Prenatal depression increases the odds of preterm birth in a prospective cohort study in Hangzhou Authors: Lulu Zhai, Chunchun Fang, Luning Wang, Jianqin Tan Published: 2026 ABSTRACT There is little evidence about the susceptible exposure window of maternal depression during pregnancy and preterm birth in China. We aimed to assess the trimester-specific association of depression during pregnancy with preterm birth, and the effect modification of social support. In 2021–2023, a prospective cohort study was conducted in Hangzhou, China. Depression was measured using the Edinburgh Postnatal Depression Scale (EPDS) during three trimesters of pregnancy. The Social Support Rating Scale (SSRS) and Life Events Scale for Pregnant Women (LESPW) were adopted to evaluate the social support and stress exposure of pregnant women. Binary logistic regression models were used to evaluate the associations of SSRS, LESPW, and EPDS with preterm birth, presenting as odds ratio (OR) and 95% confidence interval (CI). Stratified analysis by SSRS was performed. Of 3692 pregnant women, 190 (5.1%) had a preterm birth, and the prevalence of depression was 12.1%, 10.2%, and 7.4% for the 1st, 2nd, and 3rd trimester, respectively. High EPDS in the 1st trimester contributed to an increased OR of preterm birth (1.60, 95% CI 1.03, 2.40) after adjusting for potential confounders. Although stratified analyses suggested a significant association only among women with low social support (1.58, 95% CI 1.00, 2.50), the formal interaction test was not significant. Maternal depression during the first trimester seemed to elevate the odds of preterm birth. The association was not modified by social support in our study. It is recommended to establish an early screening system for depression during pregnancy and seek professional psychological counseling and treatment when necessary, to reduce the risk of preterm birth.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-56342-7. Introduction Preterm birth remains a significant global health challenge, affecting approximately 11% of live births worldwide in 2019, with an estimated 15 million infants born preterm1. This condition not only poses immediate risks to neonatal health but also has long-term implications for respiratory and overall health2. The etiology of preterm birth is multifactorial, with both biological and psychosocial factors contributing to its occurrence. Among these factors, maternal depression during pregnancy has garnered increasing attention due to its potential impact on adverse pregnancy outcomes3. Depression is one of the leading causes of disability globally4, affecting a substantial proportion of pregnant women5. Perinatal depression, encompassing both antepartum and postpartum depression, is a significant public health issue with prevalence rates varying across regions and income levels5. In China, depressive disorders are the second major cause of years lived with disability since 20106,7, and women are particularly vulnerable6. The association between prenatal depression and preterm birth has been explored in several studies, with inconsistent findings3,8–17. For instance, D’Souza et al.10 found that depressive symptoms were associated with increased risk of preterm birth, while Deepak et al.8 found no significant statistical association between antenatal depression and preterm birth. However, the evidence remains limited, especially in the context of China12,15–19. Social support and life events stress are important factors influencing maternal mental health during pregnancy and birth outcomes. Low social support has been identified as a risk factor for depression during pregnancy20–22. Adverse life events stress may also contribute to increased depression or depressive symptoms among pregnant women23. However, their potential interactions with depression and the role of these factors in preterm birth remain poorly understood. Furthermore, little attention has been given to the susceptible exposure windows. Gaps exist in previous studies, including predominant use of cross-sectional designs or single-timepoint depression assessments; lack of systematic comparison across trimesters; and relatively small sample sizes. To address these gaps, we utilized data from a prospective cohort study with repeated measures of social support, life events, and depression during three trimesters of pregnancy, and birth outcomes (e.g. preterm birth, low birth weight). Our central hypothesis is that maternal depression during pregnancy increases the risk of preterm birth, whereas high levels of social support may attenuate this association. The aim of this study is (1) to comprehensively assess the