Distinct correlation network of clinical characteristics in suicide attempters having adolescent major depressive disorder with non-suicidal self-injury

Abstract

Suicidal behavior and non-suicidal self-injury (NSSI) are common in adolescent patients with major depressive disorder (MDD). Thus, delineating the unique characteristics of suicide attempters having adolescent MDD with NSSI is important for suicide prediction in the clinical setting. Here, we performed psychological and biochemical assessments of 130 youths having MDD with NSSI. Participants were divided into two groups according to the presence/absence of suicide attempts (SAs). Our results demonstrated that the age of suicide attempters is lower than that of non-attempters in participants having adolescent MDD with NSSI; suicide attempters had higher Barratt Impulsiveness Scale (BIS-11) impulsivity scores and lower serum CRP and cortisol levels than those having MDD with NSSI alone, suggesting levels of cortisol and CRP were inversely correlated with SAs in patients with adolescent MDD with NSSI. Furthermore, multivariate regression analysis revealed that NSSI frequency in the last month and CRP levels were suicidal ideation predictors in adolescent MDD with NSSI, which may indicate that the increased frequency of NSSI behavior is a potential risk factor for suicide. Additionally, we explored the correlation between psychological and blood biochemical indicators to distinguish suicide attempters among participants having adolescent MDD with NSSI and identified a unique correlation network that could serve as a marker for suicide attempters. Our research data further suggested a complex correlation between the psychological and behavioral indicators of impulsivity and anger. Therefore, our study findings may provide clues to identify good clinical warning signs for SA in patients with adolescent MDD with NSSI.

Fig. 1. Heatmaps showing the Pearson correlation coefficients between all clinical indicators in all 130 participants with MDD and NSSI (FDR adjusted P < 0.05).

The parameters used for the correlation analyses in the heatmaps are listed in Supplementary Table 2. A Pairwise comparisons of all behavioral indexes and biochemical indicators are shown in color bars on the top, with the Pearson correlation coefficient R color-coded according to the color bar on the right. All 63 indicators included basic characteristics and biochemical and psychological indicators (see Supplementary Table 2). B A total of 438 correlations were detected (FDR adjusted P < 0.05) in the NSSI + SA group, and only 210 correlations were revealed in the NSSI group, with 174 correlations shared by the two groups. Distinct correlation profiles of the clinical characteristics of the NSSI + SA (C) and NSSI groups (D). MDD Major depressive disorder, NSSI + SA adolescent major depressive disorder with non-suicidal self-injury and suicide attempts, NSSI adolescent major depressive disorder with non-suicidal self-injury.

Fig. 2. Correlation heatmaps of the psychological and biochemical indicators.

Heatmaps presenting the Pearson correlation coefficients between the psychological and biochemical indicators in the NSSI + SA (A) and NSSI groups (B) among the participants with adolescent MDD and NSSI. Psychological indicators that significantly correlated with at least one biochemical indicator in either one or both groups are shown. The R values are color-coded based on the color bar on the right. *P < 0.05 and **P < 0.01. The asterisk in black indicates the specific correlations within each group, while the asterisk in gray represents the correlations shared between the two groups (P < 0.05). MDD Major depressive disorder, NSSI + SA adolescent major depressive disorder with non-suicidal self-injury and suicide attempts, NSSI adolescent major depressive disorder with non-suicidal self-injury.

Fig. 3. Scatter plots depicting the representative correlations between the psychological and biochemical indicators in the NSSI + SA and NSSI groups among participants with adolescent MDD and NSSI.

Distinct linear correlations are observed between CRP level and mature defense (A), CRP level and emotional neglect (B), CRP level and anger control-out (C), glucose concentration and physical neglect (D), LDL values and anger temperament (E), and ACTH level and anger expression-out (F). R and P stand for the Pearson correlation coefficient and the corresponding P value, respectively. MDD Major depressive disorder, NSSI + SA adolescent major depressive disorder with non-suicidal self-injury and suicide attempts, NSSI adolescent major depressive disorder with non-suicidal self-injury.

Fig. 4. Association networks of the NSSI + SA and NSSI groups.

Undirected graphs showing the common (A, B) and unique (C, D) characteristics of the association networks of the NSSI + SA and NSSI groups among participants with adolescent MDD and NSSI (FDR adjusted P < 0.05). The Pearson correlation coefficients are color-coded according to the color bar on the bottom right. MDD Major depressive disorder, NSSI + SA adolescent major depressive disorder with non-suicidal self-injury and suicide attempts, NSSI adolescent major depressive disorder with non-suicidal self-injury.

Fig. 5. Undirected graph of the differential association networks illustrating the differences between the common characteristics of the NSSI + SA and NSSI groups in participants with adolescent MDD and NSSI (FDR adjusted P < 0.05).

Differences are color-coded according to the color bar at the bottom right. MDD Major depressive disorder, NSSI + SA adolescent major depressive disorder with non-suicidal self-injury and suicide attempts, NSSI adolescent major depressive disorder with non-suicidal self-injury.