Childhood trauma and recent suicide risk in major depressive disorder: exploring the mediating effect of resilience and cognitive emotion regulation strategies

Abstract

Background

The mechanisms through which childhood trauma contributes to increased suicide risk in individuals diagnosed with major depressive disorder (MDD) have not been comprehensively explored. This study aimed to investigate the role of resilience and cognitive emotion regulation strategies in the link between childhood trauma and recent suicide risk in MDD patients.

Methods

The study included 136 MDD patients and 112 healthy participants. Standardized instruments were utilized to collect demographic and clinical data, including the 9-item Patient Health Questionnaire (PHQ-9), the 7-item Generalized Anxiety Disorder Scale (GAD-7), the Childhood Trauma Questionnaire-Short Form (CTQ-SF), the Cognitive Emotion Regulation Questionnaire (CERQ), and the Connor-Davidson Resilience Scale (CD-RISC). Structural equation modeling (SEM) was employed, and mediation analysis was conducted.

Results

Significant correlations were observed between recent suicide risk in MDD patients and various factors, including childhood trauma (emotional abuse, emotional neglect; positive correlation), maladaptive strategies (positive correlation), and resilience traits (tenacity, strength, optimism; negative correlations). An ordinal logistic regression identified gender and maladaptive strategies as significant predictors of suicide risk. Most importantly, the bootstrap test showed that the direct effect (β = 0.114, 95% CI: -0.103–0.362) of childhood trauma on recent suicide risk was not significant, but the indirect (β = 0.224, 95% CI: 0.099–0.461) and total effects (β = 0.338, 95% CI: 0.180–0.552) of childhood trauma on recent suicide risk were significant.

Conclusions

Building on cross-sectional data, our path analysis supports a model in which childhood trauma exerts its influence on recent suicide risk in MDD patients indirectly through resilience and emotion regulation, rather than through direct effects. In terms of suicide risk prevention and intervention for MDD patients with childhood trauma, the application of emotion regulation-oriented intervention measures is recommended.

Clinical trial number

Not applicable.

Introduction

Major depressive disorder (MDD), a common mental health condition, significantly disrupts psychosocial functioning and drastically reduces quality of life [1]. Suicide is a frequently occurring symptom of MDD and is generally considered to be the most severe clinical manifestation. According to a recent meta-analysis, an annual prevalence rate of 8% for suicide attempts (SA) and a lifetime prevalence rate of 31% were observed in individuals diagnosed with MDD [2]. Investigating the underlying mechanisms of suicide in MDD patients is critical for recognizing vulnerable populations that require prioritized intervention and preventive measures. Nevertheless, the precise pathways remain inadequately characterized.

Developmental adversity, particularly childhood trauma, has been identified as a robust predictor of SA across MDD populations [3]. Childhood trauma, including physical/emotional abuse and physical/emotional neglect, often has a long-term negative impact on an individual’s mental health [4–6]. O’Connor et al. found that 78.7% of adults with a history of SA and 37.7% of those with suicide ideation reported childhood trauma [7]; Atti et al. showed that childhood sexual, physical, emotional abuse and physical neglect linked to SA [8]; a recent meta-analysis by Angelakis et al. demonstrated that all childhood trauma subtypes were significantly associated with elevated risks of adult suicidal ideation and SA [9]. Many studies have examined the link between childhood trauma and suicide risk in MDD patients, yielding inconsistent findings [10]. Goldberg et al. demonstrated that childhood trauma was also a clear predictor of suicidal behavior in MDD patients [3]; Brodsky et al. found that just physical and sexual abuse were associated with suicidal behaviors [11], while Kim et al. reported that emotional neglect was the only predictor of SA, not other types of childhood trauma [12].

The mechanisms mediating childhood trauma’s impact on suicidality in MDD likely involve complex interactions between a series of risk and protective factors. One such factor is called resilience, which has been broadly defined as the ability of people to adapt to tragedy, trauma, adversity, hardship, and severe stress [13]. From the perspective of positive psychology, resilience is considered as one of the protective factors, which contributes to cushion the negative effects of stressors [14, 15]. Resilience-adjusted models suggest that resilience can prevent suicidal behavior by moderating risk factors associated with suicide in the general population [16]. Recent conceptual and empirical work further clarifies that emotion‑regulation processes are central to resilience: affect‑regulation frameworks integrate stress‑coping and emotion‑regulation approaches to explain why some individuals show resilient trajectories after adversity [17]. Moreover, recent large-sample studies report that higher resilience is associated with lower depressive symptom severity and reduced suicidal ideation across diverse samples, and that resilience‑enhancement programs targeting emotion regulation can reduce symptom burden [18, 19].

As positive cognitive emotion regulation is established as a resilience indicator [20], this capacity may critically modulate suicidality. Cognitive emotion regulation strategies are ways of thinking that enable people to change and manage their emotions in response to specific events [21, 22]. Min et al. demonstrated that within the cognitive emotion regulation strategies, positive reappraisal, refocusing on planning, and rumination reduction enhanced psychological resilience in patients with MDD [23]. Emerging evidence substantiates the association between impaired emotion regulation and increased risk of SA [24]. Contemporary suicide-focused theories implicate specific domains of emotional nonacceptance, limited access to effective emotion regulation strategies, and lack of goal-directed behavior in impacting suicidal symptoms. However, the role of resilience and cognitive emotion regulation strategies in the relationship between childhood trauma and suicide in MDD patients remains unclear.

