Journal of Affective Disorders 302 (2022) 204–213 Contents lists available at ScienceDirect Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad Research paper Borderline personality traits mediate the relationship between low perceived social support and non-suicidal self-injury in a clinical sample of adolescents Iria Mendez a, b, c, Anna Sintes d, Juan Carlos Pascual b, e, f, Joaquim Puntí b, g, h, Anaís Lara i, Laia Briones-Buixassa j, Stella Nicolaou k, l, Carlos Schmidt l, Soledad Romero a, b, c, Marta Fernández d, Cristina Carmona i Farrés b, e, f, Joaquim Soler b, e, f, Pilar Santamarina- Perez a, c, Daniel Vega f, l, * a Department of Child and Adolescent Psychiatry and Psychology, 2017SGR88, Institute of Neuroscience, Hospital Clínic de Barcelona, Barcelona, Spain b Centro de Investigación Biomédica en Red de Salud Mental. CIBERSAM, Madrid, Spain c Institut d’Investigacions Biomediques August Pi i Sunyer (IDIBAPS), Barcelona, Spain d Servicio de salud mental infantojuvenil, Hospital Sant Joan de Déu, Esplugues de Llobregat, Barcelona, Spain e Department of Psychiatry, Hospital de la Santa Creu i Sant Pau, Institut d’Investigació Biomèdica-Sant Pau (IIB-Sant Pau), Barcelona, Spain f Department of Psychiatry and Forensic Medicine, Institute of Neurosciences, School of Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain g Hospital de Dia de Adolescentes. Servicio de Salud Mental. Consorci Corporació Sanitària Parc Taulí, Sabadell, Barcelona, Spain h Departamento de Psicología Clínica y de la Salud, Universitat Autònoma de Barcelona, Barcelona, Spain i Servicio de Psiquiatría y Psicología, ALTHAIA, Xarxa Assistencial i Universitària de Manresa, Barcelona, Manresa, Spain j Mental Health and Social Innovation Research Group and Centre for Health and Social Care Research, Universitat de Vic – Universitat Central de Catalunya, Catalonia, Spain k PhD Programme in Biomedicine, University of Barcelona, Spain l Psychiatry and Mental Health Department, Consorci Sanitari de l’Anoia & Fundació Sanitària d’Igualada, Hospital Universitari d’Igualada. Av. Catalunya 11, Barcelona, Igualada 08700, Spain A R T I C L E I N F O A B S T R A C T Keywords: Background: Non-suicidal self-injury (NSSI) is a serious public health concern among adolescents, especially in Borderline personality disorder clinical settings. Social support plays a critical role in the onset and maintenance of NSSI in adolescence. NSSI is Non-suicidal self-injury closely associated with borderline personality disorder (BPD), yet no previous work has analyzed the mediating Self-harm role of borderline traits in the relationship between perceived social support (PSS) and NSSI. This study aimed to adolescents address this gap. Perceived social support Methods: Participants were 228 adolescent patients (12 to 18 years old), who completed a clinical interview and self-report measures of BPD-traits, current psychological distress, emotion dysregulation and PSS. They were grouped based on the presence (vs. absence) of NSSI. Univariate and multivariate logistic regression analyses were used to identify risk factors of NSSI, and a mediation analysis was conducted to examine the intermediary role of borderline traits in the relationship between PSS and NSSI. Results: NSSI was highly prevalent in our sample (58%) and was associated with higher clinical severity. Low PSS predicted NSSI in univariate, but not multivariate regression. Mediation analyses showed that borderline traits fully accounted for the relationship between low PSS and NSSI, even when controlling for current psychological distress and gender. Limitations: Cross-sectional design through self-report assessment. Conclusions: Findings suggest that adolescents with low PSS are especially vulnerable for developing NSSI due to elevated BPD traits. In clinical settings, interventions aimed to reduce borderline symptoms may be a promising treatment option for adolescents with NSSI and low PSS. * Corresponding author at: Psychiatry and Mental Health Department, Consorci Sanitari de l’Anoia & Fundació Sanitària