Early-Onset and Delayed-Onset Postpartum Psychosis: A Case Series

Abstract

Postpartum psychosis (PPP) is a rare but high-risk psychiatric emergency, with an estimated incidence of 1–2 in 1000 births. This study focuses on describing the characteristics of episodes occurring within the first postpartum year, specifically examining clinical, etiopathogenic, and prognostic differences between immediate- or early-onset PPP (≤15 days postpartum) and delayed-onset PPP (onset after several weeks). Data were collected from ten patients diagnosed with PPP during the first postpartum year, and a retrospective descriptive analysis was conducted. Five patients experienced immediate or early decompensation and five experienced delayed onset. None of the variables analyzed showed a significant association with the timing of decompensation (p > 0.05). The majority of deliveries were vaginal (n = 8), and most patients were primiparous (n = 9). The most frequent subsequent diagnosis was schizophrenia or a psychotic spectrum disorder (n = 6). The type of partner showed a non-significant trend (p = 0.15), which may warrant further investigation. Notably, the role of partner type deserves closer examination, as it may act as a protective factor against the development of mental disorders and could inform targeted support strategies within health care systems. The lack of descriptions of time to onset periods (staging) in PPP samples could be a gap in the literature.

1. Introduction

Postpartum psychosis (PPP) is a rare but high-risk psychiatric emergency, with an estimated incidence of 1–2 cases per 1000 births [1,2]. It is characterized by a severe disturbance of mental state that can lead to significant clinical consequences for the mother, the infant, and the family [3]. Moreover, it may disrupt the mother–infant bond, potentially increasing the risk of mental illness in the offspring. PPP typically manifests as a severe affective disorder, such as bipolar disorder or a brief psychotic episode triggered by childbirth. The primary symptoms include severe confusion, disorganized thinking, mood instability, anxiety or agitation, disrupted sleep patterns, and unpredictable behaviour [3,4,5].

Although its etiology remains unclear, evidence suggests a link between the specific physiological changes in the postpartum period—particularly inflammatory, immunological, and hormonal shifts—and the onset of illness in genetically vulnerable women. Despite its low prevalence, the first few weeks postpartum represent the period of highest risk in a woman’s life for developing a psychotic episode—up to 23 times greater than at other times [4]. Psychosis with immediate–early onset (<15 days postpartum) appears to be driven largely by biological mechanisms, whereas later-onset cases (weeks postpartum) may involve a more complex and multifactorial course [5,6].

Several risk factors for postpartum psychosis have been described, including a personal history of mental illness such as bipolar disorder or prior psychotic episodes, a family history of psychosis, and discontinuation or poor adherence to psychopharmacological treatment during pregnancy. Additional relevant risk and contributing factors include perinatal stress, obstetric complications, lack of social support, history of trauma or adversity, and socioeconomic factors such as social isolation or vulnerability. The interaction between biological vulnerability and environmental stressors appears to play a critical role in postpartum decompensation [5,6,7].

Despite clinical recognition of these risk factors, assessing the risk for postpartum psychosis remains challenging. Currently, no validated screening tools or standardized protocols exist to predict the onset of the disorder, complicating efforts to prevent or plan for early intervention. Several studies have attempted to develop predictive models based on clinical scales or risk algorithms; however, the low prevalence and sudden onset of PPP hinder their validation.

In this context, it is essential to explore not only classical clinical/obstetric factors but also psychosocial variables—such as the type of intimate partnership and educational attainment—or the timing of decompensation, in order to identify patterns that may contribute to improved early detection and risk stratification.

The objective of the present study is to analyze clinical, etiopathogenic, and prognostic differences between immediate/early- and delayed-onset postpartum psychosis in a sample of patients from our centre. These distinctions are essential for guiding diagnosis, treatment, and relapse prevention. Our case series shows these characteristics in a difficult-to-obtain low prevalence sample of PPP in our specific environment in Mediterranean Europe.

2. Results

The variables analyzed included age, previous diagnosis, follow-up/outcome diagnosis, type of birth, obstetric complications, educational level, and marital status. Fisher’s exact statistic or Mann–Whitney U revealed no statistically significant associations between these variables and the timing of decompensation (p > 0.05 in all cases) (

Table 1. Data summary table.

