Trajectories of perinatal depression, anxiety, perceived stress, sleep quality, and psychological flexibility from pregnancy to six months postpartum and their relationship with the number of abortions and deliveries.
Frontiers in psychiatry · 2026
Abstract originale
BACKGROUND: Perinatal psychological distress is common and longitudinally heterogeneous, yet little is known about how obstetric history shapes symptom trajectories. We identified trajectories of depressive symptoms, perceived stress, sleep quality, psychological flexibility, and anxiety from early pregnancy to 24 weeks postpartum and examined their associations with the number of previous abortions and deliveries. METHODS: In this prospective longitudinal cohort study, 254 pregnant women were recruited at 13 weeks of gestation from a tertiary hospital in China and assessed eight times through 24 weeks postpartum using the EPDS, CPSS, PSQI, CompACT, and GAD-7. Group-based trajectory modeling identified latent trajectory classes. Logistic regression examined associations with obstetric history, adjusting for sociodemographic covariates. RESULTS: Three trajectories emerged for depressive symptoms (low-stable, 38.98%; mild-stable, 47.24%; moderate-increasing, 13.78%), perceived stress (low-stable, 31.89%; moderate-increasing, 16.54%; high-stable, 51.57%), and sleep quality (good-stable, 50.00%; mild-worsening, 38.58%; persistent poor, 11.42%); two emerged for psychological flexibility (continuously declining vs high-stable) and anxiety (low-stable, 93.70%; moderate, 6.30%). Compared with women without abortion history, those with two or more abortions had higher odds of mild-stable (OR = 5.14) and moderate-increasing depression (OR = 19.12), mild-worsening (OR = 3.06) and persistent poor sleep (OR = 5.51), and declining psychological flexibility (OR = 2.61); one abortion was associated with increased stress (OR = 2.54 and 2.16) and moderate anxiety (OR = 4.18). Number of deliveries was not associated with any trajectory. CONCLUSIONS: Previous abortions predict adverse perinatal psychological trajectories in a dose-dependent manner. Routine assessment of abortion history at the first prenatal visit could help target early screening and tailored interventions.
Parole chiave
Leggi il testo integrale originale
Testo estratto dalla pubblicazione: sono esclusi immagini, media, supplementi e parti soggette a diritti distinti.
Trajectories of perinatal depression, anxiety, perceived stress, sleep quality, and psychological flexibility from pregnancy to six months postpartum and their relationship with the number of abortions and deliveries
Background
Perinatal psychological distress is common and longitudinally heterogeneous, yet little is known about how obstetric history shapes symptom trajectories. We identified trajectories of depressive symptoms, perceived stress, sleep quality, psychological flexibility, and anxiety from early pregnancy to 24 weeks postpartum and examined their associations with the number of previous abortions and deliveries.
Methods
In this prospective longitudinal cohort study, 254 pregnant women were recruited at 13 weeks of gestation from a tertiary hospital in China and assessed eight times through 24 weeks postpartum using the EPDS, CPSS, PSQI, CompACT, and GAD-7. Group-based trajectory modeling identified latent trajectory classes. Logistic regression examined associations with obstetric history, adjusting for sociodemographic covariates.
Results
Three trajectories emerged for depressive symptoms (low-stable, 38.98%; mild-stable, 47.24%; moderate-increasing, 13.78%), perceived stress (low-stable, 31.89%; moderate-increasing, 16.54%; high-stable, 51.57%), and sleep quality (good-stable, 50.00%; mild-worsening, 38.58%; persistent poor, 11.42%); two emerged for psychological flexibility (continuously declining vs high-stable) and anxiety (low-stable, 93.70%; moderate, 6.30%). Compared with women without abortion history, those with two or more abortions had higher odds of mild-stable (OR = 5.14) and moderate-increasing depression (OR = 19.12), mild-worsening (OR = 3.06) and persistent poor sleep (OR = 5.51), and declining psychological flexibility (OR = 2.61); one abortion was associated with increased stress (OR = 2.54 and 2.16) and moderate anxiety (OR = 4.18). Number of deliveries was not associated with any trajectory.
Conclusions
Previous abortions predict adverse perinatal psychological trajectories in a dose-dependent manner. Routine assessment of abortion history at the first prenatal visit could help target early screening and tailored interventions.
1Introduction
Maternal mental health during the perinatal period is a critical determinant of both maternal well-being and offspring development (1, 2). Perinatal depression, anxiety, and stress are highly prevalent worldwide, with pooled estimates suggesting that approximately 16.3% of women in mainland China experience clinically significant depressive symptoms during the perinatal period, and that antenatal depression (19.7%) is even more common than postnatal depression (14.8%) (3). Beyond depression, anxiety symptoms, poor sleep quality, and elevated perceived stress frequently co-occur during this window, and are independently associated with adverse obstetric outcomes, impaired mother–infant bonding, and long-term child neurodevelopmental deficits (4–6). Despite this substantial burden, perinatal psychological distress remains under-recognized and under-treated in routine obstetric care (3, 7).
Longitudinal studies have shown that perinatal psychological symptoms exhibit significant interindividual heterogeneity over time, with three to five distinct trajectories of depressive symptoms identified, including stable low levels, progressive increase, progressive decrease, and persistently high levels (7–10). Similar heterogeneity has been observed for anxiety and sleep quality (5, 11, 12). This longitudinal analytical approach has revealed clinically significant subgroups that were previously masked (7). Nevertheless, most studies focus on a single outcome, typically depression or anxiety. Second, although demographic and psychosocial predictors such as social support and pregnancy-related anxiety have been studied, little is currently known about the role of obstetric history, particularly the number of previous miscarriages and live births, as a factor influencing different psychological trajectories.
Prior pregnancy loss is a common reproductive experience, affecting approximately 20% of confirmed deliveries, and is considered a persistent source of psychological stress (13, 14). Compared with women without a history of miscarriage, those with prior miscarriage report significantly higher levels of pregnancy-specific anxiety, depressive symptoms, and perceived stress in subsequent deliveries (15–17). Importantly, a growing body of evidence suggests a dose-response relationship: women who have experienced two or more miscarriages report higher levels of anxiety and depressive symptoms than those who have experienced only one (16, 18–20). In contrast, the number of deliveries appears to have a more complex and potentially bidirectional effect (21, 22). Moreover, psychological flexibility has been shown to be a protective factor for perinatal mental health (6). Lower psychological flexibility is associated with higher perinatal anxiety and depression, while Acceptance and Commitment Therapy (ACT) targeting psychological flexibility has been shown to be effective in alleviating perinatal psychological distress (6, 23). However, whether psychological flexibility follows distinct developmental trajectories during the perinatal period, and whether these trajectories are influenced by reproductive history, has not yet been systematically investigated.