d’Igualada, Hospital Universitari d’Igualada. Av. Catalunya 11, Barcelona, Igualada 08700, Spain. E-mail address: daniel.vegamo@gmail.com (D. Vega). https://doi.org/10.1016/j.jad.2022.01.065 Received 31 August 2021; Received in revised form 11 January 2022; Accepted 13 January 2022 Available online 15 January 2022 0165-0327/© 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). I. Mendez et al. Journal of Affective Disorders 302 (2022) 204–213 1. Introduction protective factor for suicide and self-harm (Kleiman and Liu 2013; Muehlenkamp et al., 2013; Nemati et al., 2020; Tham et al., 2020). More Non-suicidal self-injury (NSSI) refers to the deliberate destruction of specifically, different previous studies show that high PSS decreases the one’s own body tissue in the absence of conscious suicidal intent (Nock, odds of engaging in NSSI in adolescents from nonclinical (Wan et al., 2010; Plener et al., 2015; Vega et al., 2018). NSSI is associated with 2018) and clinical samples (Baiden et al., 2017; Miller et al., 2015). A potentially detrimental consequences, ranging from low interpersonal, recent longitudinal study with a 5-year follow-up also suggests that high academic, and daily functioning to high risk of suicide (Claes et al. 2003; PSS from friends may be particularly critical in preventing the onset of Daukantaitė et al. 2020). Adolescence constitutes a critical and vulner­ self-cutting in adolescents (Rissanen et al., 2021). able phase for developing NSSI, as indicated by higher prevalence rates It should also be noted that young people with low PSS are prone to in both community (17–38%; Brunner et al., 2014) and clinical samples use the social media as an alternative source to feel connected with (50–60 %; Kaess et al., 2013). Recent research suggests a crucial role of others (Zhang et al., 2018). However, the reality is that overuse of the social factors in the onset and maintenance of NSSI during this impor­ internet and the social media has been associated with psychological tant developmental period (Brown and Witt, 2019). Therefore, it is not distress in adolescents (Marino et al., 2018). More specifically, internet surprising that in the context of the social distancing measures imposed addiction might be a risk factor for engaging in NSSI in the presence of a during the COVID-19 pandemic, NSSI rates have increased even more, comorbid psychiatric disorder (Karaer and Akdemir, 2019; Mészáros especially in those adolescents suffering from mental disorders (Plener, et al., 2020), suggesting that virtual interaction does not improve per­ 2021; Wester et al. 2018). In this regard, a deeper understanding of how ceptions of social support in these adolescents. social factors contribute to the prevalence and severity of NSSI in It is worth noting that PSS may interact with other risk factors to adolescent psychiatric samples, as well as how they interplay with other augment or reduce the odds of engaging in NSSI in adolescents. For risk factors, is necessary to maximize prevention and treatment example, a large-scale study involving ten European countries showed programs. that adolescents with parental and peer support were at significantly In adolescents, NSSI is present in broader internalizing and exter­ lower risk of engaging in direct self-injurious behaviors after being nalizing problems (Meszaros et al., 2017; Zetterqvist, 2015), and is victimized by bullying, compared to students without social support highly associated with borderline personality disorder (BPD) (Nock (Klomek et al., 2016). Similarly, high parental support appears as a et al., 2006; Sadeh et al., 2014). Around 61% of adolescents with BPD protective factor against NSSI in adolescents who are bullied (Claes pathology have engaged in NSSI at least once, making the criterion et al., 2015). Lastly, Crowell et al. (2009) also suggested that the asso­ “recurrent NSSI and suicidal behavior” the most frequently met in ad­ ciation between NSSI and emotion dysregulation becomes stronger in olescents with BPD (Chanen et al., 2008; Kaess et al., 