association of social support, adverse life event experiences, and maternal depression during pregnancy with preterm birth, (2) to explore the potential windows of susceptibility, and (3) to reveal the potential effect modification of social supports on the depression-preterm birth association, further contributing to the evidence of developing effective prevention strategies in China. Materials and methods Study design and participants This population-based prospective cohort study was conducted in Hangzhou, China, during 2021 to 2022. Pregnant women at 8 to < 14 weeks of gestation who received prenatal check-ups at the perinatal health clinics of eight neighborhood community health service centers and one district-level maternal and child health hospital in Hangzhou, and who planned to give birth in Hangzhou, were recruited by obstetric and gynecological healthcare providers. The first survey (baseline) was conducted during early pregnancy (8 to < 14 weeks). The second and third surveys were conducted during mid-pregnancy (16 to < 24 weeks) and late pregnancy (25 to < 34 weeks), respectively, using the same electronic questionnaire. Pregnancy outcomes and delivery information were obtained from the electronic medical record system or by telephone follow-up. At each survey, the demographic information of pregnant women and their spouses was collected. Life Events Scale for Pregnant Women (LESPW), Social Support Rating Scale (SSRS), and Edinburgh Postnatal Depression Scale (EPDS) were used to collect information on life events stress, social support, and depression level of pregnant women. Three surveys were administered during the first, second, and third trimesters of pregnancy, respectively. Each survey included essentially the same set of core questions (e.g., social support, life events, and depressive symptoms). The primary reason for repeated assessments was to capture potential changes in maternal psychosocial status across different gestational stages and to identify critical windows of susceptibility during pregnancy. Data on pregnancy complications and birth outcomes were collected from the medical records. Prior to the investigation being formally conducted, we conducted a knowledge search and understanding of various precautions, called a discussion meeting of the research group, provided uniform training for the staff involved, and described the procedures, methods, and precautions involved in the investigation process, to improve and ensure the quality of the study as much as possible. According to vital statistics from the Shangcheng District of Hangzhou, the local preterm birth rate was 7.89%. Referring to previous international studies, a relative risk (RR) of 1.8 for preterm birth associated with prenatal depressive symptoms was assumed. Assuming equal sample sizes for the exposed and unexposed groups, with a two‑sided α of 0.05, a power (1 − β) of 0.90, and an anticipated overall dropout rate of 20%, the minimum required sample size was calculated to be 1345 participants using the standard formula for two independent proportions. As shown in Fig. 1, a total of 3798 participants finished three trimesters of the survey, and 81 missed birth outcomes, 25 born twins. Finally, 3692 pregnant women with singleton gestations were recruited in the present study. Maternal age in our sample ranged from 18.4 to 46.4 years.Fig. 1The flowchart of the present study. The flowchart of the present study. Gestational weeks were calculated based on the date of the last menstrual period (LMP), but were corrected according to early ultrasound estimates when discrepancies existed. The three time windows (early pregnancy: 8– < 14, middle pregnancy: 16– < 24, and late pregnancy: 25– < 34 weeks) were selected based on the routine prenatal visit schedule and the three traditional trimesters of pregnancy. We set the upper bound of the third survey at < 34 weeks (rather than 37 weeks) to ensure that all participants completed the assessment before the potential onset of preterm birth (delivery < 37 weeks), thereby preserving the temporal sequence between exposure measurement and outcome occurrence and minimizing missing data due to early delivery. Recruitment lasted for two years (2021–2022). Participation was voluntary, and participants could withdraw from the study at any time without consequences. No monetary incentives were provided; however, women who completed all three surveys and delivered at the specialized maternal and child health hospital received a gift package containing infant care products as a token of appreciation. The study was conducted according to the Declaration of Helsinki. The research procedure received