Schizophrenia is a serious psychiatric disorder of unknown etiology and is associated with high mortality and morbidity (1–3). Suicide is one of the major causes of death among individuals with schizophrenia (4). About 5.6% of patients with schizophrenia die by suicide, and about 10–50% have suicidal ideation or attempt (5, 6). With high mortality rates, the adverse effects on the family and society of this pathology are significant. Therefore, previous studies were actively looking for clinical and psychobiological risk factors for suicidal behavior, thereby reducing the occurrence of suicides and implementing suicide intervention early.

Early researches have documented the association between suicide risk and demographic characteristics and clinical features in schizophrenia, including severe depressive symptoms (7), great cognitive ability (8), male gender, young age, higher intelligence quotient and family history of suicide attempts (9), alcohol and substance use disorder, smoking, among others (4, 6), but there is an abundance of research findings that failed to confirm those associations (10). Active hallucinations, as a risk factor for suicide, also had a strong evidential basis (11, 12). In addition to the demographic and clinical associations, rising evidence has supported an important role of childhood trauma (CT) and insomnia in predicting suicide risk in schizophrenia.

It has been found that those who suffered from CT have a susceptibility factor for developing schizophrenia compared to children without CT (13, 14). Interestingly, the suicidal risk was stronger among individuals with depression who reported a history of CT (15). However, previous studies on the relationship between CT and suicide in schizophrenia have been inconsistent. Some research shows that CT has nothing to do with attempted suicide in schizophrenia (16), but other studies have found that CT is a strong predictor of suicide in schizophrenia (17–19). Further studies find that suicide attempts are associated with all CT dimensions, such as physical, sexual, psychological abuse, and neglect, except for physical neglect (20), while other studies find that physical neglect and sexual abuse are indeed associated (21, 22). Much of the literature on suicidality in patients with schizophrenia considers CT but does not specify which aspect, or only considers one aspect of CT. Therefore, it is necessary to explore the relationship between CT and suicide in schizophrenia, at the same time, to further explore what aspects of CT are more related to suicide risk.

Moreover, insomnia is also commonly seen in schizophrenia patients, but it is easily ignored in those patients (23). Numerous studies reported that insomnia, as a core symptom of depression, has been associated with an increased risk of suicidal ideation (24, 25). This relationship is also found in patients with anxiety and post-traumatic stress disorder (26). Although insomnia is closely related to suicide in other mental illnesses, there is a paucity of data on this association in patients with schizophrenia. Hence, more research is needed to understand the association between insomnia and suicide risk in schizophrenia patients.

As noted in a review by Kajeepeta et al. there is extensive evidence that the impact of CT on sleep can last into adulthood (27). A study comprised of 9,582 samples observed that the risk of insomnia differed based on the age at first exposure to adversity as well as the type of adversity (28). Moreover, a cross-sectional questionnaire survey also provided that impaired subjective sleep was significantly predicted by child maltreatment experiences (29).

In the present study, we aimed to explore the relationship between CT, clinical factors such as insomnia, and suicide ideation in schizophrenia patients, and further to identify which aspects of the trauma are most associated with suicidal ideation in these patients. Based on the available literature, we hypothesized that patients with a history of CT have stronger suicidal ideation and that the relationship between CT and suicidal ideation is mediated by impaired sleep quality.

Table 1 presents the demographic characteristics of schizophrenic patients with or without suicidal ideation. Among 83 patients with schizophrenia, 33 (39.8%) with SI and 50 (60.2%) without SI. There were significant differences in terms of smoking habits (X² = 5.331, p = 0.021), insomnia (X² = 4.169, p = 0.041) and childhood trauma (X² = 7.636, p = 0.006), especially PN (X² = 12.662, p < 0.001) between the two groups, while no differences were observed in sex, age, BMI, education level, marriage history, and drinking habits (all p > 0.05). In addition, 35 patients experienced at least one type of abuse, including physical abuse (11.4%), sexual abuse (14.3%), emotional abuse (11.4%), physical neglect (85.7%), emotional neglect (40.0%).

