INTRODUCTION
Suicide ranks tenth in the world in terms of mortality, leading to nearly 800 000 deaths each year.1 The World Health Organization reported that more than 727 000 people die by suicide every year, accounting for 1.1% of the total global deaths in 2021.2 Attempted suicide, defined as deliberate actions to end life without causing death, is thought to be 20 times more common than suicide itself. These attempts represent a major contributor to global disability, reduced life satisfaction, significant psychosocial consequences, and pose the highest risk for eventual suicide.3
Major depressive disorder (MDD) is a common and often incapacitating condition that severely impairs quality of life, disrupts interpersonal and occupational functioning, and can result in long-term disability. Major depressive disorder has been consistently identified as a key psychiatric disorder associated with heightened suicide risk,4 with evidence suggesting that the likelihood of death by suicide in individuals diagnosed with MDD is approximately 20 times more likely than that in the general population.5 Prior research estimates that >60% of suicide attempters meet criteria for MDD.6
Identifying the factors linked with suicide attempts (SAs) in individuals with MDD is essential for the development of effective prevention strategies. However, the underlying risk factors remain insufficiently defined, and clinically applicable predictive markers are limited. Suicidal behavior is complex and affected by multiple interrelated psychological, social, and physiological factors.
Although numerous studies have investigated suicide risk in MDD, most have focused on outpatient or community samples,7,8 and research specifically examining SAs among hospitalized patients is scarce.9 This gap limits the understanding of the prevalence and correlates of SAs in the inpatient setting, where patients may present with more severe symptoms or higher clinical complexity. Earlier investigations have suggested several potential contributors to SAs in MDD, including cognitive dysfunction, unemployment, psychotic symptoms, persistent or severe depressive episodes, dyslipidemia, thyroid dysfunction, and abnormal glucose metabolism.10,11 By focusing on a large cohort of hospitalized MDD patients, the present study provides novel insights into the clinical and demographic factors associated with SAs in this high-risk population.
The present study examined the frequency of attempted suicide in hospitalized MDD cases and explored clinical and demographic correlates linked with increased risk. (1) It was hypothesized that the rate of attempted suicide would be substantially elevated in these individuals relative to the general public and (2) specific demographic and symptomatic characteristics would remarkably correlate with SAs. The study’s goal is to contribute to the identification of variables associated with increased or decreased likelihood of SAs, which may help guide timely and targeted clinical assessments and interventions.
MATERIALS AND METHODS
Study Design and Participants
This cross-sectional investigation was conducted at the Fourth People’s Hospital of Hefei, a major public psychiatric institution in China. A total of 3200 hospitalized Chinese patients (16-70 years) with MDD were enrolled. Diagnoses were confirmed independently by 2 experienced psychiatrists using the Structured Clinical Interview for DSM-IV (SCID).
The recruitment period was from May 2022 to June 2024. Inclusion criteria: (1) MDD was diagnosed according to the DSM-IV, Clinical Modification (ICD-10-CM), and the corresponding codes were F32.200, F32.300, F33.200, F33.300; (2) complete clinical and laboratory data were obtained during hospitalization. It should be stated clearly that the diagnosis was confirmed with the DSM-IV SCID, while ICD-10 codes were extracted from the hospital information system.
Exclusion criteria: (1) complicated with severe neurological diseases (such as epilepsy, Parkinson’s, and Alzheimer’s disease); (2) pregnant or lactating; (3) combined with other serious mental disorders (such as personality disorder, schizophrenia, bipolar disorder); (4) complicated with major physical diseases (such as severe cardiovascular and cerebrovascular events, end-stage liver and kidney dysfunction).
This retrospective study used routinely collected clinical data and involved no additional procedures. The Ethics Committee of University of Hefei Fourth People’s Hospital determined that the study met the exemption criteria (Approval No: HFSY-IRB-YJ-MCLL-ZLF [2025-069-001]; Date: June 25, 2025). General research consent is routinely obtained from all patients at admission.
Demographic Data Collection
Clinical and demographic information of patients, including age, sex, education level, marital status, age at MDD onset, body mass index, smoking, and alcohol consumption habits, was extracted retrospectively from available medical records and admission records.
Definition of Suicide Attempt
In this study, a SA was operationally defined as a deliberate, non-fatal act of self-harm undertaken with the explicit intention of ending one’s life.12 This behavior was assessed using a direct screening question: “Have you ever attempted suicide?” A positive response was recorded as an attempted suicide.13 In cases where the participant could not provide reliable information, responses were verified through close family members, relatives, or friends.
