Prognostic factors of depression and anxiety in infertile women with cesarean scar defects

Abstract

BACKGROUND

Cesarean scar defect (CSD) is a common complication of the uterine cesarean section that leads to secondary infertility. Women experiencing CSD-associated infertility have a dual psychologic burden: Fertility-related distress compounded by guilt and anxiety stemming from past surgical decisions. However, the prevalence of clinically significant depression and anxiety in this population, compared with that in the general population, and their independent effect on assisted reproductive technology (ART) outcomes, have not been comprehensively measured.

AIM

To analyze the psychological morbidity profile of women with CSD-associated infertility and to determine independent predictors for failure of ART through multivariate regression analysis.

METHODS

A retrospective cohort study was conducted at the General Hospital of the 73^rd Group Army of the People’s Liberation Army (January 2016 to December 2024), enrolling 100 women with CSD confirmed by ultrasound or hysteroscopy and concurrent infertility who underwent assisted reproduction (ART). Baseline psychological status was assessed using the Self-Rating Anxiety Scale (SAS), Self-Rating Depression Scale (SDS), Fertility Quality of Life scale (FertiQoL) and Spousal Support Scale. CSD severity was defined according to residual myometrial thickness (RMT): Mild (≥ 3.5 mm), moderate (2.5-3.4 mm), or severe (< 2.5 mm). ART outcome data, including clinical pregnancy, live birth, and miscarriage, were collected. The type of the nursing intervention (routine vs comprehensive psychological nursing) was included as a covariable in the analyses. Independent prognostic factors for clinical pregnancy failure were identified by univariate and multivariate binary logistic regression. Subgroup analyses were performed based on CSD severity.

RESULTS

At baseline, 74% of participants had clinically significant anxiety (SAS score ≥ 50) and 68% had depression (SDS score ≥ 53); 56% had co-morbid anxiety and depression, a rate nearly doubled that observed in general infertile populations. There was a significant difference in psychological scores between pregnant and non-pregnant patients (SAS score: 54.2 ± 9.6 vs 62.7 ± 10.3, P < 0.001; SDS score: 52.1 ± 9.1 vs 61.4 ± 10.0, P < 0.001). CSD severity was a very strong stratifier of both psychological burden and ART outcomes: Clinical pregnancy rates were 74.3%, 54.1% and 17.9% for mild, moderate, and severe CSD, respectively (P < 0.001). Multivariate logistic regression identified six independent prognostic factors for ART failure: SAS score [adjusted odds ratio (aOR) per one-point increase = 0.949, 95% confidence interval (CI): 0.913-0.987, P = 0.009], FertiQoL score (aOR = 1.060, P = 0.008), RMT (aOR = 1.629, P = 0.002), CSD severity (mild vs severe: aOR = 5.167, P = 0.005), co-morbid anxiety and depression (aOR = 0.306, P = 0.029), and spousal support score (aOR = 1.044, P = 0.040). The model had good discriminatory power (area under the curve = 0.821). Nursing intervention type was significantly associated with the outcome in univariate analysis (odds ratio = 2.732), but this association did not remain independently significant after full covariate adjustment.

CONCLUSION

Depression and anxiety are common among CSD-associated infertility women and are independent prognosticators for ART failure, alongside CSD severity and fertility-related quality of life. These findings support the integration of systematic psychological screening, CSD severity–based risk stratification, and spouse-involved supportive care as essential components of ART management in this population.

Key Words: Cesarean scar defect; Uterine niche; Depression; Anxiety; Infertility; Assisted reproductive technology; Prognostic factors; Fertility Quality of Life scale; Residual myometrial thickness

Core Tip: This retrospective cohort study is the first to systematically characterize the psychological burden of cesarean scar defect (CSD)-associated infertility and quantify its independent prognostic contribution to assisted reproductive technology (ART) outcomes. Among the 100 women included, clinically significant anxiety and depression were detected in 74% and 68%, respectively—rates that nearly double those reported in general infertile populations. Multivariate logistic regression identified six independent prognostic factors for ART failure: Anxiety score, fertility-related quality of life, residual myometrial thickness, CSD severity, co-morbid anxiety and depression, and spousal support. These findings support routine psychological screening and CSD-severity-based risk stratification as essential components of ART management.

INTRODUCTION

Cesarean scar defect (CSD), also known as uterine niche or isthmocele, is a myometrial dehiscence at the prior hysterotomy site resulting from impaired wound healing after cesarean section (CS). With the worldwide CS rate now exceeding 21% and reaching as high as 50% in many high-income countries, transvaginal ultrasound (TVUS)-detected CSD has been reported in 24%-70% of women with a history of CS[1-3]. Among symptomatic patients, the most burdensome long-term complication is secondary infertility, which has an estimated prevalence of 15%-30% among affected women[4,5].

