Cumulative effects of stress-sensitivity factors on depressive symptoms and suicide risk: A prospective study

Abstract

BACKGROUND

Sensitivity to stress is essential in the onset, clinical symptoms, course, and prognosis of major depressive disorder (MDD). Meanwhile, it was unclear how variously classified but connected stress-sensitivity variables affect MDD. We hypothesize that high-level trait- and state-related stress-sensitivity factors may have different cumulative effects on the clinical symptoms and follow-up outcomes of MDD.

AIM

To investigate how stress-sensitivity factors added up and affected MDD clinical symptoms and follow-up results.

METHODS

In this prospective study, 281 MDD patients were enrolled from a tertiary care setting. High-level stress-sensitivity factors were classified as trait anxiety, state anxiety, perceived stress, and neuroticism, with a total score in the top quartile of the research cohort. The cumulative effects of stress-sensitivity factors on cognitive dysfunction, disability and functional impairment, suicide risk, and depressive and anxiety symptoms were examined using an analysis of variance with linear trend analysis. Correlations were investigated further using multiple regression analysis.

RESULTS

Regarding high-level stress-sensitivity factors, 53.40% of patients had at least one at baseline, and 29.61% had two or more. Four high-level stress-sensitivity components had significant cumulative impacts on MDD symptoms at baseline (all P < 0.001). Perceived stress predicted the greatest effect sizes of state-related factors on depressive symptoms (partial η² = 0.153; standardized β = 0.195; P < 0.05). The follow-up outcomes were significantly impacted only by the high-level trait-related components, mainly when it came to depressive symptoms and suicide risk, which were predicted by trait anxiety and neuroticism, respectively (partial η² = 0.204 and 0.156; standardized β = 0.247 and 0.392; P < 0.05).

CONCLUSION

To enhance outcomes of MDD and lower the suicide risk, screening for stress-sensitivity factors and considering multifaceted measures, mainly focusing on trait-related ones, should be addressed clinically.

Key Words: Major depressive disorder; Stress sensitivity; Depressive symptoms; Suicide risk; Cumulative effect

Core Tip: How variously classified but related stress-sensitivity factors affected major depressive disorder (MDD) was unknown. This prospective study looked at the relationship between stress-sensitivity variables and MDD in a novel way. We found that in MDD patients with high-level stress-sensitivity components, baseline symptoms were worse; state factors, especially perceived stress, had a greater influence on depressive symptoms. Patients with high trait factors at follow-up performed worse, especially in trait anxiety- and neuroticism-influenced depressive symptoms and suicide risk, respectively. Screening for stress-sensitivity factors-especially trait ones-should be prioritized to improve disease prognosis.

INTRODUCTION

Major depressive disorder: Prevalence, symptoms, and the association with stress

Major depressive disorder (MDD) is one of the most common psychiatric disorders around the world, with a lifetime prevalence of 3.4% in mainland China[1]. MDD is characterized by multidimensional symptoms, including sadness, anhedonia, loss of interest and motivation, social withdrawal, suicide ideation, and impaired cognitive function[2-5], leading to disability and functional impairment[6]. Numerous studies have demonstrated significant associations between stress and the onset, clinical manifestations, and follow-up outcomes of MDD[7-10].

Stress sensitivity: Definition, factors, and state-dependent or trait-like characteristics

Individual stress response is essential for adaptation, maintenance of homeostasis, and survival[11]. Therefore, studies have increasingly focused on the relationship between MDD and the sensitivity to stress, which suggests heightened stress sensitivity in patients with MDD[12]. In studies on depression, the term stress sensitivity, or stress reactivity, refers more specifically to the predisposition resulting from vulnerability mechanisms that are acquired or innate to react to lower levels of stress[13]. Heightened stress sensitivity in patients with MDD may be attributed to trait-related factors (e.g., personalities) and acquired contributors (e.g., stressful life events). Many studies have examined altered stress sensitivity in MDD using two strategies, i.e., investigating mood responses to negative and positive events and examining mood responses to minor stressors using ecological momentary assessment techniques such as the experience sampling method[14]. For instance, a study using the experience sampling method found that compared to healthy controls, patients with MDD exhibited significantly higher negative and lower positive affect in subjective appraisals of stress and daily life[15]. However, such methods mainly focus on subjects’ immediate response to stressors, which are also viewed as state-dependent stress sensitivity. Thus, the putative trait-like characteristics of stress sensitivity were overlooked in such studies[16].