2014). The non-supportive social contexts, further emphasizing the protective role occurrence of NSSI often precedes a BPD diagnosis and is therefore of social support. considered to be a precursor for BPD (Ghinea et al., 2019; Groschwitz While a significant body of research has demonstrated the protective et al., 2015; Homan et al., 2017). On the other hand, the presence of BPD effects of social support on NSSI in adolescents, only a few studies have symptoms (such as emotion dysregulation or identity disturbances) in­ explored the mediating mechanisms that may underlie this association creases the risk of engaging in some form of NSSI in adolescents (Stead in adolescents. Rissanen et al. (2021) reported that depression symp­ et al., 2019), and the presence of a BPD diagnosis involves greater toms mediated the effect of subjective loneliness on the initiation of severity of NSSI (Jacobson et al., 2008; Stead et al., 2019). Therefore, it self-cutting in a non-clinical sample of adolescents. Similarly, Madjar has been suggested that BPD features precede and interact with NSSI et al. (2021) found that depression symptoms fully mediated the asso­ (Reichl and Kaess, 2021). Notwithstanding the above, additional ciation between perceptions of school-related social support (e.g., from research is needed to clarify this relationship, as approximately 50% of teachers and peers) and NSSI in a non-clinical adolescent sample. Lastly, adolescents with NSSI do not meet diagnostic criteria for BPD (Ayodeji a different study in young adults revealed that deficits in emotion et al., 2015). regulation mediated the relationship between feelings of alienation in Prior research indicates that emotion dysregulation, a core feature of parent and peer relationships (i.e., low PSS) and NSSI (Yurkowski et al., BPD, is a specific risk factor for NSSI (Wolff et al., 2019). For example, 2015). emotion regulation difficulties detect the presence of NSSI in psychiatric These studies only focused on the mediating role of emotion dysre­ inpatients, regardless of sex and psychopathology (Perez et al., 2012). In gulation and depressive symptoms in the relationship between PSS and addition, adolescents commonly report engaging in self-harm as a NSSI. However, despite a strong association between NSSI and BPD (e.g., strategy to cope with emotional pain (Laye-Gindhu and Schoner­ Stead et al., 2019), no previous study has analyzed the potential medi­ t-Reichl, 2005). In this regard, NSSI could be considered as a maladap­ ating role of borderline features. In this regard, it is important to note tive response to strong negative affect because of poor emotion that unsupportive and invalidating environments during infancy and regulation skills (Chapman et al., 2006). childhood (e.g., disturbed relationships, bullying) have been suggested In addition to individual risk factors (e.g., emotion dysregulation), as a prominent risk factor for the development of BPD (Crowell et al., social factors also play a prominent role in NSSI engagement (Cipriano 2009; Fruzzetti et al., 2005). However, longitudinal studies do not et al., 2017; Taylor et al., 2018). Most people engaging in NSSI usually provide consistent evidence to support this causal relationship. For endorse social motives (e.g., to cope with social conflicts) (Hankin and instance, while a study of Wolke et al. (2012) suggests that victimisation Abela, 2011; Klonsky, 2007). For instance, interpersonal conflicts or by peers in childhood increases the risk of developing BPD symptoms in feelings of rejection by peers are frequent triggers for NSSI (Brio­ early adolescence, another study indicates reciprocal associations be­ nes-Buixassa et al., 2021; Klomek et al., 2016). Importantly, adolescents tween invalidating parenting behaviors (e.g., low caregiver warmth) who engage in NSSI often report feelings of loneliness, low social and BPD symptoms in adolescence (Stepp et al., 2014). Importantly, competence, poorer relationships with parents, and invalidating life these studies assessed constructs (e.g., low caregiver