Variables	Total Sample	Immediate/Early-Onset PPP * N = 5	Delayed-Onset PPP N = 5	Signification U; p or F; p
Age (Mean, Range)	29.1; 23–37	27.2; 23–36	31.0; 24–37	0.310
Number of children	Nine first-time mothers and one mother with four children	Five first-time mothers	Four first-time mothers and one mother of four children	0.690
Previous Diagnosis	Bipolar disorder type I: 3 Esquizofrenia Espectrum diagnosis: 3 None: 4	Bipolar disorder type I: 2 Esquizofrenia Espectrum diagnosis: 2 None: 1	Bipolar disorder type I: 1 Esquizofrenia Espectrum diagnosis: 1 None: 3	0.435
Previous Treatment	Yes = 4 Prescribed, but no adherence: 3 None: 3	Case 1: Aripiprazole + venlafaxine (no adherence) Case 3: Aripiprazole (no adherence) Case 6: Haloperidol + Quetiapine Case 7: No. Case 10: Paliperidona + aripiprazole	Case 2: Risperidone + venlafaxine Case 4: Aripiprazole (no adherence) Case 5: Lorazepam if needed Case 8 and 9: No.	Not applied
Outcome Diagnosis	Bipolar disorder type I: 4 Esquizofrenia Espectrum diagnosis: 6 None: 0	Bipolar disorder type I: 3 Esquizofrenia Espectrum diagnosis: 2	Bipolar disorder type I: 1 Esquizofrenia Espectrum diagnosis: 4	0.519
Type of Birth	Cesarean: 2 Vaginal: 8	Cesarean: 1 Vaginal: 4	Cesarean: 1 Vaginal: 4	1.000
Pregnancy complications	Yes: 7 No: 3	Yes: 4 No: 2 Case 1: IUGR Type I, smoking and podalic presentation. Case 3: UUI. Case 6: No. Case 7: Delivery mixed induction for IUGR for cholestasis. Case 9: No. Case 10: Induction for 41 weeks of pregnancy and SGA.	Yes: 3 No: 1 Case 2: Delivery mixed induction for fetal hydronephrosis level 3. Case 4: Induction for oligohydramnios. Case 5: No. Case 8: Induction for 42 weeks of pregnancy.	1.000
Educational Level	Primary: 2 Secondary: 2 Professional training: 2 Superior education: 4	Primary: 1 Secondary: 1 Professional training: 1 Superior education: 2	Primary: 1 Secondary: 1 Professional training: 1 Superior education: 2	0.736
Marital Status	Single parent: 1 Married: 5 Non-married couple: 4	Single parent: 1 Married: 1 Non-married couple: 3	Single parent: 0 Married: 4 Non-married couple: 1	0.150

* PPP: Postpartum psychosis. U: Mann–Whitney U. F: Fisher exact-test. SGA: Small for gestational age. IUGR: Intrauterine growth restriction. UUI: Urgent urinary infection.

).

Most deliveries were vaginal (n = 8), and although some obstetric complications were noted, no cases of preeclampsia were reported. The majority of patients were primiparous (n = 9). The most frequent follow-up diagnosis was schizophrenia or a psychotic spectrum disorder (n = 6). Different psychiatric specialists attended the patients at every moment of their evolution. One of the patients with bipolar disorder type I had a previous PPP in their previous delivery. Two of the mothers were migrants. The most common follow-up diagnosis (but not significant) in the delayed-onset PPP group was a non-affective psychosis. The type of partnership showed a non-significant trend (p = 0.15) that may warrant further investigation.

3. Discussion

3.1. PPP Onset Definitions

Different previous research and case series have shown a lack of agreement on the definition of early-onset or delayed-onset PPP. In fact, the postpartum onset period is also considered differently between studies and series.

A classical Scottish cohort of 54,087 births from a population of 470,000 people over a 12-year period resulted in 120 psychiatric admissions within 90 days post-delivery. The relative risk of admission for psychotic illness was very significantly higher in the first 30 days after delivery, especially in primiparae [8].