This research examined the mediating role of resilience and cognitive emotion regulation strategies in the relationship between childhood trauma and recent suicide risk among MDD patients. We hypothesized that: (1) patients with MDD would have experienced more childhood trauma than healthy participants; (2) patients with MDD would show poorer performance in resilience and cognitive emotion regulation than healthy participants; (3) childhood trauma would increase recent suicide risk in MDD patients; (4) resilience and cognitive emotion regulation strategies mediate the relationship between childhood trauma and recent suicide risk in MDD patients.

Methods

Participants

The sample size estimation was performed using G*Power (version 3.1), a statistical analysis tool. We utilized ANOVA with α = 0.05, β = 0.2 (80% power), a medium effect size (f = 0.25). Therefore, the total sample size was calculated to be 128. So, the sample size for each group was at least 64. The sample size of this study meets the minimum sample size requirement.

MDD patient group: 136 patients (male/female = 32/104) with MDD were recruited from the inpatient unit at Suzhou Guangji Hospital. We recruited inpatients rather than outpatients to ensure access to complete clinical records, to permit subjects have sufficient time to complete all clinical evaluations, and to enable immediate safety support if required. All patients met the following inclusion criteria: (1) aged 18–65 years; (2) diagnosed with MDD according to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) by two psychiatrists; (3) six or more years of education. Individuals with histories of psychosis, bipolar spectrum disorders, substance use disorders, or neurocognitive/neurological pathologies were excluded, along with those incapable of providing legally valid informed consent.

Healthy comparison group: 112 healthy participants (male/female = 34/78) were recruited through advertising in local communities in Suzhou. They were healthy volunteers aged between 18 and 65 years, free of mental illness, had six or more years of education, and were able to participate in clinical assessments. All participants including the healthy comparison group received a screening interview to identify any comorbid mental disorders. The screening interviews were mainly based on the subjects’ medical history and psychiatric examination.

The Clinical Research Ethics Committee of Suzhou Guangji Hospital granted ethical approval for this study (Approval No. 2021-017). Prior to participation, a psychiatrist or research coordinator provided a detailed explanation of the study protocol and procedures, after which written informed consent was acquired. All procedures adhered to the ethical standards set by the Institutional Review Board.

Methods

The 9-item Patient Health Questionnaire(PHQ-9)

The PHQ-9 is a 9-item self-report questionnaire commonly employed to screen for depression [25]. Its items align with the DSM-IV criteria for major depressive disorder, rated on a 0 (“not at all”) to 3 (“nearly every day”) scale. The Chinese version of the tool demonstrated excellent internal consistency (Cronbach’s α = 0.86) and robust validity [26].

Item 9 of the PHQ-9, which evaluates suicidal ideation frequency, is frequently utilized in research. It asks, “Over the past two weeks, how often have you been troubled by thoughts of being better off dead or harming yourself in some way?” Responses range from “not at all” to “nearly every day,” with intermediate options of “several days” and “more than half the days.” [27]. Previous studies have shown that responses to this item can identify individuals at elevated risk of suicidal behavior, even within high-risk populations [28]. In this study, we categorized participants based on their Item 9 scores as follows: 0 = low recent suicide risk (LRSR), 1 = moderate recent suicide risk (MRSR), and 2–3 = high recent suicide risk (HRSR).

The construct of “suicide risk” is multifaceted and complex. It encompasses a spectrum of ideation, intensity, duration, controllability, previous attempts, acquired capability for suicide, and behavioral intent. The item 9 of the PHQ-9 captures only a narrow slice of this broad construct. It primarily assesses suicidal ideation. This study is not truly measuring comprehensive suicide risk but is instead measuring the presence of suicidal ideation. Consequently, this measure underrepresents the construct of “suicide risk”.

It is important to note that the term ‘recent suicide risk’ as used in this study is an operational definition based exclusively on the frequency of suicidal ideation (as measured by PHQ-9 item 9) over the past two weeks. While this provides a valuable metric for quantifying severity of ideation in a research context, it does not constitute a comprehensive clinical assessment of suicide risk. A full clinical risk assessment would incorporate additional critical dimensions, such as the presence of a specific plan, intent, or previous suicidal behavior, which are beyond the scope of this measure.

The 7-item generalized anxiety disorder scale (GAD-7)

Spitzer et al. [29] developed the GAD-7, a 7-item self-report scale for screening anxiety over the previous two weeks. Each of the seven items aligns with DSM-IV anxiety disorder criteria and is scored from 0 (“not at all”) to 3 (“nearly every day”), with total scores ranging from 0 to 21. The Chinese version of the GAD-7 demonstrates good internal consistency (Cronbach’s α coefficient = 0.93) and acceptable criterion validity [30].

The Childhood Trauma Questionnaire-Short form (CTQ-SF)

The CTQ-SF [31] is a 28-item self-report measure (comprising 25 clinical and 3 validity items) designed to retrospectively assess childhood trauma. It evaluates five domains—emotional abuse, physical abuse, sexual abuse, emotional neglect, and physical neglect—with each subscale containing five items rated from 1 (“never true”) to 5 (“very often true”). Each subscale contains five items rated on a 1 (never true) to 5 (very often true). The Chinese version of the scale showed good reliability (Cronbach’s α = 0.79) and validity [32].