warmth) related to environments (Musser et al., 2018; Wang et al. 2020). social support, but they have not directly assessed PSS. Taken together, Social interactions and social support are vital for healthy person­ prior evidence does not allow us to establish temporal precedence in the ality development in young people (Blakemore and Mills, 2014; Orben relationship between PSS and borderline features in adolescence. et al., 2020). Perceived social support (PSS) refers to an individual’s To address these gaps, the current study aimed to examine whether perceptions of the general availability and quality of the social support low PSS predicts NSSI in a Spanish clinical sample of adolescents. For the available to them (Haber et al., 2007). Adolescents with high PSS exhibit first time to our knowledge, we also explored whether borderline per­ better mental health and less high-risk behaviors (Lai and Ma, 2016; sonality traits mediate this relationship in such a clinical sample. Based Reininger et al., 2012). Previous evidence suggests that high PSS is a on previous research, we hypothesized that adolescents who engage in 205 I. Mendez et al. Journal of Affective Disorders 302 (2022) 204–213 NSSI would report lower PSS than adolescents without NSSI. Addition­ Finally, 83 participants (or parents) rejected to participate, and 71 ally, we hypothesized that psychological distress (i.e., depression and participants were excluded (74.65% were incomplete assessments). The anxiety) and emotion dysregulation would mediate the association be­ final sample of study comprised 228 adolescents. Participants were tween PSS and NSSI. Lastly, given the well-established relationship be­ divided in two groups as a function of a history of NSSI (N= 132: NSSI tween BPD and NSSI, as well as the negative bias in the processing of group; N= 96: Non-NSSI group). Demographic characteristics are pro­ social information previously found in BPD patients (Kleindienst et al., vided in Table 1. Ethics approval was issued by the Bellvitge’s Hospital 2019; Reichenberg et al., 2017), we also hypothesized that borderline Ethics Committee (PR330/17). personality traits would appear as an important mediator in the asso­ ciation between PSS and NSSI. However, given that the present study is 2.2. Procedure cross-sectional and there is no prior evidence that firmly supports the specific ordering of the variables in this mediation model, we also First, a trained senior member of the study staff gathered clinical explored a competing hypothesis (i.e., a plausible alternative model information (see Table 1). All participants were assessed with the where PSS mediates the association between BPD features and NSSI).” Spanish version of the Mini International Neuropsychiatric Interview Kid (M.I.N.I. Kid) (Colón-Soto et al., 2005) (see Table S1). After that, 2. Methods participants answered both demographic questions and self-reported questionnaires through an online platform under the supervision of a 2.1. Participants researcher. This platform also included questions about: (i) healthy habits, (ii) usage of internet and social media, and (iii) lifetime NSSI (yes The present study was conducted between April 2019 and February or not). 2020 in partial and complete hospitalization programs from six hospitals in the county of Barcelona (Spain): Consorci Sanitari de l’Anoia, Hos­ pital Clínic de Barcelona, Hospital de la Santa Creu i Sant Pau, Hospital 2.3. Measures Sant Joan de Déu de Barcelona, Althaia Manresa, and Consorci Corpo­ ració Sanitària Parc Taulí. By means of a recruiting sampling method, 2.3.1. Multidimensional Scale of Perceived Social Support (MSPSS) (Zimet 382 consecutive patients were invited to participate. Inclusion criteria et al., 1990) comprised ages between 12 and 18 years old, and a written informed We used the Spanish version of the MSPSS (www.heardalliance.org), consent form signed by both patients and their parents or legally which is one of the most broadly used scales to rate social support. It authorized representatives. Exclusion criteria