the signalling account to a direct test, so the signalling account remains proposed and contested, and the mismatch account rests on largely correlational data.

A related observation concerns the field itself. A systematic review of 199 investigations covering 151,651 women found that the biological and psychosocial literatures have developed largely apart. It called for integrated biopsychosocial studies in large samples (Yim et al., 2015). The accounts above are, accordingly, complementary candidates rather than rivals that have been tested against one another.

Attention to Infant Cues, Memory and Executive Function

Cognitive research on postpartum depression splits into two questions. One asks how a depressed mother perceives and attends to her infant; the other asks whether her own memory and executive control are affected. The first question has a proposed route to caregiving, because a mother who misreads or under-registers her infant's signals has less to respond to. The second has a weaker link to caregiving and a more mixed evidence base.

Infant Faces and Cries

A systematic review of studies of pregnant or postnatal women concluded that depression and anxiety bias the processing of infant emotion towards the negative, and that such biases may affect mother-infant interaction (Webb & Ayers, 2015). Across its studies, affected women identified sadness more readily than controls and were less accurate at recognising happiness, and women with depression may disengage faster from infant faces. The review found very few studies of post-traumatic stress, so the conclusion rests mainly on depression and anxiety.

Individual studies add detail, and some of it narrows the claim. In a laboratory task with three groups of mothers (21 with major depression, 34 with generalised anxiety disorder and 34 controls, infants aged between 10 and 18 months), the depressed group identified happy infant faces less accurately than the control group did, with no group difference for sad faces (Arteche et al., 2011). The anxious group showed only a non-significant trend on happy faces. The groups were relatively small, the controls were tested with younger infants, and the authors called for replication. A rating study of 45 mothers three months after birth (15 depressed, 15 anxious, 15 controls) found that depressed mothers rated negative infant faces more negatively after a long exposure of 2000 ms, but not after a 100 ms exposure (Stein et al., 2010). Anxious mothers did not differ from controls, which led the authors to describe the bias as specific to depression. The faces were of unfamiliar infants, and the effect disappeared once mothers with comorbid conditions were excluded.

Pregnancy studies point to a different pattern, and the studies do not agree. In 101 pregnant women at about 11 weeks of gestation, women without depressive symptoms were slower to disengage attention from distressed infant faces, while women with depressive symptoms showed no such bias (Pearson et al., 2010). The absence of this bias in depression sits oddly beside the faster disengagement reported in the review. A prospective cohort of 49 healthy pregnant women added a link to caregiving: a stronger prenatal attentional bias towards infant distress predicted better bonding, as scored on the Postpartum Bonding Questionnaire three to six months after birth (odds ratio 0.43 for each 50 ms increase), and the authors concluded that prenatal attention probably influences bonding (Pearson et al., 2011). That analysis used complete cases only. A cross-sectional emotional Stroop study of pregnant and postpartum Swedish women (40 with antenatal and 33 with postpartum depressive disorder) found no attentional bias in untreated depression (Edvinsson et al., 2017). The authors suggested that emotional numbing might explain the postpartum group's pattern, in which reaction times to emotional words were shorter than to neutral ones.

The proposed mechanism is that blunted or negatively biased processing of infant signals reduces contingent responding, the matching of a mother's behaviour to what the infant has just done. The findings are inconsistent on both direction and presence of the bias: some studies show a negative skew, one shows a lost protective bias, and one shows none. The inconsistency partly reflects different tasks, and it remains unresolved. This work belongs with research on facial expression and on attentional bias, and the cognitive model of the perinatal period is covered in its own section.

Rumination, Preoccupation and Responsiveness

Rumination offers a second route from cognition to caregiving, through preoccupation that competes with attention to the infant. A randomized experiment within a cohort of 253 mothers of 10-month-old infants induced a five-minute worry or rumination state, against a neutral prime (Stein et al., 2012). The prime raised negative affect in all groups. Afterwards, mothers with generalised anxiety (n = 90) were less responsive to their infants' vocalizations than controls (n = 106), whereas the contrast for mothers with major depression (n = 57) was only a trend (p = .082). The authors read the result as showing that such cognitions are potentially causal for disturbed interaction. They also listed limits: the priming was short, observation periods were brief, power for remission analyses was limited, and the neutral prime may itself have been negative.

A systematic review of ten perinatal studies, eight using questionnaires and two experimental, proposed a preliminary cognitive model of perinatal rumination (DeJong et al., 2016). It found the evidence mixed. Studies were few, no meta-analysis was feasible, bonding outcomes were self-reported, the experimental findings were largely trend-level, and the longitudinal studies did not allow definite causal conclusions. The review considered an effect on parenting possible. The question stays open until larger experimental work is done.

Mentalizing

Mentalizing, the capacity to think about the infant in terms of mental states, is a third proposed route. In a clinical sample of 50 mothers with postpartum depression, whose infants were three to ten months old, pre-mentalizing modes of thinking about the infant predicted a fall in maternal sensitivity during the still-face procedure, which distresses the infant (Krink et al., 2018). Interest and curiosity, and certainty about the infant's mental states, did not predict a change. The concept is described further under mentalization.

Memory and Executive Function

Depression has to be separated from the ordinary cognitive change of the peripartum. Pooling 20 studies of 709 pregnant and 521 non-pregnant women, a meta-analysis found cognition poorer in pregnancy overall; in cross-sectional comparisons the third trimester showed reduced general cognition, memory and executive function while the first and second did not, whereas longitudinal work tracking the same women showed performance dropping from the first trimester to the second, and the authors concluded that the differences arise mainly during the first trimester and called for further study (Davies et al., 2018). Sleep loss is a second confound, taken up in the section on Sleep and Circadian Rhythms.

Against that background, the depression findings are modest and mixed. A small pilot of 28 pregnant women and 26 non-pregnant controls, tested at 34 to 36 weeks, found that the ten depressed pregnant women made more working memory errors than the 18 who were not depressed, a large difference; pregnant women overall did not differ from controls (Hampson et al., 2015). Lower estradiol predicted more errors, while cortisol and sleep disruption did not. The study had no postpartum comparison group and needs replication in a larger cohort. In a cross-sectional home-visit study of 395 parents in Pelotas, Brazil (222 women and 173 men), those scoring as depressed on the Edinburgh Postnatal Depression Scale (cutoff 13 or above) had poorer word span than those who were not, and the authors concluded that symptoms affect working and short-term memory (Pio de Almeida et al., 2012). Fewer men than women took part.

A systematic review of 42 studies of perinatal women found that depressive symptoms went with subjective cognitive complaints, a finding from six studies that rests on self-report (Sobral et al., 2026). Subjective complaints are not the same as objective impairment: for working memory, attention and cognitive flexibility the relation to symptoms was inconsistent, and for verbal and recognition memory it was not significant. The review judged objective impairment inconclusive and advised longitudinal research. Its studies were mostly cross-sectional and used heterogeneous laboratory tasks. Executive function more broadly is treated under executive function.

Neural Response to the Infant

A systematic review found that mothers' neural response to infant stimuli, stronger in healthy mothers than in non-mothers and stronger to their own infant than to an unfamiliar one, was blunted in acutely depressed mothers, with greater depression severity going with greater blunting (Bjertrup et al., 2019). The studies were heterogeneous in method, and none covered bipolar disorder or remitted unipolar disorder. Twenty-two first-time mothers of 18-month-old infants, 11 depressed and 11 not, were scanned with functional imaging in a case-control design: the comparison mothers showed greater striatal and medial thalamic activation than the depressed mothers, and their own infant's cry elicited activation only in the non-depressed group, a pattern the authors suggested may reflect disturbed networks underlying parenting deficits (Laurent & Ablow, 2012). The sample was small. The brain systems involved are set out in the section on Brain Networks and Imaging.

Testing Whether Cognition Is Cause or Consequence

The cognitive findings reviewed in earlier sections are mostly associations, and an association between negative thinking and low mood after birth fits two readings: the thinking helps produce the episode, or the episode produces the thinking. Four kinds of design bear on the question: prospective studies that begin in pregnancy, experimental manipulations, trials of cognitive training, and genetically informed comparisons. The first three come with results on both sides; the last concerns children rather than mothers. This section sets out what each has shown and how far it can carry a causal reading.

Prospective Vulnerability Studies

A measure taken before the episode cannot be a consequence of it, which is why the field leans on cohorts that recruit in pregnancy. The largest is a postal-questionnaire cohort of 12,003 pregnant women in the Avon (ALSPAC) study. Among the 8540 who were not depressed at baseline, 8.6% (95% CI 8.0 to 9.2) developed depression by 32 weeks. Onset ran at 13.4% in the highest tertile of a six-item negative self-schema score and 4.8% in the lowest, and the schema association survived adjustment (adjusted OR 1.6, 95% CI 1.27 to 2.02) and persisted from eight weeks to 33 months after birth (Evans et al., 2005). The authors read it as support for a key element of Beck's theory (Evans et al., 2005). Its limits are the ones the authors acknowledge: onset came from a brief self-report screen with no diagnostic interviews, the schema items were borrowed from a personality measure, personality was not measured separately, and residual confounding by current mood is possible.

Smaller cohorts point the same way. Among 94 healthy first-time mothers seen in the third trimester, two factors each predicted depressive symptoms eight weeks postpartum once neuroticism, education and dysphoria were controlled: self-devaluation, and autobiographical memories retrieved with little specificity. Neither predicted low mood at two weeks, a period in which, the authors note, biological factors are thought to be important (Hipwell et al., 2004). A three-wave study of 101 women, first assessed in the third trimester, found a similar delay. Brooding rumination predicted symptom change by two months but not by one month, reflective rumination predicted neither, and a negative inferential style predicted later symptoms in women who already had high scores in pregnancy but not in low-scoring women. The study relied on self-report alone, and its authors describe the cognitive risk factors as possible intervention targets (Barnum et al., 2013).

The attribution studies of the early 1980s gave partial support. Prepartum depression, attributional style, stress at delivery and stressful life events each predicted postpartum depression among 170 women tracked from the second trimester until three months after birth, and jointly they explained roughly 40% of the variance. Parity, education, income and a history of menstrual problems did not predict it (O'Hara et al., 1982). Among first-time mothers, attributional style measured in pregnancy predicted later postpartum depression only for women free of depression while pregnant, not for those already depressed at that point, and the author suggested the model may need modifying (Cutrona, 1983). A related prospective result concerns attention rather than mood: in 49 healthy pregnant women, a stronger attentional bias towards infant distress in pregnancy (a 50-ms increase) went with better bonding 3 to 6 months after birth (OR 0.43, 95% CI 0.23 to 0.81), and the authors consider that prenatal attention probably influences bonding (Pearson et al., 2011).

These studies establish temporal order, not mechanism. Each leaves open that an unmeasured trait, or mood already present in pregnancy, drives both the cognitive score and the later episode.

Experimental Manipulation

An experiment can do what a cohort cannot, by changing the cognition and watching what follows. In a randomised experiment nested in a cohort of 253 mothers of 10-month-old infants (90 with generalised anxiety disorder, 57 with major depression, 106 controls), a five-minute worry and rumination prime raised negative affect in every group relative to a neutral prime. After the prime, mothers with generalised anxiety disorder were less responsive to their infants' vocalisations than controls, and mothers with depression showed a similar trend (p = .082). The authors conclude that the cognitions are potentially causal and may disturb the mother-infant interaction (Stein et al., 2012). The priming was brief, the observation periods short, the remission analyses underpowered and the neutral prime possibly negative, so the result shows that induced rumination can alter a mother's responsiveness within minutes. It does not show that habitual rumination produces a postpartum episode.

Cognitive Training as a Causal Test

A trial that changes the cognition before the episode would be the strongest test. A randomised trial of prenatal affective cognitive training randomised 147 high-risk pregnant women to five 90-minute sessions or care as usual, and 121 (83%) were followed up. Training did not reduce postpartum depression overall (p = 0.18) or depressive symptoms (p = 0.31), so the primary outcome was not significant. It did improve positive interpretation of infant emotions. Among multiparous women only, incidence was 5% with training and 35% with usual care (OR 0.10, 95% CI 0.01 to 0.92); the authors label this parity effect exploratory and call for replication in larger trials (Bjertrup et al., 2026). Changing an interpretation bias without lowering overall incidence is, if it replicates, evidence against a simple cause-and-effect path, though one null trial of this kind cannot settle it.

A Biological Causal Test

The one clear experimental test of a cause of postpartum mood disturbance is biological rather than cognitive. In a double-blind manipulation of gonadal steroids in 16 euthymic women, 8 with a history of postpartum depression and 8 without, withdrawal produced depressive symptoms in 5 of the 8 women with a history (62.5%) and in none of the 8 without. The authors conclude that hormones are involved in a subgroup, through differential sensitivity (Bloch et al., 2000). This bears on cause rather than on cognition, and the sample was small. Its relation to the cognitive findings is covered in the biology section.

Genetically Informed Designs and Children

Designs that control family background have tested the other direction of the question, whether maternal depression harms the child. In 11,599 Norwegian families (17,830 siblings), comparing siblings within families attenuated the associations between maternal depression and child problems, which points to familial confounding. After the sibling comparison, only concurrent maternal depression remained associated with child internalising and externalising problems. Prenatal and 6-month postpartum depression were no longer associated, and the authors conclude that perinatal exposure is not especially harmful. Participation was 41% of eligible women and child problems were mother-reported (Gjerde et al., 2017). A genetically informative population-based sample was likewise used to examine maternal prenatal depressive symptoms and early psychopathology in offspring (Hannigan et al., 2018).

What Would Settle It

The prospective evidence for cognitive vulnerability is consistent in direction across cohorts, and the one training trial is null on its primary outcome, so causation is not established. A decisive answer needs designs that largely have not been run: adequately powered trials of cognitive change begun before the episode in women with and without the vulnerability, with parity prespecified rather than found afterwards; cohorts with repeated cognitive measures and diagnostic interviews, to separate trait from state; and genetically informed comparisons of the mothers themselves, as the sibling designs have so far addressed children. Until then, the question remains open.

Validity of Laboratory and Observational Measures

Findings about postpartum cognition and mother-infant interaction are only as good as the instruments behind them. Two questions apply to each instrument: whether it yields a stable score (reliability), and whether that score says anything about what mothers and infants do outside the laboratory (real-world validity). This section takes attention tasks, observational coding and maternal report in turn. Symptom screens, set against diagnostic interviews, are covered under Symptoms, Diagnosis and Measurement.

Attention-Bias Tasks

The most pointed reliability critique comes from outside the perinatal field. A theory and methods paper that reviewed 13 studies reported that the global attentional bias score from the dot-probe task, a difference score, is unreliable (Rodebaugh et al., 2016). It proposed trial-level bias signal (TL-BS) scores as a better alternative, and TL-BS reliability exceeded that of the global score. Its own empirical demonstration used 24 patients with anxiety disorders, so the authors described it as a demonstration and not a best-powered study. Peak TL-BS showed only modest stability, and regression to the mean was not conclusively shown. The authors urged researchers to test and report reliability, and to treat low reliability as an opportunity to modify and improve the underlying theory (Rodebaugh et al., 2016). The paper is a general methods contribution, not a perinatal one, and it bears on postpartum work by analogy, since that work also leans on bias scores from reaction-time tasks.

The postpartum literature on infant cues is itself mixed in method. One systematic review of pregnant or postnatal women pooled studies that used emotion interpretation tasks and others that used reaction-time disengagement tasks. Its findings by outcome rested on few studies (k = 10, 4 and 2 in different comparisons). Depressed or anxious women identified sadness more often than controls and were less accurate for happiness, and depressed women may disengage faster from infant faces (Webb & Ayers, 2015). The review concluded that such biases may affect mother-infant interaction, a hedge that stands in for a demonstrated link.

Observation of Mother and Infant

Observation is the second main source of evidence. The still-face procedure has a review and meta-analysis literature of its own, which examines its many variants (Mesman et al., 2009). The wider findings on depression and interaction draw on laboratory and home observation. A review of that work, which used second-by-second microanalytic coding and naturalistic home observation, concluded that maternal depression is linked to poorer infant affect regulation, with miscoordination and repair as the process of interest (Tronick & Reck, 2009). It further noted that infants were affected not just when their mothers were severely and acutely depressed but also when mothers had high depressive symptom levels.

A longitudinal cohort study of low-risk primiparous mothers and infants shows what such coding can and cannot carry. Depressed mothers showed less sensitive attunement and less affirming of the infant than well mothers, and more negating of the infant. Interaction at 2 months predicted cognition at 18 months, but not attachment, where adversity and postpartum depression themselves were the predictors. Severe engagement disturbance was absent in this low-risk sample (Murray et al., 1996).

Taken together, coded interaction is a sample of behaviour, taken in a particular setting. It can show differences in the quality of the exchange, and in one cohort an early sample predicted a later outcome. On the evidence cited here, it cannot stand for the whole relationship.

Informant Bias in Maternal Report

Many child-outcome studies rely on what mothers say about their children. A long-standing worry is that depressed mothers distort such ratings. A critical review of 22 published studies of depressed mothers' ratings found the claim of distortion unsupported: no study met the criteria needed to test it, and the review challenged the empirical basis of the assumption while noting that future work must address open issues (Richters, 1992). A modern psychometric analysis (moderated non-linear factor analysis) of 8,507 mother-child dyads, with youth aged 9 to 11 years, in the Adolescent Brain Cognitive Development study, found that bias related to maternal psychopathology dimensions was very small and concluded there was little psychometric evidence of maternal bias (Olino et al., 2021). That sample was not a postpartum one, so the result bears on maternal report in general and on the child-outcome literature only by extension.

Current Standing

The reliability critique remains unanswered for infant-cue tasks: none of the studies cited here reports reliability figures for them, and the one detailed critique comes from a non-perinatal literature. The informant-bias worry is smaller than often assumed, although the evidence behind that comes from a critical review and a general-population sample, not from postpartum cohorts.

The Biology of Postpartum Depression

The biological literature on postpartum depression covers hormones, neuroactive steroids, stress axes, immune and thyroid markers, monoamines and genes. Each system below is given with its standing. No single system is established as the cause, and the reviews that survey the field stress several interacting pathways that may differ between women (Payne & Maguire, 2019). A systematic review of 199 investigations of 151,651 women found the biological and psychosocial literatures largely distinct, with few integrative analyses, and called for integrated biopsychosocial studies in large samples (Yim et al., 2015).

Reproductive Hormones and Differential Sensitivity

The leading account holds that the problem lies in abnormal sensitivity to normal hormonal change, not in abnormal hormone levels. A narrative review reports that absolute hormone levels did not distinguish women with postpartum depression from those without, and that findings on hormone change were inconsistent (Schiller et al., 2015).

The strongest direct test is a double-blind experiment in 16 euthymic women, eight with a history of postpartum depression and eight without. The design manipulated gonadal steroids and then withdrew them. In a double-blind trial that withdrew gonadal steroids under leuprolide, mood symptoms appeared in 5 of the 8 women with a past postpartum depression (62.5%) and in none of the 8 women who had no such history (Bloch et al., 2000). The authors read this as evidence that hormones are involved in a subgroup of women, through differential sensitivity. The sample is very small and the finding applies only to that subgroup.

The same review concludes that hormone shifts trigger depression in a subgroup only, and that the evidence is enough to support a hormone-sensitive phenotype (Schiller et al., 2015). It also notes that estradiol studies are confounded by stress and sleep, that treatment studies are small and uncontrolled, and that animal models have low face validity. Reviews of mechanism treat the hormone-withdrawal account as one pathway among several (Payne & Maguire, 2019; Meltzer-Brody & Kanes, 2020), and the latter states that the precise aetiology is unknown. Standing: supported in a subgroup by small experiments.

Neuroactive Steroids and GABA

Allopregnanolone is a metabolite of progesterone that acts at GABA-A receptors. It supplies the rationale for the neuroactive-steroid drugs, which a narrative review describes as a promising mechanism and not an established aetiology (Meltzer-Brody & Kanes, 2020). The drugs themselves are covered under Neuroactive Steroid Drugs, Hormones and Brain Stimulation.

The animal evidence comes from mice. Pregnancy lowered particular GABA-A receptor subunits (delta and gamma2), and mice whose delta subunit was not regulated during pregnancy (Gabrd+/- and Gabrd-/-) showed depression-like and abnormal maternal behaviour and poorer pup survival, so these mutants are a possible model of postpartum depression (Maguire & Mody, 2008).

In people the picture is thinner. In a prospective exploratory cohort of 60 pregnant women with a prior mood disorder, lower second-trimester allopregnanolone predicted later postpartum depression after adjustment. Third-trimester allopregnanolone, progesterone, their ratio and their rates of change did not predict it (Osborne et al., 2017). The authors call the relationship worth exploring in a larger sample. The study was exploratory, small, mostly in white, high-income and well-educated women, and had no power calculation. A narrative review found the link between allopregnanolone levels and postpartum depression inconsistent in clinical studies, with the drugs' mechanism unclear and their durable effect not explained by receptor action alone (Walton & Maguire, 2019). Standing: a plausible mechanism whose presence as a deficit in patients is not established.

Stress Hormones

Findings on the stress axis disagree. In a longitudinal cohort of 100 pregnant women, placental corticotropin-releasing hormone at about 25 weeks of gestation predicted later postpartum depressive symptoms, with an area under the curve of .78. The outcome was a self-report questionnaire and not a clinical diagnosis, and the authors state that replication is needed (Yim et al., 2009). A prospective cohort of 1230 women found no association between mid-pregnancy placental CRH and postpartum depression, with odds ratios near 1 (Meltzer-Brody et al., 2011).

A systematic review of 47 perinatal studies found the cortisol-depression link inconsistent, with 24 studies significant and 23 not. It noted considerable methodological variation across studies (Seth et al., 2016). Standing: mixed, and open. A reviewer of the wider evidence describes the role of the HPA axis as unclear and possibly epiphenomenal (Schiller et al., 2015).

Inflammation, Thyroid and Oxytocin

A systematic review of 25 articles found few significant associations between inflammatory markers (IL-6, IL-10, TNF-alpha, IFN-gamma) and postpartum depression. C-reactive protein near delivery was possibly predictive, and the authors call for replication (Lambert & Gressier, 2019).

Thyroid autoimmunity has a firmer, though still modest, signal. A meta-analysis of five studies, with 449 antibody-positive and 2483 antibody-negative women, found that anti-thyroperoxidase antibody positivity went with higher risk (RR 1.49, 95% CI 1.11-2.00). The authors flag possible publication bias, although adjusting for it had a negligible effect on the pooled estimate, and call for further well-designed studies (Minaldi et al., 2020).

A systematic review of sixteen studies found lower plasma oxytocin associated with more depressive symptoms in eight of twelve studies of endogenous oxytocin. The effect of synthetic oxytocin was inconclusive, and heterogeneity and the small number of studies limit the conclusion (Thul et al., 2020). A systematic review of biological markers concluded that reliable biomarkers in high-risk mothers are still a priority, not a result (Serati et al., 2016). Standing: preliminary.

Monoamines

In a positron emission tomography study, 15 healthy women 4 to 6 days postpartum had monoamine oxidase A total distribution volume a mean of 43% higher than 15 healthy women who were not recently postpartum, in all seven regions measured (Sacher et al., 2010). The sample did not include depressed mothers, so the study does not show that the enzyme differs in postpartum depression. The authors propose a neurobiological model linking the elevation to monoamine lowering and postpartum blues, and the model remains untested. Standing: preliminary.

Genetics and Epigenetics

Twin and sibling data showed perinatal depression to be heritable. In a sibling design heritability was 44% for perinatal depression and 32% for depression at other times, and about one third of the genetic variance was specific to the perinatal form (Viktorin et al., 2016). A family study of 120 sibling pairs with recurrent unipolar depression found clustering for postpartum onset within four weeks, but not for onset within six months (Forty et al., 2006). Familial factors are implicated only under the narrow definition.

A genome-wide meta-analysis of 20 cohorts, with 18,770 cases and 58,461 controls, found no genome-wide significant single variant. SNP heritability was 0.14, which fits a polygenic architecture. The cell-type enrichment pattern differed from that of major depression, and the authors suggest postpartum depression may have unique genetics (Guintivano et al., 2023). They note that more samples are needed.

Epigenetic work offers a candidate marker. In a prospective cohort of 51 pregnant women with a mood-disorder history, methylation at two genes, HP1BP3 and TTC9B, predicted postpartum depression. The authors suggest heightened epigenetic sensitivity to oestrogen, and note that the prediction reversed in women who were depressed before delivery until blood count data were added to the model (Guintivano et al., 2014). A replication in 51 independent high-risk women found a comparable predictive accuracy (AUC 0.81), and the same genes predicted postpartum depression in 240 women with no psychiatric diagnosis. The link to estradiol and allopregnanolone change was rated only moderate evidence (Osborne et al., 2016). Replication in small samples makes a chance finding less likely, but it does not show that the markers work in the general population or that they are causal. Standing: preliminary. Candidate-gene results more broadly have not been replicated (Schiller et al., 2015).

Sleep and Circadian Rhythms

Sleep is one of the better-studied correlates of postpartum depression, and one of the few where trials have begun to probe direction. This section covers the association, the gap between perceived and measured sleep, the trials that test direction, the circadian evidence, and the link to postpartum psychosis.

The Association With Depression

A systematic review of 31 studies, most rated weak or moderate in quality, found that poor self-reported sleep preceded postpartum depression (Lawson et al., 2015). The same review judged the evidence insufficient for anxiety and for psychosis, and mixed for objectively measured sleep. Its authors noted wide variation in design, tools that were often not validated for the perinatal period, and small samples.

A population-based questionnaire study of 2830 women at seven weeks postpartum points the same way. Even after adjustment, poor sleep as measured by the Pittsburgh Sleep Quality Index stayed linked to depressive symptoms on the Edinburgh Postnatal Depression Scale, and of the sleep components, the strongest links with depression came from subjective sleep quality and sleep disturbances (Dørheim et al., 2009). Sleep was self-reported and the design cross-sectional, so the authors concluded only that poor sleep possibly raises risk in some women. Depression also predicted poor sleep quality in the same dataset.

That reciprocal pattern is why a narrative review describes the link between insomnia or poor sleep and postpartum depression as bidirectional and additive, while describing the evidence as still emerging (Okun, 2015).

Perceived Sleep Against Measured Sleep

The studies that paired wrist actigraphy with questionnaires add a qualification. Among 25 healthy primiparous women followed prospectively, perceived sleep disturbance, sleep fragmentation, sleep efficiency and wake after sleep onset each related to Edinburgh scores, whereas nocturnal sleep duration and naps did not (Park et al., 2013). Actigraphy did not predict scores one month later. The authors read this as pointing to sleep maintenance and perceived quality, not the quantity of sleep. They also flagged shared method variance between the two self-report scales, a small sample and no structured diagnostic interview.

Forty-four healthy, low-risk women assessed one week after birth showed a similar split. Their sleep was worse than in the third trimester on both objective and subjective measures, yet objective night-time sleep related little to postpartum mood, while subjective night-time sleep, sleep-related daytime dysfunction and daytime napping each significantly predicted it (Bei et al., 2010). The authors suggested that perceived poor sleep may matter more than measured sleep. Together the two studies explain why a sleep diary and an actigraph can disagree about the same mother: the mood-relevant quantity may be how rested she feels and how often sleep is broken, not the hours recorded.

Testing the Direction

Trials offer the only route past the correlations. In a cluster randomised trial across 49 Maternal and Child Health centres, 328 mothers of infants with a sleep problem received a sleep intervention. Infant sleep problems fell (56% against 68% at 10 months, adjusted OR 0.58), and so did maternal Edinburgh scores, by 1.4 points at 10 months and 1.7 at 12 months (Hiscock et al., 2007). Mental health on the SF-12 favoured the intervention at both time points, and mean cost per family was lower. Among the studies cited here, this is the strongest evidence for the direction from sleep to mood, though the intervention targeted the infant's sleep, so it cannot separate maternal sleep from other changes in family life.

Cognitive behavioural therapy for insomnia was tested in pregnancy in a three-site randomised trial of 179 analysed women with insomnia disorder. It reduced insomnia severity more than a control therapy (d = 0.5) and produced remission faster, but it did not differ from control on objectively measured wake time, and the decline in Edinburgh scores was small (Manber et al., 2019). The trial was unmasked, excluded psychiatric comorbidity and ran in pregnancy, so it shows that prenatal insomnia is treatable. It does not show that treating insomnia prevents postpartum depression.

Circadian Timing and Light

The circadian side rests on one small trial. In 27 pregnant women with nonseasonal major depression, five weeks of bright light (7,000 lux) outperformed a dim red light (70 lux) on depression ratings, with response in 81.3% against 45.5% and remission in 68.6% against 36.4% (Wirz-Justice et al., 2011). Expectation ratings did not differ between arms. The authors framed light as a possibly useful nonpharmacologic option. The sample was antepartum and small, so the result bears on the circadian account only indirectly and has not been replicated in this section's sources.

Sleep Loss and Postpartum Psychosis

A literature review argued that sleep loss may be a final common pathway to postpartum psychosis, alongside suggested contributions from primiparity, difficult labour, genetic predisposition and hormonal change (Sharma & Mazmanian, 2003). The review described the disorder as rare, its diagnostic status as controversial and its pathophysiology as poorly understood. The role of sleep deprivation in that condition is taken up under Postpartum Psychosis in Types of Postpartum Depression.