The cognitive emotion regulation questionnaire (CERQ)

The CERQ [33] is a 36-item self-report scale that assesses cognitive emotion regulation strategies for stressful situations. Using a 5-point Likert scale, it evaluates nine strategy subscales: five adaptive (acceptance, positive refocusing, refocusing on planning, positive reappraisal, and putting into perspective) and four maladaptive (self-blame, rumination, catastrophizing, and blaming others). The Chinese version of the CERQ has demonstrated excellent internal consistency (Cronbach’s α = 0.90) and good structural validity [34].

The Connor-Davidson resilience scale (CD-RISC)

The CD-RISC is a 25-item self-report instrument assessing resilience. The Chinese version reliably assesses tenacity, strength and optimism using a 0–4 Likert scale (0="not true at all” to 4="true nearly all of the time”), yielding total scores of 0-100, where higher values reflect greater resilience [35, 36]. Psychometric evaluation confirmed acceptable internal consistency (Cronbach’s α = 0.75), concurrent validity, and construct validity [37].

Statistical analysis

Statistical analyses were conducted using SPSS software (Version 22). Demographic characteristics and scores from all assessment scales (PHQ-9, GAD-7, CTQ-SF, CERQ, and CD-RISC) were summarized using descriptive statistics. Differences in demographic and clinical characteristics of the MDD patients and healthy participants were analyzed using multivariate analysis of variance (MANOVA) for continuous variables and Chi-square tests for categorical variables. The differences in CTQ-SF, CERQ and CD-RISC scores among MDD patients with different levels of recent suicide risk were compared using MANOVA. When performing multiple comparisons, we applied the Bonferroni correction. Pearson correlation analysis was calculated to examine the relationships among CTQ-SF, CERQ, and CD-RISC scale scores and PHQ-9 item 9 responses in patients with MDD. Ordinal logistic regression analysis was conducted to examine which factors had a significant effect on recent suicide risk in MDD patients. Structural equation modeling (SEM) analyses were performed in AMOS 22.0 to examine the relationships between childhood trauma, cognitive emotion regulation strategies, resilience, and recent suicide risk in MDD patients [38, 39]. The SEM model included childhood trauma as the independent variable and recent suicide risk as the dependent variable, with both cognitive emotion regulation strategies and resilience serving as mediating variables. Statistical significance was set at p < 0.05 (two-tailed).

Results

Demographic and clinical characteristics

There was no statistically significant difference in gender distribution between the groups (p = 0.226). The two groups exhibited statistically significant differences in age (F = 8.458) and education level (F = 27.155) based on MANOVA, along with marital status disparities identified through Chi-square tests (χ² = 5.461), all with p < 0.05. After adjusting for age, education level, and marital status, Table 1 presents the clinical characteristics of MDD patients and healthy participants. Significant between-group differences were observed in nearly all variables, except for acceptance strategies of the CERQ (p > 0.05).

Table 1.

Clinical characteristics of the MDD patients and healthy participants

	MDD group (n = 136)	Comparison group (n = 112)	F or χ2	P
PHQ-9	17.11 ± 6.65	2.26 ± 3.12	428.715	0.000**
GAD-7	12.40 ± 5.73	1.92 ± 2.90	276.733	0.000**
CTQ-SF-Emotional abuse	8.74 ± 4.12	6.17 ± 1.71	19.617	0.000**
CTQ-SF-Physical abuse	6.13 ± 2.24	5.18 ± 0.51	16.434	0.000**
CTQ-SF-Sexual abuse	5.76 ± 2.01	5.17 ± 0.84	6.289	0.013*
CTQ-SF-Emotional neglect	13.43 ± 4.96	9.97 ± 3.81	21.556	0.000**
CTQ-SF-Physical neglect	8.88 ± 3.34	7.78 ± 2.52	4.826	0.029*
CERQ- Self-blame	13.85 ± 2.87	11.68 ± 2.18	30.556	0.000**
CERQ- Acceptance	13.94 ± 2.94	14.23 ± 2.90	1.247	0.265
CERQ- Rumination	13.84 ± 3.62	10.07 ± 3.02	58.222	0.000**
CERQ- Positive refocusing,	10.76 ± 2.87	12.18 ± 2.91	10.092	0.002**
CERQ- Refocusing on planning	13.29 ± 3.62	14.12 ± 2.80	4.626	0.032*
CERQ- Positive reappraisal	11.68 ± 3.17	14.85 ± 2.85	64.117	0.000**
CERQ- Putting into perspective	11.43 ± 2.53	10.54 ± 2.42	5.305	0.022**
CERQ- Catastrophizing	12.58 ± 3.61	7.76 ± 3.10	103.234	0.000**
CERQ- Blaming others	10.75 ± 3.85	9.00 ± 3.09	13.422	0.000**
CERQ- Adaptive strategies	61.10 ± 9.97	65.91 ± 9.45	15.234	0.000**
CERQ- Maladaptive strategies	51.02 ± 9.48	38.49 ± 8.63	91.955	0.000**
CD-RISC- Tenacity	21.03 ± 8.70	32.62 ± 8.59	90.189	0.000**
CD-RISC-Strength	15.36 ± 5.53	23.23 ± 5.73	103.467	0.000**
CD-RISC- Optimism	6.68 ± 3.11	10.22 ± 2.90	67.238	0.000**
CD-RISC-Total	43.07 ± 15.81	66.07 15.91	108.063	0.000**