Table 2 shows the results of the t-test of clinical characteristics between schizophrenic patients with and without suicidal ideation. No significant differences were observed between patients with and without SI for positive or negative symptoms, nor for any of the other CTQ dimensions (All p > 0.05). Also, there were no significant differences in auditory hallucinations, aggressivity, or depressive symptoms between groups (All p > 0.05). There were no significant differences in age, sex, or BMI between the groups (all P > 0.05). Patients with suicidal ideation had higher scores of ISI (7.39 ± 4.24 vs. 5.38 ± 3.31, t = −2.422, p = 0.018) and PN (9.67 ± 3.14 vs. 7.80 ± 2.16, t = −2.983, P = 0.004) compared to individuals without suicidal ideation. These significant differences remained after controlling for sex, education level, smoking history (F = 6.634, df = 1, p < 0.05; F = 4.652, df = 1, p < 0.05), and survived after Bonferroni corrected (all p < 0.05).

The binary logistic regression analysis found that PN was negatively associated with SI in patients (OR = 5.46, 95% CI = 0.007–0.483, P < 0.05) (see Table 3). The adjusted R-square of the whole model was 0.193. Further, a stepwise multiple linear regression was performed, and the results revealed that SI severity was significantly associated with ISI scores (β = 0.272, P = 0.011) and PN scores (β = 0.257, P = 0.017), with adjusted R-square as 0.177 in this model (see Table 4).

According to the principle of mediation effect (38), both PN and ISI scores were significantly related to suicidal ideation in the schizophrenia group. Therefore, a mediation model of PN-ISI-suicidal ideation might exist. Based on the theoretical assumptions (39), constructing a mediating effect model of insomnia between PN and suicide, taking the suicidal ideation as dependent variables, using PN score as an independent variable, and using ISI score as an intermediary variable. It was found that the direct effect of PN on suicidal ideation was 1.577 (p = 0.003, 95% CI: 0.565–2.589) and the mediating effect of insomnia between PN and suicidal ideation was 1.283 (95% CI: 0.558–2.008). Besides, the total effect of the PN-ISI-suicidal ideation model was 2.050 (95% CI: 1.008–3.093). The interval of indirect effect didn't contain 0, which was statistically significant. Therefore, insomnia played a mediating role between PN and suicidal ideation in schizophrenia patients (Figure 1).

Here, we investigated the association between suicidal ideation, CT, and clinical variables. The study aimed to examine potential relationships between suicidal ideation and different dimensions of CT in patients with schizophrenia to identify a promising way to effectively prevent and ultimately reduce the incidence of suicide.

In this study, we found that schizophrenia with suicidal ideation was related to insomnia and childhood traumatic experiences. Further, binary logistic regression analysis suggested that physical neglect was an independent risk factor for suicide ideation in patients with schizophrenia. According to the result of this study, insomnia truly plays a medicating role in the PN and suicidal ideation. This indicates that PN and insomnia together regulate suicidal ideation in schizophrenia patients. To the best of our knowledge, CT experience may be a promising factor for evaluating the risk of suicide ideation in patients with schizophrenia. Li et al. reported that patients with a history of suicidal behaviors had significantly higher trauma scores than patients without such behaviors (40). Other studies also reported the relationship between CT and suicide. Roy found CT increases the risk for attempting suicide (41). In the present study, patients with CT had higher suicide ideation than those without CT, which is in accordance with previous studies, to a certain extent. CT may increase suicidality in schizophrenia through several biological or psychological mechanisms (13); this question warrants further investigation. The clinical implications of these findings involve the need to prevent suicidality in patients with schizophrenia and the accompanying suggestion that clinicians carefully assess patients' trauma experiences in childhood to provide therapeutic interventions to reduce them.