This study employs a prospective longitudinal design, involving eight repeated assessments conducted between the 13th week of pregnancy and the 24th week postpartum. By simultaneously modeling the trajectories of five psychological Outcomes: depressive symptoms, anxiety, perceived stress, sleep quality and psychological flexibility, the study will investigate the association between the number of previous miscarriages and deliveries at baseline and psychological trajectories during pregnancy. The findings are expected to provide a basis for developing risk-stratified screening programmers and personalized psychological interventions tailored to women with different reproductive histories, thereby advancing the precision of perinatal mental health care.
2Materials and methods
2.1Study design and participants
This prospective longitudinal cohort study was conducted in the obstetric outpatient department of a tertiary hospital in China. Pregnant women were consecutively recruited at their first prenatal visit between March and May 2022. Eligible participants were women aged 20 years or older, confirmed as early pregnancy by ultrasound, able to read and understand Chinese, and willing to complete the follow-up assessments. Women were excluded if they had a diagnosed severe psychiatric disorder (e.g., schizophrenia, bipolar disorder) or current use of psychotropic medication, had severe fetal abnormalities requiring termination of pregnancy, or had cognitive impairment that would preclude questionnaire completion.
The sample size was estimated using the formula for a quantitative indicator in a prevalence study with antenatal anxiety as the primary calculation indicator. According to a previous meta-analysis, the prevalence of anxiety among pregnant women was taken as π = 17.4% (24). To ensure precision, the allowable error was set at δ = 0.05, and the 95% confidence level was adopted. Accounting for an anticipated 20% attrition rate across the eight repeated assessments, the target sample size was inflated to 264 participants. Of the 298 enrolled women, 44 who completed fewer than three follow-up assessments were excluded; baseline characteristics did not differ between included and excluded participants (all P > 0.05; Supplementary Table S1).
The study protocol was approved by the Ethics Committee of Xi’an Jiaotong University (approval number: 2022-1436). Written informed consent was obtained from all participants prior to enrollment. All methods were performed in accordance with the Declaration of Helsinki and relevant guidelines.
2.2Data collection and measurement time points
Study participants completed interview questionnaires at eight time points: at 13 weeks of gestation (baseline), 20, 26, 32, and 37 weeks, as well as at 1, 6, and 24 weeks postpartum. At baseline, professionally trained research nurses collected demographic and obstetric information, including age, educational level, residence, occupation, average monthly household income, family history of depression, number of previous deliveries, and number of previous abortions (including spontaneous and induced). The number of deliveries refers to the number of times a pregnant woman has given birth to a fetus at 28 weeks’ gestation or later. Abortion refers to the termination of a pregnancy before 28 weeks’ gestation, with a fetal weight of less than 1,000 grams. At each time point, a face-to-face psychological assessment was conducted using questionnaires.
2.3Measures
2.3.1Depressive symptoms
Depressive symptoms were assessed using the Edinburgh Postnatal Depression Scale (EPDS), which consists of 10 items. Originally developed for the postpartum population, the validity of this scale for use during pregnancy has been widely validated (25, 26). Each item is scored on a 4-point scale (0–3), with a total score ranging from 0 to 30. A higher score indicates more severe depressive symptoms. The scale demonstrated good internal consistency in our sample data, with Cronbach’s α coefficients ranging from 0.806 to 0.863 across all time points (see Supplementary Table S2).
2.3.2Perceived stress
Perceived stress was assessed using the 14-item Chinese Perceived Stress Scale (CPSS) (27). Items are rated on a 5-point Likert scale (0 = never to 4 = always), with seven positively stated items (4–7, 9, 10, 13) reverse-scored. The total score ranges from 0 to 56, with higher scores indicating greater perceived stress. The scale demonstrates good internal consistency, with Cronbach’s α coefficients ranging from 0.799 to 0.878 across all time points (Supplementary Table S2).
2.3.3Sleep quality
Sleep quality over the preceding month was measured with the Pittsburgh Sleep Quality Index (PSQI) (28). The 19 self-rated items generate seven component scores that sum to a global score of 0–21, with higher scores indicating poorer sleep quality. The scale demonstrates good internal consistency, with Cronbach’s α coefficients ranging from 0.717 to 0.898 across all time points (Supplementary Table S2).
2.3.4Psychological flexibility
Psychological flexibility was measured using the 15-item short form of the Comprehensive Assessment of Acceptance and Commitment Therapy Processes (CompACT) (29). The scale comprises three dyadic subscales: Openness to Experience (items 1–3), Behavioral Awareness (items 4–8), and Valued Action (items 9–15). Items 1–3 and 8 are reverse scored, and the total score ranges from 0 to 90, with higher scores indicating greater psychological flexibility. The scale demonstrates good internal consistency, with Cronbach’s α coefficients ranging from 0.852 to 0.908 across all time points (Supplementary Table S2).
2.3.5Anxiety symptoms
Anxiety was measured with the 7-item Generalized Anxiety Disorder scale (GAD-7) (30). Each item is scored from 0 (not at all) to 3 (nearly every day), giving a total score of 0–21. Scores of 5–9, 10–14, and ≥15 indicate mild, moderate, and severe anxiety, respectively. The scale demonstrates good internal consistency, with Cronbach’s α coefficients ranging from 0.865 to 0.931 across all time points (Supplementary Table S2).