Current Standing

Sleep disturbance is a consistent correlate of postpartum depression and a probable contributor. The infant-sleep trial supports a direction from sleep to mood, and insomnia treatment in pregnancy works on insomnia itself. Whether improving a mother's own sleep prevents postpartum depression remains open.

Brain Networks and Imaging

Postpartum depression is studied in a brain that is already changing. A prospective study scanned first-time mothers before and after pregnancy and compared them with fathers and nulliparous controls. Mothers showed reductions in grey matter volume that persisted for at least 2 years after pregnancy, whereas fathers did not differ from controls. Measures of maternal attachment after birth were predicted by the shifts in grey matter volume that took place across pregnancy (Hoekzema et al., 2017). Any imaging difference in a depressed mother therefore has to be read against this normal remodelling.

Findings in Postpartum Depression

A narrative review of neuroimaging reported altered functional connectivity and activity in the brains of women with postpartum depression. It also noted that no morphometry or white matter integrity studies existed at the time, and it concluded that more research is needed (Duan et al., 2017). A later systematic review pooled 26 studies, with 762 women with postpartum depression and 953 controls, and synthesised them narratively because the studies were too heterogeneous for a meta-analysis. It reported hyperactivity of the default mode network and changes in the salience network, and concluded that these changes mirror those of major depression (Horáková et al., 2022).

The same review also found that amygdala responses to negative stimuli were attenuated in postpartum depression, and responses to positive stimuli increased, the opposite direction to the findings in major depressive disorder (Horáková et al., 2022). In a face-matching fMRI study of 14 unmedicated depressed and 16 healthy mothers, activity in the left dorsomedial prefrontal cortex was lower in the depressed group. Healthy mothers showed connectivity from that region to the amygdala, and depressed mothers showed none. The authors described the dorsomedial prefrontal-amygdala deficit as either a possible mechanism or an effect of the illness (Moses-Kolko et al., 2010).

Mirror Image or Unique Profile

The reviews disagree about how distinctive the picture is. One review concluded that default mode and salience network changes resemble those of major depression (Horáková et al., 2022). Another argued that postpartum anxiety and depression have neural profiles that are partly unique, while noting that little is known about their neural correlates (Pawluski et al., 2017). The question remains open, and the amygdala difference above is the clearest point where the two accounts diverge.

Cognitive Constructs and Their Neural Correlates

Processing of infant cues has the most direct neural evidence. A systematic review found that healthy mothers showed stronger neural responses to infants, and faster responses, than non-mothers, and that the response to the own infant was accentuated relative to an unfamiliar infant. Acutely depressed mothers responded less than healthy mothers, and greater depression severity went with more blunting. The review noted large methodological heterogeneity and no studies of bipolar disorder or remitted unipolar disorder (Bjertrup et al., 2019). In an fMRI case-control study of 11 depressed and 11 comparison mothers of 18-month-old infants, non-depressed mothers showed stronger striatal and medial thalamic activation, and a stronger occipital fusiform response to the own infant's cry than to another cry. Depressed mothers showed no activation to their own infant's cry. The authors suggested that disturbed networks may explain parenting deficits (Laurent & Ablow, 2012).

Rumination is a core construct of the cognitive account, and its link to the default mode network has not been tested in postpartum samples on the evidence reviewed here.

Monoamine oxidase A is a neurochemical rather than a network finding. A PET study found that brain binding of this enzyme was higher, by a mean of 43%, in 15 women 4 to 6 days postpartum than in 15 women not recently postpartum, across all seven regions measured. The authors proposed a neurobiological model linking the elevation to postpartum blues (Sacher et al., 2010). The wider hormonal and neurochemical evidence is covered under The Biology of Postpartum Depression.

Limits of the Evidence

The imaging literature is thin. The systematic review described samples that were small, one with only 4 women, and designs that were mostly cross-sectional, so causality cannot be inferred. It also flagged antidepressant medication as a confound, anxiety comorbidity that could not be excluded, overlapping samples from the same author groups, and no agreed timing for onset. Structural and metabolite findings were inconsistent, and the findings need replication (Horáková et al., 2022). The face-matching study included unmedicated depressed mothers only (Moses-Kolko et al., 2010).

Mother-Infant Interaction

Depression in a mother changes how she and her infant engage with one another, and the changes are measurable in the minute-by-minute record of face-to-face play. This section sets out what observational studies have found, the experiments that probed it, the proposed route from maternal mood to learning, and the evidence on attachment, direction of effect and moderators.

What Is Observed

A meta-analysis of 46 observational studies linked maternal depression to parenting that was more negative and more disengaged, with negative behaviour showing the largest association, followed by disengagement and then positive behaviour (Lovejoy et al., 2000). The association with positive behaviour was relatively weak. Effects on negative behaviour were larger when the depression was current than when it was past. Positive-behaviour effects were larger for disadvantaged women and for mothers of infants than for mothers of older children. Diagnostic interviews and self-report measures produced similar effects, and the authors concluded that the deficits are not specific to depressive disorder, with moderating factors still needing study (Lovejoy et al., 2000).

Speech offers a closer view. Depressed mothers talking to their infants in play expressed more negative affect than well mothers and focused less on the infant's own experience. They also showed a tendency to acknowledge the infant's agency less. Complexity and syntax did not differ significantly (Murray et al., 1993). In a longitudinal cohort of low-risk first-time mothers, depressed mothers were less sensitively attuned, more often negated the infant and less often affirmed the infant than well mothers. Social and personal adversity was likewise associated with more maternal interaction difficulties, and no severe disturbance of engagement appeared in this low-risk sample (Murray et al., 1996).

Narrative reviews extend the picture beyond play. Postpartum depression has been reviewed as disrupting interaction, caregiving and safety practices, though studies were discrepant on safety practices and the review rested largely on self-report from mothers (Field, 2010). A systematic review of 122 studies of untreated postpartum depression covered consequences for mothers, for infants and for the mother-child bond, and concluded that early detection is apparently important (Slomian et al., 2019). Confounding factors were hard to account for in that literature. Observed differences in infant behaviour are taken up under Child Outcomes; see also infant behaviour.

The Still-Face Experiments and Mutual Regulation

An experimental test came from mothers who simulated depressed affect in face-to-face play. Their 24 infants, aged 96 to 110 days, were counterbalanced across treatments. Compared with normal play, the simulated depression produced more protest, wariness and brief positive affect, and more looking away, and it produced fewer cycles of monitoring and play. The authors read this as a specific negative reaction to the simulation, and the negativity persisted briefly after the mothers returned to normal behaviour (Cohn & Tronick, 1983).

The mutual regulation account built on such work treats infant and mother as regulating each other's affect. In a review drawing on laboratory and home observation with second-by-second coding, maternal depression was linked to poorer infant affect regulation, with miscoordination and repair as the central construct. Even high depressive symptoms short of severe depression were associated with an effect on the infant, and male infants appeared more vulnerable. The authors concluded that treatment should be dyad-focused and attend to the infant's affect (Tronick & Reck, 2009). The wider still-face literature has been reviewed and meta-analysed (Mesman et al., 2009); that work is cited here for its scope only, and no finding from it is stated. No study cited on this page appraises how well the mutual regulation account has held up, so its standing is not stated here.

Infant-Directed Speech and Learning

In a conditioned attention experiment with 225 four-month-old infants, speech recorded from mothers with higher depressive symptoms (Beck Depression Inventory above 15) did not produce significant summation, while speech from mothers scoring 15 or below did produce significant positive summation. The authors concluded that symptomatic mothers' speech does not promote associative learning. Symptoms were self-reported, and the infants tested were those of nondepressed mothers (Kaplan et al., 1999). Combined with the content differences in depressed mothers' speech (Murray et al., 1993), this is the proposed route from maternal mood to infant learning. In the 1993 study, speech focus and communication quality mediated the link to infant cognition.

Attachment

A meta-analysis of seven studies found that infants of depressed mothers were less often securely attached on the Strange Situation than infants of non-depressed mothers. Avoidant and disorganised classifications were marginally more common. The disorganised rate rose from 17% to 28% consistently across studies. The pooled effect was initially heterogeneous, one outlier study was removed, secure and avoidant effects varied between studies, and the samples were predominantly middle-income (Martins & Gaffan, 2000). In the cohort of low-risk first-time mothers, postpartum depression and adversity each predicted attachment, whereas two-month interaction quality did not significantly predict it (Murray et al., 1996).

Direction and Moderators

Effects run in both directions. In a longitudinal path analysis of 55 married women, a difficult infant temperament predicted postpartum depression at three months, directly and also through lower parenting self-efficacy. Social support acted mainly indirectly through self-efficacy (Cutrona & Troutman, 1986).

Moderators appear in several of the studies cited here. Depression that was current, not past, carried larger effects on negative parenting, and socioeconomic disadvantage and child age moderated the positive-behaviour effects (Lovejoy et al., 2000). Social and personal adversity was associated with similar difficulties in maternal interaction (Murray et al., 1996). In the ALSPAC cohort of 9848 mothers, persistent severe depression carried worse child outcomes than non-persistent depression, with an odds ratio of 4.84 (95% CI, 2.94-7.98) for one outcome. Those outcomes concern the child, not the interaction itself, and the severe persistent group was small (Netsi et al., 2018).

Fathers show a parallel pattern. A meta-analysis of 28 studies (40 independent effect sizes) found paternal depression associated with less positive and more negative parenting. Effects were small, comparable in size to those for mothers, and similar in self-report and observational studies. Child age, father age and sample race or ethnicity moderated them (Wilson & Durbin, 2010). A review also noted that one study found no significant link between paternal depression and interaction frequency (Field, 2010).

Child Outcomes

Whether a mother's depression after birth leaves a mark on her child is among the most studied questions about the condition, and its answer is graded rather than yes or no. The associations are real but usually small, the timing of exposure is hard to isolate, and the strongest designs suggest that part of the link reflects shared family factors. This section covers the cohort and meta-analytic evidence, the antenatal and postnatal question, the confounding problem, fathers, and the trials that test whether treating the parent helps the child.

What the Cohorts and Meta-Analyses Show

The earliest prospective cohort followed first-time mothers and their infants to 18 months. Infants of depressed mothers scored lower on object concept and were more often insecurely attached, with mild behavioural difficulties more common, yet general cognitive and language development did not differ from controls (Murray, 1992). The pattern was one of selective deficits, and lower social class and male gender added to the infant effects. A later critical review of English-language abstracts from 1990 onward put the effect of postpartum depression on language and IQ as small, found behavioural effects variably supported, and judged chronic or recurrent depression a likelier driver of poor child outcomes than the postpartum episode itself (Grace et al., 2003).

Meta-analysis has sized the associations more precisely. Across 193 studies of mother-child pairs, maternal depression was linked to more internalizing problems, more externalizing problems and more negative affect in children, and to less positive affect; every association was small, and the moderation patterns varied by outcome (Goodman et al., 2011). The authors concluded that models going beyond main effects are needed. A later meta-analysis of 191 studies covering 195,751 unique mother-child dyads looked at perinatal depression and anxiety together. Social-emotional problems correlated positively with exposure (r = 0.21 antenatal, r = 0.24 postnatal), and cognitive outcomes correlated negatively (r = -0.12 antenatal, r = -0.25 postnatal). The intervals for language and motor development included zero, so those links were not established (Rogers et al., 2020). A Lancet Series review reached a compatible broad reading: perinatal parental disorder is associated with increased risk of child disturbance, but disturbance is not inevitable, and parenting quality, social and material support and the duration of the parent's disorder are potentially modifiable influences (Stein et al., 2014). Its authors called for earlier identification, intervention and prevention research, noting that the interventions studied were mostly for maternal depression.

Longer follow-up shows later effects in some groups. In a cohort followed from infancy to age 16, 41.5% of offspring of mothers with postnatal depression had had depression by 16, against 12.5% of controls (odds ratio 4.99, 95% CI 1.68 to 14.70). The authors read the risk as partly child vulnerability and partly family adversity, with lower childhood resilience linking insecure infant attachment to adolescent depression (Murray et al., 2011). In the ALSPAC cohort, persistent severe postnatal depression carried the worst outcomes: more child behaviour problems, lower mathematics grades and more depression at 18 years than in children of mothers without postnatal depression. Non-persistent depression raised behaviour problems but not mathematics grades. The persistent severe group was small, attrition was high, the thresholds were screening cut-offs rather than diagnoses, and behaviour problems were reported by mothers (Netsi et al., 2018).

Treating the mother does not automatically move child outcomes. A cluster-randomised trial among 903 depressed pregnant women in rural Pakistan, using a cognitive behaviour therapy-based intervention delivered by community health workers, lowered maternal major depression at 6 and 12 months, yet infant weight-for-age and height-for-age did not differ significantly between arms (Rahman et al., 2008). For the broader literature on how children develop, see child development.

Antenatal or Postnatal Exposure?

The label postpartum depression directs attention to the months after birth, but the exposure may begin earlier. In a prospective UK birth cohort of more than 4,500 parents and offspring, antenatal maternal depression predicted depression in the offspring at age 18 independently of postnatal depression. A postnatal association appeared where maternal education was low and was not significant where it was higher (Pearson et al., 2013). The authors suggested that treating antenatal depression may prevent offspring depression.

Antenatal depression is also common. In an Avon cohort, 13.5% of women scored above the screening threshold at 32 weeks of pregnancy and 9.1% at 8 weeks postpartum, so postnatal prevalence did not exceed antenatal prevalence (Evans et al., 2001). Because depressed mood in pregnancy often continues after birth, the two exposures overlap heavily and are hard to separate statistically. Which period matters more for the child therefore remains open.

Causation or Familial Confounding?

An association between a mother's depression and her child's problems can arise in several ways: shared genes, the prenatal environment, an interaction between the two, persistent exposure to the depressed state, or the adversity that accompanies depression. Conventional cohorts, in which each family contributes one child, cannot tell these apart.

Sibling comparison offers a partial remedy, because siblings share their family background and genes but may differ in the timing of maternal symptoms. In a Norwegian cohort of 11,599 families and 17,830 siblings, concurrent maternal depressive symptoms had the strongest association with child internalizing and externalizing problems. Associations with prenatal depression and with depression at 6 months postpartum were not significant once siblings were compared, and sibling comparison attenuated effects generally, which indicates familial confounding. The authors concluded that perinatal exposure is not especially harmful. Only 41% of eligible women took part, and child problems were mother-reported (Gjerde et al., 2017). A genetically informative, population-based sample has also been used to examine maternal depressive symptoms during pregnancy and the risk of early psychopathology in the offspring (Hannigan et al., 2018).

Taken together, the sibling design points to a smaller causal effect of perinatal exposure than the raw associations suggest, with concurrent depression mattering most in that sample. It does not show that causal effects are absent, and the question of how much of the association is causal remains open.

Fathers

Fathers are affected too. In a population cohort of more than 8,000 fathers, depression 8 weeks after birth predicted emotional and behavioural problems in children at 3.5 years (adjusted OR 2.09, 95% CI 1.42 to 3.08), and conduct problems in boys in particular. The effect persisted after maternal depression and later paternal depression were controlled, though child outcomes were mother-reported (Ramchandani et al., 2005). A systematic review of 21 prospective studies likewise found worse emotional and behavioural outcomes in offspring of depressed fathers, with stronger associations where fathers showed negative expressiveness, hostility, low involvement or marital conflict. Heterogeneity in assessment, magnitude and timing of exposure limited firm conclusions about trajectory (Sweeney & MacBeth, 2016).

Does Treating the Parent Help the Child?

The evidence is mixed. A meta-analysis of randomised trials found that psychotherapy for depressed mothers lowered maternal depression (g = 0.66) and also benefited child mental health (g = 0.40), mother-child interaction (g = 0.35) and parenting or marital distress (g = 0.67). The authors judged a child benefit indicated but noted that study quality was not optimal, outcome instruments differed considerably and more high-quality research was needed (Cuijpers et al., 2015).

A cohort inside the STAR*D trial points the same way, with a caution about its sample. Among 151 mother-child pairs with children aged 7 to 17, the mothers had depression that was not perinatal and were not randomised to remission. Children whose mothers remitted showed an 11% decrease in diagnoses, while children of non-remitting mothers showed an 8% increase, and the effect persisted after adjustment for child age, sex and confounders. The reported outcomes were measured after 3 months of maternal treatment, in a study that was still following cases (Weissman et al., 2006).

Trials in postpartum samples are less encouraging. Interpersonal psychotherapy against a waitlist improved parenting stress only, which stayed higher than in non-depressed mothers, and treatment response did not predict child outcomes (Forman et al., 2007). In a four-arm trial of 193 women with postpartum depression, all three therapies beat the control on early relationship problems and on mood at 4.5 months. Attachment, infant cognition and 5-year outcomes did not differ from control (Murray et al., 2003), and the mood benefit had gone by 9 months, with no difference in subsequent depressive episodes (Cooper et al., 2003).

The standing is mixed: pooled trials favour a child benefit, whereas the trials in perinatal samples show gains that are partial or short-lived. Interventions aimed at the parent-infant relationship are covered under Parent-Infant Interventions. Many of these child outcomes rest on maternal report, a limitation taken up in Validity of Laboratory and Observational Measures.

Breastfeeding and Infant Feeding

Breastfeeding and postpartum depression are linked in both directions, and the evidence has struggled to say which way the link mainly runs. Depressed mothers may feed differently, and feeding may in turn change mood. Separating the two needs repeated measurement across pregnancy and the months after birth, and most of the work has been observational.

Depression in pregnancy predicts shorter breastfeeding. A systematic review of 48 studies found that pregnancy depression went with a shorter duration of breastfeeding, though it did not predict breastfeeding intention or initiation. The same review found that postpartum depression and breastfeeding cessation were related in several studies, and the authors called for further prospective research (Dias & Figueiredo, 2015). Surveillance data point the same way from the other side. In a cross-sectional survey of 32,659 recent-birth women across 31 US sites, short breastfeeding (under 8 weeks) was a correlate of postpartum depressive symptoms. Those symptoms were self-reported rather than diagnosed (Bauman et al., 2020).

Does Breastfeeding Protect?

A prospective cohort of 205 women followed from pregnancy to 24 months gave some support to a protective role. More frequent breastfeeding at 3 months predicted a larger later decline in depressive symptoms. Any breastfeeding versus none showed only a trend (p = .07), and exclusive breastfeeding did not differ from formula feeding on depression. Prenatal depression predicted earlier weaning by 2.3 months, whereas depression at 3 months did not predict later breastfeeding. The authors described the relation as bidirectional, with protection possible, and asked for replication before the result is relied on. The sample was largely White and upper-middle class, depression was measured by self-report, and causality cannot be inferred (Hahn-Holbrook et al., 2013).

Intentions as a Moderator

Whether a mother meant to breastfeed changed what breastfeeding was associated with. In the ALSPAC longitudinal cohort of about 14,000 pregnancies, 65% of mothers intended to breastfeed and 80% initiated. Across the whole sample, breastfeeding and not breastfeeding differed little in depression risk. Among mothers who had planned to breastfeed, those who did had lower risk than those who did not, and those who planned to breastfeed but did not had higher risk than every other group. Among mothers who had not planned to breastfeed, those who did breastfeed had higher risk at 21 months than those who did not. The authors advised that their results be interpreted with caution, and concluded that women who want to breastfeed need expert breastfeeding support and that those who intended to but could not need compassionate support (Borra et al., 2015).

What Designs Can Settle

A narrative review of empirical studies judged conclusions about the direction of the association premature. Breastfeeding definitions varied between studies, depression measures were non-specific, and naturalistic designs limit causal inference. Some studies found formula feeding tied to more depressive symptoms and others found no association (Pope & Mazmanian, 2016). No study cited on this page reports a trial that decides the question.

Treatment While Breastfeeding

Women who are breastfeeding face a further question, which is whether treatment is compatible with feeding. A systematic review judged sertraline and paroxetine the safest serotonin reuptake inhibitors in lactation, with better neonatal safety data than other drugs in the class, and preferred them as first-line options for nursing women. Data on fluvoxamine, escitalopram and duloxetine were sparse, and few studies followed infant neurodevelopment (Orsolini & Bellantuono, 2015).

For zuranolone, a phase 1 open-label study of 15 lactating women (14 completed) found low transfer into breast milk. The relative infant dose at 30 mg was 0.357%, well under the 10% threshold, and a simulated 50 mg dose also fell below 1%. The authors concluded it may be considered an option for postpartum depression. The participants were healthy and not pregnant, 30 mg is not the approved dose, and the 50 mg figure was simulated (Deligiannidis et al., 2024).

Current Position

The direction of the link remains unresolved: depression in pregnancy appears to shorten breastfeeding, a protective effect of breastfeeding on mood is possible but unreplicated, and reviews judge the methods too inconsistent to decide. Intention to breastfeed emerges as a key moderator, since the mood risk attached to feeding differed by whether the mother had planned it.

Culture and Social Context

Where a woman gives birth, and who she is when she does, shapes both how often postpartum depression is reported and how likely she is to be treated. Most of the pooled evidence below comes from studies that used screening scales, and much of the older literature comes from high-income, English-language settings, so each estimate is read with that limit in mind.

Variation Across Countries

Reported prevalence differs enormously between countries. A literature review of 143 studies from 40 countries found rates running from almost 0% to almost 60%, and concluded that the familiar 10-15% figure is probably unrepresentative of the world (Halbreich & Karkun, 2006). Its authors proposed that cultural factors, reporting style, socioeconomic environment, social support and biological vulnerability might account for the spread, but these were proposals and not tested explanations. The review also noted that most of its studies came from Western developed countries and relied on brief, unidimensional instruments.

Meta-analyses that attach national-level predictors put numbers on the pattern. Across 291 studies of 296,284 women in 56 countries, pooled prevalence was 17.7%, with very high heterogeneity (I² = 98%); estimates ran from 3% in Singapore to 38% in Chile (Hahn-Holbrook et al., 2018). Income inequality (GINI index) explained 41% of the national variation, and the share of women working 40 hours or more a week explained 31%, with maternal and infant mortality explaining 19% and 16%. When GDP and GINI were entered together, only GINI retained a unique association. Maternity leave policy and cesarean rate were not significant predictors. The authors cautioned that self-report scales probably overestimate prevalence relative to interviews, and that larger studies reported lower prevalence, which suggests publication bias.

A larger synthesis of 565 studies, 1,236,365 women and 80 countries gave a pooled estimate of 17.22% (95% CI 16.00-18.51), again with extreme heterogeneity (I² = 99.70%) (Wang et al., 2021b). Prevalence was 15.54% in high-income countries, and 14.85% in developed against 19.99% in developing countries. Regional extremes were wide: Southern Africa 39.96% and Denmark 6.48%, with Afghanistan at 60.93%. Country development and income were associated with the heterogeneity, whereas geographic region, diagnostic scale and publication year were not. Most of the contributing studies were cross-sectional.

The picture is similar when perinatal depression is pooled across pregnancy and the postnatal period. A systematic review found common perinatal mental disorders frequent across low- and lower-middle-income countries, estimating their prevalence at 15.6% during pregnancy and 19.8% after birth, although antenatal figures came from only 9 countries and postnatal figures from 17 (Fisher et al., 2012). Pooling 140 estimates from 96 studies, a meta-regression put perinatal depression prevalence at 11.9%; prevalence ran higher in low- and middle-income countries than in high-income ones (odds ratio 1.8, 95% confidence interval 1.4 to 2.2) and higher when measured with symptom scales rather than diagnostic instruments (odds ratio 1.6), though the authors noted how few low-income countries the data contained (Woody et al., 2017).

Whether the phenomenon is recognised as illness everywhere is a separate question. A qualitative study across 15 centres in 11 countries, using focus groups and key informant interviews, found morbid unhappiness after childbirth in every centre, but its authors doubted that this could be equated with the illness category used in Western psychiatry (Oates et al., 2004). The term postnatal depression was used less in Portugal, Switzerland, Uganda and among UK Asians, and antidepressants were mentioned only in the USA. Informant numbers per centre were small and the findings cannot be generalised. The cross-national differences in rates therefore remain partly a matter of what is being measured, and the studies do not settle how much reflects true differences in illness.

Social Determinants and Disparities

Immigrant women are at higher risk. A meta-analysis of 22 studies found a pooled prevalence of postpartum depressive symptoms of 20% (95% CI 17-23%) in immigrant women, and higher odds than in non-immigrant women (unadjusted OR 2.10; adjusted OR 2.18; OR 1.63 after bias adjustment) (Falah-Hassani et al., 2015). Shorter residence, lower social support, poorer marital adjustment and insufficient household income were associated with symptoms. The authors noted evidence of publication bias and the need for clinical diagnosis to confirm the symptom findings.

Partner violence is one of the more consistent correlates. A meta-analysis of 67 papers linked domestic violence with perinatal depression, anxiety and post-traumatic stress; violence during pregnancy was associated with postnatal depression (OR 3.1), as was lifetime partner violence (OR 2.9) (Howard et al., 2013). Study heterogeneity was substantial, and the direction of causation was unclear.

Survey data from the United States show how widely symptoms vary within one country. Of 32,659 women with a recent birth surveyed across 31 sites in 2018 surveillance, 13.2% reported postpartum depressive symptoms, with site rates running from 9.7% (Illinois) up to 23.5% (Mississippi) (Bauman et al., 2020). Symptoms were more common among women aged 19 or younger than among those 35 or older (22.2% vs 10.8%), among Medicaid-insured than privately insured women (17.2% vs 10.1%), and among Black and Asian/Pacific Islander women than white women. Women reporting intimate partner violence had 33.1% prevalence against 12.5% for those who did not. The data were self-reported symptoms, not diagnoses, and carried recall and social desirability biases.

Treatment follows a different gradient. In a retrospective cohort of 29,601 New Jersey Medicaid mothers, 9% of white women began treatment, against 4% of Black and 5% of Latina women; adjusted odds of initiation were lower for Black (AOR .43) and Latina (AOR .59) women (Kozhimannil et al., 2011). The authors concluded that treatment was low overall and that the disparities warranted attention. Claims data lacked symptom severity and treatment preferences, did not require a recorded diagnosis, and omitted services outside Medicaid.

Age matters at the other end of the range. A review of 67 articles on adolescent mothers reported postpartum depression prevalence about double that of adult mothers, but called the literature limited and said randomised trials were needed (Dinwiddie et al., 2018).

Evidence on sexual minority mothers is thin. A cross-sectional survey of 49 lesbian and bisexual women who were pregnant, biological mothers or non-biological mothers found higher Edinburgh Postnatal Depression Scale scores in biological mothers than in a previously published heterosexual sample, suggesting perinatal depression is not rarer and possibly commoner in these women, with longitudinal work still needed (Ross et al., 2007). A qualitative meta-aggregation of 26 studies with 1,199 LGBTQ2S+ childbearing participants linked perinatal distress to stigma, heteronormativity and cisnormativity, and attributed it largely to systems-level inequities and exclusion (Kirubarajan et al., 2022).

Traditional Postpartum Practices

Many cultures prescribe a period of ritual care after birth. Reviewing 12 of 72 screened studies, Grigoriadis and colleagues found some evidence that postpartum rituals may guard against depression when the social support they bring is appropriate and wanted by the mother, but the included studies did not uniformly show a preventive effect of the rituals themselves (Grigoriadis et al., 2009). The review concluded that welcome support, and not the ritual itself, may be the active element. It also called the evidence limited, subject to many contextual factors, and in need of culturally sensitive systematic research. The question therefore remains open.

Stigma and Help-Seeking

Barriers to help-seeking are well described. Drawing on 40 articles published from 1966 to 2005, a qualitative systematic review found that help-seeking was not proactive among the women studied, and it named non-disclosure, myths about the condition and gaps in services as barriers; women preferred talking therapies over drug treatment, though few studies examined how acceptable treatments were (Dennis & Chung-Lee, 2006). Women's own accounts of disclosure and stigma appear in The Experience of Postpartum Depression, and the structure of services is taken up under Access to Care and Service Delivery.

Comorbidity

Postpartum depression often occurs alongside other psychiatric and medical conditions, and the associations reviewed below carry clinical weight for screening and treatment. Most of the evidence is cross-sectional or based on symptom scales, so the direction of effect is rarely established. The sections that follow cover anxiety, obsessive-compulsive disorder, post-traumatic stress, bipolar disorder, substance use, partner violence and medical conditions of pregnancy.

Anxiety Disorders

A meta-analysis of 66 studies (162,120 women, 30 countries) found that co-occurring anxiety and depression in the perinatal period was common, and it concluded that the pattern warrants clinical attention (Falah-Hassani et al., 2017). In that analysis prevalence did not differ significantly by publication year, country income or selection and attrition bias.

A separate meta-analysis of 102 studies (221,974 women, 34 countries) estimated postnatal anxiety symptoms at 15.0% (95% CI 13.7–16.4) in weeks 1 to 24, any postnatal anxiety disorder at 9.9% (95% CI 6.1–13.8) and generalised anxiety disorder at 5.7% (95% CI 2.3–9.2) (Dennis et al., 2017). Anxiety rates were higher in low- and middle-income countries than in high-income ones in that review, and the authors cautioned that published rates vary with measures, cut-offs and time periods. The two reviews disagree on whether country income matters; they analysed different outcomes (anxiety prevalence against anxiety-depression comorbidity), and the question remains open.

Obsessive-Compulsive Disorder

A meta-analysis of 29 studies reported mean obsessive-compulsive disorder prevalence of 1.08% in the general population, 2.07% in pregnancy and 2.43% postpartum, with an aggregate risk ratio of 1.79 (Russell et al., 2013). The authors concluded that risk is elevated in pregnancy and after birth, and they described the risk-ratio analysis as exploratory. The finding concerns obsessive-compulsive disorder in the postpartum period generally, not only among depressed mothers. Intrusive thoughts of harm to the infant are treated under their own heading.