In the Zurich sample of Schöpf and Rust [9], a group of 119 patients suffering from a severe psychiatric postpartum disorder were admitted for the first time in their life to a psychiatric hospital. In these samples, the onset of illness was considered if it appeared within 3 months following delivery.

The Swedish sample of Unnur Valdimarsdóttir et al. [10] included 892 women from 745,596 first-time mothers (1.2 per 1000 births) who were hospitalized for psychoses during the first 90 days. Their incidence rates peaked during the first month following childbirth, but up to 285 out of the total 892 psychosis cases (32%) were hospitalized within 7 days after childbirth and 523 (59%) within the first 14 days [10].

Other samples excluded patients if they had a chronic psychotic disorder, mania, or psychosis with onset during pregnancy, a history of previous psychosis, mania, or drug abuse outside of the postpartum period, or psychosis onset > 12 weeks postpartum [11]. Moreover, it seems that rates of psychiatric admission fell during pregnancy and increased in the early postpartum period, particularly during the first 2 weeks after birth [12].

3.2. PPP in Current Diagnosis Classifications

On the other hand, the DSM-5 does not recognize PPP as a separate diagnosis. However, DSM-5 recommends the use of the “peripartum onset specifier” to characterize mood episodes of bipolar I, bipolar II, or major depressive disorders occurring during pregnancy or the postpartum period [13]. A proposed operational definition of PPP refers to a manic, mixed, or major depressive episode with psychotic features, a psychotic disorder not otherwise specified, or a brief psychotic disorder within 4 weeks postpartum [14].

The required duration of psychotic symptoms to qualify for a diagnosis of postpartum psychosis is generally not specified or variable following the current classifications. According to the DSM-5 criteria, an episode of brief psychotic disorder with postpartum onset should last 1–30 days [13], and symptoms of mania should last at least 1 week, or any duration if hospitalization is required.

The new ICD-11 International Classification of Diseases for Mortality and Morbidity Statistics, Eleventh Revision (World Health Organization, 2025), includes a group of mental or behavioural disorders associated with pregnancy, childbirth, or the puerperium (6E20-6E2Z) [15]. It includes “syndromes associated with pregnancy or the puerperium (commencing within about 6 weeks after delivery) that involve significant mental and behavioural features, including delusions, hallucinations, or other psychotic symptoms”, and adds that “mood symptoms (depressive and/or manic) are also typically present”. Moreover, when the symptoms meet the diagnostic criteria for a specific mental disorder, that diagnosis should also be assigned. Moreover, the ICD-11 specifier “current episode perinatal” can apply to recurrent or single depressive disorder episodes, with or without psychotic symptoms. It is also possible to apply the specifier to the bipolar type I disorder, current episode manic, mixed or depressive, with or without psychotic symptoms and to the bipolar type II disorder, current episode depressive, with or without psychotic symptoms (https://icd.who.int/browse/2025-01/mms/en (accessed on 25 September 2025)).

As we cited, different previous authors proposed the inclusion of the time criteria “during the pregnancy or within 4 weeks postpartum” as in the DSM 5-TR [5,14], but others such as the ICS-11 included a wider amount of time (“commencing within about 6 weeks after delivery”) [15].

However, more detailed analysis detected symptoms usually beginning within the first 2 weeks after delivery, with approximately 65% of episodes occurring in the first three days postpartum [16]. Considering the polarity of episodes, episodes of mania or mixed states generally occur earlier than periods of depression [16].

In the Dutch sample of 51 women of Bergink et al. [17], the median onset of psychiatric symptoms occurred at 8 days after the delivery (interquartile range IQR from 5 to 14), and the median duration of the episode was of 40 days in their sample (IQR from 23 to 69) in their sample. Moreover, they detected that patients with prominent depressive symptoms in their sample had a significantly later onset of psychosis and a longer duration of the episode than patients without these depressive symptoms. On the other hand, they did not identify any other delivery-related, neonatal-related, or lactational risk factors [17].

In fact, due to the specific postpartum characteristics of PPP, different authors and experts have proposed the inclusion of PPP as a specific diagnosis to appear in future diagnostic classifications [18].