Previous studies have used a variety of phenotypes (e.g., anxiety, nervousness, fear, tension, and negative emotionality/temperament) as indicators of latent stress sensitivity with the help of relevant psychological measurement tools[17,18]. Therefore, in the present study, we included some psychopathological factors (i.e., neuroticism, trait anxiety, perceived stress, and state anxiety) that have been separately found to be representative of stress sensitivity and related to the risk[12], clinical manifestations[19] and follow-up outcomes[20] of MDD. We also focused on trait-related and state-related stress sensitivity aspects by categorizing these factors accordingly. Among the trait-related factors, neuroticism is commonly regarded as a cognitive-emotional tendency to frequently experience stress and uncontrollability due to a lower level of personal effectiveness[21]. Individuals with high levels of neuroticism tend to exhibit heightened psychological responses to stress[22]. Trait anxiety, which refers to the tendency to heightened anxiety responses, has been demonstrated to be associated with the activation of brain regions related to depression under increased stress[23]. Concerning state-related factors, perceived stress, which refers to the degree to which situations are appraised as stressful, has been reported to positively correlate with the interoceptive awareness of stress[24]. State anxiety, which refers to the anxiety state involving feelings of tension and apprehension and heightened autonomic nervous system activity, is related to physiological stress responses[25].

Aims and hypothesis of the present study

To date, little has been known about how differently categorized yet interrelated stress-sensitivity factors influence the symptoms and follow-up outcomes of MDD and whether their effects are different. Similar research questions have been explored by investigating the cumulative impact of multiple factors (including psychological ones) in complex physical diseases with multifactorial etiology[26].

The present study seeks to examine the cumulative effects of stress-sensitivity factors (neuroticism, trait anxiety, perceived stress, and state anxiety) on the clinical presentation and follow-up outcomes of MDD. These outcomes include depressive and anxiety symptoms, suicide risk, disability and functional impairment, and cognitive dysfunction. We propose the following hypotheses: (1) Elevated stress-sensitivity factors will exhibit cumulative effects on the clinical symptoms of MDD at baseline; (2) Elevated stress-sensitivity factors will also exert cumulative effects on follow-up outcomes of MDD; and (3) Trait- and state-related stress-sensitivity factors may influence different symptoms of MDD differently, both at baseline and during follow-up. This study aims to enhance our understanding of the relationship between stress sensitivity and MDD, offering valuable insights for improving psychiatric primary care interventions.

MATERIALS AND METHODS

Setting

This study was performed in the inpatient and outpatient psychiatric departments of the Second Xiangya Hospital of Central South University. This hospital, located in central China, was established in 1958. It is a Class-A tertiary comprehensive hospital with the National Clinical Research Center for Mental Disorders and the National Center for Mental Disorders. This hospital’s psychiatric specialty is one of the top three in China. As a result, the hospital sees patients from all over China, which is representative and vast and helps the study be more broadly applicable.

Design and participants

This prospective study recruited 281 patients with MDD from December 2019 to March 2023 based on the principle of convenience sampling. Patients meeting the following criteria were included: (1) Aged between 18 and 60 years; (2) Having current or previous clinical manifestations in line with a major depressive episode according to the 5^th version of the Diagnostic and Statistical Manual for Psychiatric Disorders (DSM-5); (3) With a total score of > 7 on the 17-item Hamilton Depression Rating Scale (HAMD-17)[27]; and (4) Willing to receive antidepressants prescribed by their psychiatrists for two months. Patients would be excluded if they met any of the following criteria: (1) With comorbid psychiatric disorders; (2) Currently pregnant or breastfeeding; and (3) With prior or present major medical conditions such as immune system diseases, endocrine and metabolic diseases, nervous system diseases, and a history of brain injury. Finally, a total of 206 participants were enrolled through the screening.

Upon enrollment, the participants were assessed by two trained professional psychiatrists independently based on the Structured Clinical Interview for DSM-5. Then, the participants were assessed using the HAMD-17 and the Hamilton Anxiety Rating Scale (HAMA) and were asked to complete a self-designed sociodemographic questionnaire and a variety of self-rating scales. After baseline assessments, all participants received first-line antidepressant therapy prescribed by their psychiatrists. Adjuvant therapies such as hypnotics and sedatives were used at the psychiatrists’ discretion. The psychiatrists regularly reviewed and adjusted the treatment, and patients were assessed on-site for clinical outcomes at the end of the 8-week follow-up. In total, 119 patients with MDD eventually finished the follow-up, except a few who were lost or dislodged because they switched to bipolar disorder, moved away from their initial residence, etc.

Ethical considerations

This study was approved by the Ethics Committee of the National Clinical Medical Research Center, the Second Xiangya Hospital of Central South University (MDD202011). All participants provided written informed consent before enrollment.