Post-Traumatic Stress After Childbirth

A systematic review and meta-analysis of 59 studies (N = 24,267) found post-traumatic stress disorder prevalent in both pregnancy and the postpartum period, with higher rates in high-risk samples than in community samples (Yildiz et al., 2017). Rates varied widely, and the authors suggested that postpartum rates may rise if the disorder goes untreated.

A meta-analysis of 50 studies (21,429 women, 15 countries) examined risk factors for birth-related post-traumatic stress (Ayers et al., 2016). Depression in pregnancy, a negative subjective birth experience, operative birth, fear of childbirth, a history of the disorder, dissociation, and coping difficulties and stress after birth were positively associated with it, and staff support during birth was negatively associated. Partner presence, marital status, socio-economic status and (in two studies) time since birth were not significant. The review rested on small numbers of studies for several factors, noted possible publication bias and offered only a preliminary basis for screening and prevention.

Bipolar Disorder

A narrative review reported that between 21.4% and 54% of women with postpartum depression received a bipolar disorder diagnosis, with the figure varying by the population studied (Sharma et al., 2017). Younger onset age, a first depressive episode after childbirth, immediate postdelivery onset, atypical, psychotic or mixed features and a family history of bipolar disorder were associated with bipolar postpartum depression. The review found no controlled treatment data and advised prompt identification of the bipolar form. The bipolar presentation is described further under Types of Postpartum Depression.

Substance Use

A literature review of 12 studies, whose prevalence studies were cross-sectional, found substance use common after birth and depressive symptoms frequent among users (Chapman & Wu, 2013). Postpartum alcohol use (30.1% to 49%) was lower than in women neither pregnant nor postpartum but higher than in pregnancy, and the same ordering held for drug use (4.5% to 8.5%). Depressive symptoms predicted substance use, whereas alcohol use did not significantly predict depressive symptoms and depressive symptoms did not significantly predict at-risk drinking. The review noted scarce data, small samples and no prevalence figures for substance use among women with postpartum depression, and it called for research on reducing use.

Partner Violence

A meta-analysis of 67 papers linked perinatal depression, anxiety and post-traumatic stress to domestic violence (Howard et al., 2013). Violence during pregnancy was associated with postnatal depression (OR 3.1, 95% CI 2.7–3.6), and lifetime partner violence with postnatal depression at OR 2.9 (95% CI 1.8–4.8). The authors reported a significant association but judged causal direction unclear, with heterogeneous and variable-quality studies, and they advised stronger evidence.

Medical and Obstetric Conditions

Pooling 10 cohort studies with 2,000,002 women, a meta-analysis linked gestational diabetes to postpartum depressive symptoms with a relative risk of 1.32 (95% confidence interval 1.09 to 1.60), though heterogeneity across the studies was high (Arafa & Dong, 2019). The association was stronger in prospective cohorts (RR 1.95) than in retrospective ones (RR 1.16), and the authors called for well-designed prospective studies.

A systematic review of 13 studies, 8 of which found an association, concluded that preeclampsia is a risk factor for postpartum depression and that severity is linked to it, but the studies were highly heterogeneous and the authors advised caution (Caropreso et al., 2020).

A meta-analysis of five studies found thyroid peroxidase antibody positivity associated with higher postpartum depression risk (RR 1.49, 95% CI 1.11–2.00; adjusted RR 1.41) (Minaldi et al., 2020). Possible publication bias applied, and confirmation in further well-designed studies was recommended.

Direction of Effect

Across these conditions the evidence mostly shows co-occurrence. Only the gestational diabetes meta-analysis separated prospective from retrospective designs, and the other reviews either relied on cross-sectional data or stated that causal direction was unclear. Whether these conditions precede, follow or share causes with postpartum depression therefore remains open.

Suicide and Self-Harm

Suicidal thoughts and self-harm are among the most serious concerns attached to perinatal depression. This section describes how common they are, who appears to be at greater risk, how they are assessed and what is known about prevention. It gives no details of methods.

How Common Are Suicidal Thoughts, Self-Harm and Death

Thoughts of self-harm are more common than attempts or deaths (Lindahl et al., 2005). That review also reported that suicide deaths and attempts in the perinatal period were lower than in the general female population, while depressed women faced a higher risk of suicidality than nondepressed women. Definitions and measurement differed across the studies it covered.

Rare does not mean negligible. A UK confidential enquiry covering 1997 to 1999 found that suicide was a leading cause of maternal death, and its author concluded that such deaths are rare but more frequent than had been thought (Oates, 2003). In Ontario, a population-based cohort identified 51 perinatal suicides, a rate of 2.58 per 100 000 live births that the authors judged comparable to international figures; the sample was small and the data sources incomplete (Grigoriadis et al., 2017).

Recorded suicidality has also risen. In a US commercial claims database, diagnoses of suicidal ideation around delivery went from 0.1% to 0.5% between 2006 and 2017, and diagnoses of intentional self-harm from 0.1% to 0.2% (both P < .001). The increase was larger among non-Hispanic Black, lower-income and younger individuals, and the data covered a commercially insured population only (Admon et al., 2021).

Who Is at Risk

A UK national inquiry compared women who died by suicide in the perinatal period with other women who died by suicide: the perinatal group more often had a diagnosis of depression (adjusted odds ratio 2.19) and was less often receiving active treatment (odds ratio 0.46). Service contact and treatment adherence did not differ significantly. Suicides among other women fell over the period, while the perinatal trend was flat and not significant (Khalifeh et al., 2016).

Risk also extends beyond the first weeks. In a Danish register cohort, women with a first-onset severe psychiatric disorder after childbirth had elevated mortality rate ratios, and the first year after diagnosis carried a particularly high relative risk of suicide (Johannsen et al., 2016). A review of postpartum psychosis likewise found suicide rare during the acute episode, higher later and familial (Brockington, 2017); see Types of Postpartum Depression for that condition.

Assessment

Screening can bring self-harm thoughts to light. Among women who screened positive on the Edinburgh Postnatal Depression Scale (a score of 10 or higher) in a case series of ten thousand postpartum mothers, 19.3% had self-harm ideation (Wisner et al., 2013). In a cohort nested in a trial, 9% of 4,150 screened women reported any suicidal ideation, and 4% reported it sometimes or quite often. The scale's self-harm item flagged suicidality, yet ideation at baseline did not significantly predict depression or mental health scores at 18 weeks, so a link with later outcome was judged unlikely (Howard et al., 2011). The item therefore serves as a prompt for a clinical conversation, not as a forecast of course.

Prevention

No study cited on this page reports a prevention trial specific to the perinatal period, and the inquiry authors described prevention evidence as scarce (Khalifeh et al., 2016). The general literature is the best guide available. A systematic review found that restricting access to means and school awareness programmes reduced suicide or suicidal behaviour, that evidence was insufficient for primary care screening, and that no single strategy was clearly superior (Zalsman et al., 2016). Whether these approaches transfer to new parents remains open. Providers are advised to stay vigilant and to follow up assertively (Grigoriadis et al., 2017; Khalifeh et al., 2016).

Anyone who is thinking about suicide, or who is worried about a new parent, can contact local emergency services or a crisis line; in the United States the 988 Suicide and Crisis Lifeline answers calls and texts to 988. The International Association for Suicide Prevention links to Find A Helpline, a directory of verified helplines in more than 150 countries.

Intrusive Thoughts of Harm to the Infant

Unwanted intrusive thoughts of harm to the baby are a different thing from an intention to harm and from the beliefs and commands of psychosis. An intrusive thought arrives uninvited and is experienced as alien and upsetting; the parent does not wish to act on it. The distinction matters because the evidence below belongs to different categories, and reading one as another produces false alarm. This section follows the evidence on how common the thoughts are, what they predict, and where the separate risk literature on psychosis belongs. Anyone troubled by such thoughts can speak to a clinician, and the thoughts themselves are common.

The thoughts are common after childbirth. In a prospective cohort of one hundred women recruited in pregnancy and assessed three times, including at 4 and 12 weeks postpartum, intrusive thoughts of harm to the infant were a relatively normative experience (Fairbrother & Woody, 2008). Thoughts of accidental harm were more frequent than thoughts of intentional harm and took up more of the mother's time. Thoughts of intentional harm were the more distressing of the two. Higher parenting stress went with more thoughts of intentional harm, and greater social support went with fewer. The authors found little evidence that the thoughts were linked to aggressive parenting. The authors noted that few studies had been published on the topic.

Depression raises the frequency. In a case-control comparison of 100 clinically depressed mothers and 46 nondepressed mothers of a child under 3 years, 41% of the depressed group reported thoughts of harming the infant, against 7% of the comparison group (Jennings et al., 1999). Demographic, psychosocial-stressor and psychiatric variables did not significantly predict these thoughts. The comparison has limits: the controls lacked a full diagnostic interview, the groups were not selected identically, and two areas were not assessed in the controls. The authors read the result as a possible pathway through which maternal depression affects the child; the study does not show that the thoughts led to harm.

What the thoughts predict is the question parents and clinicians most need answered. In a province-wide prospective cohort, neither unwanted intrusive thoughts of infant harm nor obsessive-compulsive disorder was significantly associated with physical aggression towards the infant, and the authors found no evidence of increased risk of harm (Fairbrother et al., 2022). Of 763 women, 388 provided data, which limits how far the result generalises. Aggression was reported by 2.9% of the sample (95% CI, 1.5% to 4.7%), below the 4% to 9% found in other samples.

Obsessive-compulsive disorder itself is more prevalent after birth. A meta-analysis of 29 studies found mean prevalence of 1.08% in the general population, 2.07% in pregnancy and 2.43% postpartum, so risk was elevated in both periods and slightly higher postpartum than in pregnancy (Russell et al., 2013). The authors described the risk-ratio analysis as exploratory.

The evidence on filicide belongs to a separate category, and it concerns psychosis, not unwanted intrusive thoughts. A review of childbearing psychoses, combined with the author's own series of 321 cases, found filicide concentrated in psychoses with depressive features: 24 cases among 535 depressive psychoses (4.5%), against less than 1% in psychoses without depression (Brockington, 2017). The author cautioned that the rates reflect lifetime risk, that some filicides may have been accidental, and that part of the evidence was an unpublished compilation of cases. These figures say nothing about mothers who have intrusive thoughts and no psychosis, and the cohort data above indicate no raised aggression for them.

A qualitative systematic review of 40 articles (1966 to 2005) found that women with postpartum depression did not proactively seek help, and it listed nondisclosure, myths and service gaps among the barriers (Dennis & Chung-Lee, 2006).

Postpartum depression is one of several psychiatric states that can follow childbirth. A narrative review covering five years of research characterised the psychiatric disorders of the puerperium as a wide variety, with postpartum depression one member of it (Brockington, 2004). The subsections below take the neighbouring states in turn and give, for each, the feature that separates it from postpartum depression as far as the evidence allows. Where the evidence supplies no workable rule, the text says so.

The Postpartum Blues

The postpartum blues are short-lived and generally regarded as normal. Pitt described them as a transitory tearfulness of the early puerperium that was commonplace on maternity wards, and set against that familiarity the fact that postnatal depression itself had been little evaluated (Pitt, 1968).

Timing is the discriminator. According to a narrative review of perinatal mental illness, the blues begin in the first week postpartum, peak between day 3 and day 5 and have largely diminished by days 10 to 12, and they are milder than depression (O'Hara & Wisner, 2014). Tearfulness or low mood that continues past that window has outlasted the blues as the review describes them. The rule is rougher than it looks, because the same review found no established criteria for the blues: reported prevalence runs from 26% to 84% according to the criteria applied (O'Hara & Wisner, 2014).

Whether the blues are only a normal passage has been questioned. Henshaw's narrative review found them more likely in women who had been depressed before, whether premenstrually or at other times, and in women who were dysphoric during the pregnancy; neuroticism was a further correlate. Both the blues and postpartum elation were positively related to later postpartum depression, and the review treated the two as an apparent index of affective vulnerability, while noting that the biological research was inconclusive (Henshaw, 2003). The later review by O'Hara and Wisner keeps both halves of the picture: the blues are a positive predictor of postpartum depression, and postpartum depression is nonetheless mostly independent of them (O'Hara & Wisner, 2014). The current position is therefore that the blues are a mild, brief state that marks raised risk and is not the usual route into the disorder. They are also described under Types of Postpartum Depression.

Fatigue and Broken Sleep

Tiredness and disturbed sleep are scored as depressive symptoms on symptom inventories, so the question is whether they still indicate depression around childbirth. The evidence available here answers it for pregnancy, not for the weeks after birth. Nylen and colleagues compared 255 pregnant women with 820 women from five community-based samples on the eight scales of the Inventory of Depression and Anxiety Symptoms, using multigroup confirmatory factor analysis on cross-sectional data (Nylen et al., 2013). A single depression factor fitted adequately in both groups, but the scales did not all behave alike. Lassitude and Appetite Gain loaded less strongly on that factor in the pregnant women than in the community women, Insomnia and Well Being loaded more strongly, and the two appetite scales were poor indicators of depression in pregnancy. The authors concluded that somatic symptoms, appetite apart, are valid and usable in assessing depression prenatally (Nylen et al., 2013).

For pregnancy, then, the workable rule is that somatic complaints are not to be discounted wholesale: tiredness counted for less as a sign of depression than it did in community women, sleeplessness for more, and appetite change for little. The study did not include postpartum women, so it does not show how fatigue and broken sleep after delivery should be weighed. Sleep after birth is the subject of Sleep and Circadian Rhythms.

Perinatal Anxiety and Obsessions of Infant Harm

Anxiety around childbirth is common in its own right. A systematic review and meta-analysis of 102 studies, covering 221 974 women in 34 countries, estimated the prevalence of postnatal anxiety symptoms at 15.0% (95% CI 13.7–16.4) from 1–24 weeks after birth, of any postnatal anxiety disorder at 9.9% (95% CI 6.1–13.8) and of postnatal generalised anxiety disorder at 5.7% (95% CI 2.3–9.2); 57 of the studies supplied postnatal data (Dennis et al., 2017). Rates were higher in low- and middle-income countries than in high-income ones. Anxiety in the perinatal period is common, the authors concluded, and appears higher in low- and middle-income countries, though they cautioned that reported prevalence varies and that the pooled studies differed in their measures, cut-offs and time windows (Dennis et al., 2017). These figures describe anxiety as an outcome of its own, which is the sense in which it is a related condition and not a variety of postpartum depression. The overlap between the two is treated under Comorbidity.

Obsessions of harm to the child appear among the puerperal psychiatric disorders in Brockington's review (Brockington, 2004). Unwanted thoughts of harm coming to the infant are not, however, a sign of disorder in themselves. A prospective cohort study followed one hundred women recruited in pregnancy across three assessments, with postpartum data at 4 and 12 weeks, using questionnaires and a semi-structured interview, and found intrusive thoughts of infant harm to be a relatively normative experience after birth (Fairbrother & Woody, 2008). Thoughts of accidental harm were more frequent and took up more time than thoughts of harming the infant intentionally, while the intentional thoughts were the more distressing. Thoughts of intentional harm went with higher parenting stress and lower social support, and the study found little evidence tying intrusive thoughts to aggressive parenting; its authors noted that few studies had been published on the question (Fairbrother & Woody, 2008). The rule a reader can apply is that the presence of such thoughts does not, taken alone, separate a disorder from ordinary postpartum experience. The topic has its own section, Intrusive Thoughts of Harm to the Infant.

Post-Traumatic Stress After Birth

Post-traumatic stress after childbirth differs from postpartum depression in being anchored to an event, the birth itself. The fullest evidence is a meta-analysis of 50 studies covering 21 429 women in 15 countries, which found birth-related PTSD associated with factors from before, during and after the birth (Ayers et al., 2016). How the woman herself experienced the birth, an operative delivery and dissociation were positive correlates, and support from staff during the birth a negative one. Earlier vulnerability counted as well: depression in pregnancy, fear of childbirth and a history of PTSD were each associated with more symptoms, as was stress after the birth. Several plausible candidates were weak or absent: the presence of a partner, marital status and socio-economic status were non-significant, age had a very small though significant association (r = -0.03), and time since birth was non-significant on the evidence of two studies only; after trim and fill, the age association became non-significant and the socio-economic association became significant (Ayers et al., 2016).

The applicable discriminator is therefore the reference point of the symptoms and the instrument that captures them: the pooled studies used trauma measures such as the Posttraumatic Diagnostic Scale, the Impact of Event Scale and the Traumatic Event Scale. The two conditions are not exclusive, since depression in pregnancy was a correlate and the meta-analysis treats co-morbidity with postpartum depression as one of its topics. Its authors offered the results as a preliminary basis for screening and prevention, and noted that publication bias is possible, that some estimates rest on small numbers of studies, that the diverse measures were hard to combine and that little of the research came from low-income countries (Ayers et al., 2016).

Impaired Bonding

Difficulty in a mother's relationship with her infant is a problem separable from her mood. Brockington's review counts disorders of the mother-infant relationship among the puerperal psychiatric disorders in their own right (Brockington, 2004), and a screening questionnaire for mother-infant bonding disorders has been published (Brockington et al., 2001). The discriminator this offers is one of assessment: bonding is screened with an instrument of its own. How depression bears on the interaction between mother and infant is taken up under Mother-Infant Interaction.

Bipolar Depression

A depressive episode after childbirth may belong to bipolar disorder and not to unipolar depression. A narrative review that searched three databases reported that between 21.4% and 54% of women with postpartum depression had a diagnosis of bipolar disorder, a range that varies with the population studied (Sharma et al., 2017). The same review identified features that point towards the bipolar form. Timing is one: onset immediately after delivery, and a first-ever depressive episode arriving after childbirth, were both associated with bipolar postpartum depression. A family history of bipolar disorder and a younger age at onset point the same way, as do atypical, psychotic or mixed features (Sharma et al., 2017).

The early weeks matter for bipolar illness more generally. In a separate narrative review, the risk of admission for bipolar disorder was higher in the first 30 days after birth than in days 30 to 60 (O'Hara & Wisner, 2014). Sharma and colleagues concluded that bipolar postpartum depression should be identified promptly, while acknowledging that no controlled treatment data exist for it (Sharma et al., 2017).

Thyroid Autoimmunity

Thyroid autoimmunity is a medical condition, defined by antibody status and not by mood, that has been linked to depression after birth. Five studies entered a systematic review and meta-analysis that set 449 women carrying anti-thyroperoxidase antibodies (TPOAb) against 2483 antibody-negative women; under a random-effects model, the positive group had a higher postpartum depression risk, with a relative risk of 1.49 (95% confidence interval 1.11 to 2.00; P = 0.008) (Minaldi et al., 2020). Heterogeneity between studies was modest (I² = 47%, P for heterogeneity = 0.11). Publication bias was possible: a trim-and-fill analysis pointed to 2 putative missing studies, although the adjusted estimate (RR 1.41, 95% CI 1.18-1.68, P = 0.0002) changed negligibly from the unadjusted one. The authors concluded that thyroid autoimmunity is associated with the risk of postpartum depression and that further well-designed studies are needed to confirm it (Minaldi et al., 2020). The result is an association with risk drawn from five studies; it does not show that thyroid autoimmunity produces the depressive symptoms, and the question awaits confirmation.

Prevention

Onset of postpartum depression follows a datable event, so the period before and just after birth offers a window for acting before an episode starts. Prevention is therefore a question of its own, separate from treating an established episode (see Treatment Evidence). The evidence below covers psychological and psychosocial programmes, a test of one proposed mechanism, medication after a previous episode, the choice to continue or stop antidepressants in pregnancy, and prophylaxis in women with bipolar disorder or earlier postpartum psychosis.

Psychological and Psychosocial Prevention

Pooled evidence favours these programmes. A Cochrane review of randomised trials, covering almost 17,000 women, found that psychosocial and psychological interventions reduced the risk of postpartum depression (average RR 0.78, 95% CI 0.66 to 0.93) (Dennis & Dowswell, 2013). Professional-based and lay-based support were both effective in that review. Risk was also reduced in trials that selected women at risk (RR 0.66), delivered contact individually (RR 0.75), involved several contacts (RR 0.78) or began after birth (RR 0.73) (Dennis & Dowswell, 2013).

A meta-analysis of 37 randomised or quasi-randomised trials reached a similar verdict, though a modest one. Preventive interventions produced a small reduction in depressive symptoms (Hedges' g = 0.18) and 27% fewer episodes. Later timing of the postpartum assessment went with smaller differences between intervention and control, and after outliers were excluded the effect on symptoms was in the small range (Sockol et al., 2013).

A systematic review and meta-analysis of 50 studies (N = 22 385), done for the US task force, found that counselling lowered the onset of perinatal depression (RR 0.61, 95% CI 0.47-0.78). Most of that evidence came from women at increased risk. Health-system interventions did not show a significant pooled effect (O'Connor et al., 2019). On that evidence the task force gave counselling, with cognitive behavioural and interpersonal therapy as its examples, a B recommendation for women at increased risk. It rated certainty and net benefit as moderate, and judged the evidence on non-counselling interventions inadequate (Krist et al., 2019).

Telephone peer support has its own trial. In a multisite Ontario trial, 701 high-risk women were randomised after 21 470 had been screened. At twelve weeks, 14% (40/297) of women given peer support scored above 12 on the Edinburgh Postnatal Depression Scale, against 25% (78/315) of controls (P<0.001). At twenty-four weeks the groups did not differ significantly (11% against 14%, P=0.11), which the authors attributed to referring women identified with clinical depression at 12 weeks for treatment (Dennis et al., 2009). The authors treated the result as possible prevention in high-risk women. They noted that the interview data were questionable, since the interview was shortened, held by telephone and not validated (Dennis et al., 2009).

Taken together, the position is that counselling-type and support-based programmes lower risk, with the clearest effects among women already at increased risk and with benefits that fade at longer follow-up.

How Prevention Might Work

A trial of prenatal affective cognitive training tested one proposed route, in which changing negative cognitive bias before birth would protect against depression. It randomised 147 high-risk pregnant women to five 90-minute sessions or care as usual. The primary outcome, postpartum depression overall, did not differ between the groups (p = 0.18), nor did depressive symptoms (p = 0.31). Positive interpretation of infant emotions was better in the training group. In multiparous women alone, incidence was 5% against 35% (OR = 0.10, 95% CI 0.01-0.92). The authors called this parity finding exploratory and asked for replication in larger trials, and follow-up reached only 83% (Bjertrup et al., 2026). The cognitive route is thus untested in any confirmed sense. Compare The Cognitive Model in the Perinatal Period and Testing Whether Cognition Is Cause or Consequence.

Pharmacological Prevention After a Previous Episode

Medication given to prevent a recurrence has little support. A Cochrane review found two small trials in 81 women with earlier postnatal depression. Nortriptyline did not differ from placebo (RR 0.96, 95% CI 0.36 to 2.59). Sertraline had a lower rate in the other trial (RR 0.14), but the interval included 1 (95% CI 0.02 to 1.07). The reviewers rated the evidence very low in quality and drew no clear conclusion, asking for larger trials (Molyneaux et al., 2018). The small sertraline trial also appears in the task force review, where it showed a lower recurrence rate (7% against 50%) and more maternal adverse effects (O'Connor et al., 2019).

Continuing or Stopping Antidepressants in Pregnancy

Two prospective cohorts point in different directions. In a naturalistic study of 201 pregnant women with a history of major depression, 68% of those who discontinued antidepressants relapsed, against 26% of those who kept taking them (HR 5.0, 95% CI 2.8-9.1). The authors concluded that pregnancy is not protective against relapse. The study was not randomised, and miscarriage and loss to follow-up reduced the sample (Cohen et al., 2006).

A later prospective cohort of 778 pregnant women with a history of depression found no significant association between antidepressant use and the risk of an episode (HR = 0.88, 95% CI 0.51-1.50). Its authors warned that the design was observational, that confounding by indication was possible, and that medication use was self-reported (Yonkers et al., 2011). Four or more earlier episodes and an episode in the 6 months before the pregnancy each predicted a new episode in that cohort (Yonkers et al., 2011).

Neither study was randomised, and the two disagree on whether stopping raises the risk. The question remains open. The ACOG guideline on perinatal psychiatric drugs uses a modified GRADE framework and issues Good Practice Points where the evidence is inadequate (American College of Obstetricians and Gynecologists, 2023).

Prophylaxis in Bipolar Disorder and After Postpartum Psychosis

Women with bipolar disorder or earlier postpartum psychosis form a separate group, with higher stakes. A meta-analysis of 37 articles (5,700 deliveries, 4,023 patients) put the postpartum relapse rate at about a third (35%, 95% CI 29-41). In bipolar disorder, relapse was 66% without medication and 23% with prophylaxis in pregnancy, and the authors concluded that prophylaxis in pregnancy appears highly protective. Data were insufficient for women with both bipolar disorder and a postpartum history (Wesseloo et al., 2016).

A clinic cohort of 70 high-risk pregnant women, followed over 2003 to 2010, compared prophylaxis timing. Among women whose only history was postpartum psychosis, those who took lithium postpartum had no relapse (0%), against 44.4% of those who declined it. Women with bipolar disorder had more relapse in pregnancy (24.4%), which led the authors to favour continuous prophylaxis for them. The authors noted that the timing of prophylaxis remains controversial and that fetal exposure carries a risk (Bergink et al., 2012).

Experimental Preventives

Low-dose esketamine given after delivery is being studied as a preventive. It is covered under Emerging Treatments.

Treatment Evidence

The evidence for treating postpartum depression is stronger for psychological treatment than for drugs, and it is weakest on how long the benefit lasts. This section sets out each family of treatment with its effect size, the certainty the reviewers assign, and the caveat that most limits it. Why the treatments work is taken up in How Treatments Work, and the newer neuroactive steroid and brain stimulation options in Neuroactive Steroid Drugs, Hormones and Brain Stimulation and Emerging Treatments.

Psychological Treatments

Psychotherapy outperforms control conditions for perinatal depression. A meta-analysis of 43 randomised trials (6,270 participants, 49 comparisons) found a pooled effect of g = 0.67 (95% CI 0.45 to 0.89), with a number needed to treat of 4.39 (Cuijpers et al., 2023). The authors concluded that psychotherapy is probably effective. Three caveats qualify the figure: heterogeneity was high (I2 = 80%), there was some publication bias, and the estimate fell to g = 0.53 after adjustment for it. Effects were larger in samples with a diagnosed depression than in those defined by a symptom cut-off, and larger for control conditions of the waiting-list type than for care as usual. Antenatal and postnatal timing, high-risk status and the type of therapy did not significantly change the effect, and cognitive behavioural therapy, interpersonal therapy and supportive therapy did not differ from one another.

Individual approaches have their own trials. A meta-analysis of 40 randomised and quasi-randomised trials, spanning pregnancy to the end of the postpartum year, found that cognitive behavioural therapy reduced depressive symptoms relative to controls, in treatment and in prevention trials, and concluded that the evidence for its efficacy is strong; study quality varied widely, and the authors called for more rigorous trials (Sockol, 2015). Interpersonal psychotherapy was tested against a waiting list in 120 postpartum women with major depression. After 12 weeks, Hamilton scores fell from 19.4 to 8.3 with therapy and from 19.8 to 16.8 on the waiting list, and Beck scores fell from 23.6 to 10.6 against 23.0 to 19.2; some women did not complete the protocol (O'Hara et al., 2000). A later review of 28 studies found that interpersonal therapy lowered depressive symptoms in prevention and treatment trials and lowered anxiety in treatment trials, although moderator results were inconsistent across outcomes (Sockol, 2018).

Behavioural activation has been tested in pregnancy. In a pragmatic trial of 163 depressed pregnant women, activation therapy lowered depressive symptoms more than treatment as usual (d = 0.34), and remission was higher (56.3% versus 30.3%). Outcomes were self-reported only, and the authors described it as the first test of the approach in pregnancy (Dimidjian et al., 2017).

Whether these gains last is the open question. A four-arm trial of 193 primiparous women with postpartum depression, followed for five years, found that all three therapies lowered Edinburgh scores more than routine care at 4.5 months, but the benefit was gone by 9 months. Subsequent episodes were similar in treated and untreated women (28% versus 27%, not significant), and 73% were assessed at five years (Cooper et al., 2003). The meta-analysis gives a somewhat longer horizon, with g = 0.41 at 6 to 12 months and a non-significant effect at 13 to 24 months (Cuijpers et al., 2023). Durability beyond the first months is therefore not shown, and no trial in the evidence reviewed here establishes a lasting preventive effect on later episodes.

Antidepressants

The trial evidence for antidepressants specific to postpartum depression is thin. A Cochrane review of 11 trials (1,016 postnatal women) concluded that selective serotonin reuptake inhibitors may be more effective than placebo, with low or very low certainty. Response was 55% versus 43% (RR 1.27, 95% CI 0.97 to 1.66) and remission 42% versus 27% (RR 1.54, 95% CI 0.99 to 2.41), and both intervals include 1. The symptom-scale difference was SMD -0.30 (95% CI -0.55 to -0.05). Acceptability and adverse effects did not differ between groups. The reviewers noted few studies, high attrition and little data on severe depression, on children and on parenting (Brown et al., 2021).