3.3. PPP Management

Management of PPP cases needs urgent assessment, and immediate treatment is recommended, preferably in an in-patient treatment setting. As most of the experts have pointed out, it is essential to assess risk for the patient, the newborn, other children or dependent people at home, and other family members. The best option could be admission to a mother–baby unit, if one is available. In the event that a specific specialized mother-baby unit or area is not available, the potential risk control need an admission to a non-adapted psychiatric ward, irrespective of the impact on child bonding [19]. The length of admission will be taken into account, considering the balance between risk and benefits. In several cases, a rapid recovery is shown [20]. It is recommended to stop, pause, or discontinue breastfeeding [21,22], but, in our opinion, in most cases we could adopt personalized, informed, consensual decisions with the mother, partner, and family.

Psychoeducation and continuous support for patients, partners, and family are recommended. The suicide and filicide risk is one of the most important safety questions in PPP cases. It is necessary to establish a safety plan and a detailed action plan. The action plan have to include the patient, the partner, and the family or support circle [21,22].

Therapeutic aspects are relevant in the management of PPP. As different authors have pointed out [21,22], PPP is a psychiatric emergency deserving immediate hospital admission and urgent psychopharmacological and psychotherapy treatments. The acute phase needs treatment with antipsychotics and/or lithium, with some cases requiring electro-convulsive therapy (ECT). On the other hand, between the different treatments studied, lithium provides the greatest evidence for the prevention of new future cases of PPP [21,22].

3.4. PPP Onset and Etiology Matters

The implications of the onset of symptoms for the etiology of PPP could be relevant, but scarce data is available [23,24,25]. As previous authors have pointed out, a rapid onset of postpartum psychosis following the delivery suggests that hormonal factors are potentially involved in the origin of symptoms, but there is a lack of robust evidence to support this theory [2]. Although in our sample, cases of late-onset PPP do not present a significantly higher number of subsequent diagnoses of non-affective psychosis, it would be interesting to include this data (time elapsed from delivery to onset) in future studies on the evolution of PPP.

3.5. PPP Risk Factors

Different described risk factors applied to our sample (see

Table 2. Risk factors for postpartum psychosis (modified from: [2,4,5,6,7]).

- Personal history of bipolar disorder, especially those with poor adherence to psychopharmacological treatment.

- Personal history of previous episodes of postpartum psychosis.

- Family history of bipolar disorder or psychotic disorder.

- Primiparity.

- Obstetric complications.

- Lack of social support.

- Lack of sleep/circadian rhythm disruption.

- Severe childhood maltreatment (+/−).

- Elevated biological stress markers.

- Hormonal and immunological factors.

- Genetic factors (?).

). One of the most repeated risk factors is a previous psychiatric diagnosis, detected in most of the cases of our sample. Also, one of the patients with bipolar disorder type I also had the antecedent of previous PPP in previous deliveries, data which was repeated in different series. The majority of patients in our sample were also primiparous (n = 9), which is a repeated PPP risk factor in different samples and studies [2,4,5,6,7].

Regarding the obstetric risk factors, most deliveries were vaginal in our sample. Moreover, some obstetric complications were noted in most of the cases (7 of 10), but no cases of preeclampsia were reported in our sample [2,4,5,6,7]. However, in relation to the psychosocial characteristics, the type of partnership did not show a significant association with the timing of decompensation in our sample (p = 0.15), although an interesting trend was observed: married women tended to present with delayed decompensation, whereas those who were not married (i.e., single mothers or in stable partnerships) were more likely to decompensate immediately. Although this result is not statistically significant—likely due to the small sample size, which limits statistical power and may mask meaningful relationships—this pattern could hold clinical relevance and serve as a basis for future hypotheses. This observation regarding marital status may be related to differences in perceived postpartum support or responsibility and should be further explored in studies with larger samples. In fact, the previous review of Alcantarilla et al. [4] did not detect conclusive results related to the influence of the relationship with husbands, mothers-in-law, or social support. In any case, association itself does not necessarily implicate causality.