Note: After adjusting for age, education level, and marital status. MDD = Major Depressive Disorder; PHQ-9 = The 9-item Patient Health Questionnaire; GAD-7 = 7-item Generalized Anxiety Disorder scale; CTQ-SF = Childhood Trauma Questionnaire-Short Form; CERQ = Cognitive Emotion Regulation Questionnaire; CD-RISC = Connor-Davidson Resilience Scale. ^* < 0.05; ^** < 0.01

Comparisons of CTQ-SF, CERQ and CD-RISC scores among MDD patients with different levels of recent suicide risk

Table 2 shows comparisons of measures of depression, anxiety, childhood trauma, cognitive emotion regulation strategies, and resilience among MDD patients with different levels of recent suicide risk. There were significant differences among MDD patients with LRSR, MRSR, and HRSR in terms of depression level (F = 46.379, p < 0.001) and anxiety level (F = 13.115, p < 0.001). On the CTQ-SF, there were significant differences among MDD patients with LRSR, MRSR, and HRSR in terms of emotional abuse (F = 4.918, p = 0.009) and emotional neglect (F = 3.776, p = 0.025). On the CERQ, there were significant differences among MDD patients with LRSR, MRSR, and HRSR in terms of self-blame (F = 4.952, p = 0.008), rumination (F = 6.277, p = 0.002), positive refocusing (F = 4.068, p = 0.019), putting into perspective (F = 5.683, p = 0.004), catastrophizing (F = 7.151, p = 0.001), and total maladaptive strategies scores (F = 10.408, p < 0.001). On the CD-RISC, there were significant differences among MDD patients with LRSR, MRSR, and HRSR in terms of tenacity (F = 8.579, p < 0.001), strength (F = 10.498, p < 0.001), optimism (F = 4.735, p = 0.010) and the total score (F = 10.149, p < 0.001).

Table 2.

Comparisons of CTQ-SF, CERQ and CD-RISC scores among MDD patients with different levels of recent suicide risk

	MDD with LRSR1 (n = 35)	MDD with MRSR 2 (n = 46)	MDD with HRSR3 (n = 55)	F	P
CTQ-SF
Emotional abuse	7.09 ± 2.52	8.72 ± 3.87	9.80 ± 4.79	4.918	0.009**a
Physical abuse	5.71 ± 1.41	6.30 ± 2.28	6.25 ± 2.61	0.825	0.441
Sexual abuse	5.31 ± 0.99	5.76 ± 1.75	6.05 ± 2.59	1.458	0.236
Emotional neglect	12.20 ± 4.50	12.74 ± 4.96	14.80 ± 4.98	3.776	0.025*b
Physical neglect	8.09 ± 3.30	8.98 ± 3.07	9.31 ± 3.54	1.474	0.233
CERQ
Self-blame	12.74 ± 3.11	13.76 ± 2.59	14.64 ± 2.74	4.952	0.008**c
Acceptance	13.31 ± 2.98	13.57 ± 2.54	14.65 ± 3.11	2.877	0.060
Rumination	12.06 ± 3.36	14.26 ± 3.29	14.62 ± 3.72	6.277	0.002**d
Positive refocusing	11.09 ± 2.50	11.48 ± 2.35	9.95 ± 3.29	4.068	0.019*e
Refocusing on planning	13.06 ± 3.29	13.65 ± 3.23	13.15 ± 4.13	0.344	0.710
Positive reappraisal	12.20 ± 2.72	12.00 ± 2.57	11.07 ± 3.78	1.734	0.181
Putting into perspective	10.31 ± 3.06	11.50 ± 2.12	12.09 ± 2.25	5.683	0.004**f
Catastrophizing	10.83 ± 2.99	12.63 ± 3.50	13.65 ± 3.69	7.151	0.001**g
Blaming others	9.86 ± 3.09	10.76 ± 3.59	11.31 ± 4.41	1.535	0.219
Adaptive strategies	59.97 ± 10.77	62.20 ± 8.05	60.91 ± 10.94	0.508	0.603
Maladaptive strategies	45.49 ± 8.40	51.41 ± 8.90	54.22 ± 9.15	10.408	0.000**h
CD-RISC
Tenacity	25.29 ± 7.47	21.48 ± 6.87	17.95 ± 9.64	8.579	0.000**i
Strength	18.06 ± 4.69	16.04 ± 4.93	13.07 ± 5.66	10.498	0.000**j
Optimism	7.57 ± 3.20	7.13 ± 2.93	5.73 ± 3.01	4.735	0.010*k
CD-RISC-Total	50.91 ± 13.20	44.65 ± 13.07	36.75 ± 17.03	10.149	0.000**l

Note: MDD = Major Depressive Disorder; LRSR = Low Recent Suicide Risk; MRSR = Moderate Recent Suicide Risk; HRSR = High Recent Suicide Risk; CTQ-SF = Childhood Trauma Questionnaire-Short Form; CERQ = Cognitive Emotion Regulation Questionnaire; CD-RISC = Connor-Davidson Resilience Scale. ^* < 0.05; ^** < 0.01