Biochemical Parameters
Venous blood samples were collected from all participants upon admission, prior to any psychiatric intervention or treatment. The primary rationale for blood collection was to examine potential biological markers associated with suicide risk.
Statistical Analysis
The normality of continuous variables was tested using the Shapiro-Wilk test and Q-Q plots. For continuous variables that met the assumption of normality, data were expressed as mean ± SD, and group comparisons were conducted using independent samples t-tests. For continuous variables that did not follow a normal distribution, data were presented as median (Q1, Q3), and the Mann-Whitney U-test was applied for between-group comparisons.
Categorical variables were presented as frequencies and percentages, and comparisons between groups were performed using the chi-square test. Effect sizes were also calculated to estimate the magnitude of group differences. For continuous variables analyzed with the Mann-Whitney U-test, Cliff’s δ with 95% CIs was reported; for variables analyzed with t-tests, Hedges’s g was reported.
For categorical variables, the Cramér’s V coefficient was used to quantify association strength. Variables (sex, education level, marital status, presence of psychotic features, alcohol use, and age of onset) found to be correlated with SAs (P < .05) in univariate analyses were evaluated via multivariate logistic regression to identify variables independently associated with SAs. Prior to conducting multivariate logistic regression, collinearity diagnostics were performed to evaluate potential multicollinearity among independent variables.
The Hosmer-Lemeshow goodness-of-fit test was used to evaluate the fitness of the model. All data were analyzed using SPSS v28.0, (IBM SPSS Corp.; Armonk, NY, USA) with a two-tailed P-value < .05 considered statistically significant.
RESULTS
Participant Characteristics
Overall, 3200 participants were enrolled in the analysis, comprising 962 males and 2238 females. The mean age of the patients was 44.54 years (SD =17.08), ranging from 16 to 70 years. As for educational level, there were 403 illiterates (12.59%), 618 with primary school education (19.31%), 731 with junior middle school education (22.84%), 682 with senior high school education (21.31%), 716 with a university degree (22.38%), and 50 with a postgraduate degree (1.56%).
The mean disease duration was 6.13 years (SD=7.65), ranging from 2 weeks to 50 years. The proportion of smokers among all patients was 7.41% (237/3200), and the proportion of alcohol drinkers was 5.44% (174/3200). Of the 3200 MDD patients, 7.78% (249/3200) had psychotic symptoms.
The results are presented in Table 1.
Among the 3200 inpatients with MDD, 664 (20.75%) had a history of SAs (SA group), and 2536 (79.25%) had no history of SAs (NA group). Compared with the NA group, the SA group included more females (P=.045), had a higher prevalence of symptoms of psychosis (P < .001) and alcohol consumption (P < .001), younger age (P < .001), and earlier age of onset (P < .001). Significant differences were also observed in marital status (P < .001) and education level (P=.012) (Table 1).
Influencing Factors of Suicide Attempts in Major Depressive Disorder Patients
Variables that showed statistical significance in the univariate analysis (Table 1) were included in a multivariate logistic regression model to identify independent predictors of SAs. Collinearity diagnostics showed that there was moderate collinearity between “age” (Variance Inflation Factor (VIF)=5.64) and “age of first onset” (VIF=5.15). In the regression model, the P-value for age at onset was .001, and the P-value for age was .497.
Given that “age at first onset” was statistically significant and more clinically relevant, “age” was eventually dropped from the regression model. After adjustment, the VIF values of the remaining variables were less than 2, indicating that the collinearity problem was relieved. Additionally, the model’s goodness of fit was evaluated using the Hosmer-Lemeshow test, which yielded a chi-square value of 6.49 (df=8, P=.618), indicating that the model fit the data well.
Binary logistic regression analysis revealed several significant factors associated with SAs among MDD patients. Female patients were more likely to have attempted suicide compared to males (OR=1.436, 95% CI: 1.163-1.772, P < .001). Unmarried individuals had a higher risk of SAs compared to those who were married (OR=1.658, 95% CI: 1.232-2.232, P < .001).