included clinically sig­ consists of 12 items, which assess three sources of personal support: (i) nificant behavior problems that impede the assessment, failure to family (e.g., “my family really tries to help me”), friends (“my friends complete all phases of the evaluation, and IQ below 70. For the NSSI really try to help me”), and significant others (e.g., “there is a special group, participants reporting NSSI on less than five days in the past year person who is around when I am in need”). A higher score in this scale were also excluded from the study. indicates higher PSS. Both scores from the original and Spanish version of the MSPSS show good internal reliability (α = .88 and α = .90 Table 1 Sociodemographic characteristics of the sample. NSSI (N = 132) Non-NSSI (N = 96) Analysis Mean SD Mean SD Z/t p Age 15.46 1.30 15.17 1.62 -1.10 .270 Lifetime Suicide Attempts .73 1.76 .20 .72 -4.40 <.001 Social media use Item 1 2.46 1.07 1.85 .86 -4.75 <.001 Item 2 1.86 .91 1.69 .86 -1.48 .141 Item 3 1.88 .99 1.60 .84 -2.19 .029 Item 4 1.59 .89 1.44 .75 -1.41 .161 Item 5 2.05 1.04 1.78 .82 -2.13 .034 n % n % χ2 p Sex (female) 104 68 49 32 19.32 <.001 Academic Performance Bad 48 36.36 18 18.75 9.97 .007 Regular 36 27.27 26 27.08 Good 48 36.36 52 54.17 Internet use > 20 hours per week 47 35.61 25 26.04 2.35 .125 <10% of time for study 57 43.51 27 29.03 5.41 .020 Daily hours of Sleep < 6 hours 38 28.79 8 8.33 14.43 <.001 Meals per day < 4 food intakes 63 47.7 27 28.1 8.94 .003 Alcohol abuse Never 66 50 72 75 14.74 .001 1-3 days 45 34.1 15 15.6 ≥ 4 days 21 15.9 9 9.4 Lifetime Suicide Attempts ≥ 1 attempts 52 39.50 11 11.50 21.69 <.001 Note. Problematic social media use was captured by the following items: Item 1 = ‘do you prefer to connect with people over the Internet than in person?’; Item 2 = ‘has your academic performance been adversely affected by social network usage?’; Item 3 = ‘do you suffer from sleep disorders due to aspects related to mobile phones?’; Item 4 = ‘do you feel restless when you don’t receive messages or calls?’; Item 5 = ‘how often do you say things on your cell phone that you wouldn’t say in person?’. Participants self-reported their academic performance (bad, regular, and good). Internet use refers to the total number of hours per week spent surfing on the Internet, and the % of that time which was dedicated to study tasks. Alcohol abuse episodes in the last year (never, 1 to 3 days, more than 4 days). 206 I. Mendez et al. Journal of Affective Disorders 302 (2022) 204–213 respectively) (Calderón et al., 2021; Zimet et al., 1990). Suitable psy­ investigate between-group differences (NSSI vs. Non-NSSI), we con­ chometric properties have been shown in different populations, ducted Pearson’s chi-squared tests for categorical variables, including adolescents (Canty-Mitchell and Zimet, 2000). independent-samples t-tests for dimensional variables that met the assumption of normality, and Mann-Whitney U tests for dimensional 2.3.2. The borderline personality questionnaire (BPQ) (Fonseca-Pedrero variables that violated the assumption of normality. et al., 2011) This questionnaire assesses BPD traits or symptoms, based on the 2.4.1. Logistic regression analysis DSM-IV criteria. It consists of nine subscales: Impulsiveness, Affective The outcome was a dichotomized variable indicating the presence (≥ Instability, Abandonment, Relationship, Self-Image, Suicide/Self-Muti­ 5 days) or absence of NSSI in the past year, and the explanatory variables lation, Emptiness, Intense Anger, and Quasi-Psychotic States. BPQ is a were age, gender, perceived social support (MSPSS-Family, MSPSS- good screening tool to detect BPD in outpatient youth (Chanen et al., Friends, MSPSS-Significant Others, MSPSS-Total), borderline traits 2008). In the current study, we used the total BPD-score as a measure of (BPQ), current psychological distress (DASS-21), and difficulties in general BPD features, but we excluded the Suicide/Self-Mutilation emotion regulation (DERS-subscales). We first ran univariate logistic subscale to avoid a potential confounding