An earlier double-blind trial gave 87 postnatal women (61 completed) fluoxetine or placebo, with six sessions or one session of cognitive-behavioural counselling. Fluoxetine and fuller counselling each helped, and all four groups improved. The drug-by-counselling interaction was not significant, so combining them seemed to offer no advantage; 30% did not complete treatment (Appleby et al., 1997). A 2026 network meta-analysis of 18 trials (N = 1,722) ranked fluoxetine (SMD -0.95, 95% CI -1.43 to -0.47) among five agents that improved symptom severity over placebo, alongside saffron, esketamine, brexanolone and zuranolone. Its authors rate the evidence as low to moderate certainty, with sparse networks and exploratory rankings, and advise cautious interpretation (Xiao et al., 2026).

One further caveat concerns antidepressant trials in general, not postpartum trials specifically. A comparison of 74 FDA-registered trials of 12 antidepressants with their published reports found that selective publication inflated apparent efficacy: 94% of published results were positive against 51% in the FDA's reviews, and published effect sizes were 32% larger overall. The authors left the cause of the bias undetermined (Turner et al., 2008). None of the studies cited here tests whether the same distortion affects postpartum trials.

Antidepressants are standard treatment, yet the evidence specific to postpartum depression remains thin and of low certainty.

Medication in Pregnancy and While Breastfeeding

Guidance on psychiatric drugs in the perinatal period is graded. A clinical practice guideline applies a modified GRADE framework to pregnant and postpartum individuals, and it issues Good Practice Points where the evidence is inadequate (American College of Obstetricians and Gynecologists, 2023). Whether to continue an antidepressant through pregnancy is disputed. A naturalistic cohort of 201 pregnant women with a history of major depression found relapse in 68% of those who discontinued medication and 26% of those who maintained it (hazard ratio 5.0, 95% CI 2.8 to 9.1), which led the authors to conclude that pregnancy is not protective against relapse. The study was not randomised, and miscarriage and loss to follow-up reduced the sample (Cohen et al., 2006). A prospective cohort of 778 pregnant women with a depression history found no significant difference in episode risk between antidepressant users and non-users (HR 0.88, 95% CI 0.51 to 1.50), and its episode rate was lower than in the earlier cohort. That study was observational, could not attribute causality, and was open to confounding by indication and to self-reported medication use (Yonkers et al., 2011). The two cohorts disagree, no randomised comparison settles the question, and it remains open.

For lactation, a systematic review judged sertraline and paroxetine the safest serotonin reuptake inhibitors for the neonate and the preferred first-line choices for nursing women, with little data on fluvoxamine, escitalopram and duloxetine. Few studies followed infant neurodevelopment (Orsolini & Bellantuono, 2015). Infant exposure through milk is covered in Breastfeeding and Infant Feeding.

Antidepressant Exposure in Pregnancy

The other half of the continue-or-stop decision concerns what the drugs might do to the fetus. Pooling 23 studies of prenatal antidepressant exposure, one meta-analysis linked exposure to preterm delivery, a shorter gestation, lower birth weight and lower Apgar scores at one and five minutes, while the association with spontaneous abortion fell short of significance (P = .055) (Ross et al., 2013). The group differences were small, their clinical significance has to be weighed, and the review drew only on English-language studies; when exposed infants were compared with those of depressed mothers who took no medication, the birth-weight association was no longer significant, whereas the associations with gestational age, preterm delivery and Apgar scores remained (Ross et al., 2013). For birth weight, that last comparison is consistent with confounding by indication: women who take antidepressants differ from those who do not in the depression that prompted the prescription, so the illness and the drug are hard to pull apart. A claims-database cohort of 145 702 exposed and 3 032 745 unexposed pregnancies found crude associations with neurodevelopmental disorders reaching up to a doubling of risk, but these vanished after full adjustment and in comparisons of exposed with unexposed siblings, which hold shared family background constant (any disorder, HR 0.97, 95% CI 0.88-1.06), suggesting that antidepressant use itself probably does not raise the risk (Suarez et al., 2022). Exposure in that cohort was inferred from prescription fills and so may be misclassified, and the sibling analyses were too small to examine individual drugs (Suarez et al., 2022). Any such exposure risk is therefore weighed against the risk that depression left untreated poses to mother and child.

Does Treatment Reach the Child?

Treating the mother does not automatically improve outcomes for the child. A meta-analysis of nine trials in depressed mothers found that psychotherapy lowered maternal depression (g = 0.66) and was also linked to better child mental health (g = 0.40), mother-child interaction (g = 0.35) and parenting or marital distress (g = 0.67), though study quality was not optimal and outcome instruments differed considerably (Cuijpers et al., 2015). In a trial of interpersonal therapy against a waiting list, treatment improved parenting stress only, which stayed above that of non-depressed mothers, and treatment response did not predict child outcomes, so the authors argued that treatment should also target the mother-infant relationship (Forman et al., 2007). The child findings are set out in Child Outcomes, and interventions aimed at the relationship in Parent-Infant Interventions.

How Treatments Work

Showing that a treatment lowers depression is a different task from showing why it does. This section takes each treatment in turn and sets out the mechanism proposed for it, the test that has been run, and how much evidence there is that the proposed process carries the benefit. Outcome evidence is covered under Treatment Evidence, and the hormonal and stimulation treatments are covered under Neuroactive Steroid Drugs, Hormones and Brain Stimulation.

Behavioural Activation

Behavioural activation is the one treatment in this literature with a mediation test. A pragmatic trial compared it with treatment as usual in 163 depressed pregnant women (86 and 77 per arm) and found lower depressive symptoms with activation (d = 0.34), alongside higher remission (56.3% vs 30.3%) (Dimidjian et al., 2017). The women in the activation arm also showed greater gains in activation (d = 0.69) and in environmental reward (d = 0.54), and early change in each of the two statistically mediated the depression outcome (Dimidjian et al., 2017). Two limits apply. The outcomes were all self-report measures, and the authors present the trial as the first test of behavioural activation in pregnancy, so the finding rests on a single pregnancy sample (Dimidjian et al., 2017).

Interpersonal Psychotherapy

For interpersonal psychotherapy, the mechanisms are hypotheses. A theory paper proposed four interpersonal routes of change: enhancing social support, reducing interpersonal stress, processing emotion and improving interpersonal skills. It stated that little is known about how the therapy works and that these candidate mediators need testing in trials (Lipsitz & Markowitz, 2013). The review reports no trial that has tested them.

Cognitive Routes

The cognitive account locates the problem in negative beliefs and biased processing, so training aimed at those processes is a direct test of it. One randomised trial of prenatal affective cognitive training (five 90-minute sessions over five weeks) enrolled 147 high-risk pregnant women and compared it with care as usual. The trial's primary outcome was not significant: neither the overall rate of postpartum depression (p = 0.18) nor depressive symptoms (p = 0.31) differed with the intervention (Bjertrup et al., 2026). The training did shift how women interpreted infant emotions, in a more positive direction than care as usual (Bjertrup et al., 2026). Among multiparous women only, incidence was lower after training (5% vs 35%), but the authors treat the parity effect as exploratory and call for replication in larger trials, with follow-up at 83% (Bjertrup et al., 2026). A change in the targeted process without a change in depression is the opposite of what a mediator account needs, so this trial does not support a cognitive mechanism for depression, although a single trial does not rule one out.

Beliefs about motherhood are a candidate target. In two online cross-sectional studies of first-time mothers (136 and 288), attitudes toward motherhood were positively related to depression and anxiety symptoms, and the relation to depression held beyond general dysfunctional attitudes, social support and marital satisfaction (β = 0.15) (Sockol et al., 2014). The authors suggest a possible perinatal-specific risk factor. The design cannot show causation, the samples were small and mostly white and affluent, and a longitudinal replication is still needed (Sockol et al., 2014).

Sleep-Focused Treatment

In a three-site, unmasked trial of cognitive behavioural therapy for insomnia in pregnancy (179 women analysed), insomnia severity fell more than in the control condition (d = 0.5), and remission came sooner (Manber et al., 2019). In women who did not meet criteria for depression at entry, the decline in Edinburgh Postnatal Depression Scale scores was small but significantly greater than with control, and objective wake time did not differ from control (Manber et al., 2019). The trial therefore shows that the treatment changes insomnia. This section does not claim that the change in sleep carries the effect on mood; that question is taken up under Sleep and Circadian Rhythms.

Neuroactive Steroid Drugs

The proposed mechanism for these drugs is pharmacological. A narrative review links GABA signalling and allopregnanolone to the pathophysiology of the disorder and describes modulating neuroactive steroids as a promising mechanism, while noting that the precise etiology is unknown (Meltzer-Brody & Kanes, 2020). Some of the supporting biology comes from mice. Pregnancy lowered the delta and gamma2 subunits of GABA(A) receptors and reduced tonic and phasic inhibition, and mice without delta-subunit regulation in pregnancy (Gabrd+/- and Gabrd-/-) showed depression-like and abnormal maternal behaviour and poorer pup survival, which led the authors to propose them as a possible model of postpartum depression (Maguire & Mody, 2008).

The human evidence is thinner. In a 60-woman exploratory cohort with a prior mood disorder, lower second-trimester allopregnanolone predicted later postpartum depression after adjustment (a 63% reduction per ng/mL), but third-trimester allopregnanolone, progesterone and the allopregnanolone to progesterone ratio did not. The authors judged the relationship worth exploring in a larger sample (Osborne et al., 2017). A later review found allopregnanolone lower in postpartum depression in some studies and no different from controls in others, and it called the mechanism unclear. It also noted that how long the benefit of brexanolone lasts in trials raises the question of whether receptor action accounts for it (Walton & Maguire, 2019).

Where the Evidence Stands

Mediation has been shown once: early change in activation and environmental reward carried the effect of behavioural activation in pregnancy, on self-report measures. For interpersonal psychotherapy, sleep treatment and the neuroactive-steroid drugs, the mechanism is proposed and has not been tested in this population; in the one trial of cognitive training, interpretation of infant emotions changed while depression did not. The question remains open.

Neuroactive Steroid Drugs, Hormones and Brain Stimulation

Biological treatment of postnatal depression did not begin with the drugs developed for it. Controlled trials in this population date from the 1990s. A double-blind trial with four treatment cells enrolled 87 women with non-psychotic postnatal depression and found fluoxetine superior to placebo over 12 weeks of treatment, although 30% of the women did not complete it (Appleby et al., 1997); that trial is discussed under Treatment Evidence. A year earlier, a double-blind placebo-controlled trial in 61 women with major postnatal depression had reported that transdermal oestrogen outperformed placebo (Gregoire et al., 1996). Roughly two decades separate those trials from the approval of brexanolone for postpartum depression in mid-March 2019 (Scott, 2019). The subsections below take the two neuroactive steroid drugs, the oestrogen line and the two brain stimulation methods in turn.

Brexanolone

Brexanolone is given as an intravenous infusion over 60 hours (Meltzer-Brody et al., 2018a). The drug profile that recorded its approval frames it as a neuroactive steroid, tied to allopregnanolone, that acts by modulating GABA A receptors (Scott, 2019). The first controlled trial discussed the drug in connection with peripartum hormone fluctuations (Kanes et al., 2017), an idea examined under The Biology of Postpartum Depression.

That first test was small. In a phase 2 double-blind randomised trial, 21 women hospitalised with severe postpartum depression, each with a Hamilton Rating Scale for Depression (HAM-D) total of 26 or more, received brexanolone (n = 10) or placebo (n = 11). By hour 60 the placebo group's HAM-D total had fallen by 8.8 points and the brexanolone group's by 21.0, with standard errors of 2.8 and 2.9 respectively; this gave a between-group difference of -12.2 points (95% confidence interval from -20.77 to -3.67; p = 0.0075) and an effect size of 1.2 (Kanes et al., 2017). Adverse events were reported by 4 of 10 women on brexanolone and 8 of 11 on placebo, somnolence by 2 and 0, and follow-up ended at day 30; the pivotal programme was still in progress when the trial was reported (Kanes et al., 2017).

Two double-blind phase 3 trials followed. They randomised 246 women with postpartum depression in the USA, 138 in study 1 and 108 in study 2, out of 375 screened, and followed them to day 30 (Meltzer-Brody et al., 2018a). In study 1 the 17-item HAM-D score fell by 19.5 points in the arm labelled BRX60 and by 17.7 points in the BRX90 arm, against 14.0 points with placebo (differences of -5.5, p = 0.0013, and -3.7, p = 0.0252); in study 2 the BRX90 arm fell by 14.6 points and the placebo arm by 12.1 (difference -2.5, p = 0.0160) (Meltzer-Brody et al., 2018a). Women given placebo therefore improved by 12.1 to 14.0 points, and the margin the drug added was 2.5 to 5.5 points. Adverse events overall were similar to placebo and dizziness was more frequent with brexanolone in study 1, but the report also lists serious adverse events, namely suicidal ideation and an intentional overdose attempt in one BRX60 patient in study 1, and altered consciousness and syncope, considered treatment related, in one BRX90 patient in study 2; the authors concluded that brexanolone produced significant and clinically meaningful reductions in HAM-D score at 60 h and suggested it is a novel drug with the potential to improve treatment options (Meltzer-Brody et al., 2018a).

Approval came in mid-March 2019 (Scott, 2019). The design of the pivotal trials fixes the limits of what they show: a treatment that takes a 60-hour infusion, outcomes tracked no further than day 30, and serious events that included altered consciousness and syncope (Meltzer-Brody et al., 2018a).

Pooled analyses have not agreed on how much the drug adds. A network meta-analysis of 18 randomised trials (N = 1722) placed brexanolone among the drugs showing signals of improvement over placebo in symptom severity and found it the only one with a statistically significant improvement in remission, while all-cause discontinuation gave it a potentially unfavourable acceptability signal; the authors asked for cautious interpretation, citing sparse networks and evidence of low-to-moderate certainty (Xiao et al., 2026). A Cochrane systematic review of 6 randomised trials in 674 women reached a more guarded verdict. Applying GRADE, it found no clear difference between intravenous brexanolone and placebo in response, in remission or in depression severity, each at low certainty and with substantial heterogeneity in the intravenous analyses; adverse events did not differ (moderate certainty), and acceptability, measured by dropout, was lower with the drug (Wilson et al., 2025). Every trial in that review was industry-sponsored and run in the USA, and the reviewers cautioned that their conclusions may change as studies accumulate (Wilson et al., 2025). On the Cochrane reading, the advantage seen in the pivotal trials is not established for response or remission, and the question remains open.

Zuranolone

Zuranolone is an oral drug taken for 14 days (Deligiannidis et al., 2023). The Cochrane review groups it with brexanolone and ganaxolone as neurosteroid positive allosteric modulators of the GABA A receptor (Wilson et al., 2025).

The first phase 3 trial randomised 153 women with postpartum depression at 27 US sites, with scores of 26 or higher on the 17-item Hamilton scale (HAMD-17), to 30 mg of zuranolone or placebo for 2 weeks. By day 15, zuranolone had lowered the HAMD-17 score by 17.8 points against 13.6 for placebo, a difference of −4.2 (95% confidence interval −6.9 to −1.5; p = .003) with an effect size of 0.53; the separation was already present at day 3 (difference −2.7, p = .03) and persisted at day 45 (difference −4.1, p = .003) (Deligiannidis et al., 2021). Response (72% vs 48%; OR 2.63, p = .005) and remission (45% vs 23%; OR 2.53, p = .01) favoured the drug, as did the Montgomery-Åsberg Depression Rating Scale (−4.6; p = .02) and the Hamilton Rating Scale for Anxiety (−3.9; p = .006), whereas maternal functioning on the Barkin index differed only at day 45 (Deligiannidis et al., 2021). The effect was similar across the demographic and clinical subgroups examined; one serious adverse event occurred in each group, and somnolence was reported by 15% on zuranolone and 11% on placebo (Deligiannidis et al., 2021). The authors flagged a relatively high placebo response, secondary endpoints unadjusted for multiplicity, a US sample of limited generalisability, and no knowledge of the effect beyond day 45 or of safety during breastfeeding, and they described the drug as a potential new treatment that was tolerated (Deligiannidis et al., 2021).

A second double-blind phase 3 trial tested 50 mg/day for 14 days in 196 women with severe postpartum depression, 98 per arm, of whom 170 (86.7%) completed the 45-day study. The least-squares mean change on the 17-item Hamilton scale at day 15 was -15.6 with zuranolone and -11.6 with placebo (difference -4.0, 95% CI -6.3 to -1.7); the drug also led at days 3, 28 and 45 and on Clinical Global Impressions severity at day 15, and the authors called it a potential rapid oral treatment (Deligiannidis et al., 2023). Across the two trials the margin over placebo was 4.2 and 4.0 points, on a scale where the placebo groups themselves improved by 13.6 and 11.6 points.

Zuranolone received its first approval, for postpartum depression, in August 2023 (Heo, 2023).

The pooled evidence is firmer here than for brexanolone. The Cochrane review judged with moderate certainty that zuranolone increases response and remission relative to placebo, found lower depression severity as well, and concluded that benefit is probable (Wilson et al., 2025). It also found more maternal adverse events with the drug, no clear difference in acceptability (low certainty) and better maternal functioning (low certainty). The gaps are those of the trials: there were few of them, none ran beyond 45 days, none compared the drug with an active treatment, and none reported quality of life (Wilson et al., 2025). The network meta-analysis described above likewise counted zuranolone among five drugs that beat placebo on symptom severity and among three that did so on response rate, with more adverse events than placebo. That analysis also found fluoxetine better than placebo, with a standardised mean difference of -0.95 (95% confidence interval -1.43 to -0.47) and a higher response rate whose estimate the authors considered imprecise; evidence against active comparators was limited, and the rankings were treated as exploratory (Xiao et al., 2026).

Use during breastfeeding rests on one small study. A phase 1 open-label study of 15 healthy lactating women, 14 of whom completed it, estimated the relative infant dose (RID), its index of transfer into breast milk, at 0.357% for a 30 mg dose, far below the 10% threshold the study applied (Deligiannidis et al., 2024). The 30 mg dose is not the approved one, the figure for 50 mg (0.984%) was simulated, and effects on infants were not evaluated (Deligiannidis et al., 2024); the study is covered more fully under Breastfeeding and Infant Feeding.

A clinical practice guideline on drug treatment of perinatal mental health conditions gives graded recommendations under a modified GRADE framework and uses Good Practice Points where the evidence is inadequate (American College of Obstetricians and Gynecologists, 2023). The standing of the two drugs can be stated narrowly. For zuranolone there is moderate-certainty evidence of benefit over placebo within a 45-day window; for brexanolone the pivotal trials showed larger falls in HAM-D scores, which the Cochrane synthesis did not confirm for response or remission (Wilson et al., 2025). Three questions are open. Whether the gains last beyond the trial window is unknown, and so is how either drug compares with an established antidepressant, because no trial used an active comparator (Wilson et al., 2025). What use during breastfeeding means for the infant is the third: the first zuranolone trial left safety during breastfeeding unknown, and the lactation study did not evaluate infants (Deligiannidis et al., 2021; Deligiannidis et al., 2024).

Oestrogen

The hormonal line is older than the neuroactive steroid drugs, and it has not been confirmed. In the 1996 trial, 61 women with major postnatal depression received transdermal 17 beta-oestradiol at 200 micrograms daily (n = 34) or placebo (n = 27) for 3 months. The groups started level on the Edinburgh Postnatal Depression Scale, at 21.8 and 21.3; by month 1 the oestrogen group scored 13.3 and the placebo group 16.5, the trial put the treatment effect at 4.38 points (95% CI 1.89 to 6.87), and recovery on placebo was slower, at four months or more (Gregoire et al., 1996). None of the patient or episode characteristics tested predicted response, concurrent antidepressant use included. The authors judged oestrogen effective and asked for more studies, noting that the minimum effective dose, the shortest necessary duration and the mechanism of action were all unknown (Gregoire et al., 1996).

The later test was a pilot. A double-blind, placebo-controlled pilot randomised trial of transdermal estradiol (100 mcg/day of 17β-estradiol for 6 weeks) analysed 12 women with postpartum depression, 6 per arm, and found no significant difference from placebo on its primary outcomes (Li et al., 2020). Response was recorded in 5 women on estradiol and 2 on placebo and remission in 4 and 1, with p = 0.24 for each comparison. Lower scores with estradiol appeared on the Beck Depression Inventory at week 6 (-10.5, p = 0.049) and on the Hamilton scale at weeks 3 and 6 (-8.5, p = 0.040; -9.6, p = 0.026), but these differences did not survive correction; the Beck difference at week 3 was not significant (-9.2, p = 0.074), the Edinburgh scale showed no effect of group (p = 0.328), and plasma estradiol did not differ significantly between the arms (Li et al., 2020). The authors reported under-recruitment, three dropouts from the placebo arm and an underpowered sample, and concluded that the trial was feasible and that any sign of efficacy was preliminary and confined to secondary outcomes (Li et al., 2020).

Two small trials more than two decades apart, at different doses, give one positive result and one null result on the primary outcomes. Whether oestrogen treats postnatal depression remains unconfirmed.

Electroconvulsive Therapy and Magnetic Stimulation

Electroconvulsive therapy (ECT) has observational rather than randomised support in the postpartum period. A population-based cohort drawn from the Swedish National Quality Register for ECT matched 185 women treated within 6 months of delivery to 185 comparison patients treated outside the postpartum period. Response, defined as a Clinical Global Impressions Improvement score of 1 or 2 at 1 week after ECT, was recorded in 87.0% of the postpartum group and 73.5% of the comparison group (p = 0.001); more severe symptoms at baseline predicted response, and the other predictors tested did not (Rundgren et al., 2018). Because no group went without ECT, the study compares two sets of treated patients and cannot show how many would have improved untreated. Its authors read it as support for ECT in severe postpartum depression or psychosis (Rundgren et al., 2018).

The rest of the postpartum literature is thin. A systematic review that searched PubMed until September 2014 located 8 studies and 8 case reports, so that half of its material was case reports, and added a case report of its own; ECT was effective in all the studies it found. The reviewers noted how little is known about ECT after childbirth and suggested that combining it with joint mother-baby hospitalisation could help (Gressier et al., 2015).

Cognitive side effects are the usual concern with ECT, and the evidence on them concerns depressed patients in general, not women treated after childbirth. A meta-analysis of 84 studies and 2981 depressed patients given ECT compared standardized cognitive test scores before and after treatment. Relative to pretreatment, performance fell significantly on 72% of variables in the 0 to 3 days after ECT, all but one confidence interval included zero or showed a positive effect at 4 to 15 days, and no negative effects were observed beyond 15 days, by which point anterograde memory, processing speed, working memory and parts of executive function had improved past baseline (Semkovska & McLoughlin, 2010). Electrode placement and stimulus waveform did not moderate outcome after the first 3 days. The authors concluded that deficits fall mainly within 3 days, and they noted that standardized tests of retrograde amnesia were lacking (Semkovska & McLoughlin, 2010). No source cited here tests whether the same time course holds in a postpartum sample.

Transcranial magnetic stimulation in its repetitive form (rTMS) has been tested in randomised trials. A systematic review and meta-analysis that identified 81 studies and included 14 randomised controlled trials, with 884 women with postpartum depression, found larger reductions than control on the Hamilton Depression Rating Scale and the Edinburgh Postnatal Depression Scale, together with better cognitive function (Peng et al., 2020). From the trial evidence available, the authors concluded that rTMS could bring improvements in both depressive symptoms and cognitive function for women with postpartum depression, a question they had framed at the outset as uncertain (Peng et al., 2020).

Emerging Treatments

The treatments below are newer or less settled than those covered under Treatment Evidence and Neuroactive Steroid Drugs, Hormones and Brain Stimulation. Each carries a standing in this section: early (a few trials or short follow-up), mixed (results that differ by subgroup or quality), or untested in postpartum depression. Several rest on evidence from pregnancy or from depression in general, and that limit is stated where it applies.

Ketamine and Esketamine Around Delivery

Standing: early and promising, with short follow-up. A double-blind, placebo-controlled trial in five hospitals in China enrolled 364 mothers with prenatal depression (an Edinburgh Postnatal Depression Scale score of 10 or more). Mothers given a 40-minute infusion of esketamine at 0.2 mg/kg had far fewer major depressive episodes by 42 days postpartum: 6.7% (12 of 180) against 25.4% (46 of 181) with placebo, a relative risk of 0.26 (95% CI 0.14 to 0.48), with a number needed to treat of 5 (Wang et al., 2024). Median Edinburgh scores were 3 points lower at 7 and 42 days, and median Hamilton scores 4 points lower.

The costs were real but transient. Neuropsychiatric adverse events occurred in 45.1% of the esketamine group and 22.0% of the placebo group, including dizziness (26.4% against 9.3%) and hallucinations (3.3% against none). Persistent pain at 42 days was less common with esketamine (35.2% against 47.5%, P = 0.02). Exclusive breastfeeding, hospital stay and maternal or neonatal complications did not differ significantly (Wang et al., 2024). The authors' own caveats limit the reading: women with prepregnancy mood disorders were excluded, prenatal depression was mostly mild, adverse events may have been under-reported, the trial had no data safety monitoring board, and whether the benefit lasts beyond 42 days is unknown.

A meta-analysis of 2,916 patients having caesarean delivery pooled thirteen randomised trials and one retrospective study (six of ketamine, eight of esketamine). Ketamine or esketamine lowered the risk of postpartum depression at 1 and 4 weeks and lowered Edinburgh scores, with moderate certainty of evidence. Adverse events were more frequent with the drugs, notably dizziness, diplopia, hallucination and headache. In subgroup analyses, high dosage, delivery by patient-controlled intravenous analgesia and esketamine alone each showed a significant reduction in incidence and Edinburgh scores at one and four weeks (Li et al., 2024). The review called for high-quality studies, including in other countries, and none of the studies cited here follows women beyond the first weeks. The position is therefore open: the effect is consistent across the two sources in the short term, and its durability and safety balance remain untested outside these samples.

Psilocybin

Standing: untested in postpartum depression. A narrative review sets out a rationale for psilocybin in the postpartum period, built on maternal disconnection and the possibility of reconnection, and concludes that benefit is possible but that pilot studies should come first. It reports that psilocybin did not differ from escitalopram on the QIDS-SR in a small trial in depression, and that no postpartum safety data and no breastfeeding studies exist. The trials of psilocybin in depression it draws on were single-centre phase 2 trials with small samples, and it treats the results of an open-label pilot as likely to be biased (Jairaj & Rucker, 2022). No study cited on this page is a trial in postpartum depression, so nothing here is a finding about this population.

Digital and Internet-Delivered Treatment

Standing: early to mixed, depending on the format. In a two-group randomised trial in Australia, 43 postnatal women with diagnosed depression received either MumMoodBooster, a six-session internet cognitive behavioural programme, or usual care. At 12 weeks, 79% (15 of 19) of the programme group had remitted against 18% (4 of 22) of the usual-care group, a large effect on the Beck Depression Inventory-II (d = .83), and 86% completed every session (Milgrom et al., 2016). Gains did not extend to every outcome: anxiety, partner relationship and parenting competence did not differ significantly, and stress differed at 9 weeks but not at 12. The sample was small, participants were not blinded after allocation, usual care was heterogeneous, and self-guided delivery was not tested.

A systematic review with a preliminary meta-analysis of 11 randomised trials (2,424 postpartum women in six countries) found that mobile health interventions lowered Edinburgh scores by a mean of 1.09 points (95% CI 0.79 to 1.39) compared with controls. The interventions spanned four types, including telephone, text-message and smartphone-application formats, and the authors treated the result as preliminary, a possible adjunct to routine care that needs better trials (Zhou et al., 2022).

Who Delivers Care

Standing: early but strong on access. A four-arm pragmatic noninferiority trial randomised 1,230 pregnant and postpartum adults in the United States and Canada to behavioural activation delivered by nonspecialist or specialist providers, by telemedicine or in person. At 3 months, mean Edinburgh scores were 9.27 with nonspecialists and 8.91 with specialists, and 9.15 by telemedicine and 8.92 in person; both contrasts met the noninferiority criterion, and there was no provider-by-modality interaction (P = 0.93). Telemedicine participants attended more sessions (6.55 versus 5.07, P < 0.001). The in-person arms were smaller, the pandemic disrupted in-person recruitment, and only 3-month outcomes were reported (Singla et al., 2025). The authors read the result as compelling evidence for task-sharing and telemedicine as routes to access. Service models are discussed under Access to Care and Service Delivery.

Exercise

Standing: mixed. A meta-analysis of 13 randomised trials (1,734 eligible postpartum mothers) found that exercise reduced depressive symptoms with a moderate effect (SMD = -0.44, 95% CI -0.75 to -0.12). Heterogeneity was high (I2 85%), six trials had unclear and seven high risk of bias, and publication bias was possible. Effects did not differ significantly by population, by co-intervention or, at P = 0.05, by exercise context. Exercise duration and intensity were rarely assessed. The authors supported exercise but advised cautious interpretation (Pritchett et al., 2017).

Omega-3 Supplementation

Standing: mixed, and different for pregnancy and postpartum use. A meta-analysis of 18 trials in 4,052 perinatal women found a small overall benefit over placebo (SDM = -0.236, 95% CI -0.463 to -0.009, P = .042) with considerable heterogeneity (I2 = 88.58). Evidence quality was low overall, reporting was poor, and unblinding through a fishy aftertaste was possible. Effects were negligible in pregnancy and larger postpartum, but the postpartum subgroup effects were not significant and few postpartum trials existed. The authors advised against use in pregnancy and judged postpartum add-on use promising (Mocking et al., 2020).