Measures

Clinical symptoms and follow-up outcomes: Symptoms of depression and anxiety were assessed separately by HAMD-17 and HAMA. In the HAMD-17[28], 10 items were rated on a Likert scale (0-4), and others were rated on a Likert scale (0–2). The total score of HAMD-17 ranges from 0 to 54, with a score of less than 8 indicating no depression, 8-17 indicating mild depression, 18-24 indicating moderate depression, and ≥ 25 indicating severe depression. MDD patients with follow-up data were divided into a group with clinical remission (HAMD-17 < 8) and a group without clinical remission (HAMD-17 > 7) after eight weeks of treatment. The items of HAMA were rated on a Likert scale (0-4). The total score of HAMA ranges from 0-56, with a score of < 18 indicating mild anxiety, 18-24 indicating mild to moderate anxiety, and 25-30 indicating moderate to severe anxiety[29].

The Beck Scale for Suicide Ideation (BSSI) was used to assess the suicide risk of the participants[30]. The BSSI includes 19 items rated on a Likert scale (0-2). The total score ranges from 0-38, with higher scores indicating higher suicide risk.

The self-rated Sheehan Disability Scale (SDS) assessed functional impairment and disability in three major life domains: Work/studies, social life/Leisure activities, and family life/home responsibilities. In this scale, the items are visually presented as a horizontal line marked with both numbers (0 to 10) and verbal anchors (0 = not at all, 1–3 = mildly, 4–6 = moderately, 7–9 = markedly, and 10 = extremely). The total score ranges from 0 to 30, with higher scores indicating more severe functional impairment[31].

The Perceived Deficits Questionnaire five-item version (PDQ-5) was used to measure cognitive dysfunction. Items are rated on a Likert scale (0-4), and their total score ranges from 0 to 20, with higher scores indicating more severe cognitive dysfunction[32].

Stress-sensitivity factors: Trait-related stress-sensitivity factors included neuroticism and trait anxiety: (1) Neuroticism: 12 items from the NEO Five-Factor Inventory were used to assess neuroticism. The total score ranges from 12 to 60, with higher scores indicating a higher level of neuroticism[33]; and (2) Trait anxiety: Trait anxiety was assessed using the 20-item Trait Anxiety Inventory (T-AI) from the State-Trait Anxiety Inventory, with higher scores indicating greater trait anxiety[34].

State-related stress-sensitivity factors included state anxiety and perceived stress: (1) State anxiety: State anxiety was assessed using the 20-item State Anxiety Inventory (S-AI) from the State-Trait Anxiety Inventory[34], with higher scores indicating greater state anxiety; and (2) Perceived stress: The 10-item Perceived Stress Scale (PSS) was used, with higher total scores indicating higher levels of perceived stress[32].

In the present study, patients with high-level stress-sensitivity factors were defined as those with scores in the highest quartile of the research cohort on each scale (i.e., Neuroticism score ≥ 48, T-AI score ≥ 65, S-AI score ≥ 62, and PSS score ≥ 28).

Statistical analyses

A priori analyses were performed using the G*Power 3.1 software during the study's design phase to determine the necessary sample size. A two-tailed test with a test level α of 0.05, a test power (1-β) of 0.9, and the number of predictors were set in G*Power 3.1 to ensure that an effect size of 0.3 (medium) was obtained, as linear multiple regression analysis is necessary to analyze the correlation of high-level stress-sensitivity factors and clinical outcomes. This resulted in a total sample size of 57 for the study. Participants in the top quartile of the research cohort on each scale (Neuroticism, T-AI, S-AI, and PSS) had high stress-sensitivity factors. It was intended to recruit 326 MDD patients for the assessment, assuming a 30% dropout rate. Regretfully, the novel coronavirus pneumonia outbreak hindered subject enrollment, and only 281 depressed individuals were eventually enrolled. Ultimately, 206 people were enrolled in the baseline study, and 119 participants were included in the follow-up analysis. A two-tailed test with a test level 0.05 and an effect size of 0.3 (medium) was used for post hoc analysis in G*Power 3.1. This investigation's test power (1-β) was determined to be 0.99.

The independent-sample t-test for quantitative variables or the χ² test for qualitative variables was used to analyze differences in demographic or clinical characteristics in the overall MDD group and the follow-up subgroup, patients with and without follow-up data, and patients with and without clinical remission as appropriate.

A multistep procedure was utilized to investigate the cumulative effects of stress-sensitivity factors[26]. The initial step involved identifying participants with high levels of stress-sensitivity factors, defined as those falling within the highest quartile of the research cohort on each scale (Neuroticism, T-AI, S-AI, and PSS). Subsequently, the association between each stress-sensitivity factor and one or more manifestations of MDD in the hypothesized direction was examined using independent-sample t-tests. The false discovery rate was controlled to account for multiple comparisons, with a significance threshold set at corrected (P < 0.05). Additionally, collinearity among variables was assessed using Pearson’s correlation coefficient (r > 0.8) or the variance inflation factor > 5[27]. Taking the factors that pass the above tests into the One-way analysis of variance with linear trend analysis in the total MDD group at baseline and the subgroup at the end of the follow-up period, with the number of all high-level stress-sensitivity factors at baseline as the independent variable and only the number of high-level trait-related factors at the follow-up. The cumulative effect sizes were determined as small (0.01-0.05), medium (0.06-0.13), and large (≥ 0.14) based on the Partial η²[35]. Finally, linear multivariable regression analysis was used to the clinical symptoms and follow-up outcomes with particularly prominent cumulative effects of stress-sensitivity factors. The data analyses were conducted using SPSS Version 20 and R 4.3.2, with statistical significance set at P value < 0.05 (two-tailed).