effect in the analysis. regression on all predictors, producing odds ratios (OR) presented with 95 % confidence intervals. Predictors that showed a statistically signif­ 2.3.3. Brief Version of the Difficulties in Emotion Regulation Scale (DERS- icant effect in the univariate analyses were entered in a multivariate 18) (Victor and Klonsky, 2016) logistic regression analysis, using a backward stepwise selection to We used the Spanish translation of this scale (Briones-Buixassa et al., obtain the final model. Before running the multiple regression analysis, 2021). This 18-item self-report measure assesses 6 subscales of emotion we checked for multicollinearity by calculating the variation inflation regulation difficulties: (i) lack of emotional awareness, (ii) lack of factor (VIF) for all predictors entered in the model. All VIF values were emotional clarity, (iii) nonacceptance of emotions, (iv) inability to below 3, indicating that there are no multicollinearity problems (Hair engage in goal-directed behavior when feeling emotional, (v) engage­ et al., 2010). Receiver operating characteristic (ROC) curve analysis was ment in impulsive behavior when feeling emotional, and (vi) inability to also conducted to determine the capacity of the final model to access emotion regulation strategies. The total DERS-score presents a discriminate between patients with and without a history of NSSI. high internal reliability, and the six scales show moderate to high in­ ternal reliability (Victor and Klonsky, 2016). 2.4.2. Mediation analysis We finally conducted a mediation analysis to assess whether 2.3.4. Depression anxiety stress scales (DASS-21) (Arturo et al., 2005) borderline traits (BPQ) account for the relationship between perceived The DASS-21 is a self-report instrument that assesses current psy­ social support and NSSI and, alternatively, whether perceived social chological distress. It measures the frequency of 21 negative emotional support may account for the relationship between borderline traits and symptoms related to Depression, Anxiety, and Stress during the previous NSSI. The “Mediation” package in R (Tingley et al., 2014) was used to week. The Spanish version presents moderate to good internal reliability test the theorized mediation relationship with 10000 bootstrapped for each of the three subscales, respectively (Arturo et al., 2005). Ad hoc samples. Unstandardized indirect effects were computed for each of questionnaire assessing internet addiction and mobile (Beranuy et al., 10000 bootstrapped samples, and the 95% confidence interval was 2009). Participants were asked to indicate their Internet and mobile computed by determining the indirect effects at the 2.5th and 97.5th usage. They responded to the following 5 items by choosing one of four percentiles. Unstandardized indirect effects were computed using the answers: (‘almost never’, ‘sometimes’, ‘most times’, ‘almost always’): (1) difference between coefficients approach (c – c´), that is, the difference ‘do you prefer to connect with people over the Internet than in person?’, between the total and direct effect. To ensure that the indirect effects (2) ‘has your academic performance been adversely affected by social estimated with this approach are not affected by a change of scale in the network usage?’, (3) ‘do you suffer from sleep disorders due to aspects logistic regression models (i.e., our mediation is based on logistic related to mobile phones?’, (4) ‘do you feel restless when you don’t regression), we repeated all mediation analyses by standardizing the receive messages or calls?’, (5) ‘how often do you say things on your cell coefficients before estimating the indirect effect (Rijnhart et al., 2019). phone that you wouldn’t say in person?’. These questions were obtained from two previous questionnaires in the Spanish language (Beranuy 3. Results et al., 2009). Only participants with NSSI (NSSI group) completed the following 3.1. Sample characteristics scales: Participants were between 12 and 18 years old at the time of consent 2.3.5. Non-suicidal self-injury disorder scale (NSSIDS) (Victor et al., (Mage = 15.34, SD= 1.45), and were mostly females (67.1%) (Table 1). 