Bright Light

Standing: early, and not tested after childbirth. The one trial cited here enrolled 27 pregnant women with nonseasonal major depression, so it concerns antepartum, not postpartum, depression. Five weeks of 7,000-lux light for 1 hour a day beat a 70-lux dim red placebo: response was 81.3% against 45.5% and remission 68.6% against 36.4%. Expectation ratings did not differ between arms (Wirz-Justice et al., 2011). The authors judged light therapy a possibly useful nonpharmacological approach; the sample was small, and no replication is cited here.

Parent-Infant Interventions

Relieving a mother's low mood and repairing the relationship with her infant are separate aims, and the trial record suggests that the first does not guarantee the second. This section covers what happened when treatment was aimed at the mother-infant relationship, as well as at the mother's mood. The overall answer is that no relationship-focused intervention yet has reliable trial support.

The Problem: Mood Treatment Alone

A randomised trial with a comparison group tested interpersonal psychotherapy against a waitlist in 176 women, 120 of them depressed and 56 not. Treatment improved parenting stress and nothing else the study measured, and stress after treatment, though lower than before, stayed above the level of the non-depressed mothers. Response to treatment did not predict child outcomes (not significant). The authors concluded that treatment should also target the mother-infant relationship (Forman et al., 2007).

A four-arm trial of 193 women with postpartum depression, assessed at 4.5, 18 and 60 months, found a similar pattern. Cognitive-behavioural, psychodynamic and counselling treatments each reduced early relationship problems relative to a control condition (relative risk 0.46, 0.57 and 0.63 respectively). Attachment, infant cognition, infant behaviour management and five-year outcomes did not differ from control, and the investigators described the evidence of benefit as limited and short-term, with longer treatment possibly needed. Social adversity was possibly unbalanced between the treatment groups, so further models controlled for it (Murray et al., 2003).

What the Reviews Find

A systematic review of nineteen studies of postnatal depression interventions (1999 to 2014) found that dyad-focused interventions tended to show larger effects on maternal mood than other interventions, yet mood improvement was not sufficient for better child outcomes and related inconsistently to dyad outcomes. Most studies scored inadequately on the review's quality tool and samples were small, so the authors judged that no specific intervention could yet be recommended (Tsivos et al., 2015).

Seven randomised trials of parenting interventions were reviewed for mothers of infants who had depressive symptoms; pooling six of them (573 participants), the meta-analysis found that the parent-child relationship did not change significantly (d = 0.028, 95% confidence interval -0.30 to 0.31). Most child development outcomes were also non-significant, and the authors called the evidence insufficient and larger trials warranted. Video feedback did improve maternal sensitivity in the short term (d = 0.82), but the pooled result was unchanged when peer-support or alternative-treatment controls were removed. The included studies were small, drawn from Western high-income countries, and none started treatment in pregnancy (Rayce et al., 2020).

A broader systematic review of perinatal depression treatment (thirty-six trials in 40 articles) was more optimistic. Some treatments improved parenting and child outcomes, with large effects in places: interaction guidance and group support for postpartum depression improved parenting, for example. Other meta-analytic effects were non-significant, among them interpersonal psychotherapy on attachment and cognitive-behavioural therapy on parenting stress. The authors called the findings promising for specific outcomes and asked for randomised trials using common measures (Letourneau et al., 2017).

The reviews therefore differ in tone but not in substance. Each finds scattered positive effects on particular parenting or relationship measures, and the two most pooled analyses find no reliable effect on the relationship itself.

A Direct Test: Video-Feedback Therapy

Video-feedback therapy was tested against progressive muscle relaxation in 144 women with persistent postnatal depression (72 per group). The therapy comprised 11 home-based sessions and two boosters, and child outcomes were assessed at age 2 years. The groups did not differ on cognitive development (adjusted difference -1.01, p = 0.63), language development (1.33, p = 0.63), behaviour problems (-1.77, p = 0.19) or attachment security (0.02, p = 0.58); all were non-significant. Both groups had remitted and child outcomes in both fell within non-clinical norms, which leaves little room for a difference to appear. Six serious adverse events occurred, none treatment-related, and the authors noted that mechanisms could not be ascertained (Stein et al., 2018).

How Such Interventions Are Meant to Work

The rationale is that a dyad-focused treatment would act on the cognitive processes described earlier in this article, helping a mother read her infant's cues more accurately and loosening a preoccupation with her own state. This is a rationale, not a result: no study cited in this section tests whether those processes mediate any effect, and the one direct trial reported that mechanisms could not be ascertained (Stein et al., 2018).

Current Standing

The evidence stays unsettled. Treating maternal mood alone leaves relationship and child outcomes largely unchanged (Forman et al., 2007), and the interventions aimed at the dyad have produced short-term gains in places without a reliable effect that holds up across trials (Rayce et al., 2020; Tsivos et al., 2015). No relationship-focused intervention has reliable trial support yet, and the question is open until larger, adequately powered trials report.

Access to Care and Service Delivery

The Treatment Cascade

A systematic review that pooled weighted mean rates across studies of antenatal and postpartum depression found shortfalls at every step from identification to remission (Cox et al., 2016). The authors framed this as a Perinatal Depression Treatment Cascade and concluded that it leaves many opportunities to improve management (Cox et al., 2016). For postpartum depression, 30.8% of cases were identified, 15.8% were treated, 6.3% received adequate treatment and 3.2% reached remission (Cox et al., 2016). The antenatal figures were 49.9% identified, 13.6% treated, 8.6% adequately treated and 4.8% in remission, so the postpartum period fared worse than pregnancy at every step except treatment, where it was slightly higher (Cox et al., 2016). The review searched PubMed and EMBASE through March 2015 and admitted English-language articles only, which limits how far the rates extend to later years or other languages (Cox et al., 2016).

The treatment cascade

Each bar is a share of all women with depression in that period, not of the step above it.

Women with depression100%Identified30.8%Receive any treatment15.8%Receive adequate treatment6.3%Reach remission3.2%

Of every 1,000 women with depression after delivery, about 308 are identified, 158 receive any treatment, 63 receive adequate treatment and 32 reach remission.

Percentages are the weighted mean rates reported in the systematic review by Cox et al. (2016). The counts per 1,000 women are those rates applied to a round number for illustration.

Who Receives Care and Why Women Do Not Come Forward

Receipt of care differs by race and ethnicity. In a retrospective cohort of 29,601 New Jersey Medicaid mothers, 9% of white women began treatment, against 4% of Black and 5% of Latina women, and the adjusted odds of initiation were lower for Black (AOR = .43) and Latina (AOR = .59) women (Kozhimannil et al., 2011). Follow-up care was also less likely for both groups, and Latina women were less likely to continue care, whereas the difference for Black women in continued care was only marginal (Kozhimannil et al., 2011). Treatment was low overall, and the authors judged the disparities to warrant attention (Kozhimannil et al., 2011). The claims data held no information on symptom severity, prior history or treatment preferences, did not require a depression diagnosis, and omitted services outside Medicaid, so the true rates of illness in each group were not identified (Kozhimannil et al., 2011).

Qualitative work addresses why women stay silent. A systematic review of 40 articles published between 1966 and 2005 identified nondisclosure, myths and service gaps as barriers, and concluded that women did not proactively seek help; the same review found a preference for talking therapies over pharmacological treatment, although few studies had examined treatment acceptability (Dennis & Chung-Lee, 2006). A later meta-ethnography of 24 UK qualitative studies, all concerned with postnatal depression, found that recognising the problem and continuity of care promoted help seeking, whereas stigma, an unclear role for the health professional and negative experiences of childbirth care reduced it (Button et al., 2017). Women sought help more readily from a familiar health professional than from an unfamiliar one, and the authors concluded that continuity of care and staff training are needed (Button et al., 2017). Whether seeking help improves outcomes for women with less severe distress remains unclear (Button et al., 2017).

Ways to Widen Access

Task-sharing and telemedicine have the most direct trial evidence. In a pragmatic four-arm noninferiority trial of 1,230 pregnant and postpartum adults in the United States and Canada, behavioral activation delivered by non-specialist providers was noninferior to delivery by specialists on the Edinburgh Postnatal Depression Scale at three months, and telemedicine was noninferior to in-person delivery (Singla et al., 2025). Participants in the telemedicine arms attended more sessions (6.55 versus 5.07, P < 0.001), while session attendance did not differ between provider types (Singla et al., 2025). The authors described the evidence for task-sharing and telemedicine as compelling for expanding access (Singla et al., 2025). The in-person arms were smaller, the pandemic disrupted in-person recruitment, only three-month outcomes were reported, COVID-19 led the investigators to loosen the noninferiority margin for the telemedicine versus in-person comparison from 10% to 13%, and the sample was mostly university-educated (Singla et al., 2025).

A cluster randomised trial in rural Pakistan tested a cognitive behaviour therapy-based programme delivered by community-based primary health workers to 903 depressed pregnant women in 40 Union Council clusters (Rahman et al., 2008). Major depression was less frequent in the intervention group at six months (23% versus 53%; adjusted OR 0.22) and at twelve months (27% versus 59%; adjusted OR 0.23) (Rahman et al., 2008). Infant weight-for-age and height-for-age did not differ significantly between groups (Rahman et al., 2008). The authors concluded that integration into resource-poor health systems is possible (Rahman et al., 2008).

Peer support offers a lower-intensity route. A multisite randomised trial in Ontario enrolled 701 high-risk postpartum women, screened from 21,470, and delivered telephone support from peers who had completed a four-hour training session (Dennis et al., 2009). At 12 weeks, 14% of the peer-support group scored above 12 on the Edinburgh Postnatal Depression Scale, against 25% of controls (P < 0.001), and over 80% of recipients were satisfied (Dennis et al., 2009). By 24 weeks the groups no longer differed (11% versus 14%, P = 0.11), which the authors attributed to referring women identified with clinical depression at 12 weeks for treatment, and loneliness did not differ (Dennis et al., 2009). The authors judged telephone peer support a possible preventive approach for high-risk women; the interview used to check outcomes was shortened, telephone-based and not validated (Dennis et al., 2009).

Severe Illness and Mother and Baby Units

For severe illness, joint admission of mother and infant to a mother and baby unit is the specialised service model. A systematic review of 23 studies, covering 5,023 participants, 215 of them controls, linked admission to maternal improvement and described the evidence as encouraging but its conclusions as preliminary (Gillham & Wittkowski, 2015). The studies were of variable methodological quality, none was a randomised controlled trial, most cohorts lacked control groups, and the measures varied with little consistency (Gillham & Wittkowski, 2015). Whether such units outperform community care therefore remains open, pending controlled evidence.

Worked Example

This example follows one result, a positive Edinburgh Postnatal Depression Scale screen, to show what it does and does not tell a clinician. The accuracy figures come from an individual participant data meta-analysis of 58 studies and 15,557 women, which tested the scale against interview-based diagnoses of major depression (Levis et al., 2020).

Step 1: The Accuracy at Three Cut-Offs

The meta-analysis reported sensitivity and specificity for each cut-off score. At a cut-off of 11 or higher, sensitivity was 0.81 and specificity 0.88. At 13 or higher, sensitivity fell to 0.66 and specificity rose to 0.95. At 10 or higher, sensitivity was 0.85 and specificity 0.84 (Levis et al., 2020). Raising the cut-off therefore trades missed cases for fewer false alarms. The authors concluded that 11 or higher was the optimal balance and that 13 or higher was the more specific choice (Levis et al., 2020).

Step 2: Turning Accuracy into Counts

Sensitivity and specificity describe the test. A woman holding a positive result wants something different, the chance that she has major depression, and that depends on how common the condition is among the women screened. The arithmetic below is illustrative and comes from no study. It assumes 10,000 women screened, of whom 12% (1,200) have major depression. The 12% was chosen for round numbers. It is close to the 11.9% pooled prevalence of perinatal depression in one meta-analysis, but that pooled figure combined estimates made by any method, and scales gave higher estimates than diagnostic instruments, so it is not a reference-standard rate (Woody et al., 2017).

At a cut-off of 11 or higher, 972 of the 1,200 women with major depression screen positive and 228 are missed. Of the 8,800 women without major depression, 1,056 screen positive and 7,744 screen negative. In all, 2,028 women screen positive, and 972 of them have major depression, a positive predictive value of 47.9%. The negative predictive value is 7,744 of 7,972, or 97.1%.

At a cut-off of 13 or higher, 792 women are true positives and 440 are false positives. That gives a positive predictive value of 64.3%, with 408 cases missed and a negative predictive value of 95.3%.

At a cut-off of 10 or higher, 1,020 women are true positives and 1,408 are false positives. The positive predictive value is 42.0%, with 180 cases missed and a negative predictive value of 97.6%.

Figure 1

Positive Predictive Value of the Edinburgh Scale at Three Cut-Offs

Bar chart of the positive predictive value at an assumed prevalence of 12 percent: 42.0 percent at a cut-off of 10 or higher, 47.9 percent at 11 or higher and 64.3 percent at 13 or higher. 0% 50% 100% 42.0% 47.9% 64.3% 10 or higher 11 or higher 13 or higher Edinburgh scale cut-off
Note. Illustrative arithmetic at an assumed prevalence of 12%, using the sensitivity and specificity reported by Levis et al. (2020). Original schematic.

Step 3: Reading the Result

At the recommended cut-off, roughly half of the positive screens in this illustration are not major depression. That is why a positive screen leads to an assessment and not to a diagnosis. The negative predictive value is high at every cut-off shown, but a negative screen still leaves some missed cases, 228 of the 1,200 at a cut-off of 11 or higher. For major depression prevalence between 5% and 25%, the real meta-analysis found that a cut-off of 11 or higher gave negative predictive values of 93% to 99% and positive predictive values of 26% to 69% (Levis et al., 2020). The illustration's figures sit inside those ranges, and the true values for any one clinic depend on the local prevalence of major depression. The demonstration in Screening for Postpartum Depression lets the reader vary that prevalence and watch the positive predictive value move.

Current Directions

Research on postpartum depression is moving on several fronts at once: larger genetic samples, drugs developed for the perinatal period, trials of treatments delivered by non-specialists or by video, causal tests of cognitive targets, and a reassessment of what the perinatal brain looks like before depression is added. Most of these findings are recent and rest on few trials, so the open questions are as informative as the results.

Genetics at Scale

A genome-wide association meta-analysis pooled 20 cohorts, with 18,770 postpartum depression cases and 58,461 controls in the full sample. The European-ancestry analysis used 18 cohorts (17,339 cases, 53,426 controls). No single nucleotide polymorphism reached genome-wide significance in either the European or the trans-ancestry analysis, and the authors concluded that their results confirm postpartum depression as a polygenic and heritable phenotype (Guintivano et al., 2023). SNP-based heritability was estimated at 0.14 (SE = 0.02). Genetic correlations were positive with major depressive disorder, bipolar disorder, anxiety and PTSD, and the pattern of cell-type enrichment, across 27 human tissues and 265 cell types, differed from that of major depressive disorder, with inhibitory and cholinergic neurons implicated for postpartum depression. The authors concluded that postpartum depression may have its own genetics, and that more samples are needed (Guintivano et al., 2023). Whether it is a distinct genetic entity therefore remains open.

Neuroactive-Steroid Drugs After Approval

A Cochrane review of six randomised trials (674 women, all in the USA) judged that zuranolone probably outperforms placebo on response and remission, with moderate certainty, and also produces more maternal adverse events. Intravenous brexanolone did not differ from placebo on response, remission or severity, with low certainty, and showed worse acceptability through dropout. The reviewers noted that every study was industry-sponsored, that follow-up lasted at most 45 days, that no active comparators were used, and that conclusions may change as trials accumulate (Wilson et al., 2025).

A network meta-analysis of 18 trials (N = 1722) placed five drugs above placebo on symptom severity: saffron, fluoxetine, esketamine, brexanolone and zuranolone. Brexanolone alone was significant for remission. The authors judged the evidence to be of low-to-moderate certainty, with sparse networks and exploratory SUCRA rankings, and advised cautious interpretation (Xiao et al., 2026). The two syntheses agree that benefit is probable but unsettled, and that harms and acceptability need weighing alongside efficacy.

Esketamine at Delivery

In a double-blind placebo-controlled trial across five hospitals in China, 364 mothers with prenatal depression received esketamine at 0.2 mg/kg. A major depressive episode at 42 days postpartum occurred in 6.7% of the esketamine group and 25.4% of the placebo group (relative risk 0.26, 95% CI 0.14 to 0.48), a number needed to treat of 5. Neuropsychiatric adverse events were more frequent with esketamine (45.1% v 22.0%) but transient, and breastfeeding did not differ. The authors suggested considering low-dose esketamine in prenatal depression, while noting that mostly mild depression, the exclusion of prepregnancy mood disorders, possible under-reporting of adverse events and unknown persistence beyond 42 days constrain generalisation (Wang et al., 2024).

A meta-analysis of 2916 cesarean patients (13 randomised trials and one retrospective study) found lower postpartum depression risk at 1 and 4 weeks, and lower Edinburgh Postnatal Depression Scale scores, with ketamine or esketamine than with control. The authors judged the drugs to appear effective, with moderate certainty, but dizziness, diplopia, hallucination and headache were more common, and the authors called for high-quality studies in other countries (Li et al., 2024). Durability and safety beyond the early weeks are the unresolved questions.

Cognitive Training as a Causal Test

Prenatal affective cognitive training was tested in a randomised trial of high-risk pregnant women (147 randomised, 121 followed up, 83%). Five 90-minute sessions over five weeks did not change postpartum depression overall (p = 0.18) or depressive symptoms (p = 0.31), so the primary outcome was not met. Positive interpretation of infant emotions was higher after training. In multiparous women alone, incidence was 5% against 35% with care as usual (OR = 0.10, 95% CI 0.01 to 0.92). The authors treated the parity effect as exploratory and called for replication in larger trials (Bjertrup et al., 2026). The trial remains a test of whether shifting a cognitive bias changes later depression, and its answer is, so far, not demonstrated.

Task-Sharing and Telemedicine

A four-arm pragmatic noninferiority trial of 1,230 pregnant and postpartum adults in the US and Canada compared nonspecialist with specialist delivery of behavioral activation, and telemedicine with in-person delivery. Three months after randomisation, mean EPDS scores were 9.27 for nonspecialists against 8.91 for specialists, and 9.15 for telemedicine against 8.92 for in-person care, and both comparisons met noninferiority. The authors concluded that task-sharing and telemedicine offer compelling evidence for widening access. Limits included smaller in-person arms, 3-month outcomes only, pandemic disruption of in-person recruitment, a noninferiority margin for telemedicine versus in-person delivery that was loosened from 10% to 13% because of COVID-19, and a mostly university-educated sample (Singla et al., 2025).

The Maternal Brain as a Moving Baseline

In a prospective pre-post study, first-time mothers showed gray matter volume reductions across pregnancy relative to controls, whereas fathers did not differ from controls. The reductions persisted for at least 2 years, and the gray-matter volume changes predicted measures of postpartum maternal attachment (Hoekzema et al., 2017). That work does not address depression, so these findings mark the baseline against which perinatal depression is measured rather than evidence about it.

Subjective Cognitive Complaints

A systematic review of 42 studies of perinatal women found that depressive symptoms were positively related to self-reported cognitive complaints. Links with working memory, attention and cognitive flexibility were inconsistent, and verbal and recognition memory showed no significant link. The authors judged objective impairment inconclusive and advised longitudinal research, citing mostly cross-sectional designs, self-report symptom measures, heterogeneous non-ecological tasks and rare assessment of anxiety (Sobral et al., 2026). Whether the complaints track measurable deficit remains open.

Discussion

What Is Settled

Several conclusions are now secure. Depression around childbirth is common, and much of it begins before the birth: in a screening series of ten thousand postpartum mothers, only 40.1% of the screen-positive women reported onset after delivery, against 33.4% in pregnancy and 26.5% before it (Wisner et al., 2013). The same series found that 22.6% of primary diagnoses among screen-positive women were bipolar disorders, and its authors concluded that bipolar and unipolar illness need to be told apart (Wisner et al., 2013); severe illness with psychosis is a separate emergency, described under its own type. A brief self-report scale detects the condition with known error rates: at a cut-off of 11 or higher the Edinburgh scale had a sensitivity of 0.81 and a specificity of 0.88, and, at that cut-off, positive predictive values ranged from 26% to 69% for major depression prevalences of 5-25% (Levis et al., 2020), which is why a positive screen leads to an assessment and not to a diagnosis. Psychological treatment helps: across 43 randomised trials, psychotherapy outperformed control conditions with a moderate to large effect (g = 0.67) that lasted from six to twelve months, although heterogeneity was high and there was some publication bias (Cuijpers et al., 2023).

What Remains Open

Whether postpartum depression is a distinct disorder is unresolved. The evidence on its validity conflicts, and one review concluded that it is probably a complex, heterogeneous phenotype (Di Florio & Meltzer-Brody, 2015). A study that would settle the question would compare early-onset postpartum cases with depression at other times on course, familiality and biology in the same women, which the field has done only in parts.

Whether universal screening improves outcomes, and not only detection, is also open. A systematic review of randomised trials found a single postpartum trial, at high risk of bias, and no trials in pregnancy, and judged that guidelines recommending screening should be reconsidered (Thombs et al., 2014). The later Canadian guideline, resting on the same small trial, recommended conditionally against routine questionnaire screening on very low-certainty evidence (Lang et al., 2022), while the United States bodies advise screening. The bodies read the same thin trial evidence differently, and an adequately powered trial of screening followed by assured treatment is what the question needs.

How much of the association with child outcomes is causal is the third open question. In a Norwegian cohort of 11,599 families, comparing siblings exposed to different levels of maternal depression attenuated the associations, so that prenatal and six-month postpartum symptoms were no longer associated with child problems and only concurrent maternal depression remained, with child problems reported by the mothers themselves (Gjerde et al., 2017). That genetically informed result, stronger on confounding than conventional cohorts, points to a smaller causal effect of perinatal exposure than the raw associations suggest.

Whether cognition causes or follows the episode has been tested directly only once. In a randomised trial of prenatal affective cognitive training in 147 high-risk pregnant women, the training changed the interpretation of infant emotions but did not reduce postpartum depression overall (p = 0.18, non-significant); a lower incidence in multiparous women (5% against 35%) was, in the authors' words, exploratory and in need of replication in larger trials (Bjertrup et al., 2026). The prospective vulnerability findings are consistent in direction, but with the one trial null on its primary outcome, a causal role for cognition is not established. The biological account is stronger but narrower: when a double-blind experiment simulated pregnancy and then withdrew gonadal steroids, mood symptoms appeared in five of eight women with a past postpartum depression and in none of eight without one (Bloch et al., 2000), evidence for a hormone-sensitive subgroup that rests on a very small sample.

The durability of the neuroactive-steroid drugs is the fifth question. An independent Cochrane review of six trials judged with moderate certainty that zuranolone improves depression compared with placebo, with more maternal adverse events, and found with low certainty no clear difference between intravenous brexanolone and placebo in response or remission; all the trials were industry-sponsored, run in the United States, followed women for at most 45 days and made no comparison with an active treatment (Wilson et al., 2025). Trials against standard antidepressants with longer follow-up are what would settle their place.

Finally, treating the mother's mood does not by itself repair the relationship with the infant. For mothers of infants who had depressive symptoms, parenting interventions did not significantly improve the mother-infant relationship in a meta-analysis (d = 0.028), whose authors judged the evidence insufficient and asked for larger trials (Rayce et al., 2020). Across these disputes, the strongest designs available (individual participant data, sibling comparison, randomised tests of cause and the independent review) give smaller or more qualified effects than the studies that preceded them.

Common Misconceptions

Postpartum depression always starts after the baby is born.
Many episodes identified after delivery began earlier. Among women who screened positive after birth in a series of ten thousand mothers, 33.4% reported onset during pregnancy and 26.5% before pregnancy, against 40.1% after the birth (Wisner et al., 2013). The manual's specifier accordingly merges onset in pregnancy with onset after delivery, a choice critics have argued should be reversed (Sharma & Mazmanian, 2014). The belief persists because the condition is named after the postpartum period and is usually detected then.
Postpartum depression and the baby blues are the same thing.
The blues are a brief mood disturbance of the first days after delivery, generally regarded as a normal adjustment, whereas postpartum depression is a depressive episode that persists. The two are related rather than identical: a review found the blues more likely in women with dysphoria in pregnancy, a history of depression or high neuroticism, and treated the blues as an apparent index of vulnerability to later depression (Henshaw, 2003), and in a community cohort of 853 women the blues predicted postpartum depression and anxiety disorders in the first three months (Reck et al., 2009). The confusion arises because both begin after birth and both involve low mood.
A high score on the Edinburgh scale means a woman has postpartum depression.
The scale is a screen. In an individual participant data meta-analysis, positive predictive values at a cut-off of 11 or higher ranged from 26% to 69% for major depression prevalences of 5-25%, so that a substantial share of positive screens were not major depression on interview (Levis et al., 2020). A positive screen is a reason for a clinical assessment, which is where the diagnosis is made. The belief exists because a score is a number and looks more definitive than it is.
A mother who has frightening thoughts about harm coming to her baby is a danger to the baby.
Unwanted intrusive thoughts of harm to the infant were a relatively normative experience in the weeks after birth in a prospective cohort of one hundred women, and there was little evidence linking them to aggressive parenting (Fairbrother & Woody, 2008). In a later province-wide cohort, neither such thoughts nor obsessive-compulsive disorder was associated with physical aggression towards the infant (Fairbrother et al., 2022). The thoughts are common, and a mother troubled by them can speak to a clinician; in the cohort studies they were not linked to aggression towards the infant, and the evidence on filicide concerns psychosis.

Glossary

Allopregnanolone.
A metabolite of progesterone that acts at GABA-A receptors and supplies the rationale for the neuroactive-steroid drugs; the human evidence linking its levels to postpartum depression is inconsistent.
Antenatal depression.
Depression during pregnancy; in one cohort, depression scores were at least as high in pregnancy as after the birth.
Brexanolone.
An intravenous formulation of allopregnanolone given as a 60-hour infusion and approved for postpartum depression in March 2019; an independent review found, with low certainty, no clear difference from placebo in response or remission.
Edinburgh Postnatal Depression Scale.
A ten-item self-report screening questionnaire for depression after childbirth that takes about five minutes; a score at or above a cut-off is a positive screen, not a diagnosis.
Infant-directed speech.
Speech addressed to an infant; in one conditioning experiment, the infant-directed speech of mothers with depressive symptoms failed to promote associative learning in four-month-olds.
Interpersonal psychotherapy.
A psychological treatment built on role transitions and interpersonal problems that reduced depressive symptoms in randomised trials of postpartum depression; its proposed mechanisms of change remain largely untested.
Intrusive thoughts.
Unwanted thoughts that enter the mind against the person's wishes; after birth, intrusive thoughts of harm coming to the infant were a relatively normative experience in one cohort and were not associated with aggression towards the infant in another.
Mother and baby unit.
A psychiatric inpatient unit that admits a mother together with her infant; a systematic review of 23 studies found encouraging but preliminary evidence of maternal improvement.
Paternal postpartum depression.
Depression in fathers during the partner's pregnancy or the months after birth, affecting about one in ten fathers in a meta-analysis and correlated with depression in the mother.
Perinatal depression.
Depression during pregnancy or after the birth, the frame the current DSM classification uses in place of a strictly postpartum one.
Peripartum onset.
The DSM-5-TR specifier attached to a depressive episode that begins during pregnancy or in the four weeks after delivery; critics have argued for separating the two onsets and lengthening the window.
Positive predictive value.
The proportion of positive screens that are true cases; it depends on prevalence as well as on the test, and for the Edinburgh scale at a cut-off of 11 or higher ranged from 26% to 69% across prevalences of 5% to 25%.
Postpartum blues.
Transitory tearfulness and mood change in the first days after delivery, commonplace and generally regarded as normal, which nonetheless predicted later postpartum depression in cohort studies.
Postpartum psychosis.
A rare psychiatric emergency of rapid onset after delivery, which reviews place mainly on the bipolar spectrum, with an isolated postpartum-only form whose status remains open.
Rumination.
Repetitive, self-focused thinking about distress; in a prospective perinatal study, its brooding form predicted symptom change to two months after birth while reflection did not.
Screening.
Systematic questioning of people without a diagnosis to find those who need assessment; whether universal perinatal depression screening improves outcomes, and not only detection, is disputed.
Sensitivity.
The proportion of true cases that a test detects; for the Edinburgh scale at a cut-off of 11 or higher it was 0.81 in an individual participant data meta-analysis.
Specificity.
The proportion of people without the condition whom a test correctly classifies as negative; for the Edinburgh scale at a cut-off of 11 or higher it was 0.88.
Still-face procedure.
A face-to-face laboratory procedure for studying how infants respond when a parent's usual responsiveness is withdrawn, with its own review and meta-analysis literature; in related experiments, infants became more negative when mothers simulated depressed affect.
Treatment cascade.
The sequence from identification through treatment and adequate treatment to remission; for postpartum depression, weighted mean rates in a systematic review fell from 30.8% identified to 3.2% in remission.
Zuranolone.
An oral neuroactive steroid taken for two weeks and first approved in August 2023; an independent review judged with moderate certainty that it improves depression against placebo, with more adverse events and no follow-up beyond 45 days.