All post-hoc analyses were done using Bonferroni: ^a1vs.² p = 0.071, ¹ vs.³ p = 0.002, ² vs.³ p = 0.178; ^b1vs.² p = 0.621, ¹ vs.³ p = 0.015, ² vs.³ p = 0.036; ^c1vs.² p = 0.107, ¹ vs.³ p = 0.002, ² vs.³ p = 0.119; ^d1vs.² p = 0.006, ¹ vs.³ p = 0.001, ² vs.³ p = 0.609; ^e1vs.² p = 0.533, ¹ vs.³ p = 0.062, ² vs.³ p = 0.007; ^f1vs.² p = 0.032, ¹ vs.³ p = 0.001, ² vs.³ p = 0.228; ^g1vs.² p = 0.022, ¹ vs.³ p = 0.000, ² vs.³ p = 0.141; ^h1vs.² p = 0.003, ¹ vs.³ p = 0.000, ² vs.³ p = 0.116; ⁱ¹vs.² p = 0.041, ¹ vs.³ p = 0.000, ² vs.³ p = 0.034; ^j1vs.² p = 0.086, ¹ vs.³ p = 0.000, ² vs.³ p = 0.005; ^k1vs.² p = 0.518, ¹ vs.³ p = 0.006, ² vs.³ p = 0.022; ^l1vs.² p = 0.062, ¹ vs.³ p = 0.000, ² vs.³ p = 0.009

Associations between scores of CTQ-SF, CERQ, CD-RISC and the score of item 9 of PHQ-9 in MDD patients

Table 3 reveals the association between scores of CTQ-SF, CERQ, CD-RISC and the score of item 9 of PHQ-9 in MDD patients. Significant correlations were found between emotional abuse (r = 0.261, p < 0.01), emotional neglect (r = 0.226, p < 0.01), maladaptive strategies (r = 0.374, p < 0.01), tenacity (r = -0.367, p < 0.01), strength (r = -0.398, p < 0.01) and optimism (r = -0.247, p < 0.01) with the score of item 9 of PHQ-9 in MDD patients. After controlling for demographic and clinical covariates (age, gender, education level, marital status, anxiety level, and depression level), there were still significant correlations between tenacity (r = -0.286, p < 0.01), strength (r = -0.295, p < 0.01) and optimism (r = -0.250, p < 0.05) and the score of item 9 of PHQ-9 in MDD patients (see Table 4).

Table 3.

Associations between scores of CTQ-SF, CERQ, CD-RISC and the score of item 9 of PHQ-9 in MDD patients

	item 9 of PHQ-9
CTQ-SF
Emotional abuse	0.261**
Physical abuse	0.003
Sexual abuse	0.147
Emotional neglect	0.226**
Physical neglect	0.120
CERQ
Adaptive strategies	0.022
Maladaptive strategies	0.374**
CD-RISC
Tenacity	-0.367**
Strength	-0.398**
Optimism	-0.247**

Note: PHQ-9 = The 9-item Patient Health Questionnaire; CTQ-SF = Childhood Trauma Questionnaire-Short Form; CERQ = Cognitive Emotion Regulation Questionnaire; CD-RISC = Connor-Davidson Resilience Scale. ^* < 0.05; ^**< 0.01

Table 4.

Associations between scores of CTQ-SF, CERQ, CD-RISC and the score of item 9 of PHQ-9 after adjusting for confounding variables in MDD patients

	item 9 of PHQ-9
CTQ-SF
Emotional abuse	0.044
Physical abuse	-0.015
Sexual abuse	0.058
Emotional neglect	0.176
Physical neglect	0.039
CERQ
Adaptive strategies	-0.077
Maladaptive strategies	0.123
CD-RISC
Tenacity	-0.286**
Strength	-0.295**
Optimism	-0.250*

Note: After adjusting for age, gender, education level, marital status, anxiety level and depression level; PHQ-9 = The 9-item Patient Health Questionnaire; CTQ-SF = Childhood Trauma Questionnaire-Short Form; CERQ = Cognitive Emotion Regulation Questionnaire; CD-RISC = Connor-Davidson Resilience Scale. ^* < 0.05; ^** < 0.01

Ordinal logistic regression analysis was performed to determine predictors of recent suicide risk among individuals diagnosed with MDD. As shown in Table 5, gender (OR = 0.386, 95% CI: 0.160–0.927, p = 0.033) and maladaptive strategies (OR = 1.058, 95% CI: 1.008–1.111, p = 0.033) had significant predictive effects on recent suicide risk in MDD patients.

Table 5.

Risk factors for recent suicide risk in MDD patients

Variable	B	S.E.	Wald	P	OR (95%CI)
Age	-0.008	0.025	0.107	0.743	0.992(0.945–1.042)
Gender	-0.953	0.447	4.536	0.033*	0.386(0.160–0.927)
Education level	0.016	0.061	0.065	0.798	1.016 (0.901–1.145)
Marital status	-0.215	0.598	0.130	0.719	0.806 (0.250–2.605)
Childhood trauma	0.148	0.437	0.114	0.735	1.159 (0.493–2.728)
Adaptive strategies	0.016	0.026	0.377	0.539	1.016 (0.965–1.070)
Maladaptive strategies	0.056	0.025	5.118	0.024*	1.058 (1.008–1.111)
Tenacity	-0.026	0.041	0.412	0.521	0.974 (0.899–1.055)
Strength	-0.070	0.067	1.084	0.298	0.932 (0.818–1.064)
Optimism	-0.077	0.089	0.751	0.386	0.926 (0.778–1.102)

Note: ^* < 0.05; ^** < 0.01

Mediation analysis using structural equation models (SEM)

Figure 1 presents the direct and indirect pathways linking childhood trauma to recent suicide risk in MDD patients, mediated by resilience and cognitive emotion regulation strategies. The indices of goodness-of-fit of the model reflected an acceptable fit, with χ²/df = 2.292, GFI = 0.817, AGFI = 0.753, TLI = 0.743, CFI = 0.787, RMR = 1.590, and RMSEA = 0.098. Table 6 shows that childhood trauma negatively predicts resilience (p < 0.05) and positively predicts maladaptive strategies (p < 0.05). It also shows that maladaptive strategies positively predict recent suicide risk (p < 0.05), while resilience negatively predicts maladaptive strategies (p < 0.05) and positively predicts adaptive strategies (p < 0.05).