CSD-associated infertility has a multifactorial pathophysiology. Retention of blood and inflammatory secretions within the niche creates a pro-inflammatory endoluminal microenvironment that impairs sperm function, disrupts implantation signaling, and promotes endometritis[1,2]. These structural changes modify uterine contractility patterns, which may interfere with gamete and embryo transport[4,5]. More concerningly, progressive thinning of residual myometrial thickness (RMT) at the defect site correlates with increased risk of uterine rupture, placenta accreta spectrum, and hemodynamic complications in subsequent pregnancies[6,7], underscoring the need for meticulous surgical and reproductive planning. For affected women, assisted reproductive technology (ART) remains the expeditious route to conception; however, clinical pregnancy rates in this population (30%-54%) are consistently lower than those observed in the general infertile population[8-11].

In addition to the biomedical aspects of CSD, a largely neglected psychological burden amplifies reproductive dysfunction. Infertility per se is a serious psychosocial stressor, with anxiety and depression rates of 40%-60% and 25%-35%, respectively, among affected women[12-15]. Women with CSD-associated infertility may experience an additional and qualitatively different psychological burden: The belief that a previous obstetric choice—cesarean delivery—caused or contributed directly to their current reproductive failure brings guilt, self-blame, and anticipatory grief not present in women with primary infertility or non-CSD-related secondary infertility. The extent of this compound psychological state has not been measured systematically to date.

The bidirectional link between psychological distress and ART prognosis is well established. Chronic anxiety and depression trigger hypothalamic-pituitary-adrenal (HPA) axis activation while inhibiting hypothalamic-pituitary-gonadal (HPG) axis activity and gonadotropin-releasing hormone (GnRH) pulsatility[16,17]. During the stress response, elevated cortisol and pro-inflammatory cytokines [interleukin (IL)-1β and tumor necrosis factor-α (TNF–α)] create an endometrial microenvironment that is suboptimal for embryo implantation[16-18]. Dube et al[18] and Lu et al[19] identified significant inverse associations between ART success rates and the severity of pre-treatment anxiety/depression, suggesting that psychological status may represent a modifiable prognostic factor.

However, despite this evidence base, no published study has to date either characterized the prevalence and severity of depression and anxiety in this psychologically distinct and biologically burdened population of women with concurrent CSD and infertility, or quantified their unique contribution—along with CSD-specific anatomical variables—to ART prognosis. What remains challenging, however, is determining the comparative and conjoint prognostic utility of psychological, anatomical, and clinical variables in this cohort, as such insights would inform risk stratification, personalized counseling, and tailored treatment planning.

We therefore sought to: (1) Describe the prevalence and severity of depression and anxiety in women with infertility due to CSD; (2) Compare psychological profiles and ART outcomes according to CSD severity strata; (3) Model independent prognostic factors for ART failure using multivariate logistic regression; and (4) Explore, using subgroup analyses, whether the prognostic contribution of psychological distress differs across CSD severity strata.

MATERIALS AND METHODS

Study design and ethical approval

This retrospective cohort study was conducted at the Army 73rd Group Military Hospital from January 2016 to December 2024. The study was approved by the Institutional Ethics Committee (approval No. 73JYY2026215137); individual informed consent was waived according to national guidelines for retrospective, anonymized studies. All data were de-identified prior to analysis. The study is reported following the STROBE guidelines for observational research.

Participants

The inclusion criteria were: (1) Documented infertility (failure to conceive after ≥ 12 months of regular unprotected intercourse); (2) CSD confirmed by TVUS or hysteroscopy; (3) Age between 20 and 45 years old; (4) Completion of at least one full ART treatment cycle; and (5) Completion of all required baseline psychological assessments. The exclusion criteria were: (1) Other significant fertility-threatening conditions (e.g., bilateral tubal occlusion or severe oligospermia requiring donor sperm); (2) Diagnosed psychiatric disorder prior to study entry (schizophrenia, bipolar disorder, or active psychosis); (3) Treatment interruption for more than one cycle; (4) Incomplete or irretrievable medical records; and (5) Prior participation in another psychological-intervention trial.

CSD classification

To determine CSD severity, TVUS-measured RMT at the defect site was recorded as: Mild (RMT ≥ 3.5 mm), moderate (RMT 2.5-3.4 mm), or severe (RMT < 2.5 mm), consistent with published consensus thresholds[5,7]. TVUS measurements were performed as described by two experienced sonographers who were blinded to clinical outcomes; the intraclass correlation coefficient for RMT measurement between the two sonographers was 0.91.

Data collection and variables

Sociodemographic characteristics [age, body mass index (BMI), years of education, and employment status] and reproductive history (type and duration of infertility, previous ART cycles and CS history) were extracted from chart review. CSD morphological parameters (diameter, depth, and RMT) were obtained from routine ultrasound reports.

At ART initiation, psychological and social assessments were performed before the start of any intervention. All four scales were administered at the first ART consultation visit, prior to ovarian stimulation or any ART procedure; participants were blinded to all ART outcomes during the assessment. The validated instruments used included: (1) Self-Rating Anxiety Scale (SAS): 20 items; score ≥ 50 indicates anxiety, with cut-offs of 50-59 for mild, 60-69 for moderate, and ≥ 70 for severe; (2) Self-Rating Depression Scale (SDS): 20 items; score > 53 indicates depression, with cut-offs of 53-62 for mild, 63-72 for moderate, and ≥ 73 for severe; (3) Fertility Quality of Life scale-Core Module (FertiQoL-Core): This validated 36-item Core FertiQoL module comprises four subscales (emotional, mind-body, relational, and social); the treatment module was not administered, as fewer than half of participants had started active treatment by baseline; higher scores indicate better fertility-related quality of life[15]; domain-specific subscale scores are provided in the Supplementary Table 1; and (4) Spousal Support Scale: Adapted from the Social Support Rating Scale (SSRS), with selected and rephrased items focusing on emotional, instrumental, and informational spousal support within the fertility context; higher scores indicate greater partner support. Internal consistency was acceptable (Cronbach’s α = 0.83) in our sample.

