Perinatal depression (PND) encompasses depressive episodes experienced by women during pregnancy and up to 1 year after delivery (1). This clinically relevant disorder affects an estimated 15%–20% of women in China (2). The occurrence of PND varies across different stages of pregnancy. A comprehensive meta-analysis examining depression rates during pregnancy from 21 centers indicated that the incidence of depression in the first trimester ranged from approximately 2.2% to 12.6%, the second trimester exhibited rates between 10.7% and 14.8%, and rates of depression during the third trimester ranged from 7.4% to 12.6% (3). Notably, around 70% of pregnant women experiencing depression reported moderate levels of severity (4).

Maternal depression has been consistently linked to adverse perinatal outcomes. These outcomes encompass an increased likelihood of poor adherence to medical care, compromised nutrition (either insufficient or excessive gestational weight gain), diminished interpersonal and financial resources, as well as heightened risks of smoking and substance abuse and their associated consequences (5, 6). In the postpartum period, persistent depression has been found to impede maternal recovery, impair breast milk production, reduce the intensity of contractions, and increase the risk of postpartum hemorrhage (7).

Women experiencing severe depressive disorder during the perinatal period may exhibit extremely negative thinking patterns and, when subjected to certain triggers, may even develop suicidal tendencies or harm their infants (8). The incidence of suicides resulting from postpartum depressive disorder is alarmingly high in economically underdeveloped countries, reaching up to 20%, while economically developed countries report rates ranging from 5% to 14% (9). This represents a significant cause of maternal mortality worldwide. Furthermore, the effects of maternal depression on newborns include an increased likelihood of preterm birth, low birth weight, higher rates of diarrhea and infectious diseases, and impaired cognitive development (10, 11). Additionally, maternal depression during pregnancy raises the risk of depression in offspring during adulthood, with a risk 4.6 times higher than that in the general population (12). It can also contribute to the development of depression in the spouse (13). Although maternal depression is prevalent and associated with significant social dysfunction, it is only recently that multiple and concurrent risk factors have been identified. Identifying and addressing this issue could potentially reduce the number of suicides among perinatal women experiencing depression, as well as mitigate the adverse effects of untreated maternal depression on the cognitive and behavioral development of their children.

Perinatal depression is a multifaceted condition believed to arise from interactions involving the neuroendocrine hypothalamic-pituitary-adrenal axis, genetics, epigenetics, and environmental and social factors (1). Prior research has identified various risk factors associated with PND. These include race, age, primiparity, unplanned pregnancy, poor physical health, a history of mental illness, low social support within the family, and marital disharmony (14–19). Currently, the diagnosis of PND primarily relies on comprehensive assessments encompassing medical history, psychiatric examination, psychological evaluation, and physical examination, as there are no specific laboratory markers available. Many symptoms of PND lack specificity, such as insomnia, irritability, and self-doubt regarding the adjustment to the role of being a new mother. Consequently, accurately detecting PND can be challenging. Therefore, employing scales and objective indicators to assess the risk of PND, identify influencing factors, and actively intervene to reduce the incidence of PND holds significant importance.

In China, the number of births in 2022 reached 9.56 million. With the recent implementation of the three-child policy, the population of pregnant women is expected to further expand. Consequently, there will likely be a rise in the number of women who are of advanced age and at a higher risk of developing PND, ultimately impacting the incidence of this condition. Thus, the main objective of this study is to determine the incidence of PND and identify its risk factors at various stages of the perinatal period, followed by establishing a model to predict the occurrence of PND.

Our study yields three key findings. Firstly, we observed variations in the incidence of depression across the six time points of perinatal period. And, the rate of PND is significantly higher postpartum compared to antepartum, reaching its peak in the first week after delivery. Secondly, there are differences in the related factors of prenatal and postpartum depression. Thirdly, work status emerged as a significant determinant in the occurrence of PND throughout the entire perinatal period.

Our study revealed a decline in the incidence of prenatal depression. The incidence rates of depression were approximately 15.5%, 9.7%, and 5.2% during the first, second, and third trimesters, respectively. Furthermore, the incidence of postpartum depression was observed to be 19.4%, 17.5%, and 8.9% at 1 week, 6 months, and 12 months postpartum. This could be attributed to female patients gradually adapting to their current circumstances. Therefore, greater attention should be paid to the mental health status of patients during early pregnancy and the postpartum period. In our study, the incidence rate of postpartum depression was observed to be 33.81%, surpassing the incidence rate of prenatal depression, which was 25.57%. This finding contradicts the research results of Gelaye et al. (20) who conducted a comprehensive analysis of 51 articles on antepartum depression and 53 articles on postpartum depression. They concluded that in low- and middle-income countries, the incidence of antepartum and postpartum depression was estimated at 25.8% (95% CI, 22.8%–29.0%) and 19.7% (95% CI, 16.9%–22.8%) respectively (20). However, it is important to acknowledge that these rates may vary based on geographical and socioeconomic factors.