approval from the ethics and research committees of Hangzhou Linping District Maternal and Child Health Care Hospital in June 2021 (approval number: LLSC-KYKT-2021-0030-A). All participants gave written informed consent prior to data collection. Depression assessment during pregnancy The Edinburgh postpartum depression Scale (EPDS), compiled by Cox in 1987, was originally specially prepared for postpartum depression and has the advantages of simplicity and easy understanding24. A large number of studies have shown that the EPDS has complete measurement equivalence for both prenatal and postpartum pregnant women, and can also be used as an effective tool for screening prenatal depression25,26. It contains a total of 10 items, all assigned 0 to 3 points, divided into 4 levels: 0 points, 1 point, 2 points, 3 points represent “never”, “occasionally”, “often”, “always”, and so on. The scale is divided into three subscales, namely, the emotion deficit scale, anxiety scale, and depression scale. The higher the score, the more serious the depression degree. When the score > 13, it indicates depression or depressive symptoms25,27. For the EPDS, the Cronbach’s alpha coefficient for the total scale is 0.759,for the subscales, it is 0.774 and 0.743. The Spearman–Brown split-half reliability coefficient (unequal length) for the total scale is 0.757; for the subscales, it is 0.776 and 0.73228. Social support and life events during pregnancy In this study, the social support rating scale contains 10 items in three dimensions: objective support, subjective support, and utilization of support obtained by individuals. A higher score indicates more social support for an individual. Its internal consistency coefficient ranges from 0.89 to 0.94, and the test–retest reliability is 0.9229. We defined it as high level when the value > median. The Life Events Scale for Pregnant Women (LESPW) was adopted to evaluate the stress of pregnant women during pregnancy, which was specially designed for pregnant women and involved family life, work, study, social relations, and other aspects. A total of 53 items, each item is divided into “yes” and “no” two levels, and weighted scores were calculated. The higher the score, the more life events experienced by pregnant women, indicating the higher the level of stress during pregnancy30. With a cutoff of 375, the total score was split into two groups: “high” and “low” levels30. The Prenatal LESPW has an internal consistency coefficient of 0.96 and a test–retest reliability of 0.36–0.9230. Birth outcomes Birth outcomes included preterm birth (delivery before 37 completed weeks of gestation1,31), term birth (delivery at ≥ 37 weeks of gestation), stillbirth/fetal death (intrauterine fetal death at ≥ 20 weeks of gestation), and early termination of pregnancy (elective or therapeutic termination prior to fetal viability) in this study. Demographics Data on demographics were collected from the questionnaires and medical records as follows: The age of pregnant woman (survey date minus birth date), pre-pregnancy BMI, parity (primiparous or multiparous), gestational week, monthly household income (< 4000, 4001–6000, 6001–8000, 8001–10,000, and > 10,000 CNY), present occupation (unemployed, worker, farmer, business/service industry, individuals/freelancers, and public officials), education level (high school and below, junior college, undergraduate, and graduate and above), yes or no for smoking during pregnancy, drinking during pregnancy, Pregnancy complications included gestational anemia, gestational hypertension, and gestational diabetes mellitus. The number of smokers or drinkers during pregnancy was lower, so these two variables were not considered in subsequent analysis in the present study. Statistical analysis The statistical analysis was conducted in three steps. Step I: exploratory data analysis was performed. Continuous data are presented as mean and standard deviation (SD), or median and interquartile range (IQR), and the student’s t test or Mann–Whitney U test was used to compare the differences between the preterm group and the on-preterm group appropriately. Categorical data are presented as numbers and percentages, and Pearson’s chi-squared test was adopted to compare the differences between the two groups. Then, Spearman correlation analysis was applied to determine the correlation among three scale scores during three trimesters. Step II: univariate and multivariable binary logistic regression models were used to evaluate the associations between SSRS, LESPW, and EPDS during each trimester and preterm birth, presenting as odds ratio (OR) and 95% confidence intervals (CI). In the multivariable