For the purposes of data categorization, the nursing intervention was defined as a structured program consisting of 8 weekly individualized sessions, each lasting 60 minutes. Each session incorporated: (1) Standard cognitive behavioral therapy-based cognitive restructuring of negative appraisals related to infertility; (2) Mindfulness-based stress reduction with guided practice (1 session a week); and (3) Progressive muscle relaxation training, plus one monthly couple counseling session. Session attendance was recorded as a proxy fidelity measure; mean adherence among participants randomized to this arm was 87.5% (≥ 6 of 8 sessions completed). A limitation is that no formal independent assessment of fidelity (e.g., session recording or observer coding) was done, in contrast to standard care nursing—defined as conventional patient education, clinical monitoring, and generic emotional support provided by ward nurses, which served as the existing treatment control.

Primary ART outcomes were clinical pregnancy (gestational sac visible by TVUS at 5-6 weeks), live birth, and miscarriage. Secondary outcomes included total number of oocytes retrieved and number of top-quality embryos.

Statistical analysis

Continuous variables are reported as the mean ± SD; categorical variable as n (%). Independent-samples t-tests (continuous) and χ²/Fisher exact tests (categorical) were used for between-group comparisons. Univariate binary logistic regression was used to screen candidate prognostic variables of clinical pregnancy; variables with P < 0.10 were entered into the multivariate analysis. Independent prognostic factors were identified using multivariate binary logistic regression with backward stepwise elimination (retention threshold P < 0.05). Adjusted odds ratios (aORs) and 95% confidence intervals (CIs) of these multivariable analyses were calculated. Model discrimination was assessed by receiver operating characteristic analysis, and calibration by the Hosmer-Lemeshow test. Collinearity was evaluated using variance inflation factors (threshold < 5). Formal interactions between psychological scores and CSD severity were tested by likelihood-ratio comparisons of models with and without product terms (psychological score × CSD severity). Subgroup analyses were conducted by CSD severity strata. All tests were two-tailed; P < 0.05 was considered to indicate statistical significance. All analyses were conducted using SPSS version 26.0 (IBM, Armonk, NY, United States) and R version 4.2.0.

RESULTS

Baseline characteristics

One hundred women who met the eligibility criteria were enrolled, 51 (51.0%) of whom achieved clinical pregnancy after ART. Table 1 presents baseline demographic, reproductive, CSD morphological, psychological, and ART laboratory parameters. The two groups (pregnant vs non-pregnant) did not differ in age, BMI, education, employment status, ovarian reserve markers [antral follicle counts (AFCs) and anti-müllerian hormone (AMH)], or ART laboratory parameters such as oocytes retrieved and the number of high-quality embryos generated (all P > 0.05). There were significant between-group differences in infertility duration (P = 0.012), previous ART attempts (P = 0.017), CSD diameter (P = 0.046), RMT (P < 0.001), and CSD severity distribution (P < 0.001), as well as all four baseline psychological scale scores (all P < 0.001) and spousal support (P < 0.001).

Table 1 Baseline characteristics of study participants by assisted reproductive technology outcome (pregnant vs non-pregnant), n (%)/mean ± SD.

Variable	Pregnant (n = 51)	Non-pregnant (n = 49)	All (n = 100)	P value
Demographic characteristics
Age (years)	33.8 ± 4.1	35.3 ± 4.6	34.5 ± 4.4	0.079
BMI (kg/m²)	23.3 ± 2.6	23.9 ± 3.1	23.6 ± 2.9	0.290
Years of education	13.8 ± 2.4	13.2 ± 2.7	13.5 ± 2.6	0.248
Employed	38 (74.5)	34 (69.4)	72 (72.0)	0.556
Infertility characteristics
Infertility duration (years)	2.8 ± 1.5	3.7 ± 2.0	3.2 ± 1.8	0.012
Type: Primary	18 (35.3)	22 (44.9)	40 (40.0)	0.317
Prior ART attempts, mean	1.4 ± 0.9	1.9 ± 1.1	1.6 ± 1.0	0.017
CSD characteristics
CSD diameter (mm)	8.1 ± 2.0	9.0 ± 2.5	8.5 ± 2.3	0.046
Residual myometrium (mm)	3.6 ± 1.0	2.8 ± 1.1	3.2 ± 1.1	< 0.001
CSD severity-mild	26 (51.0)	9 (18.4)	35 (35.0)	< 0.001
CSD severity-moderate	19 (37.3)	18 (36.7)	37 (37.0)	0.954
CSD severity-severe	6 (11.8)	22 (44.9)	28 (28.0)	< 0.001
Baseline psychological scores
SAS score	54.2 ± 9.6	62.7 ± 10.3	58.3 ± 10.5	< 0.001
SDS score	52.1 ± 9.1	61.4 ± 10.0	56.6 ± 10.2	< 0.001
FertiQoL score	56.8 ± 8.4	47.9 ± 8.8	52.5 ± 9.3	< 0.001
Spousal support score	78.4 ± 9.2	68.3 ± 11.4	73.5 ± 11.2	< 0.001
ART laboratory parameters
AFC	11.4 ± 3.8	10.8 ± 4.1	11.1 ± 3.9	0.434
AMH (ng/mL)	2.8 ± 1.4	2.5 ± 1.6	2.7 ± 1.5	0.321
Oocytes retrieved (n)	12.1 ± 3.2	11.4 ± 3.7	11.8 ± 3.4	0.297
High-quality embryos (n)	5.6 ± 1.9	5.0 ± 2.2	5.3 ± 2.1	0.145