Our study identified several significant predictors of prenatal depression, including being a housewife, a monthly income less than 5,000 CNY, being unmarried and thyroid dysfunction. Low household income has been identified as an important factor influencing prenatal depression. Previous research has shown that both hypo- and hyperthyroidism can lead to mood disorders, with clinical hypothyroidism being present in 1%–4% of patients with affective disorders (21). It is worth noting that abnormal thyroid function not only increases the risk of depression in the mother during pregnancy but also in their offspring (22). The underlying pathophysiology explaining the emotional and behavioral disorders in patients with hypothyroidism remains uncertain. However, it has been observed that nuclear receptors for triiodothyronine are present in the brain, particularly in regions such as the amygdala and hippocampus, which play a significant role in emotion regulation (22). Therefore, monitoring thyroid function during pregnancy is crucial. Previous study had shown that women with unintended pregnancies had higher odds of experiencing current stress, current depressive symptoms, and initiating prenatal care after the first trimester compared to women with intended pregnancies (23). However, in our research, there was no significant relationship between unintended pregnancies and prenatal depression. Furthermore, several studies have reported associations between gestational diabetes mellitus (GDM) and prenatal depression (24–26). Conversely, depression identified in early pregnancy may increase the risk of subsequent GDM development (26). However, in Lara-Cinisomo et al.’s (27) study, no significant association between prenatal depression and diabetes status was found. The same result was observed in our study.

In our study, several factors were found to significantly predict postpartum depression, including housewife status, monthly income less than 5,000 CNY, unmarried status, primipara, and unintended pregnancy. These findings are consistent with previous research that has identified various risk factors for postpartum depression, such as a history of mental illness, high living burden, poor socioeconomic status, adverse delivery outcomes, and postpartum sleep disturbances (20, 28). Notably, our study also revealed that unintended pregnancy was an important risk factor for postpartum depression, which is supported by Gastaldon et al.’s (29) study indicating a 50% increased risk of postpartum depression in women with unintended pregnancies. However, contrary to expectations, thyroid dysfunction was not found to be statistically significant in predicting postpartum depression, which aligns with the conclusions drawn by Keshavarzi et al. (30). Additionally, Tachibana et al.’s (19) study highlighted that antenatal risk factors, including “a perceived lack of family cohesion,” being a primipara, and receiving treatment for physical or psychiatric illness during pregnancy, could be utilized as indicators for predicting postpartum depression at 20 weeks. Furthermore, clinical studies have demonstrated that experiencing perinatal pain is associated with an increased risk of postpartum depression, while the use of epidural analgesia can potentially reduce this risk (31). Regrettably, these aspects were not included in our study.

In our study, being a housewife emerged as a significant factor influencing depression throughout the entire perinatal period. Across all six time points, the incidence of depression among housewives consistently exceeded that of non-housewives. This finding might be attributed to the demanding responsibilities of childcare that housewives undertake after childbirth. However, in another study by Tachibana et al. (19), “leaving or losing one’s job” was identified as a marginally significant predictor of postnatal depression among Japanese women. This discrepancy could potentially be attributed to differences in social contexts. In China, being a housewife often signifies long-term unemployment, lack of financial resources, limited social support, and lower family status.

In summary, each period of the perinatal period is associated with specific risk factors for depression. It is crucial to focus on different high-risk factors based on the patient’s gestational week and postpartum period to timely identify postnatal depression. This highlights the significance of our research.

However, our study has several limitations. Firstly, as mentioned earlier, we selected all eligible pregnant women from a single hospital, which may not fully represent the current situation in Ningbo, Zhejiang province, China. Additionally, we did not include various factors such as twins, premature infants, perinatal pain, laboratory examinations, and genetic factors. Women with a history of mental health problems are also a major limitation that compromises the generalization of the results. Previous research has confirmed that perinatal genes play a role as risk factors for postpartum depression (31). For example, one study identified DNA methylation in antenatal TTC9B and HP1BP3 genes, which can predict with 80% accuracy whether a woman will develop depression in the postpartum period (32, 33). Therefore, future research should encompass a comprehensive exploration of the risk factors of depression, incorporating symptoms, signs, laboratory tests, and relevant biological markers. Specifically, studies should aim to integrate genetic data, clinical information, and environmental factors to construct a more holistic understanding of the etiology of postpartum depression. By adopting a multi-faceted approach, future research has the potential to identify more accurate predictors of depression and develop targeted interventions to improve the mental health of mothers during the perinatal period.

In summary, depression incidence varies across the perinatal period, with distinct factors affecting prenatal and postpartum depression. Employment status significantly influences PND rates. Our XGBoost model effectively predicts PND, offering a valuable tool for early detection and targeted intervention.