RESULTS

Clinical and demographic characteristics

A total of 206 patients with MDD were included in the analyses, and 119 (57.77%) completed the assessment at the end of the 8-week follow-up period. As shown in Table 1, no significant differences were observed in demographics and treatment regimens between the overall MDD group at baseline and the follow-up subgroup. Demographic and clinical information between patients with and without follow-up data also had no difference mostly (Supplementary Table 1). Most of the participants were females (71.8%) and unmarried (75.7%). After 8 weeks of antidepressant treatment, the participants demonstrated significant improvement in the total score of HAMD-17, HAMA, BSSI, SDS, and PDQ-5 (all P < 0.01).

Table 1 Demographic and clinical information of patients with major depressive disorders.

Items	Overall MDD group (n = 206)	Subgroup based on follow-up data (n = 119)	Sig.
Age (years)	23 (10)	22 (9)	0.132
Years of education	14.52 ± 2.55	14.33 ± 2.66	0.530
BMI (kg/m2)	22.02 ± 4.34	22.43 ± 4.78	0.424
Gender (female), No. (%)	148 (71.8)	87 (73.1)	0.806
Marriage (unmarried), No. (%)	156 (75.7)	94 (79.0)	0.713
Ethnicity (Han), No. (%)	187 (90.8)	109 (91.6)	0.803
History of drinking, No. (%)	39 (18.9)	22 (18.5)	0.983
History of smoking, No. (%)	31 (15.0)	17 (14.3)	0.947
Depressive episodes	2 (1)	2 (1)	0.493
HAMD-17 score at baseline	21.24 ± 5.28	21.94 ± 5.05	0.557
HAMD-17 score at the end of follow-up1	NA	10.88 ± 7.12	< 0.001a
HAMA score at baseline	21.38 ± 7.54	21.51 ± 7.58	0.883
HAMA score at the end of follow-up1	NA	11.69 ± 7.74	< 0.001a
BSSI score at baseline	8.07 ± 7.68	7.57 ± 7.52	0.568
BSSI score at the end of follow-up1 (n = 66)	NA	5.67 ± 7.10	0.004a
SDS score at baseline	18.50 ± 5.96	19.01 ± 6.07	0.466
SDS score at the end of follow-up1	NA	11.91 ± 7.57	< 0.001a
PDQ-5 score at baseline	11.88 ± 3.54	11.69 ± 3.46	0.630
PDQ-5 score at the end of follow-up1	NA	8.27 ± 4.58	< 0.001a
Antidepressants at baseline
SSRIs, No. (%)	66 (32)	43 (36.1)	0.451
SNRIs, No. (%)	20 (9.7)	12 (10.1)	0.913
SARIs, No. (%)	98 (47.6)	62 (52.1)	0.432
Others2, No. (%)	14 (6.8)	7 (5.9)	0.747
Antidepressants at the end of follow-up
SSRIs, No. (%)	NA	67 (32.5)	NA
SNRIs, No. (%)	NA	18 (8.7)	NA
SARIs, No. (%)	NA	86 (41.7)	NA
Others2, No. (%)	NA	21 (17.6)	NA
Neuroticism score at baseline	44.92 ± 5.97	45.05 ± 5.70	0.919
T-AI score at baseline	59.98 ± 7.06	60.15 ± 6.80	0.824
PSS score at baseline	24.76 ± 4.96	24.83 ± 5.28	0.899
S-AI score at baseline	56.64 ± 8.67	56.47 ± 8.06	0.858

^aP < 0.05.

¹Comparison of clinical manifestations between baseline and the end of the 8-week follow-up.

²Others include Noradrenergic and Specific Serotonergic Antidepressants and Norepinephrine-Dopamine Reuptake Inhibitors.

BMI: Body mass index; HAMD-17: 17-item Hamilton Depression Scale; HAMA: Hamilton Anxiety Scale; BSSI: Beck Scale for Suicide Ideation; SDS: Sheehan Disability Scale; PDQ-5: Perceived Deficits Questionnaire five-item version; SSRIs: Selective Serotonin Reuptake Inhibitors; SNRIs: Serotonin-Norepinephrine Reuptake Inhibitors; SARIs: Serotonin Receptor Antagonists and Reuptake Inhibitors; T-AI: Trait Anxiety Inventory of State-Trait Anxiety Inventory; PSS: Perceived Stress Scale; S-AI: State Anxiety Inventory of State-Trait Anxiety Inventory; NA: Not available.