2.4Statistical analysis
Descriptive statistics were used to characterize the sample. The number of abortions and deliveries at baseline was self-reported by maternal participants and categorized as 0, 1, or ≥2. Descriptions were provided for each group using frequencies and percentages for categorical variables, and Chi-square tests were used for comparisons between groups. The group-based trajectory modeling (GBTM), a semi-parametric finite mixture modeling approach, was applied to identify latent subgroups following distinct developmental courses for each of the five psychological outcomes (EPDS, CPSS, GAD-7, PSQI, and CompACT) (31, 32). GBTM accommodates intermittently missing data under maximum likelihood estimation, using all available observations. Time was coded as gestational/postpartum week. Models with one to six trajectory classes were estimated for each outcome using a censored normal distribution, with polynomial functions (intercept, linear, quadratic terms) tested for each class. The optimal number of classes was determined by: (1) lower Bayesian Information Criterion (BIC) and Akaike Information Criterion (AIC) values; (2) entropy approaching 1.0; (3) average posterior probability (avePP) of assignment ≥ 0.70 for each class; and (4) each class comprising at least 5% of the sample with clinical interpretability (32).
Following trajectory identification, associations between baseline obstetric history (number of previous abortions: 0, 1, ≥ 2; number of previous deliveries: 0, 1, ≥ 2) and trajectory class membership were examined using binary or multinomial logistic regression, reporting odds ratios (ORs) with 95% confidence intervals (CIs). Two models were fitted: Model 1 was unadjusted; Model 2 adjusted for age, education level, average monthly household income, residence, occupation, and family history of depression. All analyses were performed using R version 4.3.3. Two-sided P values <0.05 were considered statistically significant.
3Results
3.1Participant characteristics
The participant flow is shown in Figure 1. A total of 254 pregnant women were included in the trajectory analysis. Grouped by the number of previous abortions, 128 women (50.39%) reported no history of abortion, 86 (33.86%) reported one, and 40 (15.75%) reported two or more; grouped by the number of previous deliveries, 127 (50.00%) were in their first pregnancy, 99 (38.98%) had been pregnant once before, and 28 (11.02%) twice or more (Table 1). Women with a history of abortion were significantly less likely to hold a bachelor’s degree or above (P = 0.005); a family history of depression was also more frequent among women with prior abortions, although this difference did not reach statistical significance (P = 0.076). Compared with first-pregnancy women, women with one or more previous deliveries were significantly older (P = 0.001), less educated (P = 0.003), and exercised less frequently (P = 0.001). Notably, satisfaction with the parental, marital, and mother-in-law relationships decreased significantly with an increasing number of both previous abortions and previous deliveries. No significant group differences were observed in average monthly household income, residence, occupation, mode of conception, or pregnancy complications (all P > 0.05).
Table 1Baseline characteristics of participants grouped by number of abortions and deliveries.
Characteristics The number of abortions The number of deliveries
0 1 ≥ 2 P value 0 1 ≥ 2 P value
N = 128 N = 86 N = 40 N = 127 N = 99 N = 28
Age, n (%)
< 30 72 (56.25) 43 (50.00) 14 (35.00) 0.063 95 (74.80) 28 (28.28) 6 (21.43) 0.001
≥ 30 56 (43.75) 43 (50.00) 26 (65.00) 32 (25.20) 71 (71.72) 22 (78.57)
Education, n (%)
High school or below 34 (26.56) 39 (45.35) 19 (47.50) 0.005 33 (25.98) 46 (46.46) 13 (46.43) 0.003
Bachelor or above 94 (73.44) 47 (54.65) 21 (52.50) 94 (74.02) 53 (53.54) 15 (53.57)
Average monthly household income, n (%)
< 6,000 84 (65.62) 57 (66.28) 28 (70.00) 0.876 82 (64.57) 68 (68.69) 19 (67.86) 0.799
≥ 6,000 44 (34.38) 29 (33.72) 12 (30.00) 45 (35.43) 31 (31.31) 9 (32.14)
Residence, n (%)
Urban 85 (66.41) 60 (69.77) 26 (65.00) 0.827 86 (67.72) 68 (68.69) 17 (60.71) 0.723
Rural 43 (33.59) 26 (30.23) 14 (35.00) 41 (32.28) 31 (31.31) 11 (39.29)
Occupation, n (%)
Unemployed 33 (25.78) 29 (33.72) 17 (42.50) 0.111 33 (25.98) 36 (36.36) 10 (35.71) 0.211
Employed 95 (74.22) 57 (66.28) 23 (57.50) 94 (74.02) 63 (63.64) 18 (64.29)
Family history of depression, n (%)
No 125 (97.66) 79 (91.86) 36 (90.00) 0.076 123 (96.85) 92 (92.93) 25 (89.29) 0.194
Yes 3 (2.34) 7 (8.14) 4 (10.00) 4 (3.15) 7 (7.07) 3 (10.71)
Mode of conception, n (%)
Natural 126 (98.44) 84 (97.67) 40 (100.00) 0.621 126 (99.21) 96 (96.97) 28 (100.00) 0.315
Assisted 2 (1.56) 2 (2.33) 0 (0.00) 1 (0.79) 3 (3.03) 0 (0.00)
Exercise frequency / week, n (%)
≥ 5 times 21 (16.41) 19 (22.09) 5 (12.50) 0.858 35 (27.56) 9 (9.09) 1 (3.57) 0.001
3–4 times 29 (22.66) 20 (23.26) 11 (27.50) 27 (21.26) 22 (22.22) 11 (39.29)
1–2 times 30 (23.44) 20 (23.26) 9 (22.50) 29 (22.83) 22 (22.22) 8 (28.57)
Rarely 48 (37.50) 27 (31.40) 15 (37.50) 36 (28.35) 46 (46.46) 8 (28.57)
Complications, n (%)
No 119 (92.97) 71 (82.56) 35 (87.50) 0.062 117 (92.13) 84 (84.85) 24 (85.71) 0.205
Yes 9 (7.03) 15 (17.44) 5 (12.50) 10 (7.87) 15 (15.15) 4 (14.29)
Parental relationship satisfaction
M (IQR) 9.50 (8.00, 10.00) 9.00 (8.00, 10.00) 9.00 (7.75, 10.00) 0.018 10.00 (8.00, 10.00) 9.00 (8.00, 10.00) 9.00 (7.00, 10.00) 0.046
Marital relationship satisfaction
M (IQR) 9.00 (8.00, 10.00) 9.00 (8.00, 9.75) 8.50 (7.75, 10.00) 0.003 9.00 (8.00, 10.00) 9.00 (8.00, 10.00) 9.00 (7.75, 10.00) 0.016
Mother-in-law relationship satisfaction
M (IQR) 9.00 (8.00, 10.00) 8.00 (6.25, 9.00) 8.00 (6.00, 9.25) 0.007 8.00 (7.50, 10.00) 8.00 (7.00, 9.00) 8.00 (7.00, 9.00) 0.037
Categorical variables were analyzed using the Chi−square test, and continuous variables using the Mann–Whitney U test.