2017) As shown in Table 1, the presence (vs. absence) of NSSI was associ­ The NSSIDS is a self-report measure that assesses whether partici­ ated with a worse academic performance, less hours of sleep, more pants meet criteria for the DSM-5 NSSI disorder. We used a Spanish alcohol abuse, lower food intake, and more problematic use of the translation to test whether the ‘criterion A’ was met (i.e., five or more internet and the social media. Specifically, participants in the NSSI days in the past year in which participants engaged in NSSI). group showed a significantly stronger preference to connect with others via the internet and online social networks rather than in person, and 2.3.6. Inventory of statements about self-injury (ISAS) (Klonsky and they recognized sleep disturbances due to social media use more Glenn, 2009) frequently (Table 1). In addition, only 62.1% (n=82) of participants in This inventory assesses lifetime frequency of 12 NSSI behaviors (e.g., the NSSI group met the full criteria for the NSSI disorder (based on the cutting) and 13 NSSI functions. These functions are grouped into two NSSIDS). This result is similar to previous findings in clinical samples of separate factors that index interpersonal (e.g., interpersonal influence) adolescents (e.g., Washburn et al., 2015). and intrapersonal NSSI functions (e.g., affect-regulation). We adminis­ Compared to participants without NSSI, those engaging in NSSI tered the Spanish version of the ISAS (Vega et al., 2017). exhibited more depressive and anxiety disorders, and higher suicide risk according to the MINI-KID classification. Moreover, although most 2.4. Statistical analyses participants in the non-NSSI group had 1 or 2 diagnoses, most partici­ pants in the NSSI group exhibited three or more clinical diagnoses All analyses were conducted in RStudio (R Core Team, 2021). To (Table S1). Importantly, the average number of lifetime suicide attempts 207 I. Mendez et al. Journal of Affective Disorders 302 (2022) 204–213 was higher in the NSSI group than in the non-NSSI group. Specifically, Table 3 there were significantly more patients reporting one or more suicide Univariate and multivariate logistic regression analyses examining predictors of attempts in the NSSI group vs the non-NSSI group (see Table 1). NSSI. Collectively, these results indicate that the presence of NSSI entails Univariate Analyses Multivariate Analyses greater clinical severity and higher suicide risk. OR (95% CI) P-value OR (95% CI) P Adolescents who engage in NSSI reported that the mean age of onset Age 1.15 (.96-1.38) .12 was 12.76 years (SD = 2.11). The most frequent types of NSSI were self- Sex 3.56 (2.00- < .001 1.96 (.97-3.93) .05 cutting (79.4%), hitting (74.8%), and scratching (71%) oneself. Results 6.35) MSPSS from the ISAS revealed that adolescents from the NSSI group incurred in Family .91 (.87-.95) < .001 NSSI mainly with an intrapersonal motivation (intrapersonal vs. inter­ Friends .95 (.92-.99) < .05 personal NSSI-function; Mintrapersonal = 3.41, SD = 1.43; Minterpersonal = Significant others .99 (.95-1.03) .54 .88, SD = .85; Z = -8.30, P < .001). MSPSS Total .98 (.96-.99) < .01 Table 2 depicts between-group differences (NSSI vs. Non-NSSI) in the BPQ 1.11 (1.08- < .001 1.07 (1.03- < .001 1.15) 1.11) self-report measures. Participants with NSSI showed lower scores in DASS-21 1.09 (1.06- < .001 1.05 (1.02- < .01 perceived social support and higher scores in borderline traits, emotion 1.11) 1.08) dysregulation, and psychological distress, which further suggests higher DERS-18 clinical severity in the presence of NSSI. Awareness 1.14 (1.04- < .01 1.24) Clarity 1.22 (1.12- < .001 3.2. Logistic regression analysis 1.32) Goals 1.22 (1.12- < .001 Univariate logistic regression analyses showed that participant´s 1.33) Impulse 1.21 (1.13- < .001 gender, perceived social support (except the "significant other" MSPSS 1.31) subscale), borderline traits (BPQ), current