P < 0.05. BMI: Body mass index; CSD: Cesarean scar defect; ART: Assisted reproductive technology; SAS: Self-Rating Anxiety Scale (≥ 50 = anxiety); SDS: Self-Rating Depression Scale (≥ 53 = depression); FertiQoL: Fertility Quality of Life scale; AFC: Antral follicle count; AMH: Anti-Müllerian hormone.

Prevalence and severity of depression and anxiety

Clinically significant anxiety (SAS ≥ 50) or depression (SDS ≥ 53) was detected in 74 participants (74.0%) and 68 participants (68.0%), respectively. These rates were significantly greater than published estimates for general infertile populations (40%-60% for anxiety; 25%-35% for depression). Fifty-six participants (56.0%) had co-morbid anxiety and depression, a prevalence more than twice that expected in infertile women without CSD-related complications. Table 2 details the severity distributions and their associations with ART outcomes. Across all severity categories of both SAS and SDS, higher psychological morbidity was associated with lower pregnancy rates (all P < 0.05). The proportions of moderate-to-severe anxiety and moderate-to-severe depression were significantly higher in non-pregnant women compared to pregnant ones (P = 0.003 and P = 0.013, respectively).

Table 2 Prevalence and severity distribution of anxiety and depression by assisted reproductive technology outcome, n (%).

Category	All (n = 100)	Pregnant (n = 51)	Non-pregnant (n = 49)	P value
SAS
Normal (< 50)	26 (26.0)	19 (37.3)	7 (14.3)	0.008
Mild anxiety (50-59)	32 (32.0)	20 (39.2)	12 (24.5)	0.107
Moderate anxiety (60-69)	28 (28.0)	10 (19.6)	18 (36.7)	0.054
Severe anxiety (≥ 70)	14 (14.0)	2 (3.9)	12 (24.5)	0.003
Any anxiety (≥ 50)	74 (74.0)	32 (62.7)	42 (85.7)	0.007
SDS
Normal (< 53)	32 (32.0)	24 (47.1)	8 (16.3)	0.001
Mild depression (53-62)	35 (35.0)	19 (37.3)	16 (32.7)	0.624
Moderate depression (63-72)	24 (24.0)	7 (13.7)	17 (34.7)	0.012
Severe depression (≥ 73)	9 (9.0)	1 (2.0)	8 (16.3)	0.013
Any depression (≥ 53)	68 (68.0)	27 (52.9)	41 (83.7)	< 0.001
Co-morbid anxiety + depression
Both SAS ≥ 50 and SDS ≥ 53	56 (56.0)	19 (37.3)	37 (75.5)	< 0.001

P values from χ² test (pregnant vs non-pregnant). Cut-offs: Self-Rating Anxiety Scale (SAS) score ≥ 50 = any anxiety; Self-Rating Depression Scale (SDS) score ≥ 53 = any depression. Co-morbid anxiety + depression = both SAS score ≥ 50 and SDS score ≥ 53 simultaneously. SAS: Self-Rating Anxiety Scale; SDS: Self-Rating Depression Scale.

ART outcomes stratified by CSD severity

The ART outcome rates stratified by CSD severity are presented in Table 3. A clear gradient of decreasing pregnancy rate with increasing severity of CSD was seen: 74.3% (mild), 54.1% (moderate), and 17.9% (severe) (P < 0.001 across strata). Live birth rates showed a similar gradient (62.9%, 40.5%, and 10.7%, respectively). The miscarriage rates did not differ significantly across CSD severity strata; however, the number of miscarriages in the severe CSD subgroup (40.0%) was too small to allow meaningful statistical comparison. These data support CSD severity as a major anatomical determinant of ART prognosis, and provide the stratified framework within which the independent contribution of psychological factors must be interpreted.

Table 3 Assisted reproductive technology outcomes stratified by cesarean scar defect severity.