Prevalence and overlap of high-level stress-sensitivity factors

The prevalence of high-level stress-sensitivity factors is illustrated in Figure 1A. Among the 206 patients with MDD, 110 (53.40%) presented with at least one high-level stress-sensitivity factor, 29.61% exhibited two or more high-level factors, and 8.25% had all four high-level factors (Figure 1B).

Figure 1 The proportion of stress-sensitivity factors in patients with major depressive disorders. A: The proportion of patients with high- or low-level stress-sensitivity factors; B: The proportion of patients with different numbers of high-level stress-sensitivity factors; C: Venn diagram revealing the overlap between high-level stress-sensitivity factors in patients. PSS: Perceived Stress Scale; T-AI: Trait Anxiety Inventory of State-Trait Anxiety Inventory; S-AI: State Anxiety Inventory of State-Trait Anxiety Inventory.

The overlap of high-level stress-sensitivity factors is illustrated in Figure 1C. Thirty patients exhibited both high-level neuroticism and trait anxiety, while twenty-eight displayed both high-level perceived stress and state anxiety. The proportion of overlapping within-category factors (neuroticism and trait anxiety, or perceived stress and state anxiety) was not significantly different from the proportion of overlapping between-category factors (neuroticism and state anxiety, neuroticism and perceived stress, trait anxiety and perceived stress, or trait anxiety and state anxiety) (all P > 0.05). Ninety-six patients (46.6%) exhibited no high-level stress-sensitivity factors and were not included in the diagram. The size of different areas was not exactly proportional to the number of patients in the corresponding area.

As presented in Supplementary Table 2 in the appendix, patients who did not achieve clinical remission at the end of the 8-week follow-up (HAMD-17 total score ≥ 8) exhibited a significantly higher proportion of baseline high-level trait-related factors (i.e., the proportion of patients with both high-level Neuroticism and T-AI) compared to those who achieved remission through follow-up (P = 0.023).

Associations among stress-sensitivity factors, clinical manifestations, and follow-up outcomes of MDD

As shown in Table 2, in the overall MDD group at baseline, significant associations with at least four clinical manifestations were observed for all four factors. For instance, patients with a high T-AI score at baseline exhibited significantly higher HAMD-17, HAMA, BSSI, SDS, and PDQ-5 scores than those with a low T-AI score (all P < 0.01). However, only two trait-related factors, i.e., neuroticism and trait anxiety, survived the false discovery rate correction at the end of the 8-week follow-up (all P < 0.05).

Table 2 Comparison of clinical manifestations at baseline and follow-up outcomes between major depressive disorders patients with high-level and low-level stress-sensitivity factors.

Items	Groups	HAMD-17 at baseline	HAMD-17 at the end of follow-up	HAMA at baseline	HAMA at the end of follow-up	BSSI at baseline	BSSI at the end of follow-up (n = 66)	SDS at baseline	SDS at the end of follow-up	PDQ-5 at baseline	PDQ-5 at the end of follow-up
Neuroticism	Low-level	20.88 ± 5.42	9.54 ± 6.56	20.75 ± 7.65	10.31 ± 7.31	6.92 ± 7.04	4.23 ± 5.77	17.28 ± 6.01	10.90 ± 7.34	11.35 ± 3.47	7.34 ± 4.20
	High-level	22.10 ± 4.86	14.11 ± 7.46	22.83 ± 7.12	14.97 ± 7.87	10.74 ± 8.46	8.55 ± 8.64	21.35 ± 4.74	14.31 ± 7.68	13.11 ± 3.40	10.49 ± 4.75
	Sig.	0.070	0.002a	0.024a	0.003a	0.003a	0.044a	< 0.001a	0.035a	0.001a	0.001a
Trait anxiety	Low-level	20.08 ± 5.05	9.17 ± 6.38	20.00 ± 6.94	9.95 ± 6.46	6.44 ± 6.21	4.19 ± 5.13	17.52 ± 6.18	10.79 ± 7.12	11.51 ± 3.61	7.38 ± 4.17
	High-level	24.76 ± 4.33	15.97 ± 6.88	25.55 ± 7.84	16.83 ± 8.98	13.04 ± 9.46	9.61 ± 9.86	21.49 ± 3.97	15.23 ± 7.99	13.02 ± 3.06	10.91 ± 4.79
	Sig.	< 0.001a	< 0.001a	< 0.001a	< 0.001a	< 0.001a	0.038a	< 0.001a	0.005a	0.008a	< 0.001a
Perceived stress	Low-level	20.08 ± 5.13	10.26 ± 7.05	20.00 ± 7.01	10.67 ± 7.45	6.76 ± 7.15	4.84 ± 5.80	17.48 ± 5.97	11.28 ± 7.35	11.36 ± 3.50	7.78 ± 4.59
	High-level	24.07 ± 4.54	12.44 ± 7.17	24.72 ± 7.80	14.23 ± 7.99	11.27 ± 8.04	7.32 ± 9.09	21.00 ± 5.16	13.47 ± 8.00	13.15 ± 3.31	9.50 ± 4.40
	Sig.	< 0.001a	0.132	< 0.001a	0.023	< 0.001a	0.253	< 0.001a	0.155	0.001a	0.063
State anxiety	Low-level	20.41 ± 5.13	10.23 ± 6.91	20.33 ± 7.44	10.59 ± 6.93	6.82 ± 6.89	5.20 ± 6.08	17.59 ± 6.04	11.38 ± 7.39	11.52 ± 3.59	7.80 ± 4.48
	High-level	23.98 ± 4.87	13.11 ± 7.50	24.81 ± 6.89	15.42 ± 9.23	12.19 ± 8.71	6.75 ± 9.11	21.50 ± 4.58	13.70 ± 8.04	13.08 ± 3.09	9.87 ± 4.66
	Sig.	< 0.001a	0.064	< 0.001a	0.017	< 0.001a	0.418	< 0.001a	0.162	0.007a	0.038