Figure 2 presents the score distributions of the five psychological measures across the eight assessment points. EPDS scores were highest and most widely dispersed at baseline (13 weeks of gestation) and declined to lower, stable levels from mid-pregnancy onward. CPSS scores remained broadly stable around moderate levels throughout. PSQI scores increased gradually across pregnancy, peaked in the early postpartum period, and declined slightly thereafter. CompACT scores were stably distributed around the mid-range. and GAD-7 scores remained low overall, with a mild elevation in late pregnancy. At every time point, the distributions were right-skewed, indicating a subgroup of women with persistently elevated symptoms.
3.2Trajectories of depressive symptoms (EPDS)
Model fit indices for one- to six-class GBTM solutions are presented in Supplementary Table S3. The three-class model was selected as optimal, balancing fit improvement (AIC = 7715.20, BIC = 7768.26), acceptable entropy (0.74), avePP (0.85–0.92), and class sizes (all >5%). The three trajectories were: Class 1 (38.98%), characterized by persistently low depressive symptoms (low-stable); Class 2 (47.24%), with mild symptoms that remained relatively flat across the perinatal period (mild-stable); and Class 3 (13.78%), characterized by moderate baseline symptoms (EPDS score of approximately 14 at 13 weeks of gestation), with symptoms first worsening and then improving from late pregnancy through the postpartum period (moderate-increasing) (Figure 3A).
3.3Trajectories of perceived stress (CPSS)
Supplementary Table S4 presents the fit statistics for the CPSS trajectory models in classes 1 to 6; the three models are the optimal solutions. This approach achieves a balance among fit (AIC = 8,867.63, BIC = 8,920.69), entropy (0.70), avePP (0.83–0.92), and class size (all >5%). These three trajectories are: Class 1 (31.89%), characterized by persistently low levels of psychological stress (low-stable); Class 2 (16.54%), characterized by an increasing trend in stress (moderate-increasing); and Class 3 (51.57%), characterized by persistently high levels (high-stable) (Figure 3B).
3.4Trajectories of sleep quality (PSQI)
Supplementary Table S5 presents the fit statistics for the PSQI trajectory models in classes 1 to 6; the three models are the optimal solutions. This approach achieves a balance among fit (AIC = 5801.44, BIC = 5854.50), entropy (0.75), avePP (0.87–0.90), and class size (all >5%). Class 1 (50.00%) maintained good sleep quality throughout the entire pregnancy (PSQI < 5; ‘good-stable’); Class 2 (38.58%) exhibited mild sleep difficulties that gradually worsened during pregnancy and persisted into the postpartum period (PSQI < 8; ‘mild-worsening’); Class 3 (11.42%) exhibited persistent poor sleep quality throughout pregnancy (PSQI ≥ 8) (‘persistent poor’) (Figure 3C).
3.5Trajectories of psychological flexibility (CompACT)
Two trajectories of psychological flexibility emerged (Supplementary Table S6). The model achieved a balance between fit (AIC = 10,676.67, BIC = 10,715.58), entropy (0.70), avePP (0.89, 0.93), and class size (all > 5%). Class 1 (57.87%) exhibited low baseline flexibility (CompACT < 55), with a trend of continuous decline, indicating a persistently worsening pattern of psychological flexibility (‘continuously declining’). Class 2 (42.13%) maintained high psychological flexibility throughout the perinatal period (CompACT > 60, ‘high-stable’) (Figure 3D).
3.6Trajectories of anxiety symptoms (GAD-7)
Two anxiety trajectories were identified (Supplementary Table S7). The model achieved a balance between fit (AIC = 6357.32, BIC = 6436.23), entropy (0.90), avePP (0.89, 0.98), and class size (all > 5%). Class 1 (93.70%) remained at a low and stable level at all time points (GAD-7 score < 5 points), representing stable mild anxiety. Class 2 (6.30%) followed an inverted U-shaped course: symptoms rose from moderate baseline levels to a peak in the third trimester, then declined after delivery, but remained at a moderate level throughout (moderate anxiety) (Figure 3E).
3.7Associations of obstetric history with trajectory membership
3.7.1Depressive symptoms
In a multinomial logistic regression analysis using the low-stable class as the reference (Table 2), women with a history of one previous abortion had a significantly higher odds of membership in the mild-stable class (OR = 1.99, 95% CI: 1.09–3.66, P = 0.026), whereas the OR for assignment to the moderate-increasing class was not statistically significant (P = 0.055). Women with a history of ≥ 2 abortions had a significantly higher odds of membership in the mild-stable class (OR = 5.14, 95% CI: 1.79–14.78, P = 0.002) and a significantly higher odds for being classified into the moderate-increasing class (OR = 19.12, 95% CI: 5.26–69.44, P < 0.001). The latter association was also significant in the unadjusted model. Furthermore, the association between the two became stronger as the number of previous abortions increased. There was no statistically significant association between the number of deliveries and the trajectories of any type of depressive symptoms.
Table 2Multinomial logistic regression models of the association between the number of abortions and deliveries with mild-stable and moderate-increasing depression, relative to low-stable depressive symptoms.
Model a Group Mild-stable depression Moderate-increasing depression
ORs (95% CI) P value ORs (95% CI) P value
Number of abortions
Model 1 0 1.00 (Ref.) 1.00 (Ref.)
1 2.04 (1.13, 3.70) 0.018 3.14 (1.19, 8.30) 0.020
≥ 2 5.23 (1.85, 14.80) 0.002 20.53 (5.97, 70.67) <0.001
Model 2 0 1.00 (Ref.) 1.00 (Ref.)
1 1.99 (1.09, 3.66) 0.026 2.68 (0.98, 7.35) 0.055
≥ 2 5.14 (1.79, 14.78) 0.002 19.12 (5.26, 69.44) <0.001
Number of deliveries
Model 1 0 1.00 (Ref.) 1.00 (Ref.)
1 1.07 (0.61, 1.89) 0.810 1.28 (0.55, 2.96) 0.568
≥ 2 1.25 (0.50, 3.16) 0.635 2.31 (0.71, 7.54) 0.165
Model 2 0 1.00 (Ref.) 1.00 (Ref.)