CSD severity	n	Clinical pregnancy rate (%)	Live birth rate (%)	Miscarriage rate (%)	P value1
Mild (RMT ≥ 3.5 mm)	35	26/35 (74.3)	22/35 (62.9)	4/26 (15.4)	Ref
Moderate (RMT 2.5-3.4 mm)	37	20/37 (54.1)	15/37 (40.5)	5/20 (25.0)	0.046
Severe (RMT < 2.5 mm)	28	5/28 (17.9)	3/28 (10.7)	2/5 (40.0)	< 0.001
Overall	100	51/100 (51.0)	40/100 (40.0)	11/51 (21.6)	-

¹P value vs mild cesarean scar defect group (χ² test).

Clinical pregnancy: Gestational sac confirmed by transvaginal ultrasound at 5-6 weeks. RMT: Residual myometrial thickness; CSD: Cesarean scar defect.

Univariate logistic regression analysis

The univariate logistic regression results for all candidate prognostic variables are provided in Table 4. Determinants highly associated with clinical pregnancy (P < 0.10 entry threshold for multivariate analysis) were: SAS score (OR = 0.935, P < 0.001), SDS score (OR = 0.932, P < 0.001), FertiQoL score (OR = 1.087, P < 0.001), spousal support score (OR = 1.075, P < 0.001), co-morbid anxiety and depression (OR = 0.202, P < 0.001), RMT (OR = 1.861, P < 0.001), CSD diameter (OR = 0.854, P = 0.024), CSD severity (mild vs severe: OR = 11.21, P < 0.001), infertility duration (OR = 0.799, P = 0.017), prior ART attempts (OR = 0.685, P = 0.024), high-quality embryos (OR = 1.208, P = 0.032), and nursing intervention type (comprehensive vs standard: OR = 2.732, P = 0.019). BMI was not significant (P = 0.438).

Table 4 Univariate logistic regression analysis of candidate prognostic factors for clinical pregnancy.

Variable	β	OR (95%CI)	SE	Wald (df = 1)	P value
Psychological variables
SAS score (per 1 point)	-0.067	0.935 (0.904-0.967)	0.016	17.24	< 0.001
SDS score (per 1 point)	-0.071	0.932 (0.901-0.963)	0.016	19.89	< 0.001
FertiQoL score (per 1 point)	0.083	1.087 (1.047-1.128)	0.020	17.61	< 0.001
Spousal support score	0.072	1.075 (1.037-1.114)	0.019	14.23	< 0.001
Co-morbid anxiety & depression	-1.601	0.202 (0.088-0.462)	0.421	14.47	< 0.001
CSD characteristics
Residual myometrium (mm)	0.621	1.861 (1.382-2.506)	0.152	16.73	< 0.001
CSD diameter (mm)	-0.158	0.854 (0.744-0.980)	0.070	5.09	0.024
CSD severity: Mild vs severe	2.417	11.21 (3.72-33.79)	0.565	18.27	< 0.001
Infertility & ART variables
Infertility duration (years)	-0.225	0.799 (0.665-0.959)	0.094	5.74	0.017
Prior ART attempts (n)	-0.378	0.685 (0.494-0.949)	0.167	5.12	0.024
High-quality embryos (n)	0.189	1.208 (1.017-1.435)	0.088	4.62	0.032
Demographic variables
Age (years)	-0.071	0.932 (0.858-1.012)	0.042	2.86	0.091
BMI (kg/m²)	-0.039	0.962 (0.871-1.062)	0.050	0.60	0.438
Nursing intervention (comprehensive vs none)	1.005	2.732 (1.188-6.282)	0.427	5.54	0.019

All variables with P < 0.10 were entered into multivariate analysis. Dependent variable: Clinical pregnancy (1 = pregnant, 0 = non-pregnant). Reference for cesarean scar defect severity: Severe group. BMI: Body mass index; OR: Odds ratio; CI: Confidence interval; CSD: Cesarean scar defect; SAS: Self-Rating Anxiety Scale; SDS: Self-Rating Depression Scale; FertiQoL: Fertility Quality of Life scale; ART: Assisted reproductive technology.

Multivariate logistic regression: Independent prognostic factors

After multivariate backward elimination, six variables remained independently predictive (Table 5). In descending order of effect size, these were: (1) CSD severity (mild vs severe: aOR = 5.167, 95%CI: 1.628-16.40, P = 0.005); (2) RMT per mm increase (aOR = 1.629, 95%CI: 1.192-2.225, P = 0.002); (3) Co-morbid anxiety and depression (aOR = 0.306, 95%CI: 0.106-0.886, P = 0.029); (4) SAS score per point (aOR = 0.949, 95%CI: 0.913-0.987, P = 0.009); (5) FertiQoL score per point (aOR = 1.060, 95%CI: 1.016-1.106, P = 0.008); and (6) Spousal support score per point (aOR = 1.044, 95%CI: 1.003-1.086, P = 0.040). Variables that entered the multivariate analysis but were eliminated by backward selection included infertility duration (P = 0.067), age, BMI, number of previous ART cycles, high-quality embryos (P = 0.097), and nursing intervention type (P = 0.214), none of which reached the retention threshold of P < 0.05. Interestingly, while nursing intervention type was significant on univariate analysis, its effect was attenuated to non-significant after full adjustment for psychological scores and CSD severity, suggesting that its observed univariate association was largely accounted for by these covariates. The final model demonstrated good discrimination (area under the curve = 0.821, 95%CI: 0.744-0.899; Figure 1) and excellent calibration (Hosmer-Lemeshow χ² = 6.84, P = 0.612).