models, we adjusted for maternal age, pre-pregnancy BMI, gestational weight gain, delivery mode, and complications during pregnancy, including anemia during pregnancy, gestational hypertension, and gestational diabetes, based on previous literature and directed acyclic graph (Fig. S1). We used a p-value of 0.2 as the cutoff for covariate adjustment to avoid missing potential confounders, which is a standard epidemiological practice to ensure important variables are not excluded simply because their univariate association with the outcome did not reach statistical significance (p < 0.05). In addition, stratified analyses by SSRS group (above median = high, at or below median = low) were carried out to reveal the moderating effect on LESPW/EPDS-preterm birth association. The variance inflation factor (VIF) was calculated in the adjusted model to avoid multicollinearity (VIF < 5) as well. Step III: a series of sensitive analyses was also performed using different cut-offs (i.e., 10 scores for EPDS, and the 75th percentile value of SSRS) to generate binary groups, to verify the robustness of the results presented. All the analyses were performed using “bruceR”, “tydiverse”, “car”, and “ggplot2” packages based on R 4.2.2 (R Core Team, Boston, USA). Statistical significance was defined as a two-tailed p-value of < 0.05. Results The demographics of the participants in this study are depicted in Table 1. Among 3692 pregnant women, the average age (SD) was 29.44 (4.15) years, and 190 (5.1%) had preterm birth. In the preterm group, maternal age was higher (30.43 vs 29.38 years), gestational weight gain was lower (11.12 vs 12.84 kg), and there were more cesarean delivery births (61.6 vs 38.6%), compared to the non-preterm group.Table 1The demographics of the participants.VariablesOverall (N = 3692) ≥ 37 weeks (N = 3502) < 37 weeks (N = 190)p-valueAge (year, mean (SD))29.44 (4.15)29.38 (4.13)30.43 (4.43)0.001Weight before pregnancy (kg, mean (SD))55.93 (8.91)55.91 (8.91)56.14 (8.87)0.729Pre-BMI (kg/m2, mean (SD))21.73 (3.16)21.72 (3.16)21.99 (3.27)0.13Gestational weight gain (kg, mean (SD))12.75 (4.42)12.84 (4.40)11.12 (4.44) < 0.001Delivery mode (N (%)) < 0.001 Vaginal delivery2223 (60.2)2150 (61.4)73 (38.4) Cesarean delivery1469 (39.8)1352 (38.6)117 (61.6)Parity (N (%))0.488 Primiparous2042 (55.3)1942 (55.4)100 (52.6) Multiparous1650 (44.7)1560 (44.6)90 (47.4)Complications during pregnancy (N (%))566 (15.3)529 (15.1)37 (19.5)0.079Sleep time per day (hour, N (%))0.882 < 623 (0.6)21 (0.6)2 (1.1) 6–7429 (11.6)404 (11.5)25 (13.2) 7–81416 (38.4)1343 (38.3)73 (38.4) 8–91392 (37.7)1323 (37.8)69 (36.3) > 9432 (11.7)411 (11.7)21 (11.1)Occupation (N (%))0.496 Unemployed800 (21.7)767 (21.9)33 (17.4) Worker355 (9.6)333 (9.5)22 (12) Farmer17 (0.5)15 (0.4)2 (1.1) Business/server1326 (35.9)1257 (35.9)69 (36) Individual827 (22.4)781 (22.3)46 (24) Public official367 (9.9)349 (10.0)18 (9.5)Monthly household income (CNY, N (%))0.928 ≤ 4000120 (3.3)114 (3.3)6 (3.2) 4001–6000411 (11.1)391 (11.2)20 (11.6) 6001–8000501 (13.6)472 (13.5)29 (15.3) 8001–10,000750 (20.3)715 (20.4)35 (18.4) ≥ 10,0001910 (51.7)1810 (51.7)100 (52.6)Education levels (N (%))0.688 High school and below1312 (35.5)1241 (35.4)71 (37.4) Junior college1096 (29.7)1036 (29.6)60 (31.6) Undergraduate1137 (30.8)1086 (31.0)51 (26.8) Graduate and above147 (4.0)139 (4.0)8 (4.2)Data were presented with mean (standard deviation, SD) or number (percentage, %). Complications during pregnancy includes gestational anemia, gestational hypertension, and gestational diabetes mellitus. Table 1The demographics of the participants.VariablesOverall (N = 3692) ≥ 37 weeks (N = 3502) < 37 weeks (N = 190)p-valueAge (year, mean (SD))29.44 (4.15)29.38 (4.13)30.43 (4.43)0.001Weight before pregnancy (kg, mean (SD))55.93 (8.91)55.91 (8.91)56.14 (8.87)0.729Pre-BMI (kg/m2, mean (SD))21.73 (3.16)21.72 (3.16)21.99 (3.27)0.13Gestational weight gain (kg, mean (SD))12.75 (4.42)12.84 (4.40)11.12 (4.44) < 0.001Delivery mode (N (%)) < 0.001 Vaginal delivery2223 (60.2)2150 (61.4)73 (38.4) Cesarean delivery1469 (39.8)1352 (38.6)117 (61.6)Parity (N (%))0.488 Primiparous2042 (55.3)1942 (55.4)100 (52.6) Multiparous1650 (44.7)1560 (44.6)90 (47.4)Complications during pregnancy (N (%))566 (15.3)529 (15.1)37 (19.5)0.079Sleep time per day (hour, N (%))0.882 < 623 (0.6)21 (0.6)2 (1.1) 6–7429 (11.6)404 (11.5)25 (13.2) 7–81416 (38.4)1343 (38.3)73 (38.4) 8–91392 (37.7)1323 (37.8)69 (36.3) > 9432 (11.7)411 (11.7)21 (11.1)Occupation (N (%))0.496 Unemployed800 (21.7)767 (21.9)33 (17.4) Worker355 (9.6)333 (9.5)22 (12) Farmer17 (0.5)15 (0.4)2 (1.1) Business/server1326 (35.9)1257 (35.9)69 (36) Individual827 (22.4)781 (22.3)46 (24) Public