Non-suicidal self-injury (NSSI) has been defined as the deliberate and self-inflicted destruction of body tissues without suicidal intent (1). The prevalence of NSSI among the clinical samples of adults ranged from 13 to 37% (2–4). Typical NSSI behaviors include cutting, scraping skin, skin-picking, self-hitting and biting, burning, and tying (5, 6). NSSI usually begins in adolescence. It has the highest prevalence during adolescence and early adulthood, but it can manifest at any age (4, 7, 8). Although by definition, NSSI occurs without an intent to commit suicide, it is strongly associated with suicidal thoughts and behaviors (8, 9).

NSSI has received increasing attention over the past several decades (10) with growing evidence suggesting that it is a transdiagnostic symptom commonly associated with psychiatric disorders including mood disorders (11–13). While self-harm is a diagnostic criterion of borderline personality disorder, studies have shown that mood disorders and NSSI frequently co-exist (14, 15). Individuals who engaged in NSSI exhibited elevated levels of anxiety and depression compared to those who did not engage (16, 17). A meta-analysis showed that individuals with mood disorders exhibited more than twice the odds (odds ratio = 2.09) of engaging in NSSI compared to those without such disorders (18).

Emotion dysregulation (ED) is a factor that could mediate the relationship between mood disorders and NSSI behavior (19). ED refers to “an individual's ability to modify an emotional state so as to promote adaptive, goal-oriented behaviors” (20). Therefore, ED refers to the failure to change the reactivity of emotions or the unacceptance and devaluing of emotions (21). ED has been proposed as a critical component in the development and maintenance of mood disorders (19). NSSI is often used to regulate affect to reduce or escape from an aversive or negative affective state (22, 23). It may provide relief from emotional distress (24). Thus, NSSI is often intended to avoid the negative emotional experiences associated with mood disorders.

Previous studies demonstrated the association of mood disorders with NSSI and ED (13, 19). However, little is known about whether the patterns of NSSI behavior and ED differ by mood disorder subtypes such as bipolar disorder (BD) and major depressive disorder (MDD). In addition, differences in NSSI behavior and ED between patients with bipolar I disorder (BD-I) and bipolar II disorder (BD-II) remain unclear. Considering that BD-I, BD-II, and MDD differ clinically in terms of long-term illness trajectory (25–27), they might show different patterns of ED and NSSI behavior.

Childhood trauma is generally accepted as a risk factor for NSSI (28, 29) and mood disorders (30). Some studies have suggested that depression (31) and ED (24, 32) might play a potential role in the association between NSSI and childhood trauma. However, previous studies investigating the relationship between NSSI and childhood trauma mainly targeted adolescents (29). Few studies involved adult clinical samples.

The aim of this study was to determine differences in NSSI patterns and association with ED according to mood disorder subtypes (i.e., BD-I, BD-II, and MDD). The relationship between childhood trauma and NSSI and the possible mediating effects of ED on the relationship were also explored.

NSSI is a widespread phenomenon without diagnostic boundaries. However, previous studies did not explore the differences across mood disorder subgroups. The current study investigated NSSI and ED patterns among patients with BD-I, BD-II, and MDD. We additionally analyzed the association of NSSI and ED with childhood trauma in patients with mood disorders.

Approximately 54% of the participants in our study reported a lifetime history of NSSI. The rate of NSSI in our study was comparable to that of a previous study (13). A previous study involving patients seeking treatment at a general practice clinic in the United States (U.S.) reported a higher prevalence of NSSI in patients with mood disorders than in those with other psychiatric disorders (43 vs. 20%, respectively) (13). The prevalence of NSSI was especially high in subjects with bipolar disorder (up to 52%) (13).

This was the first study that compared the rates and patterns of NSSI in patients with BD-I, BD-II, and MDD. Notably, patients with BD-I showed substantially lower rates of lifetime NSSI behavior than the other groups. The rate of lifetime NSSI behavior was the highest in the BD-II group. Previous studies evaluating the rate of NSSI behavior in patients with psychiatric diagnoses reported mixed results regarding NSSI rates in patients with BD (18, 41). The contrasting results of NSSI frequency between BD-I and BD-II patients might have contributed to the mixed results. The patterns of NSSI behavior also differed among the three groups. Patients diagnosed with BD-II manifested cutting as the most common method of NSSI, which is potentially associated with a high degree of tissue damage compared to the other methods.

Consistent with the higher rate of NSSI behavior in patients with BD-II and MDD, difficulties in ED were more severe in those groups. BD-II was associated with profound ED, similar to MDD, rather than BD-I. A meta-analysis revealed that ED was common in patients with BD, but the differences in ED between patients with BD-I and BD-II were not clear (42). A single study reported the absence of differences in DERS scores between patients with BD-I and BD-II (43). However, the small sample size affected the results. Previous studies reported significant ED in patients with mood disorders compared to the general population, which was pervasive across diverse mood states including manic, depressive, and euthymic conditions (44, 45). In particular, mood dysregulation is more severe in BD-II than in BD-I (46, 47), which corroborates our study findings. Decreased ED correlated with a decrease in depression and anxiety (42), suggesting an association between negative affect and persistent ED. Depressive episodes are known to be more frequent in patients with BD-II than BD-I (27). The more frequent depressive episodes and mood swings observed in patients with BD-II often resemble severe ED found in borderline personality disorder (48, 49). A neuroimaging study also suggested differences in mood regulation circuitry between patients with BD-I and BD-II (50).