Figure 1 Receiver operating characteristic curve for the multivariate logistic regression model predicting clinical pregnancy. The model included Self-Rating Anxiety Scale score, Fertility Quality of Life Scale score, residual myometrial thickness, cesarean scar defect severity (mild vs severe), co-morbid anxiety and depression, and spousal support score. Area under the curve = 0.821 (95% confidence interval: 0.744-0.899). The diagonal dashed line represents chance-level discrimination. The optimal threshold (sensitivity 78.4%, specificity 73.5%) is indicated by the filled circle. AUC: Area under the curve; CI: Confidence interval.

Table 5 Multivariate binary logistic regression: Independent prognostic factors for clinical pregnancy.

Variable	β	aOR (95%CI)	SE	Wald (df = 1)	P value
Model AUC = 0.821 (95%CI: 0.744-0.899); Hosmer-Lemeshow P = 0.612
SAS score (per 1 point)	-0.052	0.949 (0.913-0.987)	0.020	6.76	0.009
FertiQoL score (per 1 point)	0.058	1.060 (1.016-1.106)	0.022	7.00	0.008
Residual myometrium (mm)	0.488	1.629 (1.192-2.225)	0.160	9.30	0.002
CSD severity: Mild vs severe	1.642	5.167 (1.628-16.40)	0.589	7.77	0.005
Co-morbid anxiety & depression	-1.183	0.306 (0.106-0.886)	0.541	4.78	0.029
Spousal support score	0.043	1.044 (1.003-1.086)	0.021	4.20	0.040

Method: Multivariate binary logistic regression with backward stepwise elimination (retention threshold P < 0.05); only variables retained in the final model are shown. Model area under the curve = 0.821 (95% confidence interval: 0.744-0.899); Hosmer-Lemeshow P = 0.612. Reference for cesarean scar defect severity: Severe group. Variables entering multivariate analysis but eliminated by backward selection (P ≥ 0.05): Age, body mass index, infertility duration (P = 0.067), prior assisted reproductive technology attempts, high-quality embryos (P = 0.097), and nursing intervention type (P = 0.214). AUC: Area under the curve; FertiQoL: Fertility Quality of Life scale; aOR: Adjusted odds ratio; CI: Confidence interval; SAS: Self-Rating Anxiety Scale; CSD: Cesarean scar defect.

Subgroup analysis: Psychological scores and ART outcomes by CSD severity

To examine whether the prognostic contribution of psychological factors differed across CSD severity strata, subgroup analyses were carried out (Table 6). Among non-pregnant women, the pattern was consistent across all severity strata: SAS and SDS scores were higher and FertiQoL scores were lower in this group compared with pregnant women (all P < 0.05). The absolute differences in SAS scores between pregnant and non-pregnant women were 8.4 points (mild CSD), 8.8 points (moderate CSD), and 7.4 points (severe CSD), indicating a clinically comparable magnitude of psychological disadvantage across all anatomical severity levels. Furthermore, a severity-dependent escalation of psychological burden across CSD strata was confirmed by Cochran-Armitage trend tests, which showed a consistent shift toward higher-severity anxiety and depression categories with increasing CSD severity (Figure 2; SAS trend P = 0.003; SDS trend P = 0.001).

Figure 2 Psychological severity categories across cesarean scar defect severity strata (n = 100). A: Self-Rating Anxiety Scale (SAS) score severity distribution (normal/mild/moderate/severe) by cesarean scar defect (CSD) severity; increasing CSD severity was associated with a shift toward higher anxiety categories (SAS, P = 0.003); B: Self-Rating Depression Scale (SDS) score severity distribution by CSD severity; increasing CSD severity was associated with a shift toward higher depression categories (SDS, P = 0.001). CSD: Cesarean scar defect; SAS: Self-Rating Anxiety Scale; SDS: Self-Rating Depression Scale; RMT: Residual myometrial thickness.

Table 6 Subgroup analysis: Psychological scores and assisted reproductive technology outcomes stratified by cesarean scar defect severity and pregnancy outcome, mean ± SD.

Subgroup	SAS	SDS	FertiQoL	Pregnancy rate (%)	P (SAS/SDS)
Mild CSD (RMT ≥ 3.5 mm, n = 35)
Pregnant (n = 26)	52.8 ± 9.0	50.4 ± 8.6	58.9 ± 7.9	74.3	-
Non-pregnant (n = 9)	61.2 ± 9.8	59.1 ± 9.2	49.6 ± 8.4	25.7	0.028
Moderate CSD (RMT 2.5-3.4 mm, n = 37)
Pregnant (n = 20)	55.3 ± 9.8	52.8 ± 9.3	56.4 ± 8.8	54.1	-
Non-pregnant (n = 17)	64.1 ± 10.4	62.3 ± 10.1	46.8 ± 9.1	45.9	0.009
Severe CSD (RMT < 2.5 mm, n = 28)
Pregnant (n = 5)	58.0 ± 11.2	55.6 ± 10.4	53.2 ± 9.0	17.9	-
Non-pregnant (n = 23)	65.4 ± 10.8	65.0 ± 9.8	44.2 ± 9.6	82.1	0.018
P value (across CSD severity groups, non-pregnant only)
Between-group (non-pregnant)	0.048	0.039	0.044	-	One-way ANOVA