Two of the mothers in our sample were migrants, and people who have migrated or refugee mothers suffer from different psychosocial difficulties, which are sometimes severe and could be another risk factor [26,27]. In fact, prevalence of psychotic disorders in low- and middle-income countries could be greater, due to lack of adequate nutrition, obstetric complications, poverty, war/refugee situations, or other highly stressful psychosocial situations, but there is scarce information. For example, in a sample of Bangladeshi mothers, psychotic disorder represented 4.5% of the sample [28]. Moreover, migrant women with schizophrenia have specific management and treatment needs [29].

3.6. PPP Predictive Models

As we cited in the introduction, predictive models based on clinical scales or risk algorithms are difficult to develop in the case of PPP [30]. On the other hand, it is an ongoing area of research in perinatal mental health. In the case of postpartum depression, different predictive models have been developed [31,32,33,34]. In every case, these promising models for postpartum depression could be hard to translate to PPP because of their low prevalence and heterogeneous characteristics.

In a study in Australia, the authors included administrative health data for all inpatient live births in the state of Queensland (2009–2014), including mothers with one or more indicators of mental health problems during pregnancy (n = 75,054 births). They used multiple machine learning methods to predict hospital admission in the 12 months following delivery. The primary diagnosis was recorded as an ICD-10 psychotic, bipolar, or depressive disorder [30]. Amongst the strongest 20 predictors of postpartum psychiatric admission as indicated by the boosted trees model, we find the “classical” PPP risk factors to be mental or nervous disorder impacting pregnancy (OR 1.96; p < 0.001), prenatal psychotic and bipolar disorders diagnosis (OR 5.54; p < 0.001), prenatal unipolar mood disorder admission (OR 8.84; p < 0.001), prenatal psychotic and bipolar disorders admission (OR 3.44; p < 0.001), or (any) maternal psychiatric admission prior to pregnancy (OR 3.28; p < 0.001). However, they detected other psychiatric risk factors, such as prenatal anxiety disorder diagnosis (OR 1.97; p < 0.001), a personal history of other mental or behavioural disorders (OR 2.46; p < 0.001) and prenatal substance use disorders admission (OR 5.04; p < 0.001) or prenatal smoking (negative elastic net − 0.041; OR 0.79; p = 0.064). They included therapeutic aspects, because crisis situation/event counselling (OR 8.35; p < 0.001) or cognitive and/or behavioural therapy (OR 2.25; p = 0.031) also represent risk factors [30].

Moreover, this Australian study also detected “classic” obstetric risk factors like labour and delivery complicated by intrapartum hemorrhage (OR 3.03; p = 0.010), but also other non-classical factors such as the association of neuraxial block during labour and delivery (OR 5.59; p = 0.002), and neuraxial block alone (OR 1.59; p = 0.020), or breastfeeding (OR 1.22; p = 0.003). Other interesting risk factors involve maternal age (OR 1.03; p = 0.003) and psychosocial–cultural aspects, such as indigenous status (OR 2.12, p < 0.001). Organizational variables included in the model are relevant, because to be enrolled as a private patient is a risk factor (OR 1.61; p = 0.002), but to be attended by a primary prenatal health care provider/general practitioner is a protective factor (negative elastic net—0.067; OR 0.70; p = 0.009). The authors highlighted the usefulness of these predictive methods to detect mothers at risk in order to prevent future admission or facilitate more timely admission [30].

Other models also included these aspects. The team of Dr. Simone N. Vigod and collaborators analyzed a population-based cohort study of 1433 mothers with schizophrenia in Ontario, Canada (2003–2011) [35]. Admitted women (n = 275, 19% of the sample) included fewer adolescents, fewer people receiving a prenatal ultrasound before 20 weeks gestation, and lower income people, compared to non-admitted women. Admitted people also had higher rates of pre-delivery psychiatric comorbidities and mental health service use. Amongst the factors independently associated with postpartum admission they found older age (<20 vs. ≥35 years), low income, and mental health service use to be factors during the pregnancy. These include risk factors such as greater time in admission, and protective factors such as assistance to outpatient mental health care or the presence of a consistent mental health care provider during pregnancy [35].

Length of stay is also a relevant question, and women with short stay postpartum admissions (<72 h) seems to be a clinically distinct group that may benefit from targeted intervention [36].