^aP values that survived false discovery rate correction for multiple testing.

HAMD-17: 17-item Hamilton Depression Scale; HAMA: Hamilton Anxiety Scale; BSSI: Beck Scale for Suicide Ideation; SDS: Sheehan Disability Scale; PDQ-5: Perceived Deficits Questionnaire five-item version.

Cumulative effects of high-level stress-sensitivity factors on MDD clinical symptoms and follow-up outcomes

The results of the one-way analysis of variance with linear trend analysis are presented in Table 3 and Figure 2. At baseline, the cumulative effects of all four high-level stress-sensitivity factors on the total scores of HAMD-17, HAMA, BSSI, SDS, and PDQ-5 were significant in the overall MDD group (all P < 0.001). There was a gradual increase in the severity of clinical manifestations at baseline with the increase in the number of high-level stress-sensitivity factors (Figure 2).

Figure 2 Gradual increase in severity of clinical manifestations at baseline with an increasing number of high-level stress-sensitivity factors in patients with major depressive disorders. The P values are obtained from the analysis of variance with linear trend analysis. All P < 0.001. Error bars represent the 95% confidence intervals. HAMD-17: 17-item Hamilton Depression Scale; HAMA: Hamilton Anxiety Scale; BSSI: Beck Scale for Suicide Ideation; SDS: Sheehan Disability Scale; PDQ-5: Perceived Deficits Questionnaire five-item version.

Table 3 Association of the number of high-level stress-sensitivity factors with clinical manifestations and follow-up outcomes in major depressive disorders.

Items	Time	All factors Partial eta-squared (η2)	Sig.	PSS + S-AI Partial eta-squared (η2)	Sig.	Neuroticism + T-AI partial eta-squared (η2)	Sig.
HAMD-17	Baseline	0.158	< 0.001a	0.153	< 0.001a	0.085	< 0.001a
	At the end of follow-up	NA	NA	NA	NA	0.204	< 0.001a
HAMA	Baseline	0.145	< 0.001a	0.129	< 0.001a	0.072	< 0.001a
	At the end of follow-up	NA	NA	NA	NA	0.155	< 0.001a
BSSI	Baseline	0.160	< 0.001a	0.117	< 0.001a	0.135	< 0.001a
	At the end of follow-up (n = 66)	NA	NA	NA	NA	0.156	0.002a
SDS	Baseline	0.240	< 0.001a	0.132	< 0.001a	0.145	< 0.001a
	At the end of follow-up	NA	NA	NA	NA	0.072	0.003a
PDQ-5	Baseline	0.138	< 0.001a	0.065	< 0.001a	0.066	< 0.001a
	At the end of follow-up	NA	NA	NA	NA	0.143	< 0.001a

^aP < 0.05.

The effect size is classified as small (0.01-0.05), medium (0.06-0.13), and large (≥ 0.14) based on Partial η². HAMD-17: 17-item Hamilton Depression Scale; HAMA: Hamilton Anxiety Scale; BSSI: Beck Scale for Suicide Ideation; SDS: Sheehan Disability Scale; PDQ-5: Perceived Deficits Questionnaire five-item version; NA: Not available.