1 0.87 (0.45, 1.66) 0.665 0.80 (0.30, 2.17) 0.943
≥ 2 1.04 (0.38, 2.81) 0.663 1.47 (0.38, 5.74) 0.581
aModel 1 did not adjust for covariates; Model 2 adjusted for age, education level, average monthly household income, residence, occupation and family history of depression.
3.7.2Perceived stress
In a multivariate-adjusted multinomial logistic regression analysis using the low-stable psychological stress group as the reference (Table 3), women with a history of one previous abortion had a significantly higher odds of membership in the moderate-increasing class (OR = 2.54, 95% CI: 1.06–6.06, P = 0.035), while the odds were significantly higher for those classified in the high-stable class (OR = 2.16, 95% CI: 1.09–4.27, P = 0.027). These findings were also statistically significant in the unadjusted model. However, for women with a history of ≥ 2 previous abortions, no statistically significant association was observed with any classes of psychological stress. There was no statistically significant association between the number of deliveries and the trajectory of any type of psychological stress; however, the OR estimates were all less than 1 after adjusting for covariates.
Table 3Multinomial logistic regression models of the association between the number of abortions and deliveries with moderate-increasing and high-stable stress, relative to low-stable psychological stress.
Model a Group Moderate-increasing stress High-stable stress
ORs (95% CI) P value ORs (95% CI) P value
Number of abortions
Model 1 0 1.00 (Ref.) 1.00 (Ref.)
1 2.73 (1.18, 6.33) 0.019 2.31 (1.21, 4.42) 0.011
≥ 2 1.73 (0.55, 5.45) 0.347 2.15 (0.94, 4.92) 0.069
Model 2 0 1.00 (Ref.) 1.00 (Ref.)
1 2.54 (1.06, 6.06) 0.035 2.16 (1.09, 4.27) 0.027
≥ 2 1.51 (0.46, 4.95) 0.501 1.92 (0.80, 4.58) 0.142
Number of deliveries
Model 1 0 1.00 (Ref.) 1.00 (Ref.)
1 0.98 (0.43, 2.21) 0.958 1.55 (0.86, 2.82) 0.148
≥ 2 1.22 (0.36, 4.16) 0.748 1.43 (0.56, 3.67) 0.457
Model 2 0 1.00 (Ref.) 1.00 (Ref.)
1 0.89 (0.32, 2.17) 0.718 0.84 (0.81, 3.36) 0.169
≥ 2 0.69 (0.31, 4.64) 0.792 0.90 (0.56, 4.57) 0.385
aModel 1 did not adjust for covariates; Model 2 adjusted for age, education level, average monthly household income, residence, occupation and family history of depression.
3.7.3Sleep quality
Number of previous abortions was significantly associated with the mild-worsening sleep trajectory (Table 4): one abortion (OR = 2.03, 95% CI: 1.11–3.72, P = 0.022) and two or more abortions (OR = 3.06, 95% CI: 1.32–7.09, P = 0.009); compared with women with no history of miscarriage, the odds of experiencing a decline in sleep quality were increased in both groups. A significant association was also found with a persistently poor sleep trajectory: two or more abortions (OR = 5.51, 95% CI: 1.73–17.59, P = 0.004), whereas one abortion showed no statistically significant association. Number of deliveries was not significantly associated with sleep trajectories.
Table 4Multinomial logistic regression models of the association between the number of abortions and deliveries with mild-worsening and persistent poor sleep quality, relative to good-stable group.
Model a Group Mild-worsening sleep quality Persistent poor sleep quality
ORs (95% CI) P value ORs (95% CI) P value
Number of abortions
Model 1 0 1.00 (Ref.) 1.00 (Ref.)
1 2.05 (1.14, 3.71) 0.017 2.11 (0.81, 5.50) 0.128
≥ 2 3.09 (1.36, 6.99) 0.007 5.85 (1.97, 17.34) 0.001
Model 2 0 1.00 (Ref.) 1.00 (Ref.)
1 2.03 (1.11, 3.72) 0.022 2.04 (0.73, 5.63) 0.170
≥ 2 3.06 (1.32, 7.09) 0.009 5.51 (1.73, 17.59) 0.004
Number of deliveries
Model 1 0 1.00 (Ref.) 1.00 (Ref.)
1 1.00 (0.57, 1.76) 0.992 1.01 (0.42, 2.42) 0.982
≥ 2 0.75 (0.30, 1.87) 0.545 1.20 (0.35, 4.17) 0.774
Model 2 0 1.00 (Ref.) 1.00 (Ref.)
1 0.89 (0.46, 1.69) 0.713 0.84 (0.30, 2.35) 0.738
≥ 2 0.69 (0.26, 1.84) 0.464 0.90 (0.21, 3.88) 0.883
aModel 1 did not adjust for covariates; Model 2 adjusted for age, education level, average monthly household income, residence, occupation and family history of depression.
3.7.4Psychological flexibility
Two trajectories of psychological flexibility were identified: a high-stable and continuously declining groups. Results from a binary logistic model adjusted for covariates revealed a positive association between the number of abortions and a continuously declining psychological flexibility (Table 5). Compared with women with no history of abortion, those who had one abortion (OR = 2.44, 95% CI: 1.35–4.49, P = 0.004) and ≥ 2 abortions (OR = 2.61, 95% CI: 1.20–5.99, P = 0.019) had a significantly higher odds of continuously declining class. Similarly, there was no statistically significant association between the number of deliveries and psychological flexibility.
Table 5Binary logistic regression models of the association between the number of abortions and deliveries with continuously declining psychological flexibility and moderate anxiety, relative to high-stable psychological flexibility group and stable mild anxiety group.
Model a Group Continuously declining psychological flexibility Moderate anxiety
ORs (95% CI) P value ORs (95% CI) P value
Number of abortions
Model 1 0 1.00 (Ref.) 1.00 (Ref.)
1 2.48 (1.41, 4.43) 0.002 5.48 (1.62, 25.02) 0.012
≥ 2 2.64 (1.26, 5.83) 0.012 3.38 (0.60, 18.93) 0.146
Model 2 0 1.00 (Ref.) 1.00 (Ref.)