Consistent with the significant differences in total DERS scores, patients with BD-I showed better emotion regulation than those manifesting BD-II and MDD on all subscales of the DERS, and the difference reached a significant level in four domains (non-acceptance of emotion response, lack of emotional clarity, limited access to emotion regulation strategies, and difficulties engaging in goal-directed behavior). In contrast to our study, no previous studies analyzed the differences in DERS subscales between mood disorder subtypes. In a previous meta-analysis, subjects with BD showed significantly higher DERS subscale scores except for the awareness subscale score compared to healthy controls (42). Compared to borderline personality disorder, BD was associated with significantly lower scores on all the DERS subscales. A previous study investigating the latency profiles of DERS reported that the awareness subscale did not correlate with other subscale scores (51). Understanding the ED profiles across mood disorder subtypes will facilitate treatment strategies for these populations. Further studies are needed to confirm the study findings.

Although NSSI implies non-suicidal intent, it is associated with an increased risk of suicidal attempts (8, 9). The higher the rate of NSSI behavior, the more frequent the use of methods with the potential for a high degree of tissue damage, and greater ED in patients with BD-II might contribute to the increased suicide risk in patients with BD-II. A recent clinical study with the largest-ever sample size (52) and a meta-analysis (53) confirmed the higher prevalence of suicide attempts in patients with BD-II. In a recent prospective study of patients diagnosed with BD and MDD (54, 55), suicide attempts were more frequently observed in patients with BD than in those with MDD mainly because of the higher duration of high-risk illness, i.e., more frequent depressive episodes. Subjects in the MDD group in our study showed NSSI rates comparable to those in the BD-II group, reinforcing the association between NSSI behavior and recurrent depression.

In our study, the CTQ scores were also higher in patients with BD-II and MDD than in patients with BD-I. However, in the sub-scale analyses, such differences were significant only for the emotional neglect subscale. Emotional neglect is arguably the most subjective and difficult to define among the forms of abuse (56). A prior study reported no differences in the rate of childhood traumatic experience of patients with BD-I and BD-II (57). Further studies are needed to confirm our findings.

In accordance with previous studies (8, 58), the NSSI group was younger and included proportionately higher numbers of females than the non-NSSI group. The CTQ and DERS scores differed depending upon the diagnosis, with NSSI being more prevalent in the group with higher CTQ and DERS scores. These results support the mediating role of ED on the association between childhood trauma and NSSI. As hypothesized, the SEM results revealed that ED mediated the relationship between childhood trauma and NSSI. When evaluating the model fit using total subjects including all diagnostic groups, a single indicator was not acceptable (RMSAEA), probably due to heterogeneity between the diagnostic groups. Subgroup analyses showed a better fit except for the BD-II group for unknown reasons. Diagnostic characteristics may also affect the model fit. The correlation between CT and NSSI has been confirmed in studies enrolling various subjects as well as clinical samples (29). A single study explored the mediating role of ED in the relationship between CT and NSSI, although it involved adolescent inpatients (59). A study of adult clinical samples has yet to be reported. The current study further emphasized the importance of ED as a mediating factor by limiting the patient group to those with mood disorders. ED is a major independent risk factor for NSSI (21, 60). The functional aspects of ED in NSSI have been studied, especially in terms of behavioral theory. Based on the behavioral model, the positive reinforcement function (e.g., to feel something) and negative reinforcement function (e.g., to relieve depression or uncomfortable internal experiences) of NSSI might be relevant to individuals with ED (22, 61). In this respect, individuals who experience negative emotions may use NSSI as a coping strategy. A group skills training program in dialectical behavior therapy is one possible treatment option that appears to be effective in decreasing emotional reactivity and improving psychological wellbeing in patients with bipolar disorder (62). A further study is needed to identify better strategies to decrease ED and NSSI in patients with mood disorders.

This study had several limitations. First, all measures were evaluated using self-reported questionnaires. In particular, the retrospective study may be associated with recall bias. However, a previous study showed that recall bias accounted for <1% of the reported variance in measures of childhood abuse (63). Second, since the data included patients with mood disorders only, it was not clear if the current study results apply only to mood disorders or the findings could be generalized to other clinical samples. Third, this study was performed as a cross-sectional and single-center study of patients who were treated in a single hospital. Therefore, prospective and multicenter studies are needed before the results can be generalized. Fourth, SEM showed that childhood trauma and ED only explained part of the mechanism of NSSI. Further studies that include diverse factors that contribute to the development of NSSI are needed.

In conclusion, NSSI is common across diverse mood disorders subtypes. Patients with BD-I had the lowest prevalence of NSSI and significantly less ED compared to those diagnosed with BD-II and MDD. The association between NSSI and childhood trauma was also mediated by ED. Thus, it could be beneficial to promote emotional coping skills in patients with mood disorders and a history of childhood trauma to prevent NSSI.

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 Institutional Review Board (IRB) of Samsung Medical Center. The patients/participants provided their written informed consent to participate in this study.

JB designed the study and wrote the protocol. JB and KH obtained the funding. DL, YC, and JA collected and managed the data. SY managed the literature searches and analyses. SY, DL, and HJ performed the statistical analysis, and SY and JB wrote the first draft of the manuscript. All authors contributed to the final manuscript and have approved it.

This work was supported by the Korea Medical Device Development Fund grant funded by the Korean government (the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, the Ministry of Health & Welfare, and the Ministry of Food and Drug Safety) [Project Numbers NTIS 9991006915 and KMDF_PR_20200901_0250] and the Original Technology Research Program for Brain Science through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT [No. 2019M3C7A1030624].

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.