Fig. 1.

Model of childhood trauma as a predictor of recent suicide risk in MDD patients, mediated by resilience and cognitive emotion regulation strategies.

Note: EA = emotional abuse; PA = physical abuse; SA = sexual abuse; EN = emotional neglect; PN = physical neglect; ACC = acceptance; REF = positive refocusing; PLA = refocusing on planning; REA = positive reappraisal; PER = putting into perspective; SEL = self-blame; RUM = rumination; CAT = catastrophizing; BLA = blaming others

Table 6.

Relationships between multiple variables in the structural equation model

Path	B	95% CI	P
childhood trauma→recent suicide risk	0.114	-0.065 ~ 0.429	0.311
childhood trauma→resilience	-0.343	-0.552~-0.115	0.012 *
resilience→recent suicide risk	-0.213	-0.443 ~ 0.038	0.095
childhood trauma→adaptive strategies	0.252	-0.012 ~ 0.534	0.063
adaptive strategies→recent suicide risk	-0.089	-0.316 ~ 0.159	0.359
childhood trauma→maladaptive strategies	0.405	0.137 ~ 0.635	0.013*
maladaptive strategies→recent suicide risk	0.312	0.079 ~ 0.525	0.016*
resilience→adaptive strategies	0.577	0.308 ~ 0.745	0.010*
resilience→maladaptive strategies	-0.273	-0.469~-0.086	0.015*

Note: ^* < 0.05; ^** < 0.01

To test the theoretical model, 95% confidence intervals (CI) for mediation effects were estimated with 5000 bootstrap samples. Significance was indicated when the 95% CI did not contain zero. Table 7 shows the results of the standardized effects and 95% CI for direct and indirect associations between childhood trauma and recent suicide risk via bootstrap. The bootstrap test shows that the direct effect (β = 0.114, 95% CI: -0.103 ~ 0.362) of childhood trauma on recent suicide risk was not significant but the indirect (β = 0.224, 95% CI: 0.099 ~ 0.461) and total effects (β = 0.338, 95% CI: 0.180 ~ 0.552) of childhood trauma on recent suicide risk were significant.

Table 7.

The paths and effect analysis

Effect	B	SE	95% CI
Direct effect	0.114	0.120	-0.103 ~ 0.362
Indirect effect	0.224	0.082	0.099 ~ 0.461
Total effect	0.338	0.086	0.180 ~ 0.552

Note: ^* < 0.05; ^** < 0.01

Discussion

As far as we know, this is the first study to investigate the mediating role of resilience and cognitive emotion regulation strategies in the relationship between childhood trauma and recent suicide risk among MDD patients. Our study found that (1) MDD patients experienced more childhood trauma, performed poorer resilience and used less positive and more negative emotion regulation strategies than healthy participants; (2) MDD patients with different levels of recent suicide risk had differences in the characteristics of childhood trauma, resilience, and cognitive emotion regulation strategies adopted; (3) the direct effect of childhood trauma on recent suicide risk was not significant, but the indirect and total effects of childhood trauma on recent suicide risk were significant.

MDD is among the most frequently observed psychiatric conditions in individuals who die by suicide. Approximately 14.5% of MDD patients exhibit at least one suicide attempt over a five-year observation period, according to epidemiological estimates [40]. A recent meta-analysis conducted in China revealed lifetime prevalence of suicidal ideation, suicide plan, and SA in MDD patients to be 53.1%, 17.5%, and 23.7%, respectively [41]. Considering the significant burden of suicide linked to depression, pinpointing potential risk and protective factors and devising targeted preventive and interventive strategies is essential. According to the stress-diathesis model for suicide risk, severe traumatic experiences represent distal risk factor for suicide risk in adults MDD patients [10].The association between childhood trauma and lifetime suicide risk has already been well documented [42]. A recent meta-analysis has reported that individuals with a history of childhood sexual, physical, or emotional abuse show a two- to three-fold higher risk for SA and suicidal ideation than those without such trauma [9]. Our findings revealed that while individuals with MDD exhibited elevated scores across all types of childhood trauma, a significant correlation with recent suicide risk was observed only in terms of emotional abuse and emotional neglect. Consistent with Luciano et al.’s study, suicide risk was independently linked to emotional abuse and neglect, whereas no such association was observed with physical abuse or neglect [10]. The psychological pain theory of suicide suggested that suicidal individuals are experiencing psychological pain or suffering and that suicide may be, at least in part, an attempt to escape from this suffering. Psychache is the hurt, anguish, or ache that takes hold in the mind such as the pain of excessively felt shame, guilt, anxiety, loneliness [43]. Emotional abuse and neglect are potent, direct generators of this type of pain (shame, humiliation, hopelessness), whereas physical abuse might be experienced more as fear or anger. What’s more, emotional abuse can instill feelings of worthlessness (perceived burdensomeness), and emotional neglect can instill feelings of isolation (thwarted belongingness). These feelings are more likely to trigger their motivation for suicide according to the interpersonal-psychological theory of suicide (IPTS) [44]. However, the underlying mechanisms contributing to suicide development in MDD patients with childhood trauma remain unclear [9].