Values shown are the mean ± SD (psychological scores) or percentage (pregnancy rates). P values from independent-samples t-test [pregnant vs non-pregnant within each cesarean scar defect (CSD) stratum]. P values for between-group comparison across CSD severity strata (non-pregnant only) by one-way ANOVA. RMT: Residual myometrial thickness; SAS: Self-Rating Anxiety Scale; SDS: Self-Rating Depression Scale; FertiQoL: Fertility Quality of Life; CSD: Cesarean scar defect.

DISCUSSION

Formal likelihood-ratio interaction tests for the product term (psychological score × CSD severity) were conducted within the multivariate logistic regression framework; no interaction term reached statistical significance (all P > 0.10), indicating that the prognostic effect of psychological distress did not differ significantly across CSD severity strata. Accordingly, these findings are presented as descriptive subgroup explorations rather than formal interaction analyses. Importantly, the severe CSD stratum (n = 28; only 5 of whom became pregnant) was critically underpowered for subgroup regression, with an events-per-variable ratio well below the recommended minimum of 10, rendering stratum-specific effect estimates statistically unreliable. Descriptive comparisons within this stratum are reported for completeness but should be interpreted with particular caution. Nonetheless, a consistent pattern of psychological disadvantage was observed across all three severity strata, with clinically significant psychological distress associated with failure to achieve clinical pregnancy even among women with mild CSD, in whom anatomical barriers to implantation are modest. This cross-stratum pattern supports the independent prognostic relevance of psychological status across the full spectrum of CSD severity, while acknowledging that definitive subgroup-specific conclusions—particularly for the severe stratum—await adequately powered future studies.

To our knowledge, this study offers the first systematic assessment of depression and anxiety prevalence, as well as their independent prognostic contribution to ART outcomes, in women with CSD-associated infertility. Three main findings extend the existing literature.

The 74% prevalence of clinically significant anxiety and 68% prevalence of depression observed in this cohort far exceeds the 40%-60%[13,14] and 25%-35%[20-22] figures reported in systematic reviews of general infertile populations. The 56% co-morbidity between anxiety and depression is especially remarkable and clinically relevant: Co-morbid presentations are associated with greater severity, poorer treatment response, and more pronounced neuroendocrine dysregulation than either disorder alone[16,19]. The pathways underlying this increased burden seem to be additive. Standard infertility-related distress—injury to self-concept, social stigma, marital tension, and existential uncertainty regarding reproductive futurity—is augmented in CSD women by CSD-specific psychological content: Guilt due to perceived past decisional error (the choice of CS), body-image disturbance linked to an anatomically impaired uterus, chronic physical symptoms (abnormal uterine bleeding), and heightened anxiety about rupture of the uterus in future pregnancies[23]. One grounded theory study exploring women’s lived experience with ART established self-reconstruction under infertility and treatment burden as the core psychosocial process, highlighting the need for narrative and identity-informed therapeutic approaches[23]. Formal characterization of these CSD-specific psychological themes will help guide targeted therapeutic content in future qualitative research[23,24].

After adjusting for CSD anatomical severity and all other covariables, independent prognostic determinants of ART failure in the multivariate model included SAS score, co-morbid anxiety and depression, FertiQoL score, and spousal support. These findings have mechanistic coherence. Both anxiety and depression activate the HPA axis and inhibit the function of the HPG axis: Increased cortisol reduces GnRH pulsatility, lowers bioavailability of gonadotropins, and impairs both follicular maturation and endometrial receptivity[16,17]. At the stress-immune interface, pro-inflammatory cytokines (IL-1β and TNF–α) are released, which disrupt the biochemical cascade of embryo implantation and block leukemia inhibitory factor and IL-11 signaling[16-18]. Behavioural pathways, such as reduced treatment adherence, poorer lifestyle behaviours (e.g., sleep disruption, poor nutrition, and physical inactivity), and withdrawal from medical engagement, further compound the biological impact of psychological distress on ART outcomes[24-27].

Importantly, the attenuation of the nursing intervention variable to non-significance after full adjustment may suggest that its observed univariate benefit operates, at least in part, through improvements in anxiety, fertility-related quality of life, and spousal support—the psychological and relational constructs measured by these scales. However, given the non-randomized assignment of nursing intervention type and the wide CI around the adjusted estimate, this interpretation should be regarded as hypothesis-generating only. Randomized trials with adequate power and fidelity monitoring are needed to formally test whether structured psychological nursing confers prognostic benefit independent of its effects on measured psychological scores. This finding supports the clinical necessity of identifying and treating psychological morbidity, regardless of whether formal structured nursing programmes are available[27,28]. Most importantly, psychosocial distress is not limited to female partners: Male anxiety and depression are also independently predictive of poorer IVF outcomes, and couple-level psychological screening has been proposed as a more accurate prognostic tool than female-only assessment[29].