1 2.44 (1.35, 4.49) 0.004 4.18 (1.23, 20.95) 0.034
≥ 2 2.61 (1.20, 5.99) 0.019 2.14 (0.36, 12.60) 0.383
Number of deliveries
Model 1 0 1.00 (Ref.) 1.00 (Ref.)
1 1.20 (0.71, 2.05) 0.500 2.02 (0.70, 6.21) 0.198
≥ 2 1.47 (0.64, 3.53) 0.377 0.75 (0.04, 4.62) 0.791
Model 2 0 1.00 (Ref.) 1.00 (Ref.)
1 1.15 (0.62, 2.15) 0.667 0.97 (0.28, 3.49) 0.960
≥ 2 1.51 (0.60, 3.95) 0.392 0.26 (0.01, 2.14) 0.273
aModel 1 did not adjust for covariates; Model 2 adjusted for age, education level, average monthly household income, residence, occupation and family history of depression.
3.7.5Anxiety
After adjusting for covariates in the binary logistic model, the results showed a positive association between the number of abortions and anxiety (Table 5). Compared with women with no history of abortion, women with one abortion had a significantly higher likelihood of moderate anxiety (OR = 4.18, 95% CI: 1.23–20.95, P = 0.034). However, the association between ≥ 2 abortions and moderate anxiety was not statistically significant (OR = 2.14, 95% CI: 0.36–12.60, P = 0.383). Similarly, there was no statistically significant association between the number of deliveries and anxiety during pregnancy.
4Discussion
To our knowledge, this is the first study to simultaneously characterize the longitudinal trajectories of five interrelated psychological domains: depressive symptoms, anxiety, perceived stress, sleep quality, and psychological flexibility, across the full perinatal continuum from 13 weeks of gestation to 24 weeks postpartum, and to link baseline reproductive history to these trajectories in a Chinese clinical sample. This study found significant heterogeneity across all five psychological domains, with two to three distinct developmental trajectories identified for each psychological dimension, confirming that overall average trends mask clinically significant subgroups. There was a significant dose-dependent association between the number of previous abortions and adverse trajectories in depressive symptoms, sleep quality, and psychological flexibility. Women who had two or more previous abortions were more likely to experience severe depression, poor sleep quality, and reduced psychological flexibility. In addition, having had one abortion was also associated with increased perceived stress and anxiety during pregnancy. These associations persisted after adjusting for covariates, whereas the number of previous deliveries was not significantly associated with any of the trajectories.
This study identified three trajectories of depression: low-stable (38.98%), mild-stable (47.24%), and moderate-increasing (13.78%), which are similar to the patterns reported in cohort studies from Canada, Norway, and China (8–11, 33). These studies all found that within populations dominated by low-symptom groups, there is a smaller high-risk subgroup. Notably, Yu et al. found that more than half of the women who screened positive for perinatal depression first did so in early pregnancy (10), which aligns with our observation that the high-risk class already showed moderate symptoms at 13 weeks of gestation. This convergence reinforces calls for depression screening to begin in the first trimester rather than being deferred to the postpartum period.
This study also found a dose-dependent association between the number of previous abortions and adverse psychological trajectories. Previous cross-sectional and cohort studies have confirmed that a history of abortion is associated with increased anxiety, depressive symptoms, and perceived stress in subsequent deliveries (15–17). The Screening for Pregnancy Endpoints (SCOPE) cohort study indicated that women with a history of two previous abortions had a higher risk of depression than those with a history of only one abortion (16). A study of Chinese women with recurrent pregnancy loss also reported that the greater the number of abortions, the more severe the depression and sleep difficulties (20). The findings of this study suggest that cumulative pregnancy loss not only predicts symptom exacerbation at a specific point in time but also indicates a trajectory of persistent worsening or persistent psychological distress throughout the perinatal period (from 13 weeks of gestation to 24 weeks postpartum). This suggests that the psychological sequelae of recurrent abortions are not transient adaptive responses but rather constitute a persistent vulnerability that shapes the entire process of perinatal adaptation. Additionally, women with a history of one or more abortions were significantly more likely to exhibit a sustained decline in psychological flexibility, suggesting that prior pregnancy and pregnancy losses may impair the ability to accept difficult internal experiences and act in accordance with personal values, thereby contributing to perinatal psychological distress (29, 34). This finding suggests that recurrent pregnancy loss may foster experiential avoidance and lead to cognitive fusion due to fear of recurrence, thereby progressively narrowing behavioral patterns throughout pregnancy (34).
The results of this study found that women who had experienced two or more abortions were more likely to experience persistent poor sleep quality during the perinatal period. This is consistent with evidence from previous studies indicating that poor sleep quality in late pregnancy predicts persistent postpartum sleep problems (5), as well as the significant association between sleep disorders and mood disorders in the perinatal population (12, 35). Furthermore, with regard to perceived stress, having only one abortion was significantly associated with the moderate-increasing and high-stable stress trajectories, whereas a higher number of abortions was not statistically associated with perceived stress. This result may be due to the extremely small sample size, which limited statistical power in the subgroup with ≥ 2 abortions, or it may reflect the development of certain adaptive habits or proactive coping mechanisms among women who have experienced multiple abortions (16, 19). For anxiety, one previous abortion was associated with more than four-fold higher odds of following the moderate (inverted U-shaped) trajectory, whereas the estimate for two or more abortions was imprecise owing to the small size of this class (6.30%). The peak of anxiety in the third trimester observed in this class is consistent with the escalation of fear of childbirth and concerns about fetal well-being as delivery approaches (15, 17), and its postpartum decline suggests at least partial symptom resolution once a live birth is achieved. The null findings for number of deliveries across all five domains are compatible with the mixed parity literature. Primiparity has been linked to higher perinatal anxiety and depression in some studies (21), whereas multiparity has been associated with steeper postpartum symptom increases in others (22). This fundings indicate that cumulative loss history carries greater prognostic weight than the number of deliveries itself in this population.