Angelakis et al. supposed that individuals exposed to childhood trauma had a reduced capacity to buffer the effects of life stresses or adverse life events, which may lead to diminished resilience and an increased susceptibility to suicidality [9]. Resilience is a defense mechanism that individuals can develop against adversities [45]. Çapar et al. [46] found that individuals with psychiatric disorders who displayed lower resilience were more susceptible to suicide, lending support to the hypothesis that resilience serves as a protective factor against suicide risk [47]. Among adolescents and young adults, Latha et al. observed that resilience served as a moderator of lifetime violent events and attempted suicide, even when accounting for pre-existing depression [48]. Our findings are consistent with prior research. High psychological resilience has been associated with greater optimism in the face of adversity, which in turn facilitates problem-solving [46].

Cognitive emotion regulation, which refers to ways of thinking people adopt to alter and manage their emotions in response to specific events [24], is considered a potential indicator of resilience. Previous studies found that patients with SA reported greater emotion regulation difficulties compared to those with suicidal ideation [49]. Research suggests that consistent use of maladaptive emotion regulation strategies, such as rumination and self-blame, may increase suicide risk [50–52]. Kees van Heeringen et al. [53] identified that a greater risk of suicide was linked to alterations in brain regions responsible for the cognitive regulation of emotions in adults with depression. Furthermore, emotion dysregulation is recognized as a significant contributor to suicide in individuals exposed to childhood trauma [54, 55]. The trans-diagnostic model linking childhood trauma to psychopathology suggests that emotion regulation difficulties mediate the relationship between childhood trauma exposure and trans-diagnostic psychopathology [56].

Previous studies which have primarily focused on identifying suicide risk factors often overlooked the potential interrelationships between these factors. In reality, multiple risk factors may interact synergistically to increase the likelihood of suicidal ideation and attempts. Our findings showed that childhood trauma had a non-significant direct effect on recent suicide risk, but its indirect and total effects were significant. In our study, childhood trauma did not directly cause recent suicide risk in a straightforward way. In practical terms, this means that not everyone who experiences childhood trauma will necessarily develop suicidal risk. Other factors can influence this link. In our study, its harmful effect was almost entirely indirect—it works by undermining a person’s resilience and promoting the use of maladaptive coping strategies. Childhood trauma had a significant negative effect on resilience. This meant trauma could erode a person’s innate ability to bounce back from adversity, making them more vulnerable. Trauma directly led people to use more maladaptive coping mechanisms but didn’t directly prevent people from using adaptive coping strategies. The use of maladaptive coping strategies was a direct, significant predictor of higher recent suicide risk. Resilience had a significant protective effect, reducing the use of maladaptive strategies and increasing the use of adaptive strategies. These findings were not completely congruent with previous studies. Chen et al. demonstrated that emotional abuse, emotional neglect, and physical neglect were indirectly associated with suicidal ideation through resilience, especially for emotional neglect and physical neglect, as their direct associations with suicidal ideation became insignificant after adjusted for the mediation of resilience [57]. Mohammadzadeh et al. revealed that perception of childhood trauma had no direct relation to suicidal ideation in males who use heroin [54]. Conversely, Wu et al.‘s findings revealed a direct effect of childhood trauma on recent suicidal ideation [39]. A recent meta-analysis further indicated that childhood maltreatment’s association with SA was consistent among individuals with no, common, or severe mental health conditions, suggesting a direct and trans-diagnostic impact of childhood trauma on suicide risk [9]. There are many causes for the inconsistent of results across our and previous studies. In our study, we recruited patients with depression, while many previous studies had not distinguished the diseases that patients suffered from or based on a general population. We used item 9 of the PHQ-9 to assess suicide risk, which only focuses on the frequency of suicidal ideation, while other studies assessed the severity of suicide, suicide plans, SA, and suicide methods. Furthermore, we only examined recent suicide risk during the two weeks preceding the assessment.

Our analysis revealed a significant mediating effect of resilience and cognitive emotion regulation strategies on the childhood trauma–recent suicide risk link in MDD patients, consistent with existing literature. Overall, in our study, the indirect effect value of childhood trauma on recent suicidal risk was relatively small, and the overall effect value was moderate. For every unit increase in childhood trauma, we expected a 0.338 unit increased in suicide risk when considering all its direct and indirect pathways. Crucially, 66% of trauma’s total effect (0.224 / 0.338) was indirect. This highlighted that the majority of the risk was not from the trauma itself, but from the downstream consequences on a person’s psychological resources and coping behaviors. Roy et al. provided evidence that resilience, as a protective factor, attenuates the association between childhood trauma and suicidal behavior risk [58]. Tripp et al. found that resilience is more appropriately viewed as a mediator within the causal relationship between childhood trauma and suicidal behaviors, rather than functioning as a moderator of this association [59]. Chen et al. reported that resilience played a significant mediating role in the relationship between childhood maltreatment and suicidal ideation in Chinese children and adolescents [57]. When examining resilience through its dimensions, emotion regulation was of the most prominent mediators [57]. Mohammadzadeh et al. found that the relationship between the perception of childhood trauma and suicidal ideation is indirectly influenced through the mediating role of specific dimensions of emotion regulation difficulties and cognitive emotion regulation strategies [54]. It has been reported that those who adopted good emotion regulation strategies could handle unpleasant situations better [60]. Maladaptive behaviors, including suicide, may function as a strategy to modulate shame arising from childhood trauma or to cope with negative affect, distressing internal experiences, and reduced distress tolerance [54].