In the multivariate model, CSD severity (mild vs severe: AOR = 5.167) and RMT (aOR = 1.629 per mm) were two of the strongest independent predictors. The gradient in pregnancy rates across severity strata (74.3% to 17.9%) is consistent with previous cohort studies[30-33] and mirrors the cumulative negative impact of severe CSD on endometrial receptivity (chronic inflammation and fluid accumulation), implantation mechanics (gravity defect), and maintenance of pregnancy (risk for rupture through thin RMT). CSD morphologic features—such as niche depth, width, and RMT—are independently linked to pregnancy outcomes in women with prior cesarean delivery[34]. The association between CSD and secondary infertility, as well as reduced ART outcomes, has been further verified by systematic reviews of the uterine niche, whilst surgical repair could optimally restore fertility potential in selected patients[35]. This association has also been corroborated by case series and mechanistic reviews demonstrating that chronic intrauterine inflammation, exacerbated by abnormal/invasive cervical mucus, leads to impaired reproductive function and failed pregnancies[36]. Against this anatomical background, ART surveillance data suggest that overall clinical pregnancy rates achieved after IVF remain heterogeneous cross centers and patient populations[37], highlighting the need for a better understanding of modifiable prognostic markers in high-risk populations, including women with CSDs. Importantly, our subgroup analyses showed that psychological distress had a significant and consistent prognostic effect within each of the CSD severity stratum, including among women with mild CSD, in whom anatomical barriers are modest. This cross-stratum consistency argues against the notion that psychological factors are relevant only when anatomy is relatively preserved, and instead supports an independent prognostic role for psychological status across the full spectrum of anatomical disease severity.

The spousal support score was an independent predictor of clinical pregnancy (aOR = 1.044 per point, P = 0.040) and differed significantly between pregnant and non-pregnant women (78.4 ± 9.2 vs 68.3 ± 11.4, P < 0.001). This finding is consistent with the broader infertility literature, in which partner support buffers psychological distress, increases treatment adherence, and may positively modulate neuroendocrine function via psychosocial pathways[15,20,38,39]. For patients with CSD, in whom guilt and self-blame may be uniquely salient, spousal understanding and non-judgmental support may carry particular value. Within a broader ART support paradigm, routine spousal support assessment, coupled with couple-centered psychoeducational interventions where indicated, should be implemented[28].

This study has several limitations. First, the retrospective, nonrandomized design may introduce selection bias and restrict causal inference. Second, data were collected up to the index ART cycle but did not capture outcomes from subsequent cycles (e.g., cumulative live birth rates), limiting conclusions about long-term reproductive prognosis. Third, psychological assessment was completed at a single pre-treatment baseline and therefore did not capture intra-treatment fluctuations in distress that might have an independent impact on outcomes during any of three distinct procedural phases (ovarian stimulation, oocyte retrieval, or embryo transfer). Future studies should address this gap using longitudinal psychological assessment designs. Fourth, although the SAS and SDS have been extensively validated as self-report tools, they carry the risk of social desirability and recall bias, which may lead to underreporting of psychological involvement in this clinical context. Future prospective work should therefore consider structured clinical interviews (e.g., the MINI International Neuropsychiatric Interview or SCID-5) as an alternative.

Because the type of nursing intervention was not randomized, one cannot ascribe the observed univariate nursing effect to the intervention itself rather than to preexisting patient characteristics. The severe CSD subgroup was critically underpowered: Only 5 of 28 patients (17.9%) achieved clinical pregnancy, yielding fewer than 10 events in this stratum. With an events-per-variable ratio far below the recommended minimum of 10, formal subgroup regression analyses in this stratum lack sufficient statistical power and are prone to unstable effect estimates. Accordingly, the severe CSD subgroup findings in Table 6 must be interpreted as exploratory and hypothesis-generating only; no definitive subgroup-specific conclusions regarding independent prognostic factors can be drawn from this stratum. Future multicenter studies recruiting larger severe CSD cohorts are required to validate these observations.

The single-centre design may limit the generalisability of our findings to other institutions, where patient demographics, ART protocols, and supportive clinical resources may differ. Although ovarian reserve markers were evaluated (AFC and AMH) were evaluated and did not differ between groups, their interpretation has well-recognised limitations: Normative reference ranges for AMH and AFC vary by assay platforms[40] and patient populations, and diminished ovarian reserve—even when present—is itself associated with adverse reproductive outcomes, including recurrent pregnancy loss, representing a potential confounder we could not completely fully account for given the retrospective design[41]. Biological mediators of the psychological-ART outcome relationship (cortisol, inflammatory cytokines, and natural killer cell activity) were not measured. The FertiQoL treatment module was not administered, meaning that quality of life during treatment phase was not assessed.

CONCLUSION

The findings of this study support risk stratification at ART initiation that integrates both anatomical and psychological domains. We suggest the systematic use of validated instruments for psychological screening in all affected women. Future multicenter RCTs are recommended to investigate whether targeted psychological interventions—specifically addressing the modifiable and potentially treatable stressors unique to CSD, particularly guilt over previous cesarean decisions and the body-image distress associated with a standing postoperative scar—can improve both mental health and ART prognosis.