In another retrospective, population-based study in British Columbia, Canada, the researchers analyzed all 189,530 deliveries between 2015 and 2019, including all hospital admissions. Rates, characteristics, and risk factors for postpartum psychiatric hospital admissions were analyzed. There were 812 postpartum psychiatric admissions (~4/1000 deliveries). Although the most common postpartum admissions were depression (n = 233, 28.7%) and substance-related disorders (n = 234, 28.8%), hospitalizations for psychosis and bipolar disorder appeared later than expected for psychosis (median of 111 days postpartum, IQR = 32 to 213), and bipolar disorder (median of 118 days, IQR = 42 to 230) [37].

In a global perspective, women with schizophrenia show sex-specific health needs that differ according to stage of life [38,39], and the perinatal period is always a challenging period for every mother, especially for mothers with schizophrenia. As Dr. Mary V. Seeman pointed out: “The challenges are inherent in the illness—schizophrenia—but also in the associated stigma, poverty, social isolation, and adverse life circumstances” [40].

Finally, we propose to include detailed description of the onset of symptoms (early or delayed) and follow-up diagnosis of PPP in future studies in order to better define the illness and outcomes.

4. Materials and Methods

4.1. Method

An observational, descriptive, and retrospective study was conducted based on data from patients treated between 2017 and 2024 at Parc Taulí Hospital, all of whom had been diagnosed with a psychotic episode within the first postpartum year, according to DSM-5 clinical criteria [13]. The Parc Taulí Hospital is a General Hospital attending a 450.000 population of the “Vallés Occidental Nord” County in the province of Barcelona (Catalonia, Spain). It is the only obstetric department in this specific area, with a mean of 2.300 deliveries per year. We have a specific High Risk Obstetric Unit and a Perinatal Mental Functional Unit included in the psychiatric department of the same hospital.

Patients were selected from data records and perinatal databases available at the Functional Perinatal Mental Health Unit from our hospital. Personal psychiatric history, including previous diagnoses and pharmacological treatments, was collected. Psychosocial factors such as maternal age at delivery, obstetric complications, type of delivery, parity, level of education, and marital status (married, in a stable relationship, or single parent) were also considered. All information was obtained from electronic medical records.

4.1.1. Inclusion Criteria

• Clinical diagnosis of an acute psychotic episode, with or without affective features, in accordance with DSM-5 criteria.
• Onset of symptoms during the postpartum period, with a clearly defined classification of immediate or early onset (within the first 15 days) or delayed onset (several weeks later).
• Availability of complete medical records, including clinical assessments, sociodemographic data, and follow-up documentation.

4.1.2. Exclusion Criteria

• Psychotic episodes not clearly related to the postpartum period, or cases where the onset date of symptoms could not be accurately determined.
• Presence of severe neurological disorders or other somatic conditions that could explain the psychotic symptoms.
• Insufficient clinical data or incomplete medical records that prevent adequate comparison.

4.2. Statistical Analysis

A retrospective descriptive analysis was conducted on 10 cases of women who experienced a psychotic episode during the postpartum period. Five patients presented with immediate or early decompensation (within 15 days), and five with delayed decompensation (in the following weeks). Descriptive statistics were used to analyze the data, and contingency tables were constructed to examine the relationship between the timing of decompensation and selected variables. Associations between variables were assessed using Mann–Whitney U or Fisher exact-test, with a p-value < 0.05 considered statistically significant. The statistical analyses were performed with the software SPSS, version 23 (SPSS Inc, Chicago, IL, USA).

5. Limitations

Given the small sample size, the results were interpreted with caution, emphasizing the identification of descriptive trends that might be clinically relevant and could inform future research.

6. Final Remarks

Despite the observed clinical differences related to the timing of symptom onset, we remain far from identifying specific associations that would enable the recognition and prevention of postpartum psychosis in our environment. Given the rarity of this condition, larger sample sizes are likely required to obtain statistically significant findings. It is worth emphasizing the potential relevance of exploring the role of partner type more deeply, as this variable may serve as a protective factor against the development of mental disorders and could inform the implementation of targeted support strategies by health care systems.

The present study supports the need to analyze not only clinical variables but also psychosocial factors—such as relational context and living conditions—in order to better understand the variability in symptom emergence during this critical period in a woman’s life.