official367 (9.9)349 (10.0)18 (9.5)Monthly household income (CNY, N (%))0.928 ≤ 4000120 (3.3)114 (3.3)6 (3.2) 4001–6000411 (11.1)391 (11.2)20 (11.6) 6001–8000501 (13.6)472 (13.5)29 (15.3) 8001–10,000750 (20.3)715 (20.4)35 (18.4) ≥ 10,0001910 (51.7)1810 (51.7)100 (52.6)Education levels (N (%))0.688 High school and below1312 (35.5)1241 (35.4)71 (37.4) Junior college1096 (29.7)1036 (29.6)60 (31.6) Undergraduate1137 (30.8)1086 (31.0)51 (26.8) Graduate and above147 (4.0)139 (4.0)8 (4.2)Data were presented with mean (standard deviation, SD) or number (percentage, %). Complications during pregnancy includes gestational anemia, gestational hypertension, and gestational diabetes mellitus. The demographics of the participants. Data were presented with mean (standard deviation, SD) or number (percentage, %). Complications during pregnancy includes gestational anemia, gestational hypertension, and gestational diabetes mellitus. Table 2 shows the different distribution of three scales scores during early, middle, and late pregnancy. Depression symptoms are assessed by the EPDS. Stressful events are assessed by the LESPW. Social supports are assessed by the SSRS. The EPDS scores decreased gradually as pregnancy continued, from 7 (4, 10) of the first trimester, to 5 (2, 9) of the second trimester, and to 4 (1, 8) of the third trimester (Fig. S2). The prevalence of depression or depression symptoms (EPDS scores > 13) was 12.1%, 10.2%, and 7.4% as well. The prevalence rate was higher (17.4%) in the preterm group than that in the non-preterm group (11.8%). We could observe that total SSRS scores slightly increased from early to late pregnancy. On the contrary, total LESPW scores attenuated mildly from the 1st to the 3rd trimester (Table S1). Figure S3 presents the detailed Spearman correlation coefficients among the total scores of EPDS, SSRS, and LESPW in three trimesters of pregnancy. EPDS was negatively correlated to SSRS (rs ranged from − 0.50 to − 0.30), and positively correlated to LESPW (rs ranged from 0.16 to 0.28).Table 2The distribution of the EDPS, SSRS, and LESPW.VariablesOverall (N = 3692) ≥ 37 weeks (N = 3502) < 37 weeks (N = 190)p-valueEPDS group (> 13, (N (%))) 1st trimester445 (12.1)412 (11.8)33 (17.4)0.021 2nd trimester375 (10.2)358 (10.2)17 (8.9)0.571 3rd trimester275 (7.4)262 (7.5)13 (6.8)0.744SSRS group (> median, (N (%))) 1st trimester1978 (53.6)1875 (53.5)103 (54.2)0.857 2nd trimester1985 (53.8)1887 (53.9)98 (51.6)0.535 3rd trimester1891 (51.2)1778 (50.8)113 (59.5)0.019LESWP group (> 375, (N (%))) 1st trimester587 (15.9)558 (15.9)29 (15.3)0.806 2nd trimester620 (16.8)592 (16.9)28 (14.7)0.436 3rd trimester477 (12.9)452 (12.9)25 (13.2)0.920Data were presented with number and percentage (%); p-value was calculated by Pearson’s Chi-squared test.EPDS, Edinburgh Postnatal Depression Scale; SSRS, The Social Support Rating Scale; LESPW, Life Events Scale for Pregnant Women. Table 2The distribution of the EDPS, SSRS, and LESPW.VariablesOverall (N = 3692) ≥ 37 weeks (N = 3502) < 37 weeks (N = 190)p-valueEPDS group (> 13, (N (%))) 1st trimester445 (12.1)412 (11.8)33 (17.4)0.021 2nd trimester375 (10.2)358 (10.2)17 (8.9)0.571 3rd trimester275 (7.4)262 (7.5)13 (6.8)0.744SSRS group (> median, (N (%))) 1st trimester1978 (53.6)1875 (53.5)103 (54.2)0.857 2nd trimester1985 (53.8)1887 (53.9)98 (51.6)0.535 3rd trimester1891 (51.2)1778 (50.8)113 (59.5)0.019LESWP group (> 375, (N (%))) 1st trimester587 (15.9)558 (15.9)29 (15.3)0.806 2nd trimester620 (16.8)592 (16.9)28 (14.7)0.436 3rd trimester477 (12.9)452 (12.9)25 (13.2)0.920Data were presented with number and percentage (%); p-value was calculated by Pearson’s Chi-squared test.EPDS, Edinburgh Postnatal Depression Scale; SSRS, The Social Support Rating Scale; LESPW, Life Events Scale for Pregnant Women. The distribution of the EDPS, SSRS, and LESPW. Data were presented with number and percentage (%); p-value was calculated by Pearson’s Chi-squared test. EPDS, Edinburgh Postnatal Depression Scale; SSRS, The Social Support Rating Scale; LESPW, Life Events Scale for Pregnant Women. Figure 2 shows the associations of preterm birth with depression (i.e., EPDS > 13) during different trimesters of pregnancy. Maternal depression during the 1st trimester of pregnancy was associated with an increased OR of preterm birth (adjusted OR = 1.60, 95% CI 1.03, 2.40). Preterm birth was not independently associated with LESPW and SSRS, both in continuous and categorical forms (data not shown).Fig. 2Associations of preterm birth with EPDS during different trimesters of pregnancy. Data were presented as estimated odds ratio and 95% confidence intervals calculated by binary logistic regression models. Crude