Our findings support three concrete clinical actions. First, the number of previous abortions is inexpensive, routinely collected information that can be used at the first prenatal visit to flag women at elevated risk for worsening depression, deteriorating sleep, declining psychological flexibility, and persistent stress. Given the dose-response gradient observed, women with two or more prior abortions warrant particular attention. Second, because the high-risk depressive symptom class already presented moderate symptoms at 13 weeks of gestation, psychological screening for this subgroup should begin in the first trimester and be repeated across pregnancy and into the postpartum period, rather than relying on a single assessment (7, 10). Third, the identification of declining psychological flexibility as a loss-related vulnerability points to a modifiable intervention target. Acceptance and Commitment Therapy, which directly targets psychological flexibility, has demonstrated feasibility and preliminary efficacy in perinatal populations (6, 23), and may be especially suitable for women with repeated pregnancy loss. The persistence of poor sleep and elevated stress well into the postpartum period further suggests that support should extend beyond delivery and incorporate sleep-focused and stress-management components (5, 12).
Strengths of this study include: a prospective study design with eight repeated measurements spanning from early pregnancy to the postpartum period; simultaneous modeling of five psychological domains within a unified trajectory framework; and the use of validated Chinese scales. Nevertheless, several limitations of this study should be acknowledged. First, the single-center design and sample size limited the precision of estimates for small trajectory classes (e.g., the moderate anxiety class at 6.30%), and substantial sample loss at the end of the postpartum follow-up is reflected in the wide confidence intervals. Therefore, larger-scale, multicenter studies are needed to confirm the association between miscarriage and various psychological dimensions. Second, psychological outcomes relied on self-reports rather than clinical diagnostic interviews. Third, the study did not distinguish between spontaneous and induced abortions, which may have different psychological consequences. Fourth, despite adjustments for covariates, residual confounding due to unmeasured factors cannot be ruled out. Fifth, follow-up ended at 24 weeks postpartum, and trajectories may evolve further during the first postpartum year (12). Finally, although the average posterior probabilities for all retained classes exceeded the recommended threshold of 0.70, the entropy values for some models were moderate (0.70–0.75), suggesting that there may be some uncertainty in classification. Therefore, future research should consider study designs involving large samples and long-term follow-up, as well as validating the trajectories of each psychological dimension in independent samples.
5Conclusions
Perinatal psychological adjustment varies across five domains. Recurrent abortions serve as a strong, dose-dependent predictor of worsening depressive symptoms, decreased sleep quality, and reduced psychological flexibility in pregnant women from early pregnancy through the early postpartum period. Routine assessment of abortion history at the first prenatal visit is a practical and low-cost strategy that can help identify, at an early stage, pregnant women who may benefit from enhanced psychological monitoring and targeted interventions during the perinatal period.
References
1Fisher J Cabral de Mello M Patel V Rahman A Tran T Holton S .
Prevalence and determinants of common perinatal mental disorders in women in low- and lower-middle-income countries: a systematic review . Bull World Health Organ . (2012 ) 90 :139g–49g. doi: 10.2471/blt.11.091850 22423165
PMC3302553
2O'Hara MW Wisner KL .
Perinatal mental illness: definition, description and aetiology . Best Pract Res Clin Obstetrics Gynaecology . (2014 ) 28 :3 –12 . doi: 10.1016/j.bpobgyn.2013.09.002 24140480
PMC7077785
3Nisar A Yin J Waqas A Bai X Wang D Rahman A .
Prevalence of perinatal depression and its determinants in Mainland China: a systematic review and meta-analysis . J Affect Disord . (2020 ) 277 :1022–37. doi: 10.1016/j.jad.2020.07.046 33065811
4Bei B Coo S Trinder J .
Sleep and mood during pregnancy and the postpartum period . Sleep Med Clinics . (2015 ) 10 :25 –33 . doi: 10.1016/j.jsmc.2014.11.011 26055670
5Tomfohr LM Buliga E Letourneau NL Campbell TS Giesbrecht GF .
Trajectories of sleep quality and associations with mood during the perinatal period . Sleep . (2015 ) 38 :1237–45. doi: 10.5665/sleep.4900 25845691
PMC4507729
6Waters CS Annear B Flockhart G Jones I Simmonds JR Smith S .
Acceptance and commitment therapy for perinatal mood and anxiety disorders: a feasibility and proof of concept study . Br J Clin Psychol . (2020 ) 59 :461–79. doi: 10.1111/bjc.12261 32715513
7Baron E Bass J Murray SM Schneider M Lund C .
A systematic review of growth curve mixture modelling literature investigating trajectories of perinatal depressive symptoms and associated risk factors . J Affect Disord . (2017 ) 223 :194 –208 . doi: 10.1016/j.jad.2017.07.046 28763638
PMC5592733
8Hong L Le T Lu Y Shi X Xiang L Liu M .
Distinct trajectories of perinatal depression in Chinese women: application of latent growth mixture modelling . BMC Pregnancy Childbirth . (2022 ) 22 :24 . doi: 10.1186/s12884-021-04316-0 35012496
PMC8751241
9Yan P Liu X Xu J .
Trajectories and predictors of symptoms of depression in Chinese women from early pregnancy to the early postpartum period . J Obstetric Gynecologic Neonatal Nurs JOGNN . (2022 ) 51 :577–89. doi: 10.1016/j.jogn.2022.07.008 35987260
10Yu M Li H Xu DR Wu Y Liu H Gong W .
Trajectories of perinatal depressive symptoms from early pregnancy to six weeks postpartum and their risk factors-a longitudinal study . J Affect Disord . (2020 ) 275 :149–56. doi: 10.1016/j.jad.2020.07.005 32734901
11Ahmed A Bowen A Feng CX Muhajarine N .
Trajectories of maternal depressive and anxiety symptoms from pregnancy to five years postpartum and their prenatal predictors . BMC Pregnancy Childbirth . (2019 ) 19 :26 . doi: 10.1186/s12884-019-2177-y 30642277
PMC6332639
12Lin-Lewry M Tzeng YL Li CC Lee GT Lee PH Chen SR .
Trajectories of sleep quality and depressive symptoms in women from pregnancy to 3 months postpartum: a prospective cohort study . J Sleep Res . (2023 ) 32 :e13918. doi: 10.1111/jsr.13918 37128654
13Farren J Jalmbrant M Falconieri N Mitchell-Jones N Bobdiwala S Al-Memar M .
Posttraumatic stress, anxiety and depression following miscarriage and ectopic pregnancy: a multicenter, prospective, cohort study . Am J Obstetrics Gynecology . (2020 ) 222 :367.e1–.e22. doi: 10.1016/j.ajog.2019.10.102 31953115
14Quenby S Gallos ID Dhillon-Smith RK Podesek M Stephenson MD Fisher J .