On the subdomain level, logistic regression analyses indicated that among CT dimensions, only physical neglect was a predictor of current suicidal ideation, but no difference in the exposure to sexual, emotional, or physical abuse. This shows the importance of evaluating and considering the effect of each type of trauma on its own. It has been found that PN is a stronger risk factor for suicide in patients with schizophrenia, compared to other types of trauma such as emotional abuse or neglect (21), which can damage their social function, then worse coping strategies related to negative consequences such as suicidal behaviors (42, 43). This variability of the association between several types of CT and suicidal risk was also found in a prospective study (44). PN of child maltreatment should be considered important risks to suicide. The awareness of the serious long-term consequences of PN should encourage better identification of those at risk and the development of effective interventions to protect children from violence.

The other main findings in our study are that insomnia predicted suicidal ideation independently from any type of CT. Our findings are in line with the growing literature suggesting a significant association of sleep disturbances with suicidal behaviors. A retrospective cohort study showed that insomnia increases the risk of suicide in patients with mental disorders (45). An 8-year longitudinal study of schizophrenia-spectrum disorders also confirmed this hypothesis (46). While those findings support the association between insomnia symptoms and suicide-related risk, the specific mechanisms by which insomnia symptoms confer risk for suicide are still unknown. Several psychological and physiological mechanisms have been proposed (47, 48). Specifically, they reported that negative cognitive appraisals, perceived social isolation, and unhelpful emotion regulation strategies may be potential mechanisms. It is crucial to take the development of targeted psychological interventions for those who have sleep problems.

Moreover, schizophrenia patients have presented correlations among PN, ISI, and suicidal ideation. The PN-ISI-suicidal ideation intermediate model has proved that PN and insomnia can affect suicidality together in schizophrenia patients. Findings from this study have specifically suggested a potential role of the PN-ISI-suicidal ideation loop in the pathogenesis of schizophrenia. Since no study to data explored this loop in schizophrenia patients, more further studies are warranted to verify our preliminary findings.

The current study also has several limitations. First of all, the Childhood Trauma Scale is a subjective questionnaire, which is a retrospective study, resulting in biased reporting. Second, the small sample size not only represents the difficulties in conduction research in this patient group but also affects the robustness of the statistical analyses with a risk of type-II errors which calls for future replication in larger subject samples. And the sample size of specific subdomains in CT was too small to demonstrate weaker associations between two variables. It is possible patients that physical neglect is more likely to report PN over other abuses, which would bias results. One study showed that (PN) and emotional neglect (EN) were most reported, and sexual abuse (SA) and physical abuse (PA) were the least reported (21). Therefore, the results should be taken as preliminary data. Further investigation with a large study population is required to identify the subdomains in CT that might have been overlooked due to poor statistical power and confirm the possible associations that were observed in the present study. Thirdly, the study was cross-sectional. No temporal relationships could be established. Finally, the lack of a normal or clinical control group was a further limiting factor to clarify whether these results are specific to schizophrenia.

In conclusion, schizophrenia with SI reported higher exposure to CT than those without SI. The results point toward childhood physical neglect to be of specific importance to schizophrenia, which may be an area for future prevention and clinical attention. At the same time, our study put forward the intermediate model of PN-insomnia-suicidal ideation while demonstrating insomnia as a full intermediary between PN and suicidal ideation in schizophrenia patients. It may be, then, that addressing adverse childhood experiences and insomnia in these patients at higher risk for suicide because of having suffered CT and insomnia could result in reducing suicides.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Medical Ethics Committee of the Affiliated Kangning Hospital of Wenzhou Medical University. The patients/participants provided their written informed consent to participate in this study.

YC, WT, and WX conceptualized and designed the study. YZ, XF, and WT recruited the participants and completed the screening assessments. BT, KF, YC, and NW analyzed the data and performed the statistical analysis. YZ, XF, and KZ wrote the first draft of the manuscript. All authors revised the manuscript and approved the final manuscript.

This work was supported by the Science and Technology Program of Wenzhou (Y20190478).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.