The presence of psychotic symptoms (OR=1.550, 95% CI: 1.154-2.081, P=.004) and alcohol use (OR=2.312, 95% CI: 1.619-3.301, P < .001) were also significantly associated with increased risk. Additionally, higher education level and older age at first onset were associated with a decreased likelihood of SAs. Patients with a master’s degree or higher educational attainment had a significantly lower likelihood of SAs compared to illiterate patients (OR=0.210, 95% CI: 0.071-0.619, P=.005).
For each year increase in age of onset associated with a 1.6% decrease in the odds of SAs (OR=0.984, 95% CI: 0.975-0.992, P < .001) (Table 2).
DISCUSSION
Suicide attempts among patients with MDD are a global public health issue, with prevalence rates differing across countries. For instance, 33.7% of patients with unipolar depression or bipolar disorder in France reported having attempted suicide at least once.14 Rates as high as 35% have been reported in India,15 38% in the UK,16 and 36.3% in the USA.17 Among Asian populations, the prevalence is relatively lower but still notable—19.8% in Korea18 and 16.9% in Thailand.19 Collectively, these data highlight the substantial global burden of SAs among individuals with MDD. In light of this global trend, it is essential to investigate the prevalence and variables associated with SAs in clinical populations.
Based on available information, this is the first large-scale, cross-sectional analysis designed specifically to assess the frequency and variables associated with attempted suicide in hospitalized Chinese MDD patients. In this study, the observed SA rate was 20.75%.
This study found that female patients with MDD were significantly more likely to report a history of SAs compared to male patients, which is consistent with prior research.20,21 From a biological perspective, hormonal differences between sexes play an important role in regulating mood and stress responses.22 Major sex hormones such as estrogen and testosterone can affect neurotransmitter systems like serotonin, norepinephrine, Gamma-Aminobutyric Acid (GABA), and dopamine, which are closely related to the risk of depression.23 Fluctuations in estrogen levels during the menstrual cycle, pregnancy, and perimenopause may lead to emotional instability, making women more likely to experience depressive symptoms during these periods.24 In contrast, testosterone may have some protective effects on men’s stress responses. Neuroimaging studies have found that women often show stronger activity or altered connectivity in brain regions involved in emotion regulation, such as the prefrontal cortex and limbic system, which may make them more vulnerable to stress. In depressed women, weakened functional connectivity between the prefrontal cortex and the amygdala has been observed, reducing their ability to manage negative emotions and increasing the risk of impulsive suicidal behavior.25 On a sociocultural level, women are more likely to experience stress from role conflicts or emotional imbalance within marriage or family relationships.26 An animal study also suggested that chronic social stress could damage the integrity of the blood-brain barrier in brain regions like the prefrontal cortex in female rodents, leading to anxiety- and depression-like behaviors.27 Together, these findings—from molecular mechanisms to behavioral outcomes—highlight the increased vulnerability of women to depression and suicidal behavior.
In addition to gender, marital status was also significantly associated with suicide risk. This study found that being unmarried increases the risk of SAs among patients with MDD. Marriage is often considered an important source of social support, offering emotional, financial, and crisis-related stability that can help reduce the psychological distress and loneliness caused by depression.
In contrast, single individuals may lack close personal relationships and reliable support when facing stress or negative emotions. Married people are also more likely to delay or restrain suicidal thoughts, while single individuals may be more prone to act impulsively without others around to intervene in time, which could further increase the risk of suicide.
In the present analysis, MDD with psychotic features, characterized by the presence of hallucinations and/or delusions, was found to be significantly predictive of attempted suicide. In individuals with MDD, 11.1% of those presenting with psychotic symptoms had attempted suicide, compared to 6.9% of those without such symptoms. These findings are in agreement with those of Dold et al,28 which demonstrated that individuals with psychotic depression show more prominent depressive symptomatology, a heightened risk of suicidal behavior, and resistance to treatment relative to those with non-psychotic MDD.
Recent studies have reported that psychotic MDD patients face a 2-fold increase in the possibility of attempting suicide, both during acute episodes and across their lifetime, relative to their non-psychotic MDD.29,30 These findings suggest that psychotic MDD may involve different neurobiological mechanisms that increase the risk of suicide. For example, dysfunction of the hypothalamic-pituitary-adrenal axis—such as higher cortisol levels after dexamethasone suppression in patients with psychotic MDD—may lead to stronger stress responses and poor emotion regulation, which can raise the likelihood of suicidal behavior.31 In addition, reduced dopamine-β-hydroxylase activity in psychotic MDD may cause excess dopamine, which could worsen symptoms like agitation, impulsivity, or delusional guilt—factors closely linked to suicide risk. Magnetic resonance imaging studies have also shown that patients with psychotic MDD have smaller amygdala volumes compared to non-psychotic patients, suggesting altered processing of fear and threat.