model was the univariate model; adjusted model controlled maternal age, pre-pregnancy BMI, gestational weight gain, delivery mode, and complications during pregnancy including anemia during pregnancy, gestational hypertension, and gestational diabetes. Associations of preterm birth with EPDS during different trimesters of pregnancy. Data were presented as estimated odds ratio and 95% confidence intervals calculated by binary logistic regression models. Crude model was the univariate model; adjusted model controlled maternal age, pre-pregnancy BMI, gestational weight gain, delivery mode, and complications during pregnancy including anemia during pregnancy, gestational hypertension, and gestational diabetes. In the stratified analysis, high EPDS during the 1st trimester contributed to an increased risk of preterm birth (crude OR: 1.70, 95% CI 1.05, 2.68; adjusted OR: 1.58, 95% CI 1.00, 2.50) in the low SSRS group. A non-significant link was observed in the high SSRS group (Fig. 3). Although stratified analyses suggested a significant association only among women with low social support, the formal interaction test was not significant (p for interaction > 0.05). We combined EPDS and SSRS into a new variable with four levels: low-low, high-low, low–high, and high-high, and found a similar trend (Fig. S4). Compared to the low-low group, high EPDS and low SSRS during the 1st trimester contributed to an elevated OR of preterm birth (1.65, 95% CI 1.01, 2.64).Fig. 3Associations of preterm birth with EPDS during different trimesters of pregnancy stratified by SSRS group. Data were presented as estimated odds ratio and 95% confidence intervals calculated by binary logistic regression models. Crude model was the univariate model; adjusted model controlled maternal age, pre-pregnancy BMI, gestational weight gain, delivery mode, and complications during pregnancy including anemia during pregnancy, gestational hypertension, and gestational diabetes. Associations of preterm birth with EPDS during different trimesters of pregnancy stratified by SSRS group. Data were presented as estimated odds ratio and 95% confidence intervals calculated by binary logistic regression models. Crude model was the univariate model; adjusted model controlled maternal age, pre-pregnancy BMI, gestational weight gain, delivery mode, and complications during pregnancy including anemia during pregnancy, gestational hypertension, and gestational diabetes. Discussion In this prospective cohort study conducted in Hangzhou, China, we found that maternal depression during the first trimester of pregnancy was associated with an increased risk of preterm birth. Social support did not significantly modify this association. Preterm birth and maternal depression are both significant public health concerns, affecting the well-being of both mothers and their offspring1–4,6,8,10,18,19,31–34. The prevalence of perinatal depression varies across countries and regions, with recent estimates suggesting a global prevalence of 26.3%35. In China, the prevalence of antenatal and postnatal depression is comparable to global estimates. Our study found that the prevalence of depression or depressive symptoms during pregnancy was 12.1%, 10.2%, and 7.4% in the first, second, and third trimesters, respectively. These findings are consistent with previous studies12,16,17. The association between prenatal depression and preterm birth has been reported in several studies8–17 . However, the findings remain inconsistent, and evidence from China is limited. Zhang et al.16,18,19 reported that women with prenatal depressive symptoms in third trimester were associated with higher risk of preterm birth (OR = 1.40, 95% CI 1.02, 1.93). Li et al.12 found a null statistically significant association between prenatal depression and preterm birth. Furthermore, few previous studies have explored whether and which exposure windows, like the first, second, or third trimester of pregnancy depression, might play an important role in preterm birth. Recently, Zhou et al.17 conducted a prospective multicenter cohort study involving 2053 participants in Shanghai, China, and found that a higher EPDS score (≥ 10) in the third trimester was associated with increased risks of preterm birth (adjusted OR = 1.08, 95% CI 1.01, 1.16). In our study, a higher EPDS score (≥ 13) in the first trimester was associated with increased risks of preterm birth (adjusted OR = 1.64, 95% CI 1.06, 2.46). We systematically analyzed potential reasons for this discrepancy, including differences in: (a) study population and occupational distribution (Shanghai vs. Hangzhou), (b) regional prenatal care pathways and intervention protocols; (c) covariate