Miscarriage matters: the epidemiological, physical, psychological, and economic costs of early pregnancy loss . Lancet (London England) . (2021 ) 397 :1658–67. doi: 10.1016/s0140-6736(21)00682-6 33915094
15Bergner A Beyer R Klapp BF Rauchfuss M .
Pregnancy after early pregnancy loss: a prospective study of anxiety, depressive symptomatology and coping . J Psychosomatic Obstetrics Gynaecology . (2008 ) 29 :105–13. doi: 10.1080/01674820701687521 17943588
16McCarthy FP Moss-Morris R Khashan AS North RA Baker PN Dekker G .
Previous pregnancy loss has an adverse impact on distress and behaviour in subsequent pregnancy . BJOG Int J Obstetrics Gynaecology . (2015 ) 122 :1757–64. doi: 10.1111/1471-0528.13233 25565431
17Lazarides C Moog NK Verner G Voelkle MC Henrich W Heim CM .
The association between history of prenatal loss and maternal psychological state in a subsequent pregnancy: an ecological momentary assessment (EMA) study . Psychol Med . (2023 ) 53 :855–65. doi: 10.1017/s0033291721002221 34127159
PMC9975992
18Toffol E Koponen P Partonen T .
Miscarriage and mental health: results of two population-based studies . Psychiatry Res . (2013 ) 205 :151–8. doi: 10.1016/j.psychres.2012.08.029 22985545
19Volgsten H Jansson C Svanberg AS Darj E Stavreus-Evers A .
Longitudinal study of emotional experiences, grief and depressive symptoms in women and men after miscarriage . Midwifery . (2018 ) 64 :23–8. doi: 10.1016/j.midw.2018.05.003 29864578
20Wang TT Liu YL Hou Y Li JP Qiao C .
The risk factors of progestational anxiety, depression, and sleep disturbance in women with recurrent pregnancy loss: a cross-sectional study in China . Front Psychol . (2023 ) 14 :1116331. doi: 10.3389/fpsyg.2023.1116331 37063554
PMC10102375
21Nakamura Y Okada T Morikawa M Yamauchi A Sato M Ando M .
Perinatal depression and anxiety of primipara is higher than that of multipara in Japanese women . Sci Rep . (2020 ) 10 :17060 . doi: 10.1038/s41598-020-74088-8 33051530
PMC7555526
22Canário C Figueiredo B .
Anxiety and depressive symptoms in women and men from early pregnancy to 30 months postpartum . J Reprod Infant Psychol . (2017 ) 35 :431–49. doi: 10.1080/02646838.2017.1368464 29517385
23Bonacquisti A Cohen MJ Schiller CE .
Acceptance and commitment therapy for perinatal mood and anxiety disorders: development of an inpatient group intervention . Arch Women's Ment Health . (2017 ) 20 :645–54. doi: 10.1007/s00737-017-0735-8 28600645
24Yang L Sun J Nan Y Waqas A Nisar A Wang D .
Prevalence of perinatal anxiety and its determinants in mainland China: a systematic review and meta-analysis . J Affect Disord . (2023 ) 323 :193 –203 . doi: 10.1016/j.jad.2022.11.075 36442655
25Cox JL Holden JM Sagovsky R .
Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale . Br J Psychiatry J Ment Sci . (1987 ) 150 :782–6. doi: 10.1192/bjp.150.6.782 3651732
26Cox J .
Thirty years with the Edinburgh Postnatal Depression Scale: voices from the past and recommendations for the future . Br J Psychiatry J Ment Sci . (2019 ) 214 :127–9. doi: 10.1192/bjp.2018.245 30774059
27Yang TZ Huang HT .
An epidemiological study on stress among urban residents in social transition period . Zhonghua liu xing bing xue za zhi = Zhonghua liuxingbingxue zazhi . (2003 ) 24 :760–4. doi: 10.3760/j.issn:0254-6450.2003.09.004
14521764
28Buysse DJ Reynolds CF Monk TH Berman SR Kupfer DJ .
The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research . Psychiatry Res . (1989 ) 28 :193 –213 . doi: 10.1016/0165-1781(89)90047-4 2748771
29Francis AW Dawson DL Golijani-Moghaddam N .
The development and validation of the Comprehensive assessment of Acceptance and Commitment Therapy processes (CompACT) . J Contextual Behav Sci . (2016 ) 5 :134–45. doi: 10.1016/j.jcbs.2016.05.003 42574925
30Spitzer RL Kroenke K Williams JB Löwe B .
A brief measure for assessing generalized anxiety disorder: the GAD-7 . Arch Internal Med . (2006 ) 166 :1092–7. doi: 10.1001/archinte.166.10.1092 16717171
31Nagin DS Odgers CL .
Group-based trajectory modeling in clinical research . Annu Rev Clin Psychol . (2010 ) 6 :109–38. doi: 10.1146/annurev.clinpsy.121208.131413 20192788
32Nagin DS . Group-Based Modeling of Development . Cambridge, MA:
Harvard University Press (2005 ).
33van der Waerden J Galéra C Larroque B Saurel-Cubizolles MJ Sutter-Dallay AL Melchior M .
Maternal depression trajectories and children's behavior at age 5 years . J Pediatr . (2015 ) 166 :1440 –1448.e1 . doi: 10.1016/j.jpeds.2015.03.002 25866387
34Hayes SC Strosahl KD Wilson KG . Acceptance and Commitment Therapy: The Process and Practice of Mindful Change . New York, NY:
Guilford Press (2011 ).
35Okun ML Mancuso RA Hobel CJ Schetter CD Coussons-Read M .
Poor sleep quality increases symptoms of depression and anxiety in postpartum women . J Behav Med . (2018 ) 41 :703–10. doi: 10.1007/s10865-018-9950-7 30030650
PMC6192841
Come citare questo studio
Gao Y, Sun J, Yang L, Zhang Y, Nan Y, Li X. (2026). Trajectories of perinatal depression, anxiety, perceived stress, sleep quality, and psychological flexibility from pregnancy to six months postpartum and their relationship with the number of abortions and deliveries.. Frontiers in psychiatry. DOI: 10.3389/fpsyt.2026.1973469
Consulta la pubblicazione originale ↗