The contribution of trait-related and state-related factors on MDD symptoms and follow-up outcomes was investigated in subsequent analyses (Table 3). Cumulative effects of state-related factors (perceived stress and state anxiety) and trait-related factors (neuroticism and trait anxiety) were separately significant on the total scores of HAMD-17, HAMA, BSSI, SDS, and PDQ-5 at baseline (all P < 0.001). At the end of the 8-week follow-up, a significant cumulative effect of trait-related factors (neuroticism and trait anxiety) only on clinical outcomes was found (all P < 0.01).

Furthermore, the cumulative effect sizes of state-related factors on the baseline score of HAMD-17 were most significant among the baseline manifestations of MDD (Table 3, partial η² = 0.153, P < 0.001), larger than that of trait-related factors on HAMD-17 (partial η²_state = 0.153, partial η²_trait = 0.085). Multivariable regression analysis showed that among the state-related factors, only the score of PSS had a significant predictive effect on the score of HAMD-17 at baseline (Table 4, standardized β = 0.195, P < 0.05). At the end of the follow-up, the cumulative effect sizes of trait-related factors on the score of HAMD-17 and BSSI were the most significant (partial η² = 0.204 and 0.156; both P < 0.01). Further analysis of multivariable regression showed that at the end of the follow-up, the score of T-AI could predict the score of HAMD-17 (Table 4, standardized β = 0.247, P < 0.05), and the score of neuroticisms could predict the score of BSSI (standardized β = 0.392, P < 0.05).

Table 4 Multivariable regression analysis of depressive symptoms and suicide risk.

Variables	HAMD-17 at baseline	HAMD-17 at the end of follow-up	BSSI at baseline	BSSI at the end of follow-up
Neuroticism	–0.089	0.179	0.117	0.392a
Trait anxiety	0.284a	0.247a	0.219a	0.077
Perceived stress	0.195a	NA	0.056	NA
State anxiety	0.145	NA	0.057	NA
R2	0.261a	0.129b	0.139b	0.186a

^aP < 0.05.

^bP < 0.001.

Data are presented as standardized β.

HAMD-17: 17-item Hamilton Depression Scale; BSSI: Beck Scale for Suicide Ideation; NA: Not available.

DISCUSSION

The present study explored the cumulative effects of high-level stress-sensitivity factors on the clinical symptoms and follow-up outcomes in patients with MDD. The results showed that: (1) At baseline, more than half of the MDD patients had at least one high-level stress-sensitivity factor, and almost a third displayed high levels in two or more stress-sensitivity factors; (2) Compared with patients with clinical remission at the end of the 8-week follow-up, those without clinical remission showed higher proportional high-level trait-related factors; (3) At baseline, MDD patients with more high-level stress-sensitivity factors showed more severe clinical manifestations. The severity of depressive symptoms was more likely to be attributed to high-level state-related factors compared to trait-related factors. Only perceived stress could predict the severity of depressive symptoms among the two state-related factors; and (4) MDD patients with more high-level trait-related factors showed poorer outcomes at the end of follow-up, especially in the severity of depressive symptoms, which could be predicted by trait anxiety, and suicide risk which could be predicted by neuroticism.

High-level state-related stress-sensitivity factors were associated with severe depressive symptoms at baseline

The results supported the first hypothesis that MDD patients with more high-level stress-sensitivity factors had more severe clinical manifestations at baseline. This complements the findings of previous studies exploring the relationship between stress-sensitivity factors and depression[36]. Furthermore, the findings that state-related factors had a more significant effect on depressive symptoms during an MDD episode compared to trait-related factors also partially supported our third hypothesis.

Additionally, the results indicated that perceived stress could predict the severity of depressive symptoms at baseline, which was aligned with the prior scholarly work. According to Cristóbal-Narváez et al[37], there was a linear increase in the prevalence of depression with increasing perceived stress scores; specifically, a one-unit increase in the perceived stress score was associated with 1.40 times higher odds for depression (95%CI: 1.35-1.44). However, it cannot be ignored that there might be a potential overlap between perceived stress and depression. To investigate the impact of perceived stress on the severity of depression prospectively by minimizing any overlapping effects, future research may look at examining perceived stress before depression. This could serve as a guide to lessen the intensity of depressed symptoms during acute episodes.

High-level trait-related factors were associated with poor outcomes

Meanwhile, this study indicated that MDD patients with more high-level trait-related factors were more likely to have poorer follow-up outcomes, especially in depressive symptoms and suicidal ideation, which partially supported our second and third hypotheses. Although previous studies have reported that MDD patients with high neuroticism and trait anxiety tend to have poorer clinical outcomes[20,38], our study importantly demonstrated that multiple trait-related factors cumulatively influenced follow-up outcomes. More attention should be given to trait-related factors to advance the development of interventions targeting these factors to improve the prognosis for long-term treatment.