adjustment strategies; (d) depression cut-off values (cut-off ≥ 10 vs. cut-off ≥ 13); and (e) sample size and statistical power across trimesters (n = 2053 vs. n = 3692). These differences may contribute to the divergent findings regarding the specific trimester during which prenatal depressive symptoms exert the strongest effect on preterm birth risk. We acknowledge that these differences are exploratory and emphasize the need for multi-center prospective studies with standardized protocols to resolve this discrepancy. In addition, we found that high maternal age, less gestational weight gain, and cesarean section delivery were the risk factors for preterm birth. We have adjusted for these potential confounders in the multivariable models and found that the association of higher EPDS in the first trimester with increased risks of preterm birth was robust. Some potential mechanisms were involved in these associations. The activated immune system may serve as the biologic cause for the development of depressive symptoms and their associations with preterm birth36,37. The dysregulation of the hypothalamic–pituitary–adrenal axis might also contribute to the relationship between prenatal depression and preterm birth36. In this study, we did not find evidence for a protective effect of social support on preterm birth in any trimester. The observed higher proportion of high social support in the preterm group during the third trimester (Table 2) should not be interpreted as a harmful effect; rather, it may reflect chance variation, reverse causality, or residual confounding. Future studies with larger sample sizes are needed to clarify the role of social support across different trimesters. Our study found that the association between first-trimester depression and preterm birth was stronger among women with low social support. Although stratified analyses suggested a significant association only among women with low social support, the formal interaction test was not significant (p for interaction > 0.05). Therefore, we found no robust statistical evidence for a moderating effect of social support. The stratified results should be interpreted with caution, as they may be influenced by reduced statistical power or other confounding factors. Previous studies have highlighted the importance of social support in reducing the risk of perinatal depression20–22. Whether social support may act as a protective factor against the negative impact of depression on pregnancy outcomes still needs to be verified by more large‑sample cohort studies. Several limitations of our study should be acknowledged. First, despite the prospective design, we cannot establish causation due to the observational nature of the study. Second, the reliance on self-reported data through e-questionnaires may introduce information bias, such as recall bias. Third, although we adjusted for several potential confounders, other unmeasured confounders may still exist, such as marital status, use of assisted reproductive technology, intrahepatic cholestasis of pregnancy, premature rupture of membranes, placenta previa, and placental abruption. However, in line with previous studies suggesting that prenatal depression screening and intervention might help prevent preterm birth38,39, our study added to the body of data that social support showed limited evidence of a protective or moderating effect. Our findings, based on data from economically developed East China (Hangzhou), are most representative of this region. However, generalization to less developed central and western areas requires caution. Future multi-center prospective cohort studies are needed to validate external applicability, incorporating objective measures—such as structured clinical interviews (e.g., SCID-5) or clinician-administered scales (e.g., HAM-D) for depression, and observed social interaction tasks or ecological momentary assessment for social support—as well as biological markers (e.g., hair or salivary cortisol, inflammatory cytokines such as IL-6, TNF-α, and CRP, and oxytocin levels) to elucidate the underlying mechanisms. Conclusion In conclusion, our study highlights the importance of addressing maternal depression during the first trimester of pregnancy to reduce the risk of preterm birth. Social support showed limited evidence of a protective or moderating effect in this cohort. Implementing early screening and intervention for depression during pregnancy is a recommended strategy to improve maternal and neonatal health outcomes. Supplementary Information Supplementary Information 1. Supplementary Information 2. Supplementary Information 1. Supplementary Information 2.