Our model shows that the link between trauma and suicide is not inevitable. By strengthening the protective factors (resilience, adaptive coping), the harmful indirect path can be weakened, offering a clear avenue for effective prevention and treatment. Individuals with a history of childhood trauma should be screened for low resilience and the use of maladaptive coping strategies, as these are the mechanisms that lead to suicide risk. This is a crucial insight for clinicians and interventionists. It suggests that simply addressing the memory of trauma might be less effective than building resilience and teaching positive coping skills. In clinical practice, we should focus on building resilience and teaching adaptive coping strategies rather than just focus on the past trauma. Clinicians should understand resilience is a set of skills that can be cultivated through targeted strategies. They should help MDD patients with childhood trauma develop distress tolerance skills, strengthen resilience resources, shift attention toward positive experiences and build regulatory skills to cope with intense emotional pain. In our study, we found that these MDD patients with high suicide risk were more likely to use maladaptive strategies such as self-blame, rumination, and catastrophizing, and were less likely to use adaptive coping strategies such as positive refocusing and putting into perspective. In clinical interventions, these coping strategies can be prioritized to deal with through cognitive adjustment techniques. Clinicians should help patients use more proactive strategies to cope with trauma and reduce the risk of suicide.

Several limitations should be noted. First, although our model is theoretically grounded and the results are statistically consistent with the proposed mediation, the cross-sectional design remains a critical limitation that prevents causal inference. The relationships observed are associations, and the direction of effects cannot be definitively established. Consequently, longitudinal studies are necessary to establish causality. Second, a primary limitation of this study is the sample size for the MDD group was relatively small relative to the complexity of the SEM analysis. With n = 136, the model may be underpowered, particularly given the number of observed variables and parameters estimated. This means that our findings, especially non-significant results, should be interpreted with caution, as there is an increased risk of Type II errors (failing to detect a true effect). Furthermore, the stability of the parameter estimates may be suboptimal. Third, the interpretation of our findings regarding suicidal risk must be qualified by the limitations of our measurement tool. Our use of a single PHQ-9 item means we have only captured a narrow and non-specific aspect of a much broader clinical phenomenon. The reported associations are therefore likely conservative estimates and may not be generalizable to individuals with active suicidal planning or intent. Future research must employ validated, multi-dimensional instruments designed specifically to assess suicide risk to confirm these findings and explore the full spectrum of suicidality. Finally, it is important to note that the fit indices for the structural equation model (TLI = 0.743; CFI = 0.787; RMSEA = 0.098) fell below the conventional thresholds for excellent fit. This suggests that the proposed model may not fully capture all the complexities of the relationships between the constructs. The less-than-optimal fit could be attributed to the model’s complexity, measurement error in self-report scales, or sample size limitations. While the significant paths provide valuable insights, the results should be interpreted with this limitation in mind. Future studies could further refine the model, for example by incorporating more nuanced latent variables or increasing the sample size, to improve model fit.

Conclusions

In conclusion, our study found that MDD patients reported more childhood trauma, lower resilience, and a pattern of less adaptive emotion regulation compared to healthy participants. Our cross-sectional findings suggest that childhood trauma is associated with recent suicide risk indirectly through its relationships with resilience and emotion regulation, rather than through a direct effect. However, the causal nature of these relationships cannot be definitively established due to the study’s design. Despite this limitation, this study provides important implications for prevention and intervention. The observed associations indicate that interventions focusing on building resilience and improving emotional regulation could be promising strategies for helping to manage suicidal risk in MDD patients with a history of childhood trauma.

Abbreviations

MDD

Major Depressive Disorder

PHQ-9

9-item Patient Health Questionnaire

GAD-7

7-item Generalized Anxiety Disorder Scale

CTQ-SF

Childhood Trauma Questionnaire-Short Form

CERQ

Cognitive Emotion Regulation Questionnaire

CD-RISC

Connor-Davidson Resilience Scale

SEM

Structural Equation Modeling

SA

Suicide Attempts

MANOVA

Multivariate Analysis of Variance

M

Mean

SD

Standard Deviation

OR

Odds Ratio

CI

Confidence Interval

Author contributions

Study Design: GY, XD, HM, QL. Investigation: QL, HM, JL, MC. Analysis and interpretation of data: HM, QL, YZ, ZT.Drafting of the manuscript: QL, HM. Critical revision of the manuscript: GY, XD. All authors reviewed the manuscript.

Funding

The Scientific and Technological Key Program of Suzhou (SYW2024008), Suzhou major diseases clinical multi-center research project (DZXYJ202414), Medical research key project of Jiangsu Provincial Health Commission (K2023015/ MR-32-25-009505).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Before participating in the research, both the participants and their legal guardians signed written informed consent forms. We hereby state that this research was performed in compliance with the ethical standards outlined in the Declaration of Helsinki and fundamental human rights principles. This study was approved by the Clinical Research Ethics Committee of Suzhou Guangji Hospital (Approval No: 2021-017) and was conducted in accordance with local laws, regulations, and institutional guidelines.

Consent for publication

Not application.

Competing interests

The authors declare no competing interests.