Key Researchers

Cheryl Tatano Beck

Nurse researcher at the University of Connecticut School of Nursing who synthesised qualitative studies of postpartum depression, pooled its predictors in a meta-analysis and developed the Postpartum Depression Screening Scale. ORCID - Wikipedia - Google Scholar - Faculty Page

John L. Cox

Psychiatrist and professor emeritus at Keele University who led the development of the Edinburgh Postnatal Depression Scale and a controlled study of the onset and prevalence of postnatal depression. ORCID

Kristina M. Deligiannidis

Psychiatrist at Zucker Hillside Hospital and the Feinstein Institutes for Medical Research, Northwell Health, who led the phase 3 trials of zuranolone in postpartum depression. ORCID

Cindy-Lee Dennis

Professor at the Lawrence S. Bloomberg Faculty of Nursing, University of Toronto, who reviewed barriers to help-seeking and interventions that prevent postpartum depression, and led a trial of telephone peer support. Wikipedia - Google Scholar - Faculty Page

Louise M. Howard

Psychiatrist at King's College London, in its psychiatry institute, who reviewed non-psychotic perinatal mental disorders and studied domestic violence and suicidal ideation in perinatal depression. ORCID - Faculty Page

Louis-Victor Marcé

(1828–1864). Physician at Bicêtre Hospital in Paris who wrote the 1858 treatise on the mental disorders of pregnant, recently delivered and nursing women. Wikipedia

Samantha Meltzer-Brody

Psychiatrist at the University of North Carolina at Chapel Hill who led the phase 3 trials of brexanolone and studied placental stress hormones and GABA signalling in postpartum depression. ORCID

Lynne Murray

Developmental psychologist at the University of Reading whose longitudinal studies followed the infants of mothers with postnatal depression into adolescence and examined mother-infant interaction and speech. Wikipedia - Faculty Page

Michael W. O'Hara

Psychologist at the University of Iowa who ran controlled prospective studies of postpartum depression, pooled its prevalence and risk factors in a meta-analysis and tested interpersonal psychotherapy in a randomised trial. Faculty Page

Brice Pitt

(1931–2021). Psychiatrist at St Mary's Hospital and Imperial College London who published an early study of depression following childbirth in 1968. Obituary

Alan Stein

Child and adolescent psychiatrist at the University of Oxford who studied how maternal depression and anxiety affect responsiveness to infants and reviewed the effects of perinatal mental disorders on the child. ORCID - Faculty Page

Katherine L. Wisner

Psychiatrist at Children's National Hospital and the George Washington University School of Medicine who led a screening study of ten thousand postpartum women and wrote on the definition of the postpartum onset specifier. ORCID

Frequently Asked Questions

What is the difference between postpartum depression and the baby blues?

The blues are a brief disturbance of mood in the first days after delivery that is generally regarded as normal, whereas postpartum depression is a depressive episode that persists. The two are linked: a review of mood disturbance in the early puerperium found the blues more likely in women with dysphoria in pregnancy, a history of depression or high neuroticism, and treated the blues as an apparent index of vulnerability to later postpartum depression (Henshaw, 2003).

Does postpartum depression start only after the birth?

No. In a series of ten thousand postpartum mothers who were screened, women who screened positive reported onset after the birth in 40.1% of cases, during pregnancy in 33.4% and before pregnancy in 26.5%, although only 59.2% completed home visits and a further 10.5% a telephone diagnostic interview (Wisner et al., 2013).

Is postpartum depression a separate diagnosis in DSM-5-TR?

It is not. DSM-5-TR records it as a depressive episode with the specifier with peripartum onset, which merges onset in pregnancy with onset after delivery; a commentary on DSM-5 argued that the two onsets should be separated and that the four-week window should be extended to six months (Sharma & Mazmanian, 2014).

How is postpartum depression screened for?

Most programmes use the Edinburgh Postnatal Depression Scale, a ten-item self-report questionnaire. In an individual participant data meta-analysis of 58 studies, a cut-off of 11 or higher gave a sensitivity of 0.81 and a specificity of 0.88, and a cut-off of 13 or higher was more specific but less sensitive, so a positive screen leads to a clinical assessment rather than a diagnosis (Levis et al., 2020).

Does postpartum depression affect thinking, memory or how a mother reads her baby?

The clearest evidence concerns the reading of infant emotion. A systematic review found that depressed and anxious mothers identified sadness in infant faces more readily and happiness less accurately than controls, and that depressed mothers may disengage faster from infant faces, biases its authors suggest may affect mother-infant interaction (Webb & Ayers, 2015). Studies of memory and executive function are fewer and smaller.

Can fathers have postpartum depression?

Yes. A meta-analysis of 43 studies with 28,004 participants estimated that about 10.4% of fathers were depressed during the partner's pregnancy or after the birth, with higher rates three to six months after birth and a correlation (r = 0.308) with depression in the mother (Paulson & Bazemore, 2010).

Does postpartum depression harm the child?

Maternal depression is associated with more emotional and behavioural problems in children, but the associations are small. A meta-analysis of 193 studies found small links with internalizing and externalizing problems and with negative and positive affect, and its authors called for models that go beyond main effects (Goodman et al., 2011). An association of this kind does not by itself show that the depression causes the problems.

How is postpartum depression treated?

Psychological therapy is the best-supported treatment. A meta-analysis of 43 randomised trials with 6270 participants found that psychotherapy outperformed control conditions for perinatal depression with a moderate to large effect that lasted six to twelve months, although the trials were heterogeneous and showed some publication bias (Cuijpers et al., 2023). Antidepressants and newer neuroactive steroid drugs are also used, with the evidence described in the treatment sections.

Support Organizations

Organizations that provide information, peer support and helplines for postpartum depression and related perinatal conditions.

Postpartum Support International — a nonprofit that runs a helpline, online and local support groups, a peer mentor programme and a directory of providers. (United States)

988 Suicide & Crisis Lifeline — free, confidential crisis support by call, text or online chat, available at any hour. (United States)

PANDAS Foundation — a registered charity offering a bookable helpline call, a messaging support service and support groups for parents with perinatal mental illness. (United Kingdom)

PANDA — a registered charity with a national helpline and information for expecting and new parents and their families. (Australia)

References

Abramson, L. Y., Metalsky, G. I., & Alloy, L. B. (1989). Hopelessness depression: A theory-based subtype of depression. Psychological Review, 96(2), 358-372. https://doi.org/10.1037/0033-295X.96.2.358

Admon, L. K., Dalton, V. K., Kolenic, G. E., Ettner, S. L., Tilea, A., Haffajee, R. L., Brownlee, R. M., Zochowski, M. K., Tabb, K. M., Muzik, M., & Zivin, K. (2021). Trends in suicidality 1 year before and after birth among commercially insured childbearing individuals in the United States, 2006-2017. JAMA Psychiatry, 78(2), 171-176. https://doi.org/10.1001/jamapsychiatry.2020.3550

American College of Obstetricians and Gynecologists. (2018). ACOG Committee Opinion No. 757: Screening for perinatal depression. Obstetrics & Gynecology, 132(5), e208-e212. https://doi.org/10.1097/aog.0000000000002927

American College of Obstetricians and Gynecologists. (2023). Treatment and management of mental health conditions during pregnancy and postpartum: ACOG Clinical Practice Guideline No. 5. Obstetrics & Gynecology, 141(6), 1262-1288. https://doi.org/10.1097/AOG.0000000000005202

American College of Obstetricians and Gynecologists' Committee on Clinical Practice Guidelines–Obstetrics. (2023). Screening and diagnosis of mental health conditions during pregnancy and postpartum: ACOG Clinical Practice Guideline No. 4. Obstetrics & Gynecology, 141(6), 1232-1261. https://doi.org/10.1097/AOG.0000000000005200

American Psychiatric Association. (2022). Diagnostic and statistical manual of mental disorders (5th ed., text rev.). American Psychiatric Association Publishing. https://doi.org/10.1176/appi.books.9780890425787

Appleby, L., Warner, R., Whitton, A., & Faragher, B. (1997). A controlled study of fluoxetine and cognitive-behavioural counselling in the treatment of postnatal depression. BMJ, 314(7085), 932-936. https://doi.org/10.1136/bmj.314.7085.932

Arafa, A., & Dong, J.-Y. (2019). Gestational diabetes and risk of postpartum depressive symptoms: A meta-analysis of cohort studies. Journal of Affective Disorders, 253, 312-316. https://doi.org/10.1016/j.jad.2019.05.001

Arteche, A., Joormann, J., Harvey, A., Craske, M., Gotlib, I. H., Lehtonen, A., Counsell, N., & Stein, A. (2011). The effects of postnatal maternal depression and anxiety on the processing of infant faces. Journal of Affective Disorders, 133(1-2), 197-203. https://doi.org/10.1016/j.jad.2011.04.015

Ayers, S., Bond, R., Bertullies, S., & Wijma, K. (2016). The aetiology of post-traumatic stress following childbirth: a meta-analysis and theoretical framework. Psychological Medicine, 46(6), 1121-1134. https://doi.org/10.1017/S0033291715002706

Barnum, S. E., Woody, M. L., & Gibb, B. E. (2013). Predicting changes in depressive symptoms from pregnancy to postpartum: The role of brooding rumination and negative inferential styles. Cognitive Therapy and Research, 37(1), 71-77. https://doi.org/10.1007/s10608-012-9456-5

Batt, M. M., Duffy, K. A., Novick, A. M., Metcalf, C. A., & Epperson, C. N. (2020). Is postpartum depression different from depression occurring outside of the perinatal period? A review of the evidence. Focus, 18(2), 106-119. https://doi.org/10.1176/appi.focus.20190045

Bauman, B. L., Ko, J. Y., Cox, S., D'Angelo, D. V., Warner, L., Folger, S., Tevendale, H. D., Coy, K. C., Harrison, L., & Barfield, W. D. (2020). Vital signs: Postpartum depressive symptoms and provider discussions about perinatal depression - United States, 2018. MMWR Morbidity and Mortality Weekly Report, 69(19), 575-581. https://doi.org/10.15585/mmwr.mm6919a2

Beck, A. T., & Bredemeier, K. (2016). A unified model of depression: Integrating clinical, cognitive, biological, and evolutionary perspectives. Clinical Psychological Science, 4(4), 596-619. https://doi.org/10.1177/2167702616628523

Beck, A. T. (2008). The evolution of the cognitive model of depression and its neurobiological correlates. American Journal of Psychiatry, 165(8), 969-977. https://doi.org/10.1176/appi.ajp.2008.08050721

Beck, C. T., & Gable, R. K. (2000). Postpartum Depression Screening Scale: development and psychometric testing. Nursing Research, 49(5), 272-282. https://doi.org/10.1097/00006199-200009000-00006

Beck, C. T. (2001). Predictors of postpartum depression: An update. Nursing Research, 50(5), 275-285. https://doi.org/10.1097/00006199-200109000-00004

Beck, C. T. (2002). Postpartum depression: A metasynthesis. Qualitative Health Research, 12(4), 453-472. https://doi.org/10.1177/104973202129120016

Bei, B., Milgrom, J., Ericksen, J., & Trinder, J. (2010). Subjective perception of sleep, but not its objective quality, is associated with immediate postpartum mood disturbances in healthy women. Sleep, 33(4), 531-538. https://doi.org/10.1093/sleep/33.4.531

Bell, A. F., & Andersson, E. (2016). The birth experience and women's postnatal depression: A systematic review. Midwifery, 39, 112-123. https://doi.org/10.1016/j.midw.2016.04.014

Bergink, V., Bouvy, P. F., Vervoort, J. S. P., Koorengevel, K. M., Steegers, E. A. P., & Kushner, S. A. (2012). Prevention of postpartum psychosis and mania in women at high risk. American Journal of Psychiatry, 169(6), 609-615. https://doi.org/10.1176/appi.ajp.2012.11071047

Bergink, V., Burgerhout, K. M., Koorengevel, K. M., Kamperman, A. M., Hoogendijk, W. J., Lambregtse-van den Berg, M. P., & Kushner, S. A. (2015). Treatment of psychosis and mania in the postpartum period. American Journal of Psychiatry, 172(2), 115-123. https://doi.org/10.1176/appi.ajp.2014.13121652

Bergink, V., Rasgon, N., & Wisner, K. L. (2016). Postpartum psychosis: Madness, mania, and melancholia in motherhood. American Journal of Psychiatry, 173(12), 1179-1188. https://doi.org/10.1176/appi.ajp.2016.16040454

Biaggi, A., Conroy, S., Pawlby, S., & Pariante, C. M. (2016). Identifying the women at risk of antenatal anxiety and depression: A systematic review. Journal of Affective Disorders, 191, 62-77. https://doi.org/10.1016/j.jad.2015.11.014

Bjertrup, A. J., Friis, N. K., & Miskowiak, K. W. (2019). The maternal brain: Neural responses to infants in mothers with and without mood disorder. Neuroscience & Biobehavioral Reviews, 107, 196-207. https://doi.org/10.1016/j.neubiorev.2019.09.011

Bjertrup, A., Rasmussen, K. H. E., Tørring, S., & Miskowiak, K. W. (2026). Targeting affective cognition during pregnancy to prevent postpartum depression: A randomized controlled trial with parity-specific effects. European Neuropsychopharmacology, 114, 113885. https://doi.org/10.1016/j.euroneuro.2026.113885

Bloch, M., Schmidt, P. J., Danaceau, M., Murphy, J., Nieman, L., & Rubinow, D. R. (2000). Effects of gonadal steroids in women with a history of postpartum depression. American Journal of Psychiatry, 157(6), 924-930. https://doi.org/10.1176/appi.ajp.157.6.924

Borra, C., Iacovou, M., & Sevilla, A. (2015). New evidence on breastfeeding and postpartum depression: The importance of understanding women's intentions. Maternal and Child Health Journal, 19(4), 897-907. https://doi.org/10.1007/s10995-014-1591-z

Brockington, I. (2004). Postpartum psychiatric disorders. The Lancet, 363(9405), 303-310. https://doi.org/10.1016/S0140-6736(03)15390-1

Brockington, I. (2017). Suicide and filicide in postpartum psychosis. Archives of Women's Mental Health, 20(1), 63-69. https://doi.org/10.1007/s00737-016-0675-8

Brockington, I. F., Oates, J., George, S., Turner, D., Vostanis, P., Sullivan, M., Loh, C., & Murdoch, C. (2001). A screening questionnaire for mother-infant bonding disorders. Archives of Women's Mental Health, 3(4), 133-140. https://doi.org/10.1007/s007370170010

Brown, J. V. E., Wilson, C. A., Ayre, K., Robertson, L., South, E., Molyneaux, E., Trevillion, K., Howard, L. M., & Khalifeh, H. (2021). Antidepressant treatment for postnatal depression. Cochrane Database of Systematic Reviews, 2021(2), CD013560. https://doi.org/10.1002/14651858.CD013560.pub2

Button, S., Thornton, A., Lee, S., Shakespeare, J., & Ayers, S. (2017). Seeking help for perinatal psychological distress: a meta-synthesis of women's experiences. British Journal of General Practice, 67(663), e692-e699. https://doi.org/10.3399/bjgp17X692549

Caropreso, L., de Azevedo Cardoso, T., Eltayebani, M., & Frey, B. N. (2020). Preeclampsia as a risk factor for postpartum depression and psychosis: a systematic review and meta-analysis. Archives of Women's Mental Health, 23(4), 493-505. https://doi.org/10.1007/s00737-019-01010-1

Chapman, S. L. C., & Wu, L.-T. (2013). Postpartum substance use and depressive symptoms: A review. Women & Health, 53(5), 479-503. https://doi.org/10.1080/03630242.2013.804025

Cohen, L. S., Altshuler, L. L., Harlow, B. L., et al. (2006). Relapse of major depression during pregnancy in women who maintain or discontinue antidepressant treatment. JAMA, 295(5), 499-507. https://doi.org/10.1001/jama.295.5.499

Cohn, J. F., & Tronick, E. Z. (1983). Three-month-old infants' reaction to simulated maternal depression. Child Development, 54(1), 185-193. https://doi.org/10.2307/1129876

Cooper, P. J., & Murray, L. (1995). Course and recurrence of postnatal depression: Evidence for the specificity of the diagnostic concept. British Journal of Psychiatry, 166(2), 191-195. https://doi.org/10.1192/bjp.166.2.191

Cooper, P. J., Murray, L., Wilson, A., & Romaniuk, H. (2003). Controlled trial of the short- and long-term effect of psychological treatment of post-partum depression: 1. Impact on maternal mood. British Journal of Psychiatry, 182(5), 412-419. https://doi.org/10.1192/bjp.182.5.412

Cox, E. Q., Sowa, N. A., Meltzer-Brody, S. E., & Gaynes, B. N. (2016). The perinatal depression treatment cascade: Baby steps toward improving outcomes. Journal of Clinical Psychiatry, 77(9), 1189-1200. https://doi.org/10.4088/jcp.15r10174

Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression: Development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry, 150(6), 782-786. https://doi.org/10.1192/bjp.150.6.782

Cox, J. L., Murray, D., & Chapman, G. (1993). A controlled study of the onset, duration and prevalence of postnatal depression. British Journal of Psychiatry, 163(1), 27-31. https://doi.org/10.1192/bjp.163.1.27

Cuijpers, P., Franco, P., Ciharova, M., Miguel, C., Segre, L., Quero, S., & Karyotaki, E. (2023). Psychological treatment of perinatal depression: a meta-analysis. Psychological Medicine, 53(6), 2596-2608. https://doi.org/10.1017/S0033291721004529

Cuijpers, P., Weitz, E., Karyotaki, E., Garber, J., & Andersson, G. (2015). The effects of psychological treatment of maternal depression on children and parental functioning: a meta-analysis. European Child & Adolescent Psychiatry, 24(2), 237-245. https://doi.org/10.1007/s00787-014-0660-6

Cutrona, C. E., & Troutman, B. R. (1986). Social support, infant temperament, and parenting self-efficacy: A mediational model of postpartum depression. Child Development, 57(6), 1507-1518. https://doi.org/10.2307/1130428

Cutrona, C. E. (1983). Causal attributions and perinatal depression. Journal of Abnormal Psychology, 92(2), 161-172. https://doi.org/10.1037/0021-843X.92.2.161

Darwin, Z., Galdas, P., Hinchliff, S., Littlewood, E., McMillan, D., McGowan, L., Gilbody, S., & Born and Bred in Yorkshire (BaBY) team. (2017). Fathers' views and experiences of their own mental health during pregnancy and the first postnatal year: A qualitative interview study of men participating in the UK Born and Bred in Yorkshire (BaBY) cohort. BMC Pregnancy and Childbirth, 17(1), 45. https://doi.org/10.1186/s12884-017-1229-4

Davies, S. J., Lum, J. A. G., Skouteris, H., Byrne, L. K., & Hayden, M. J. (2018). Cognitive impairment during pregnancy: a meta-analysis. Medical Journal of Australia, 208(1), 35-40. https://doi.org/10.5694/mja17.00131

DeJong, H., Fox, E., & Stein, A. (2016). Rumination and postnatal depression: A systematic review and a cognitive model. Behaviour Research and Therapy, 82, 38-49. https://doi.org/10.1016/j.brat.2016.05.003

Deligiannidis, K. M., Bullock, A., Nandy, I., Dunbar, J., Lasser, R., Witte, M., Leclair, B., & Wald, J. (2024). Zuranolone concentrations in the breast milk of healthy, lactating individuals: Results from a phase 1 open-label study. Journal of Clinical Psychopharmacology, 44(4), 337-344. https://doi.org/10.1097/jcp.0000000000001873

Deligiannidis, K. M., Meltzer-Brody, S., Gunduz-Bruce, H., Doherty, J., Jonas, J., Li, S., Sankoh, A. J., Silber, C., Campbell, A. D., Werneburg, B., Kanes, S. J., & Lasser, R. (2021). Effect of zuranolone vs placebo in postpartum depression: A randomized clinical trial. JAMA Psychiatry, 78(9), 951-959. https://doi.org/10.1001/jamapsychiatry.2021.1559

Deligiannidis, K. M., Meltzer-Brody, S., Maximos, B., Peeper, E. Q., Freeman, M., Lasser, R., Bullock, A., Kotecha, M., Li, S., Forrestal, F., Rana, N., Garcia, M., Leclair, B., & Doherty, J. (2023). Zuranolone for the treatment of postpartum depression. American Journal of Psychiatry, 180(9), 668-675. https://doi.org/10.1176/appi.ajp.20220785

Dennis, C.-L., & Chung-Lee, L. (2006). Postpartum depression help-seeking barriers and maternal treatment preferences: A qualitative systematic review. Birth, 33(4), 323-331. https://doi.org/10.1111/j.1523-536X.2006.00130.x

Dennis, C.-L., & Dowswell, T. (2013). Psychosocial and psychological interventions for preventing postpartum depression. Cochrane Database of Systematic Reviews, (2), CD001134. https://doi.org/10.1002/14651858.CD001134.pub3

Dennis, C.-L., Falah-Hassani, K., & Shiri, R. (2017). Prevalence of antenatal and postnatal anxiety: Systematic review and meta-analysis. British Journal of Psychiatry, 210(5), 315-323. https://doi.org/10.1192/bjp.bp.116.187179

Dennis, C.-L., Hodnett, E., Kenton, L., Weston, J., Zupancic, J., Stewart, D. E., & Kiss, A. (2009). Effect of peer support on prevention of postnatal depression among high risk women: multisite randomised controlled trial. BMJ, 338, a3064. https://doi.org/10.1136/bmj.a3064

Di Florio, A., & Meltzer-Brody, S. (2015). Is postpartum depression a distinct disorder? Current Psychiatry Reports, 17(10), 76. https://doi.org/10.1007/s11920-015-0617-6

Di Florio, A., Forty, L., Gordon-Smith, K., Heron, J., Jones, L., Craddock, N., & Jones, I. (2013). Perinatal episodes across the mood disorder spectrum. JAMA Psychiatry, 70(2), 168-175. https://doi.org/10.1001/jamapsychiatry.2013.279

Dias, C. C., & Figueiredo, B. (2015). Breastfeeding and depression: a systematic review of the literature. Journal of Affective Disorders, 171, 142-154. https://doi.org/10.1016/j.jad.2014.09.022

Dimidjian, S., Goodman, S. H., Sherwood, N. E., Simon, G. E., Ludman, E., Gallop, R., Welch, S. S., Boggs, J. M., Metcalf, C. A., Hubley, S., Powers, J. D., & Beck, A. (2017). A pragmatic randomized clinical trial of behavioral activation for depressed pregnant women. Journal of Consulting and Clinical Psychology, 85(1), 26-36. https://doi.org/10.1037/ccp0000151

Dinwiddie, K. J., Schillerstrom, T. L., & Schillerstrom, J. E. (2018). Postpartum depression in adolescent mothers. Journal of Psychosomatic Obstetrics & Gynecology, 39(3), 168-175. https://doi.org/10.1080/0167482x.2017.1334051

Disner, S. G., Beevers, C. G., Haigh, E. A. P., & Beck, A. T. (2011). Neural mechanisms of the cognitive model of depression. Nature Reviews Neuroscience, 12(8), 467-477. https://doi.org/10.1038/nrn3027

Dørheim, S. K., Bondevik, G. T., Eberhard-Gran, M., & Bjorvatn, B. (2009). Sleep and depression in postpartum women: A population-based study. Sleep, 32(7), 847-855. https://doi.org/10.1093/sleep/32.7.847

Duan, C., Cosgrove, J., & Deligiannidis, K. M. (2017). Understanding peripartum depression through neuroimaging: a review of structural and functional connectivity and molecular imaging research. Current Psychiatry Reports, 19(10), 70. https://doi.org/10.1007/s11920-017-0824-4

Earls, M. F., Yogman, M. W., Mattson, G., Rafferty, J., & Committee on Psychosocial Aspects of Child and Family Health. (2019). Incorporating recognition and management of perinatal depression into pediatric practice. Pediatrics, 143(1), e20183259. https://doi.org/10.1542/peds.2018-3259

Edmondson, O. J. H., Psychogiou, L., Vlachos, H., Netsi, E., & Ramchandani, P. G. (2010). Depression in fathers in the postnatal period: Assessment of the Edinburgh Postnatal Depression Scale as a screening measure. Journal of Affective Disorders, 125(1-3), 365-368. https://doi.org/10.1016/j.jad.2010.01.069

Edvinsson, Å., Skalkidou, A., Hellgren, C., Gingnell, M., Ekselius, L., Willebrand, M., & Sundström Poromaa, I. (2017). Different patterns of attentional bias in antenatal and postpartum depression. Brain and Behavior, 7(11), e00844. https://doi.org/10.1002/brb3.844

Evans, J., Heron, J., Francomb, H., Oke, S., & Golding, J. (2001). Cohort study of depressed mood during pregnancy and after childbirth. BMJ, 323(7307), 257-260. https://doi.org/10.1136/bmj.323.7307.257

Evans, J., Heron, J., Lewis, G., Araya, R., & Wolke, D. (2005). Negative self-schemas and the onset of depression in women: longitudinal study. British Journal of Psychiatry, 186(4), 302-307. https://doi.org/10.1192/bjp.186.4.302

Fairbrother, N., & Woody, S. R. (2008). New mothers' thoughts of harm related to the newborn. Archives of Women's Mental Health, 11(3), 221-229. https://doi.org/10.1007/s00737-008-0016-7

Fairbrother, N., Collardeau, F., Woody, S. R., Wolfe, D. A., & Fawcett, J. M. (2022). Postpartum thoughts of infant-related harm and obsessive-compulsive disorder: Relation to maternal physical aggression toward the infant. Journal of Clinical Psychiatry, 83(2), 21m14006. https://doi.org/10.4088/jcp.21m14006

Falah-Hassani, K., Shiri, R., & Dennis, C.-L. (2017). The prevalence of antenatal and postnatal co-morbid anxiety and depression: a meta-analysis. Psychological Medicine, 47(12), 2041-2053. https://doi.org/10.1017/S0033291717000617

Falah-Hassani, K., Shiri, R., Vigod, S., & Dennis, C.-L. (2015). Prevalence of postpartum depression among immigrant women: A systematic review and meta-analysis. Journal of Psychiatric Research, 70, 67-82. https://doi.org/10.1016/j.jpsychires.2015.08.010

Field, T. (2010). Postpartum depression effects on early interactions, parenting, and safety practices: A review. Infant Behavior and Development, 33(1), 1-6. https://doi.org/10.1016/j.infbeh.2009.10.005

First, M. B., Gaebel, W., Maj, M., Stein, D. J., Kogan, C. S., Saunders, J. B., Poznyak, V. B., Gureje, O., Lewis-Fernández, R., Maercker, A., Brewin, C. R., Cloitre, M., Claudino, A., Pike, K. M., Baird, G., Skuse, D., Krueger, R. B., Briken, P., Burke, J. D., Lochman, J. E., Evans, S. C., Woods, D. W., & Reed, G. M. (2021). An organization- and category-level comparison of diagnostic requirements for mental disorders in ICD-11 and DSM-5. World Psychiatry, 20(1), 34-51. https://doi.org/10.1002/wps.20825

Fisher, J., Cabral de Mello, M., Patel, V., Rahman, A., Tran, T., Holton, S., & Holmes, W. (2012). Prevalence and determinants of common perinatal mental disorders in women in low- and lower-middle-income countries: a systematic review. Bulletin of the World Health Organization, 90(2), 139-149H. https://doi.org/10.2471/BLT.11.091850

Fisher, S. D., Cobo, J., Figueiredo, B., Fletcher, R., Garfield, C. F., Hanley, J., Ramchandani, P., & Singley, D. B. (2021). Expanding the international conversation with fathers' mental health: Toward an era of inclusion in perinatal research and practice. Archives of Women's Mental Health, 24(5), 841-848. https://doi.org/10.1007/s00737-021-01171-y

Forman, D. R., O'Hara, M. W., Stuart, S., Gorman, L. L., Larsen, K. E., & Coy, K. C. (2007). Effective treatment for postpartum depression is not sufficient to improve the developing mother-child relationship. Development and Psychopathology, 19(2), 585-602. https://doi.org/10.1017/S0954579407070289

Forty, L., Jones, L., Macgregor, S., Caesar, S., Cooper, C., Hough, A., Dean, L., Dave, S., Farmer, A., McGuffin, P., Brewster, S., Craddock, N., & Jones, I. (2006). Familiality of postpartum depression in unipolar disorder: Results of a family study. American Journal of Psychiatry, 163(9), 1549-1553. https://doi.org/10.1176/ajp.2006.163.9.1549

Gavin, N. I., Gaynes, B. N., Lohr, K. N., Meltzer-Brody, S., Gartlehner, G., & Swinson, T. (2005). Perinatal depression: a systematic review of prevalence and incidence. Obstetrics & Gynecology, 106(5), 1071-1083. https://doi.org/10.1097/01.AOG.0000183597.31630.db

Gibson, J., McKenzie-McHarg, K., Shakespeare, J., Price, J., & Gray, R. (2009). A systematic review of studies validating the Edinburgh Postnatal Depression Scale in antepartum and postpartum women. Acta Psychiatrica Scandinavica, 119(5), 350-364. https://doi.org/10.1111/j.1600-0447.2009.01363.x

Gillham, R., & Wittkowski, A. (2015). Outcomes for women admitted to a mother and baby unit: A systematic review. International Journal of Women's Health, 7, 459-476. https://doi.org/10.2147/ijwh.s69472

Gjerde, L. C., Eilertsen, E. M., Reichborn-Kjennerud, T., McAdams, T. A., Zachrisson, H. D., Zambrana, I. M., Røysamb, E., Kendler, K. S., & Ystrom, E. (2017). Maternal perinatal and concurrent depressive symptoms and child behavior problems: a sibling comparison study. Journal of Child Psychology and Psychiatry, 58(7), 779-786. https://doi.org/10.1111/jcpp.12704