This may impair cognitive control and emotional recovery, making these individuals more vulnerable to self-harm during periods of acute stress.
This study also found that alcohol use was significantly associated with an increased risk of SAs among patients with MDD. Some individuals with MDD may use alcohol to relieve negative emotions or as a form of self-medication. Although this may temporarily ease emotional pain, alcohol is a central nervous system depressant that can reduce cognitive control, judgment, and self-restraint, thereby increasing the risk of impulsive suicidal behavior.32 In addition, alcohol use is closely linked to personality traits such as impulsivity and aggression, and impulsivity is one of the key predictors of suicidal behavior.33 Studies have further suggested that alcohol use, suicidal thoughts, and impulsivity may share common genetic risk factors, indicating possible shared neurobiological mechanisms.33
Finally, this study found that having a postgraduate degree or higher, as well as an older age at onset, was associated with a decreased likelihood of SAs in patients with MDD. Individuals with higher education levels often have better cognitive abilities, emotional regulation skills, and health awareness. When facing stress, they are more likely to use rational and constructive coping strategies and to seek professional help, which can reduce the risk of suicidal behavior.34,35 In addition, people with higher education usually have greater access to social support and more stable living conditions, which also contribute to better mental health.
Early-onset depression is often linked to a more severe clinical course, including higher rates of relapse, chronic symptoms, comorbid psychiatric disorders, and impaired psychosocial functioning—all of which are known risk factors for suicide. In contrast, those with later onset may benefit from greater emotional maturity, better-developed coping strategies, and stronger cognitive control, helping to protect them from impulsive and suicidal behaviors.36,37
Based on the variables associated with SAs identified in this study, several intervention strategies can be considered. For female patients, targeted psychoeducation and stress management programs may help address hormonal and psychosocial vulnerabilities. For unmarried individuals, enhancing social support networks through community programs, peer support groups, and family engagement interventions could reduce feelings of isolation and loneliness.
Patients exhibiting psychotic features may benefit from early recognition, optimized antipsychotic treatment, and regular monitoring of psychotic symptoms to prevent escalation of suicidal behavior. For patients with alcohol use, integrated interventions combining psychiatric care with substance use counseling and cognitive-behavioral therapy may help reduce impulsivity and self-harm risk. In addition, educational programs aimed at improving coping strategies, emotional regulation, and health awareness—particularly for those with lower education levels—can empower patients to seek help early and employ adaptive coping mechanisms during periods of stress.
Limitations
Although the present study provides important observations, limitations have also been reported. Firstly, its cross-sectional nature restricts the ability to infer causality between clinical or demographic variables and SAs. Future longitudinal investigations are necessary to establish more definitive links between variables associated with SAs and suicidal behavior in MDD patients.
Secondly, this study did not evaluate the intensity of depressive symptoms among participants, which may influence suicidal tendencies. Moreover, while information regarding attempted suicide was collected from interviews with patients and their families, the absence of a standardized suicide assessment instrument raises the possibility of recall bias. In addition, in this study, other established predictors of suicide (such as MDD severity, family history of suicide, traumatic experience, and addiction), as identified in previous literature, were not examined, thus constraining this ability to control for confounding influences.
Additionally, although this study included a large sample size of more than 3000 hospitalized patients with depression, its generalizability may be limited because the data were obtained from a single tertiary hospital in Hefei, Anhui Province, China city. Lastly, as the participant pool consisted exclusively of psychiatric inpatients, the findings may not be fully applicable to broader MDD populations, including outpatients or those treated in other clinical contexts.
Conclusion
This study found that 20.75% of hospitalized MDD patients had attempted suicide, highlighting the critical need for timely suicide assessments and early intervention in this population. This study identified variables associated with an increased likelihood of SAs, including being female, being single, having psychotic symptoms, and alcohol consumption. Variables associated with a decreased likelihood of SAs included having a postgraduate degree or higher and an older age at the onset of the illness.
Clinicians should consider these characteristics when identifying MDD cases at increased risk of attempted suicide, as this may assist in developing targeted prevention strategies and decreasing the likelihood of suicide in this vulnerable population.