Furthermore, suicide risk has been previously found to be influenced by both trait and state components, such as trait/state hopelessness and trait/state anxiety[21], which was inconsistent with the present study finding that only neuroticism can predict suicide risk. Meanwhile, the limited inclusion of trait or state factors in this paper may have limited the discovery of more positive results. For instance, prior research has indicated that stressful life events may influence the risk of suicide[39], a finding that should be further explored in future studies as it was not examined in this study at baseline or during follow-up.

According to the MDD stress sensitization hypothesis, patients gradually become more sensitive to stress, which lowers the degree of stress triggering episode onset with each subsequent episode[40]. Even though we found that bad outcomes of MDD were linked to high levels of trait-related characteristics, it is still crucial to look into more detailed mechanisms and better understand how these temperamental predictors affect outcomes. Previous research has found several intriguing explanations, including genetic and heritable variables, aberrant brain structure and function, neuroimmune and endocrine system dysfunction, inappropriate emotion regulation, dysfunctional cognition, and socio-environmental stresses[41-44]. The hypothalamic-pituitary-adrenal axis is a significant stress response system, and excessive responses may have an impact on depression symptoms and suicide[45].

Adapting to adversity is the process of resilience, which necessitates the integration of several core and peripheral systems[46]. Previous research suggests that patients with depression may exhibit a more favorable response to treatment when characterized by low levels of trait anxiety, high resilience, or a combination of these factors[20]. A more thorough understanding of MDD's susceptibility may be possible with research integrating stress sensitivity and resilience mechanisms and how they interact with MDD. This is an important area for future study.

The idea that psychological therapies can address temperamental vulnerability has received some preliminary support from researchers who have worked to develop psychotherapy for neurotic and trait anxiety temperaments. Barlow et al[47] devised a comprehensive cognitive-behavioral intervention strategy for the transdiagnostic treatment of emotional disorders. Numerous randomized controlled trials confirmed that patients with emotional disorders who received treatment using a unified protocol saw reductions in neuroticism, which was linked to improvements in core symptoms, functional impairment, and emotional disorder quality of life[48,49].

Selective serotonin reuptake inhibitors[50], lysergic acid diethylamide-assisted therapy[51,52], medications with anti-inflammatory effects targeted to microglia[53], virtual reality-based biofeedback[54], lavender inhalation[55], and music therapy[56] have all been found to potentially lower neuroticism and state-trait anxiety. More high-quality RCTs with more uniform study designs are required to validate these results.

In summary, the above findings underscored the screening and assessment of stress-sensitivity factors and the need for comprehensive investigations into the mechanism underlying the relationship between stress sensitivity and MDD. The findings also highlighted the importance of trait-related stress-sensitivity factors in the follow-up outcomes, with implications for advancing the development of interventions targeting these factors to achieve clinical remission and better outcomes.

Limitations

Firstly, while our study incorporated longitudinal data, its design did not examine causal relationships between stress-sensitivity factors and MDD follow-up outcomes, a capability better suited to interventional studies. Secondly, the evaluation of stress-sensitivity components relied on self-reported scales. Future research should integrate multi-dimensional objective indicators, such as neuroimaging markers, endocrine hormones, inflammatory factors, and susceptibility genes to gain a more comprehensive understanding of stress sensitivity phenotypes. Thirdly, the sample predominantly consisted of young, single Chinese women. Therefore, the findings require validation in more diverse populations and may not be generalizable to other demographic groups. Fourthly, some conceptual overlap may exist between depression and state-related stress-sensitivity characteristics. Future studies should refine their designs to minimize the potential interference of such overlaps on the results. Fifthly, the study focused on trait-related characteristics, including stress-sensitive traits like neuroticism and trait anxiety. However, the relatively short 8-week follow-up period limits a deeper understanding of how stress-sensitivity parameters may influence longer-term outcomes of MDD, such as relapse or chronic dysfunction. Further research with extended and more comprehensive follow-up periods is needed. Finally, 42.23% of participants were lost to follow-up during the 8-week period, which could have affected the findings. The clinical outcomes of patients who completed follow-up may differ from those who were lost to follow-up. Enhancing follow-up methodologies and efforts to increase follow-up rates are essential to improving the accuracy and generalizability of future studies.

CONCLUSION

To the best of our knowledge, this is the first study to investigate the cumulative effect of stress-sensitivity factors on MDD manifestations and follow-up outcomes. The varying associations between stress-sensitivity factors and different clinical manifestations highlighted the need for screening and assessment of stress-sensitivity factors in psychiatric clinical practice and call for multi-faceted and holistic approaches. Particularly, attention to trait-related factors is essential for improving follow-up outcomes and reducing the risk of suicide in MDD. Understanding the function of stress sensitivity in depression can be advanced by incorporating future research on stress-related psychological and neurobiological pathways.

ACKNOWLEDGEMENTS

We would like to express our gratitude to all participants in the study.