Godderis, R. (2013). A tricky object to classify: Evidence, postpartum depression and the DSM-IV. Journal of the History of the Behavioral Sciences, 49(2), 123-141. https://doi.org/10.1002/jhbs.21590

Goodman, S. H., Rouse, M. H., Connell, A. M., Broth, M. R., Hall, C. M., & Heyward, D. (2011). Maternal depression and child psychopathology: A meta-analytic review. Clinical Child and Family Psychology Review, 14(1), 1-27. https://doi.org/10.1007/s10567-010-0080-1

Grace, S. L., Evindar, A., & Stewart, D. E. (2003). The effect of postpartum depression on child cognitive development and behavior: A review and critical analysis of the literature. Archives of Women's Mental Health, 6(4), 263-274. https://doi.org/10.1007/s00737-003-0024-6

Gregoire, A. J. P., Kumar, R., Everitt, B., Henderson, A. F., & Studd, J. W. W. (1996). Transdermal oestrogen for treatment of severe postnatal depression. The Lancet, 347(9006), 930-933. https://doi.org/10.1016/s0140-6736(96)91414-2

Gressier, F., Rotenberg, S., Cazas, O., & Hardy, P. (2015). Postpartum electroconvulsive therapy: A systematic review and case report. General Hospital Psychiatry, 37(4), 310-314. https://doi.org/10.1016/j.genhosppsych.2015.04.009

Grigoriadis, S., Erlick Robinson, G., Fung, K., Ross, L. E., Chee, C. Y. I., Dennis, C.-L., & Romans, S. (2009). Traditional postpartum practices and rituals: Clinical implications. The Canadian Journal of Psychiatry, 54(12), 834-840. https://doi.org/10.1177/070674370905401206

Grigoriadis, S., Wilton, A. S., Kurdyak, P. A., Rhodes, A. E., VonderPorten, E. H., Levitt, A., Cheung, A., & Vigod, S. N. (2017). Perinatal suicide in Ontario, Canada: a 15-year population-based study. CMAJ, 189(34), E1085-E1092. https://doi.org/10.1503/cmaj.170088

Guintivano, J., Arad, M., Gould, T. D., Payne, J. L., & Kaminsky, Z. A. (2014). Antenatal prediction of postpartum depression with blood DNA methylation biomarkers. Molecular Psychiatry, 19(5), 560-567. https://doi.org/10.1038/mp.2013.62

Guintivano, J., Byrne, E. M., Kiewa, J., Yao, S., Bauer, A. E., et al. (2023). Meta-analyses of genome-wide association studies for postpartum depression. American Journal of Psychiatry, 180(12), 884-895. https://doi.org/10.1176/appi.ajp.20230053

Hagen, E. H. (1999). The functions of postpartum depression. Evolution and Human Behavior, 20(5), 325-359. https://doi.org/10.1016/S1090-5138(99)00016-1

Hahn-Holbrook, J., & Haselton, M. (2014). Is postpartum depression a disease of modern civilization? Current Directions in Psychological Science, 23(6), 395-400. https://doi.org/10.1177/0963721414547736

Hahn-Holbrook, J., Cornwell-Hinrichs, T., & Anaya, I. (2018). Economic and health predictors of national postpartum depression prevalence: A systematic review, meta-analysis, and meta-regression of 291 studies from 56 countries. Frontiers in Psychiatry, 8, 248. https://doi.org/10.3389/fpsyt.2017.00248

Hahn-Holbrook, J., Haselton, M. G., Dunkel Schetter, C., & Glynn, L. M. (2013). Does breastfeeding offer protection against maternal depressive symptomatology? A prospective study from pregnancy to 2 years after birth. Archives of Women's Mental Health, 16(5), 411-422. https://doi.org/10.1007/s00737-013-0348-9

Halbreich, U., & Karkun, S. (2006). Cross-cultural and social diversity of prevalence of postpartum depression and depressive symptoms. Journal of Affective Disorders, 91(2-3), 97-111. https://doi.org/10.1016/j.jad.2005.12.051

Hampson, E., Phillips, S.-D., Duff-Canning, S. J., Evans, K. L., Merrill, M., Pinsonneault, J. K., Sadée, W., Soares, C. N., & Steiner, M. (2015). Working memory in pregnant women: Relation to estrogen and antepartum depression. Hormones and Behavior, 74, 218-227. https://doi.org/10.1016/j.yhbeh.2015.07.006

Hannigan, L. J., Eilertsen, E. M., Gjerde, L. C., Reichborn-Kjennerud, T., Eley, T. C., Rijsdijk, F. V., Ystrom, E., & McAdams, T. A. (2018). Maternal prenatal depressive symptoms and risk for early-life psychopathology in offspring: genetic analyses in the Norwegian Mother and Child Birth Cohort Study. The Lancet Psychiatry, 5(10), 808-815. https://doi.org/10.1016/s2215-0366(18)30225-6

Henshaw, C. (2003). Mood disturbance in the early puerperium: a review. Archives of Women's Mental Health, 6(Suppl. 2), S33-S42. https://doi.org/10.1007/s00737-003-0004-x

Heo, Y.-A. (2023). Zuranolone: First approval. Drugs, 83(16), 1559-1567. https://doi.org/10.1007/s40265-023-01953-x

Herbert, D., Young, K., Pietrusińska, M., & MacBeth, A. (2022). The mental health impact of perinatal loss: A systematic review and meta-analysis. Journal of Affective Disorders, 297, 118-129. https://doi.org/10.1016/j.jad.2021.10.026

Heron, J., O'Connor, T. G., Evans, J., Golding, J., & Glover, V. (2004). The course of anxiety and depression through pregnancy and the postpartum in a community sample. Journal of Affective Disorders, 80(1), 65-73. https://doi.org/10.1016/j.jad.2003.08.004

Hewitt, C., Gilbody, S., Brealey, S., Paulden, M., Palmer, S., Mann, R., Green, J., Morrell, J., Barkham, M., Light, K., & Richards, D. (2009). Methods to identify postnatal depression in primary care: An integrated evidence synthesis and value of information analysis. Health Technology Assessment, 13(36), 1-145. https://doi.org/10.3310/hta13360

Hipwell, A. E., Reynolds, S., & Pitts Crick, E. (2004). Cognitive vulnerability to postnatal depressive symptomatology. Journal of Reproductive and Infant Psychology, 22(3), 211-227. https://doi.org/10.1080/02646830410001723797

Hiscock, H., Bayer, J., Gold, L., Hampton, A., Ukoumunne, O. C., & Wake, M. (2007). Improving infant sleep and maternal mental health: A cluster randomised trial. Archives of Disease in Childhood, 92(11), 952-958. https://doi.org/10.1136/adc.2006.099812

Hoekzema, E., Barba-Müller, E., Pozzobon, C., Picado, M., Lucco, F., García-García, D., Soliva, J. C., Tobeña, A., Desco, M., Crone, E. A., Ballesteros, A., Carmona, S., & Vilarroya, O. (2017). Pregnancy leads to long-lasting changes in human brain structure. Nature Neuroscience, 20(2), 287-296. https://doi.org/10.1038/nn.4458

Horáková, A., Němcová, H., Mohr, P., & Sebela, A. (2022). Structural, functional, and metabolic signatures of postpartum depression: A systematic review. Frontiers in Psychiatry, 13, 1044995. https://doi.org/10.3389/fpsyt.2022.1044995

Howard, L. M., Flach, C., Mehay, A., Sharp, D., & Tylee, A. (2011). The prevalence of suicidal ideation identified by the Edinburgh Postnatal Depression Scale in postpartum women in primary care: findings from the RESPOND trial. BMC Pregnancy and Childbirth, 11(1), 57. https://doi.org/10.1186/1471-2393-11-57

Howard, L. M., Molyneaux, E., Dennis, C.-L., Rochat, T., Stein, A., & Milgrom, J. (2014). Non-psychotic mental disorders in the perinatal period. The Lancet, 384(9956), 1775-1788. https://doi.org/10.1016/S0140-6736(14)61276-9

Howard, L. M., Oram, S., Galley, H., Trevillion, K., & Feder, G. (2013). Domestic violence and perinatal mental disorders: A systematic review and meta-analysis. PLoS Medicine, 10(5), e1001452. https://doi.org/10.1371/journal.pmed.1001452

Hutchens, B. F., & Kearney, J. (2020). Risk factors for postpartum depression: An umbrella review. Journal of Midwifery & Women's Health, 65(1), 96-108. https://doi.org/10.1111/jmwh.13067

Jairaj, C., & Rucker, J. J. (2022). Postpartum depression: A role for psychedelics? Journal of Psychopharmacology, 36(8), 920-931. https://doi.org/10.1177/02698811221093793

Jennings, K. D., Ross, S., Popper, S., & Elmore, M. (1999). Thoughts of harming infants in depressed and nondepressed mothers. Journal of Affective Disorders, 54(1-2), 21-28. https://doi.org/10.1016/s0165-0327(98)00185-2

Johannsen, B. M. W., Larsen, J. T., Laursen, T. M., Bergink, V., Meltzer-Brody, S., & Munk-Olsen, T. (2016). All-cause mortality in women with severe postpartum psychiatric disorders. American Journal of Psychiatry, 173(6), 635-642. https://doi.org/10.1176/appi.ajp.2015.14121510

Jones, I., Chandra, P. S., Dazzan, P., & Howard, L. M. (2014). Bipolar disorder, affective psychosis, and schizophrenia in pregnancy and the post-partum period. The Lancet, 384(9956), 1789-1799. https://doi.org/10.1016/S0140-6736(14)61278-2

Jones, L., Scott, J., Cooper, C., Forty, L., Gordon Smith, K., Sham, P., Farmer, A., McGuffin, P., Craddock, N., & Jones, I. (2010). Cognitive style, personality and vulnerability to postnatal depression. British Journal of Psychiatry, 196(3), 200-205. https://doi.org/10.1192/bjp.bp.109.064683

Kane, F. J., Jr., Harman, W. J., Jr., Keeler, M. H., & Ewing, J. A. (1968). Emotional and cognitive disturbance in the early puerperium. British Journal of Psychiatry, 114(506), 99-102. https://doi.org/10.1192/bjp.114.506.99

Kanes, S., Colquhoun, H., Gunduz-Bruce, H., Raines, S., Arnold, R., Schacterle, A., Doherty, J., Epperson, C. N., Deligiannidis, K. M., Riesenberg, R., Hoffmann, E., Rubinow, D., Jonas, J., Paul, S., & Meltzer-Brody, S. (2017). Brexanolone (SAGE-547 injection) in post-partum depression: a randomised controlled trial. The Lancet, 390(10093), 480-489. https://doi.org/10.1016/S0140-6736(17)31264-3

Kaplan, P. S., Bachorowski, J.-A., & Zarlengo-Strouse, P. (1999). Child-directed speech produced by mothers with symptoms of depression fails to promote associative learning in 4-month-old infants. Child Development, 70(3), 560-570. https://doi.org/10.1111/1467-8624.00041

Kendell, R. E., Chalmers, J. C., & Platz, C. (1987). Epidemiology of puerperal psychoses. British Journal of Psychiatry, 150(5), 662-673. https://doi.org/10.1192/bjp.150.5.662

Khalifeh, H., Hunt, I. M., Appleby, L., & Howard, L. M. (2016). Suicide in perinatal and non-perinatal women in contact with psychiatric services: 15 year findings from a UK national inquiry. The Lancet Psychiatry, 3(3), 233-242. https://doi.org/10.1016/S2215-0366(16)00003-1

Kirubarajan, A., Barker, L. C., Leung, S., Ross, L. E., Zaheer, J., Park, B., Abramovich, A., Yudin, M. H., & Lam, J. S. H. (2022). LGBTQ2S+ childbearing individuals and perinatal mental health: A systematic review. BJOG: An International Journal of Obstetrics & Gynaecology, 129(10), 1630-1643. https://doi.org/10.1111/1471-0528.17103

Kowalczyk, O. S., Pauls, A. M., Fusté, M., Williams, S. C. R., Hazelgrove, K., Vecchio, C., Seneviratne, G., Pariante, C. M., Dazzan, P., & Mehta, M. A. (2021). Neurocognitive correlates of working memory and emotional processing in postpartum psychosis: An fMRI study. Psychological Medicine, 51(10), 1724-1732. https://doi.org/10.1017/S0033291720000471

Kozhimannil, K. B., Trinacty, C. M., Busch, A. B., Huskamp, H. A., & Adams, A. S. (2011). Racial and ethnic disparities in postpartum depression care among low-income women. Psychiatric Services, 62(6), 619-625. https://doi.org/10.1176/ps.62.6.pss6206_0619

Krink, S., Muehlhan, C., Luyten, P., Romer, G., & Ramsauer, B. (2018). Parental reflective functioning affects sensitivity to distress in mothers with postpartum depression. Journal of Child and Family Studies, 27(5), 1671-1681. https://doi.org/10.1007/s10826-017-1000-5

Lambert, M., & Gressier, F. (2019). Inflammatory biomarkers and postpartum depression: A systematic review of literature. Canadian Journal of Psychiatry, 64(7), 471-481. https://doi.org/10.1177/0706743719828970

Lang, E., Colquhoun, H., LeBlanc, J. C., Riva, J. J., Moore, A., Traversy, G., Wilson, B., Grad, R., & Canadian Task Force on Preventive Health Care. (2022). Recommendation on instrument-based screening for depression during pregnancy and the postpartum period. CMAJ, 194(28), E981-E989. https://doi.org/10.1503/cmaj.220290

Laurent, H. K., & Ablow, J. C. (2012). A cry in the dark: depressed mothers show reduced neural activation to their own infant's cry. Social Cognitive and Affective Neuroscience, 7(2), 125-134. https://doi.org/10.1093/scan/nsq091

Lawson, A., Murphy, K. E., Sloan, E., Uleryk, E., & Dalfen, A. (2015). The relationship between sleep and postpartum mental disorders: A systematic review. Journal of Affective Disorders, 176, 65-77. https://doi.org/10.1016/j.jad.2015.01.017

Letourneau, N. L., Dennis, C.-L., Cosic, N., & Linder, J. (2017). The effect of perinatal depression treatment for mothers on parenting and child development: A systematic review. Depression and Anxiety, 34(10), 928-966. https://doi.org/10.1002/da.22687

Levis, B., Negeri, Z., Sun, Y., Benedetti, A., & Thombs, B. D. (2020). Accuracy of the Edinburgh Postnatal Depression Scale (EPDS) for screening to detect major depression among pregnant and postpartum women: systematic review and meta-analysis of individual participant data. BMJ, 371, m4022. https://doi.org/10.1136/bmj.m4022

Li, H. J., Martinez, P. E., Li, X., Schenkel, L. A., Nieman, L. K., Rubinow, D. R., & Schmidt, P. J. (2020). Transdermal estradiol for postpartum depression: Results from a pilot randomized, double-blind, placebo-controlled study. Archives of Women's Mental Health, 23(3), 401-412. https://doi.org/10.1007/s00737-019-00991-3

Li, S., Zhou, W., Li, P., & Lin, R. (2024). Effects of ketamine and esketamine on preventing postpartum depression after cesarean delivery: A meta-analysis. Journal of Affective Disorders, 351, 720-728. https://doi.org/10.1016/j.jad.2024.01.202

Lindahl, V., Pearson, J. L., & Colpe, L. (2005). Prevalence of suicidality during pregnancy and the postpartum. Archives of Women's Mental Health, 8(2), 77-87. https://doi.org/10.1007/s00737-005-0080-1

Lipsitz, J. D., & Markowitz, J. C. (2013). Mechanisms of change in interpersonal therapy (IPT). Clinical Psychology Review, 33(8), 1134–1147. https://doi.org/10.1016/j.cpr.2013.09.002

Loudon, I. (1988). Puerperal insanity in the 19th century. Journal of the Royal Society of Medicine, 81(2), 76-79. https://doi.org/10.1177/014107688808100207

Lovejoy, M. C., Graczyk, P. A., O'Hare, E., & Neuman, G. (2000). Maternal depression and parenting behavior: A meta-analytic review. Clinical Psychology Review, 20(5), 561-592. https://doi.org/10.1016/S0272-7358(98)00100-7

Maguire, J., & Mody, I. (2008). GABA(A)R plasticity during pregnancy: relevance to postpartum depression. Neuron, 59(2), 207-213. https://doi.org/10.1016/j.neuron.2008.06.019

Manber, R., Bei, B., Simpson, N., Asarnow, L., Rangel, E., Sit, A., & Lyell, D. (2019). Cognitive behavioral therapy for prenatal insomnia: A randomized controlled trial. Obstetrics & Gynecology, 133(5), 911-919. https://doi.org/10.1097/AOG.0000000000003216

Marcé, L.-V. (1858). Traité de la folie des femmes enceintes, des nouvelles accouchées et des nourrices, et considérations médico-légales qui se rattachent à ce sujet. Paris: J.-B. Baillière.

Martins, C., & Gaffan, E. A. (2000). Effects of early maternal depression on patterns of infant-mother attachment: A meta-analytic investigation. Journal of Child Psychology and Psychiatry, 41(6), 737-746. https://doi.org/10.1111/1469-7610.00661

Matthey, S. (2008). Using the Edinburgh Postnatal Depression Scale to screen for anxiety disorders. Depression and Anxiety, 25(11), 926-931. https://doi.org/10.1002/da.20415

Matthey, S., Barnett, B., Kavanagh, D. J., & Howie, P. (2001). Validation of the Edinburgh Postnatal Depression Scale for men, and comparison of item endorsement with their partners. Journal of Affective Disorders, 64(2-3), 175-184. https://doi.org/10.1016/s0165-0327(00)00236-6

Matthey, S., Henshaw, C., Elliott, S., & Barnett, B. (2006). Variability in use of cut-off scores and formats on the Edinburgh Postnatal Depression Scale: implications for clinical and research practice. Archives of Women's Mental Health, 9(6), 309-315. https://doi.org/10.1007/s00737-006-0152-x

Meltzer-Brody, S., & Kanes, S. J. (2020). Allopregnanolone in postpartum depression: Role in pathophysiology and treatment. Neurobiology of Stress, 12, 100212. https://doi.org/10.1016/j.ynstr.2020.100212

Meltzer-Brody, S., Colquhoun, H., Riesenberg, R., Epperson, C. N., Deligiannidis, K. M., Rubinow, D. R., Li, H., Sankoh, A. J., Clemson, C., Schacterle, A., Jonas, J., & Kanes, S. (2018). Brexanolone injection in post-partum depression: two multicentre, double-blind, randomised, placebo-controlled, phase 3 trials. The Lancet, 392(10152), 1058-1070. https://doi.org/10.1016/S0140-6736(18)31551-4

Meltzer-Brody, S., Howard, L. M., Bergink, V., Vigod, S., Jones, I., Munk-Olsen, T., Honikman, S., & Milgrom, J. (2018). Postpartum psychiatric disorders. Nature Reviews Disease Primers, 4, 18022. https://doi.org/10.1038/nrdp.2018.22

Meltzer-Brody, S., Stuebe, A., Dole, N., Savitz, D., Rubinow, D., & Thorp, J. (2011). Elevated corticotropin releasing hormone (CRH) during pregnancy and risk of postpartum depression (PPD). Journal of Clinical Endocrinology & Metabolism, 96(1), E40-E47. https://doi.org/10.1210/jc.2010-0978

Mesman, J., van IJzendoorn, M. H., & Bakermans-Kranenburg, M. J. (2009). The many faces of the Still-Face Paradigm: A review and meta-analysis. Developmental Review, 29(2), 120-162. https://doi.org/10.1016/j.dr.2009.02.001

Milgrom, J., Danaher, B. G., Gemmill, A. W., Holt, C., Holt, C. J., Seeley, J. R., Tyler, M. S., Ross, J., & Ericksen, J. (2016). Internet cognitive behavioral therapy for women with postnatal depression: A randomized controlled trial of MumMoodBooster. Journal of Medical Internet Research, 18(3), e54. https://doi.org/10.2196/jmir.4993

Minaldi, E., D'Andrea, S., Castellini, C., Martorella, A., Francavilla, F., Francavilla, S., & Barbonetti, A. (2020). Thyroid autoimmunity and risk of post-partum depression: A systematic review and meta-analysis of longitudinal studies. Journal of Endocrinological Investigation, 43(3), 271-277. https://doi.org/10.1007/s40618-019-01120-8

Mocking, R. J. T., Steijn, K., Roos, C., Assies, J., Bergink, V., Ruhé, H. G., & Schene, A. H. (2020). Omega-3 fatty acid supplementation for perinatal depression: A meta-analysis. The Journal of Clinical Psychiatry, 81(5), 19r13106. https://doi.org/10.4088/jcp.19r13106

Molyneaux, E., Telesia, L. A., Henshaw, C., Boath, E., Bradley, E., & Howard, L. M. (2018). Antidepressants for preventing postnatal depression. Cochrane Database of Systematic Reviews, (4), CD004363. https://doi.org/10.1002/14651858.cd004363.pub3

Moses-Kolko, E. L., Perlman, S. B., Wisner, K. L., James, J., Saul, A. T., & Phillips, M. L. (2010). Abnormally reduced dorsomedial prefrontal cortical activity and effective connectivity with amygdala in response to negative emotional faces in postpartum depression. American Journal of Psychiatry, 167(11), 1373-1380. https://doi.org/10.1176/appi.ajp.2010.09081235

Mott, S. L., Schiller, C. E., Richards, J. G., O'Hara, M. W., & Stuart, S. (2011). Depression and anxiety among postpartum and adoptive mothers. Archives of Women's Mental Health, 14(4), 335-343. https://doi.org/10.1007/s00737-011-0227-1

Munk-Olsen, T., Laursen, T. M., Meltzer-Brody, S., Mortensen, P. B., & Jones, I. (2012). Psychiatric disorders with postpartum onset: Possible early manifestations of bipolar affective disorders. Archives of General Psychiatry, 69(4), 428-434. https://doi.org/10.1001/archgenpsychiatry.2011.157

Munk-Olsen, T., Laursen, T. M., Pedersen, C. B., Mors, O., & Mortensen, P. B. (2006). New parents and mental disorders: a population-based register study. JAMA, 296(21), 2582-2589. https://doi.org/10.1001/jama.296.21.2582

Murray, L. (1992). The impact of postnatal depression on infant development. Journal of Child Psychology and Psychiatry, 33(3), 543-561. https://doi.org/10.1111/j.1469-7610.1992.tb00890.x

Murray, L., Arteche, A., Fearon, P., Halligan, S., Goodyer, I., & Cooper, P. (2011). Maternal postnatal depression and the development of depression in offspring up to 16 years of age. Journal of the American Academy of Child & Adolescent Psychiatry, 50(5), 460-470. https://doi.org/10.1016/j.jaac.2011.02.001

Murray, L., Cooper, P. J., Wilson, A., & Romaniuk, H. (2003). Controlled trial of the short- and long-term effect of psychological treatment of post-partum depression: 2. Impact on the mother-child relationship and child outcome. British Journal of Psychiatry, 182(5), 420-427. https://doi.org/10.1192/bjp.182.5.420

Murray, L., Fiori-Cowley, A., Hooper, R., & Cooper, P. (1996). The impact of postnatal depression and associated adversity on early mother-infant interactions and later infant outcome. Child Development, 67(5), 2512-2526. https://doi.org/10.1111/j.1467-8624.1996.tb01871.x

Murray, L., Kempton, C., Woolgar, M., & Hooper, R. (1993). Depressed mothers' speech to their infants and its relation to infant gender and cognitive development. Journal of Child Psychology and Psychiatry, 34(7), 1083-1101. https://doi.org/10.1111/j.1469-7610.1993.tb01775.x

Myers, S., Burger, O., & Johns, S. E. (2016). Postnatal depression and reproductive success in modern, low-fertility contexts. Evolution, Medicine, and Public Health, 2016(1), 71-84. https://doi.org/10.1093/emph/eow003

Nesse, R. M. (2000). Is depression an adaptation? Archives of General Psychiatry, 57(1), 14-20. https://doi.org/10.1001/archpsyc.57.1.14

Netsi, E., Pearson, R. M., Murray, L., Cooper, P., Craske, M. G., & Stein, A. (2018). Association of persistent and severe postnatal depression with child outcomes. JAMA Psychiatry, 75(3), 247-253. https://doi.org/10.1001/jamapsychiatry.2017.4363

Nolen-Hoeksema, S. (1991). Responses to depression and their effects on the duration of depressive episodes. Journal of Abnormal Psychology, 100(4), 569-582. https://doi.org/10.1037/0021-843X.100.4.569

Nolen-Hoeksema, S., Wisco, B. E., & Lyubomirsky, S. (2008). Rethinking rumination. Perspectives on Psychological Science, 3(5), 400-424. https://doi.org/10.1111/j.1745-6924.2008.00088.x

Nylen, K. J., Williamson, J. A., O'Hara, M. W., Watson, D., & Engeldinger, J. (2013). Validity of somatic symptoms as indicators of depression in pregnancy. Archives of Women's Mental Health, 16(3), 203-210. https://doi.org/10.1007/s00737-013-0334-2

Oates, M. (2003). Suicide: the leading cause of maternal death. British Journal of Psychiatry, 183(4), 279-281. https://doi.org/10.1192/bjp.183.4.279

Oates, M. R., Cox, J. L., Neema, S., Asten, P., Glangeaud-Freudenthal, N., Figueiredo, B., Gorman, L. L., Hacking, S., Hirst, E., Kammerer, M. H., Klier, C. M., Seneviratne, G., Smith, M., Sutter-Dallay, A.-L., Valoriani, V., Wickberg, B., Yoshida, K., & TCS-PND Group. (2004). Postnatal depression across countries and cultures: a qualitative study. British Journal of Psychiatry, 184(S46), s10-s16. https://doi.org/10.1192/bjp.184.46.s10

O'Connor, E., Rossom, R. C., Henninger, M., Groom, H. C., & Burda, B. U. (2016). Primary care screening for and treatment of depression in pregnant and postpartum women: Evidence report and systematic review for the US Preventive Services Task Force. JAMA, 315(4), 388-406. https://doi.org/10.1001/jama.2015.18948

O'Connor, E., Senger, C. A., Henninger, M. L., Coppola, E., & Gaynes, B. N. (2019). Interventions to prevent perinatal depression: Evidence report and systematic review for the US Preventive Services Task Force. JAMA, 321(6), 588-601. https://doi.org/10.1001/jama.2018.20865

O'Hara, M. W., & McCabe, J. E. (2013). Postpartum depression: Current status and future directions. Annual Review of Clinical Psychology, 9, 379-407. https://doi.org/10.1146/annurev-clinpsy-050212-185612

O'Hara, M. W., & Swain, A. M. (1996). Rates and risk of postpartum depression—a meta-analysis. International Review of Psychiatry, 8(1), 37-54. https://doi.org/10.3109/09540269609037816

O'Hara, M. W., & Wisner, K. L. (2014). Perinatal mental illness: definition, description and aetiology. Best Practice & Research Clinical Obstetrics & Gynaecology, 28(1), 3-12. https://doi.org/10.1016/j.bpobgyn.2013.09.002

O'Hara, M. W., Rehm, L. P., & Campbell, S. B. (1982). Predicting depressive symptomatology: Cognitive-behavioral models and postpartum depression. Journal of Abnormal Psychology, 91(6), 457-461. https://doi.org/10.1037/0021-843x.91.6.457

O'Hara, M. W., Schlechte, J. A., Lewis, D. A., & Wright, E. J. (1991). Prospective study of postpartum blues: Biologic and psychosocial factors. Archives of General Psychiatry, 48(9), 801-806. https://doi.org/10.1001/archpsyc.1991.01810330025004

O'Hara, M. W., Stuart, S., Gorman, L. L., & Wenzel, A. (2000). Efficacy of interpersonal psychotherapy for postpartum depression. Archives of General Psychiatry, 57(11), 1039-1045. https://doi.org/10.1001/archpsyc.57.11.1039

O'Hara, M. W., Zekoski, E. M., Philipps, L. H., & Wright, E. J. (1990). Controlled prospective study of postpartum mood disorders: Comparison of childbearing and nonchildbearing women. Journal of Abnormal Psychology, 99(1), 3-15. https://doi.org/10.1037/0021-843x.99.1.3

Okun, M. L. (2015). Sleep and postpartum depression. Current Opinion in Psychiatry, 28(6), 490-496. https://doi.org/10.1097/yco.0000000000000206

Olino, T. M., Michelini, G., Mennies, R. J., Kotov, R., & Klein, D. N. (2021). Does maternal psychopathology bias reports of offspring symptoms? A study using moderated non-linear factor analysis. Journal of Child Psychology and Psychiatry, 62(10), 1195-1201. https://doi.org/10.1111/jcpp.13394

Orsolini, L., & Bellantuono, C. (2015). Serotonin reuptake inhibitors and breastfeeding: a systematic review. Human Psychopharmacology: Clinical and Experimental, 30(1), 4-20. https://doi.org/10.1002/hup.2451

Osborne, L., Clive, M., Kimmel, M., Gispen, F., Guintivano, J., Brown, T., Cox, O., Judy, J., Meilman, S., Braier, A., Beckmann, M. W., Kornhuber, J., Fasching, P. A., Goes, F., Payne, J. L., Binder, E. B., & Kaminsky, Z. (2016). Replication of epigenetic postpartum depression biomarkers and variation with hormone levels. Neuropsychopharmacology, 41(6), 1648-1658. https://doi.org/10.1038/npp.2015.333

Osborne, L. M., Gispen, F., Sanyal, A., Yenokyan, G., Meilman, S., & Payne, J. L. (2017). Lower allopregnanolone during pregnancy predicts postpartum depression: An exploratory study. Psychoneuroendocrinology, 79, 116-121. https://doi.org/10.1016/j.psyneuen.2017.02.012

Park, E. M., Meltzer-Brody, S., & Stickgold, R. (2013). Poor sleep maintenance and subjective sleep quality are associated with postpartum maternal depression symptom severity. Archives of Women's Mental Health, 16(6), 539-547. https://doi.org/10.1007/s00737-013-0356-9

Paulson, J. F., & Bazemore, S. D. (2010). Prenatal and postpartum depression in fathers and its association with maternal depression: A meta-analysis. JAMA, 303(19), 1961-1969. https://doi.org/10.1001/jama.2010.605

Pawluski, J. L., Lonstein, J. S., & Fleming, A. S. (2017). The neurobiology of postpartum anxiety and depression. Trends in Neurosciences, 40(2), 106-120. https://doi.org/10.1016/j.tins.2016.11.009

Payne, J. L., & Maguire, J. (2019). Pathophysiological mechanisms implicated in postpartum depression. Frontiers in Neuroendocrinology, 52, 165-180. https://doi.org/10.1016/j.yfrne.2018.12.001

Pearson, R. M., Cooper, R. M., Penton-Voak, I. S., Lightman, S. L., & Evans, J. (2010). Depressive symptoms in early pregnancy disrupt attentional processing of infant emotion. Psychological Medicine, 40(4), 621-631. https://doi.org/10.1017/S0033291709990961

Pearson, R. M., Evans, J., Kounali, D., Lewis, G., Heron, J., Ramchandani, P. G., O'Connor, T. G., & Stein, A. (2013). Maternal depression during pregnancy and the postnatal period: Risks and possible mechanisms for offspring depression at age 18 years. JAMA Psychiatry, 70(12), 1312-1319. https://doi.org/10.1001/jamapsychiatry.2013.2163

Pearson, R. M., Lightman, S. L., & Evans, J. (2011). Attentional processing of infant emotion during late pregnancy and mother-infant relations after birth. Archives of Women's Mental Health, 14(1), 23-31. https://doi.org/10.1007/s00737-010-0180-4

Peng, L., Fu, C., Xiong, F., Zhang, Q., Liang, Z., Chen, L., He, C., & Wei, Q. (2020). Effects of repetitive transcranial magnetic stimulation on depression symptoms and cognitive function in treating patients with postpartum depression: A systematic review and meta-analysis of randomized controlled trials. Psychiatry Research, 290, 113124. https://doi.org/10.1016/j.psychres.2020.113124

Perry, A., Gordon-Smith, K., Jones, L., & Jones, I. (2021). Phenomenology, epidemiology and aetiology of postpartum psychosis: A review. Brain Sciences, 11(1), 47. https://doi.org/10.3390/brainsci11010047

Pio de Almeida, L. S., Jansen, K., Köhler, C. A., Pinheiro, R. T., da Silva, R. A., & Bonini, J. S. (2012). Working and short-term memories are impaired in postpartum depression. Journal of Affective Disorders, 136(3), 1238-1242. https://doi.org/10.1016/j.jad.2011.09.031

Pitt, B. (1968). "Atypical" depression following childbirth. British Journal of Psychiatry, 114(516), 1325-1335. https://doi.org/10.1192/bjp.114.516.1325

Pope, C. J., & Mazmanian, D. (2016). Breastfeeding and postpartum depression: An overview and methodological recommendations for future research. Depression Research and Treatment, 2016, 4765310. https://doi.org/10.1155/2016/4765310

Postpartum Depression: Action Towards Causes and Treatment (PACT) Consortium. (2015). Heterogeneity of postpartum depression: a latent class analysis. The Lancet Psychiatry, 2(1), 59-67. https://doi.org/10.1016/S2215-0366(14)00055-8

Pritchett, R. V., Daley, A. J., & Jolly, K. (2017). Does aerobic exercise reduce postpartum depressive symptoms? A systematic review and meta-analysis. British Journal of General Practice, 67(663), e684-e691. https://doi.org/10.3399/bjgp17x692525

Putnam, K. T., Wilcox, M., Robertson-Blackmore, E., et al. (2017). Clinical phenotypes of perinatal depression and time of symptom onset: analysis of data from an international consortium. The Lancet Psychiatry, 4(6), 477-485. https://doi.org/10.1016/S2215-0366(17)30136-0

Putnick, D. L., Sundaram, R., Bell, E. M., Ghassabian, A., Goldstein, R. B., Robinson, S. L., Vafai, Y., Gilman, S. E., & Yeung, E. (2020). Trajectories of maternal postpartum depressive symptoms. Pediatrics, 146(5), e20200857. https://doi.org/10.1542/peds.2020-0857

Radoš, S. N., Akik, B. K., Žutić, M., Rodriguez-Muñoz, M. F., Uriko, K., Motrico, E., Moreno-Peral, P., Apter, G., & Lambregtse-van den Berg, M. (2024). Diagnosis of peripartum depression disorder: A state-of-the-art approach from the COST Action Riseup-PPD. Comprehensive Psychiatry, 130, 152456. https://doi.org/10.1016/j.comppsych.2024.152456

Rahman, A., Malik, A., Sikander, S., Roberts, C., & Creed, F. (2008). Cognitive behaviour therapy-based intervention by community health workers for mothers with depression and their infants in rural Pakistan: a cluster-randomised controlled trial. The Lancet, 372(9642), 902-909. https://doi.org/10.1016/S0140-6736(08)61400-2

Ramchandani, P., Stein, A., Evans, J., O'Connor, T. G., & ALSPAC study team. (2005). Paternal depression in the postnatal period and child development: a prospective population study. The Lancet, 365(9478), 2201-2205. https://doi.org/10.1016/S0140-6736(05)66778-5

Rao, W.-W., Zhu, X.-M., Zong, Q.-Q., Zhang, Q., Hall, B. J., Ungvari, G. S., & Xiang, Y.-T. (2020). Prevalence of prenatal and postpartum depression in fathers: A comprehensive meta-analysis of observational surveys. Journal of Affective Disorders, 263, 491-499. https://doi.org/10.1016/j.jad.2019.10.030

Rasmussen, M.-L. H., Strøm, M., Wohlfahrt, J., Videbech, P., & Melbye, M. (2017). Risk, treatment duration, and recurrence risk of postpartum affective disorder in women with no prior psychiatric history: A population-based cohort study. PLOS Medicine, 14(9), e1002392. https://doi.org/10.1371/journal.pmed.1002392

Rayce, S. B., Rasmussen, I. S., Væver, M. S., & Pontoppidan, M. (2020). Effects of parenting interventions for mothers with depressive symptoms and an infant: Systematic review and meta-analysis. BJPsych Open, 6(1), e9. https://doi.org/10.1192/bjo.2019.89

Reck, C., Stehle, E., Reinig, K., & Mundt, C. (2009). Maternity blues as a predictor of DSM-IV depression and anxiety disorders in the first three months postpartum. Journal of Affective Disorders, 113(1-2), 77-87. https://doi.org/10.1016/j.jad.2008.05.003

Reed, G. M., First, M. B., Kogan, C. S., et al. (2019). Innovations and changes in the ICD-11 classification of mental, behavioural and neurodevelopmental disorders. World Psychiatry, 18(1), 3-19. https://doi.org/10.1002/wps.20611

Rezaie-Keikhaie, K., Arbabshastan, M. E., Rafiemanesh, H., Amirshahi, M., Ostadkelayeh, S. M., & Arbabisarjou, A. (2020). Systematic review and meta-analysis of the prevalence of the maternity blues in the postpartum period. Journal of Obstetric, Gynecologic & Neonatal Nursing, 49(2), 127-136. https://doi.org/10.1016/j.jogn.2020.01.001

Richters, J. E. (1992). Depressed mothers as informants about their children: A critical review of the evidence for distortion. Psychological Bulletin, 112(3), 485-499. https://doi.org/10.1037/0033-2909.112.3.485

Robertson, E., Grace, S., Wallington, T., & Stewart, D. E. (2004). Antenatal risk factors for postpartum depression: a synthesis of recent literature. General Hospital Psychiatry, 26(4), 289-295. https://doi.org/10.1016/j.genhosppsych.2004.02.006

Rodebaugh, T. L., Scullin, R. B., Langer, J. K., Dixon, D. J., Huppert, J. D., Bernstein, A., Zvielli, A., & Lenze, E. J. (2016). Unreliability as a threat to understanding psychopathology: The cautionary tale of attentional bias. Journal of Abnormal Psychology, 125(6), 840–851. https://doi.org/10.1037/abn0000184

Rogers, A., Obst, S., Teague, S. J., Rossen, L., Spry, E. A., Macdonald, J. A., Sunderland, M., Olsson, C. A., Youssef, G., & Hutchinson, D. (2020). Association between maternal perinatal depression and anxiety and child and adolescent development: A meta-analysis. JAMA Pediatrics, 174(11), 1082-1092. https://doi.org/10.1001/jamapediatrics.2020.2910

Ross, L. E., Grigoriadis, S., Mamisashvili, L., Vonderporten, E. H., Roerecke, M., Rehm, J., Dennis, C.-L., Koren, G., Steiner, M., Mousmanis, P., & Cheung, A. (2013). Selected pregnancy and delivery outcomes after exposure to antidepressant medication: A systematic review and meta-analysis. JAMA Psychiatry, 70(4), 436-443. https://doi.org/10.1001/jamapsychiatry.2013.684

Ross, L. E., Steele, L., Goldfinger, C., & Strike, C. (2007). Perinatal depressive symptomatology among lesbian and bisexual women. Archives of Women's Mental Health, 10(2), 53-59. https://doi.org/10.1007/s00737-007-0168-x

Rundgren, S., Brus, O., Båve, U., Landén, M., Lundberg, J., Nordanskog, P., & Nordenskjöld, A. (2018). Improvement of postpartum depression and psychosis after electroconvulsive therapy: A population-based study with a matched comparison group. Journal of Affective Disorders, 235, 258-264. https://doi.org/10.1016/j.jad.2018.04.043

Russell, E. J., Fawcett, J. M., & Mazmanian, D. (2013). Risk of obsessive-compulsive disorder in pregnant and postpartum women: a meta-analysis. Journal of Clinical Psychiatry, 74(4), 377-385. https://doi.org/10.4088/JCP.12r07917

Sacher, J., Wilson, A. A., Houle, S., Rusjan, P., Hassan, S., Bloomfield, P. M., Stewart, D. E., & Meyer, J. H. (2010). Elevated brain monoamine oxidase A binding in the early postpartum period. Archives of General Psychiatry, 67(5), 468-474. https://doi.org/10.1001/archgenpsychiatry.2010.32

Schiller, C. E., Meltzer-Brody, S., & Rubinow, D. R. (2015). The role of reproductive hormones in postpartum depression. CNS Spectrums, 20(1), 48-59. https://doi.org/10.1017/S1092852914000480

Scott, L. J. (2019). Brexanolone: First global approval. Drugs, 79(7), 779-783. https://doi.org/10.1007/s40265-019-01121-0

Semkovska, M., & McLoughlin, D. M. (2010). Objective cognitive performance associated with electroconvulsive therapy for depression: A systematic review and meta-analysis. Biological Psychiatry, 68(6), 568-577. https://doi.org/10.1016/j.biopsych.2010.06.009

Serati, M., Redaelli, M., Buoli, M., & Altamura, A. C. (2016). Perinatal major depression biomarkers: A systematic review. Journal of Affective Disorders, 193, 391-404. https://doi.org/10.1016/j.jad.2016.01.027

Seth, S., Lewis, A. J., & Galbally, M. (2016). Perinatal maternal depression and cortisol function in pregnancy and the postpartum period: a systematic literature review. BMC Pregnancy and Childbirth, 16(1), 124. https://doi.org/10.1186/s12884-016-0915-y

Sharma, V., & Mazmanian, D. (2003). Sleep loss and postpartum psychosis. Bipolar Disorders, 5(2), 98-105. https://doi.org/10.1034/j.1399-5618.2003.00015.x

Sharma, V., & Mazmanian, D. (2014). The DSM-5 peripartum specifier: prospects and pitfalls. Archives of Women's Mental Health, 17(2), 171-173. https://doi.org/10.1007/s00737-013-0406-3

Sharma, V., Doobay, M., & Baczynski, C. (2017). Bipolar postpartum depression: An update and recommendations. Journal of Affective Disorders, 219, 105-111. https://doi.org/10.1016/j.jad.2017.05.014

Shorey, S., Chee, C. Y. I., Ng, E. D., Chan, Y. H., Tam, W. W. S., & Chong, Y. S. (2018). Prevalence and incidence of postpartum depression among healthy mothers: A systematic review and meta-analysis. Journal of Psychiatric Research, 104, 235-248. https://doi.org/10.1016/j.jpsychires.2018.08.001

Singla, D. R., Silver, R. K., Vigod, S. N., et al. (2025). Task-sharing and telemedicine delivery of psychotherapy to treat perinatal depression: a pragmatic, noninferiority randomized trial. Nature Medicine, 31(4), 1214-1224. https://doi.org/10.1038/s41591-024-03482-w

Siu, A. L., & US Preventive Services Task Force. (2016). Screening for depression in adults: US Preventive Services Task Force recommendation statement. JAMA, 315(4), 380-387. https://doi.org/10.1001/jama.2015.18392

Slomian, J., Honvo, G., Emonts, P., Reginster, J.-Y., & Bruyère, O. (2019). Consequences of maternal postpartum depression: A systematic review of maternal and infant outcomes. Women's Health, 15, 1745506519844044. https://doi.org/10.1177/1745506519844044

Sobral, M., Ferreira, A. M., Tomás, A., Pozo-Neira, J., Coimbra, M., Fonseca, A., Osório, A., & Ganho-Ávila, A. (2026). The effects of perinatal depression and anxiety on memory, attention, and cognitive flexibility: A systematic review and meta-analysis. Archives of Women's Mental Health, 29(6), Article 142. https://doi.org/10.1007/s00737-026-01778-z

Sockol, L. E. (2015). A systematic review of the efficacy of cognitive behavioral therapy for treating and preventing perinatal depression. Journal of Affective Disorders, 177, 7-21. https://doi.org/10.1016/j.jad.2015.01.052

Sockol, L. E. (2018). A systematic review and meta-analysis of interpersonal psychotherapy for perinatal women. Journal of Affective Disorders, 232, 316-328. https://doi.org/10.1016/j.jad.2018.01.018

Sockol, L. E., Epperson, C. N., & Barber, J. P. (2013). Preventing postpartum depression: A meta-analytic review. Clinical Psychology Review, 33(8), 1205-1217. https://doi.org/10.1016/j.cpr.2013.10.004

Sockol, L. E., Epperson, C. N., & Barber, J. P. (2014). The relationship between maternal attitudes and symptoms of depression and anxiety among pregnant and postpartum first-time mothers. Archives of Women's Mental Health, 17(3), 199-212. https://doi.org/10.1007/s00737-014-0424-9

Stein, A., Arteche, A., Lehtonen, A., Craske, M., Harvey, A., Counsell, N., & Murray, L. (2010). Interpretation of infant facial expression in the context of maternal postnatal depression. Infant Behavior and Development, 33(3), 273-278. https://doi.org/10.1016/j.infbeh.2010.03.002

Stein, A., Craske, M. G., Lehtonen, A., Harvey, A., Savage-McGlynn, E., Davies, B., Goodwin, J., Murray, L., Cortina-Borja, M., & Counsell, N. (2012). Maternal cognitions and mother-infant interaction in postnatal depression and generalized anxiety disorder. Journal of Abnormal Psychology, 121(4), 795-809. https://doi.org/10.1037/a0026847

Stein, A., Netsi, E., Lawrence, P. J., Granger, C., Kempton, C., Craske, M. G., Nickless, A., Mollison, J., Stewart, D. A., Rapa, E., West, V., Scerif, G., Cooper, P. J., & Murray, L. (2018). Mitigating the effect of persistent postnatal depression on child outcomes through an intervention to treat depression and improve parenting: a randomised controlled trial. The Lancet Psychiatry, 5(2), 134-144. https://doi.org/10.1016/S2215-0366(18)30006-3

Stein, A., Pearson, R. M., Goodman, S. H., Rapa, E., Rahman, A., McCallum, M., Howard, L. M., & Pariante, C. M. (2014). Effects of perinatal mental disorders on the fetus and child. The Lancet, 384(9956), 1800-1819. https://doi.org/10.1016/S0140-6736(14)61277-0

Stewart, D. E., & Vigod, S. (2016). Postpartum depression. New England Journal of Medicine, 375(22), 2177-2186. https://doi.org/10.1056/NEJMcp1607649

Suarez, E. A., Bateman, B. T., Hernández-Díaz, S., Straub, L., Wisner, K. L., Gray, K. J., Pennell, P. B., Lester, B., McDougle, C. J., Zhu, Y., Mogun, H., & Huybrechts, K. F. (2022). Association of antidepressant use during pregnancy with risk of neurodevelopmental disorders in children. JAMA Internal Medicine, 182(11), 1149. https://doi.org/10.1001/jamainternmed.2022.4268

Sweeney, S., & MacBeth, A. (2016). The effects of paternal depression on child and adolescent outcomes: A systematic review. Journal of Affective Disorders, 205, 44-59. https://doi.org/10.1016/j.jad.2016.05.073

Thombs, B. D., Arthurs, E., Coronado-Montoya, S., Roseman, M., Delisle, V. C., Leavens, A., Levis, B., Azoulay, L., Smith, C., Ciofani, L., Coyne, J. C., Feeley, N., Gilbody, S., Schinazi, J., Stewart, D. E., & Zelkowitz, P. (2014). Depression screening and patient outcomes in pregnancy or postpartum: A systematic review. Journal of Psychosomatic Research, 76(6), 433-446. https://doi.org/10.1016/j.jpsychores.2014.01.006

Thul, T. A., Corwin, E. J., Carlson, N. S., Brennan, P. A., & Young, L. J. (2020). Oxytocin and postpartum depression: A systematic review. Psychoneuroendocrinology, 120, 104793. https://doi.org/10.1016/j.psyneuen.2020.104793

Trede, K., Baldessarini, R. J., Viguera, A. C., & Bottéro, A. (2009). Treatise on insanity in pregnant, postpartum, and lactating women (1858) by Louis-Victor Marcé: A commentary. Harvard Review of Psychiatry, 17(2), 157-165. https://doi.org/10.1080/10673220902891802

Tronick, E., & Reck, C. (2009). Infants of depressed mothers. Harvard Review of Psychiatry, 17(2), 147-156. https://doi.org/10.1080/10673220902899714

Tsivos, Z.-L., Calam, R., Sanders, M. R., & Wittkowski, A. (2015). Interventions for postnatal depression assessing the mother-infant relationship and child developmental outcomes: A systematic review. International Journal of Women's Health, 7, 429-447. https://doi.org/10.2147/IJWH.S75311

Turner, E. H., Matthews, A. M., Linardatos, E., Tell, R. A., & Rosenthal, R. (2008). Selective publication of antidepressant trials and its influence on apparent efficacy. New England Journal of Medicine, 358(3), 252-260. https://doi.org/10.1056/NEJMsa065779

US Preventive Services Task Force; Barry, M. J., Nicholson, W. K., Silverstein, M., Chelmow, D., Coker, T. R., Davidson, K. W., Davis, E. M., Donahue, K. E., Jaén, C. R., Li, L., Ogedegbe, G., Pbert, L., Rao, G., Ruiz, J. M., Stevermer, J. J., Tsevat, J., Underwood, S. M., & Wong, J. B. (2023). Screening for depression and suicide risk in adults: US Preventive Services Task Force recommendation statement. JAMA, 329(23), 2057-2067. https://doi.org/10.1001/jama.2023.9297

US Preventive Services Task Force; Curry, S. J., Krist, A. H., Owens, D. K., Barry, M. J., Caughey, A. B., Davidson, K. W., Doubeni, C. A., Epling, J. W., Jr., Grossman, D. C., Kemper, A. R., Kubik, M., Landefeld, C. S., Mangione, C. M., Silverstein, M., Simon, M. A., Tseng, C.-W., & Wong, J. B. (2019). Interventions to prevent perinatal depression: US Preventive Services Task Force recommendation statement. JAMA, 321(6), 580-587. https://doi.org/10.1001/jama.2019.0007

VanderKruik, R., Barreix, M., Chou, D., Allen, T., Say, L., & Cohen, L. S., on behalf of the Maternal Morbidity Working Group. (2017). The global prevalence of postpartum psychosis: a systematic review. BMC Psychiatry, 17(1), 272. https://doi.org/10.1186/s12888-017-1427-7

Vesga-López, O., Blanco, C., Keyes, K., Olfson, M., Grant, B. F., & Hasin, D. S. (2008). Psychiatric disorders in pregnant and postpartum women in the United States. Archives of General Psychiatry, 65(7), 805-815. https://doi.org/10.1001/archpsyc.65.7.805

Vigod, S. N., Villegas, L., Dennis, C.-L., & Ross, L. E. (2010). Prevalence and risk factors for postpartum depression among women with preterm and low-birth-weight infants: a systematic review. BJOG: An International Journal of Obstetrics & Gynaecology, 117(5), 540-550. https://doi.org/10.1111/j.1471-0528.2009.02493.x

Viktorin, A., Meltzer-Brody, S., Kuja-Halkola, R., Sullivan, P. F., Landén, M., Lichtenstein, P., & Magnusson, P. K. (2016). Heritability of perinatal depression and genetic overlap with nonperinatal depression. American Journal of Psychiatry, 173(2), 158-165. https://doi.org/10.1176/appi.ajp.2015.15010085

Vliegen, N., Casalin, S., & Luyten, P. (2014). The course of postpartum depression: A review of longitudinal studies. Harvard Review of Psychiatry, 22(1), 1-22. https://doi.org/10.1097/HRP.0000000000000013

Walton, N., & Maguire, J. (2019). Allopregnanolone-based treatments for postpartum depression: Why/how do they work? Neurobiology of Stress, 11, 100198. https://doi.org/10.1016/j.ynstr.2019.100198

Wang, L., Kroenke, K., Stump, T. E., & Monahan, P. O. (2021). Screening for perinatal depression with the Patient Health Questionnaire depression scale (PHQ-9): A systematic review and meta-analysis. General Hospital Psychiatry, 68, 74-82. https://doi.org/10.1016/j.genhosppsych.2020.12.007

Wang, S., Deng, C.-M., Zeng, Y., et al. (2024). Efficacy of a single low dose of esketamine after childbirth for mothers with symptoms of prenatal depression: randomised clinical trial. BMJ, 385, e078218. https://doi.org/10.1136/bmj-2023-078218

Wang, Z., Liu, J., Shuai, H., Cai, Z., Fu, X., Liu, Y., Xiao, X., Zhang, W., Krabbendam, E., Liu, S., Liu, Z., Li, Z., & Yang, B. X. (2021). Mapping global prevalence of depression among postpartum women. Translational Psychiatry, 11(1), 543. https://doi.org/10.1038/s41398-021-01663-6

Warfa, N., Harper, M., Nicolais, G., & Bhui, K. (2014). Adult attachment style as a risk factor for maternal postnatal depression: a systematic review. BMC Psychology, 2, 56. https://doi.org/10.1186/s40359-014-0056-x

Webb, R., & Ayers, S. (2015). Cognitive biases in processing infant emotion by women with depression, anxiety and post-traumatic stress disorder in pregnancy or after birth: A systematic review. Cognition and Emotion, 29(7), 1278-1294. https://doi.org/10.1080/02699931.2014.977849

Weissman, M. M., Pilowsky, D. J., Wickramaratne, P. J., Talati, A., Wisniewski, S. R., Fava, M., Hughes, C. W., Garber, J., Malloy, E., King, C. A., et al.; STARD-Child Team (2006). Remissions in maternal depression and child psychopathology: A STARD-Child report. JAMA, 295(12), 1389-1398. https://doi.org/10.1001/jama.295.12.1389

Wesseloo, R., Kamperman, A. M., Munk-Olsen, T., Pop, V. J. M., Kushner, S. A., & Bergink, V. (2016). Risk of postpartum relapse in bipolar disorder and postpartum psychosis: A systematic review and meta-analysis. American Journal of Psychiatry, 173(2), 117-127. https://doi.org/10.1176/appi.ajp.2015.15010124

Wilson, C. A., Robertson, L., Ayre, K., Hendon, J. L., Dawson, S., Bridges, C., & Khalifeh, H. (2025). Brexanolone, zuranolone and related neurosteroid GABAA receptor positive allosteric modulators for postnatal depression. Cochrane Database of Systematic Reviews, 2025(6), CD014624. https://doi.org/10.1002/14651858.cd014624.pub2

Wilson, S., & Durbin, C. E. (2010). Effects of paternal depression on fathers' parenting behaviors: A meta-analytic review. Clinical Psychology Review, 30(2), 167-180. https://doi.org/10.1016/j.cpr.2009.10.007

Wirz-Justice, A., Bader, A., Frisch, U., Stieglitz, R.-D., Alder, J., Bitzer, J., Hösli, I., Jazbec, S., Benedetti, F., Terman, M., Wisner, K. L., & Riecher-Rössler, A. (2011). A randomized, double-blind, placebo-controlled study of light therapy for antepartum depression. The Journal of Clinical Psychiatry, 72(7), 986-993. https://doi.org/10.4088/jcp.10m06188blu

Wisner, K. L., Moses-Kolko, E. L., & Sit, D. K. Y. (2010). Postpartum depression: A disorder in search of a definition. Archives of Women's Mental Health, 13(1), 37-40. https://doi.org/10.1007/s00737-009-0119-9

Wisner, K. L., Sit, D. K. Y., McShea, M. C., Rizzo, D. M., Zoretich, R. A., Hughes, C. L., Eng, H. F., Luther, J. F., Wisniewski, S. R., Costantino, M. L., Confer, A. L., Moses-Kolko, E. L., Famy, C. S., & Hanusa, B. H. (2013). Onset timing, thoughts of self-harm, and diagnoses in postpartum women with screen-positive depression findings. JAMA Psychiatry, 70(5), 490-498. https://doi.org/10.1001/jamapsychiatry.2013.87

Woody, C. A., Ferrari, A. J., Siskind, D. J., Whiteford, H. A., & Harris, M. G. (2017). A systematic review and meta-regression of the prevalence and incidence of perinatal depression. Journal of Affective Disorders, 219, 86-92. https://doi.org/10.1016/j.jad.2017.05.003

World Health Organization. (2024). Clinical descriptions and diagnostic requirements for ICD-11 mental, behavioural and neurodevelopmental disorders (CDDR). World Health Organization.

Xiao, Y., Xie, P., Li, D., Zhao, T., & Nie, J. (2026). Comparative efficacy and acceptability of pharmacological interventions for postpartum depression: A systematic review and network meta-analysis. Psychiatry Research, 364, 117318. https://doi.org/10.1016/j.psychres.2026.117318

Yildiz, P. D., Ayers, S., & Phillips, L. (2017). The prevalence of posttraumatic stress disorder in pregnancy and after birth: A systematic review and meta-analysis. Journal of Affective Disorders, 208, 634-645. https://doi.org/10.1016/j.jad.2016.10.009

Yim, I. S., Glynn, L. M., Dunkel-Schetter, C., Hobel, C. J., Chicz-DeMet, A., & Sandman, C. A. (2009). Risk of postpartum depressive symptoms with elevated corticotropin-releasing hormone in human pregnancy. Archives of General Psychiatry, 66(2), 162-169. https://doi.org/10.1001/archgenpsychiatry.2008.533

Yim, I. S., Tanner Stapleton, L. R., Guardino, C. M., Hahn-Holbrook, J., & Dunkel Schetter, C. (2015). Biological and psychosocial predictors of postpartum depression: Systematic review and call for integration. Annual Review of Clinical Psychology, 11, 99-137. https://doi.org/10.1146/annurev-clinpsy-101414-020426

Yonkers, K. A., Gotman, N., Smith, M. V., Forray, A., Belanger, K., Brunetto, W. L., Lin, H., Burkman, R. T., Zelop, C. M., & Lockwood, C. J. (2011). Does antidepressant use attenuate the risk of a major depressive episode in pregnancy? Epidemiology, 22(6), 848-854. https://doi.org/10.1097/ede.0b013e3182306847

Zalsman, G., Hawton, K., Wasserman, D., van Heeringen, K., Arensman, E., Sarchiapone, M., Carli, V., Höschl, C., Barzilay, R., Balazs, J., Purebl, G., Kahn, J. P., Sáiz, P. A., Lipsicas, C. B., Bobes, J., Cozman, D., Hegerl, U., & Zohar, J. (2016). Suicide prevention strategies revisited: 10-year systematic review. The Lancet Psychiatry, 3(7), 646-659. https://doi.org/10.1016/S2215-0366(16)30030-X

Zhou, C., Hu, H., Wang, C., Zhu, Z., Feng, G., Xue, J., & Yang, Z. (2022). The effectiveness of mHealth interventions on postpartum depression: A systematic review and meta-analysis. Journal of Telemedicine and Telecare, 28(2), 83-95. https://doi.org/10.1177/1357633x20917816