psychological distress (DASS- Non-acceptance 1.24 (1.15- < .001 21), and difficulties in emotion regulation (all DERS-subscales) signifi­ 1.34) cantly predicted the odds of engaging in NSSI (Table 3). However, when Strategies 1.31 (1.20- < .001 all predictors were tested simultaneously in a multivariate regression, 1.43) DERS total 1.08 (1.06- < .001 the final model obtained via backward stepwise selection showed that 1.11) only borderline traits and current psychological distress significantly AIC = 226.08 predicted NSSI, while gender was maintained in the model with a Note. OR = Odds Ratio. MSPSS = Multidimensional Scale of Perceived Social marginal significance (see Table 3). Support. BPQ = Borderline Personality Questionnaire. DASS-21 = Depression Receiver operating characteristic (ROC) analysis showed that the Anxiety and Stress Scales. DERS-18 = Brief version of the Difficulties in Emotion Area Under the Curve (AUC) of the final multivariate model was .84 Regulation Scale. (95% CI: .79-.89; Sensitivity = .92; Specificity = .62) and the OR was Table 2 Between-group differences in borderline traits, perceived social support, current psychological distress and emotion dysregulation. NSSI (N=132) Non-NSSI (N=96) Analysis Effect Size Mean SD Mean SD W/t r/d BPQ Suicide 4.74 1.29 1.19 1.62 -11.40** .78 Impulsivity 3.15 1.60 2.12 1.44 -5.00** .33 Self-Image 5.80 2.49 3.22 2.75 -6.59** .44 Emptiness 7.31 2.51 4.16 2.75 -7.80** .52 Affective Instability 7.54 2.04 5.62 2.58 -5.64** .37 Intense Anger 5.72 2.80 4.31 2.60 -3.75** .25 Quasi-Psychotic States 2.49 1.88 1.05 1.22 -6.03** .40 Abandonment 4.96 2.68 2.59 2.04 -6.52** .43 Relationships 4.33 2.42 2.74 2.03 -4.86** .32 BPQ Total 41.31 11.81 25.80 11.69 -9.83** 1.32 MSPSS Family 18.64 6.85 22.47 6.30 -4.56** .30 Friends 18.92 7.47 21.30 6.68 -2.58* .17 Significant Other 21.22 7.08 21.80 7.46 .89 MSPSS Total 58.79 17.47 65.57 16.26 -3.08* .20 DASS-21 Depression 15.31 4.98 8.38 5.70 -8.10** .54 Anxiety 13.29 5.64 7.24 5.18 -7.22** .48 Stress 13.57 4.60 9.31 5.14 -5.98** .40 DASS total 42.17 13.13 24.94 14.42 -7.84** .52 DERS-18 Awareness 9.73 2.86 8.57 3.22 -2.71* .18 Clarity 10.14 3.26 7.76 3.59 -4.90** .33 Goals 12.07 3.02 9.77 3.66 -4.74** .32 Impulse 9.95 3.98 7.04 3.56 -5.29** .35 Non-acceptance 9.40 4.00 6.29 3.34 -5.66** .38 Strategies 10.17 3.54 6.71 3.35 -6.68** .44 DERS total 61.45 12.80 46.15 13.94 -8.59** 1.15 Note. BPQ = Borderline Personality Questionnaire. MSPSS = Multidimensional Scale of Perceived Social Support. DASS-21 = Depression, Anxiety and Stress Scale. DERS-18 = Brief version of the Difficulties in Emotion Regulation Scale. W = Mann-Whitney test. t = Student’s t-test. r = effect size measure for the Mann-Whitney test. d = Cohen’s d effect size. * p < .01, **p < .001. 208 I. Mendez et al. Journal of Affective Disorders 302 (2022) 204–213 20.33 (95% CI: 9.45-43.73, p < .001), indicating a high capacity to explored as potential mediators. discriminate between patients with and without NSSI (see Fig. 1). Finally, as depicted in Figure S1, the mediation results we obtained using the standardized regression coefficients are consistent with the 3.3. Mediation analysis results we obtained using the unstandardized coefficients. In addition, given the cross-sectional design of our data, we analyzed alternative Given that the effect of PSS on NSSI disappeared when borderline models with BPD-traits as the predictor and PSS as the mediator, but the traits was included in the regression, we conducted two mediation an­ indirect effect was not statistically significant (see the Supplementary alyses to examine whether this variable could account for the relation­ Material). ship between: (i) MSPSS-Family and NSSI, and (ii) MSPSS-Friends and NSSI. That is, whether the relationship between low perceived social 4. Discussion support and NSSI could be explained by a negative association of PSS with borderline traits. Given that gender and current psychological In the current study, we evaluated the association between perceived distress (DASS-Total) significantly (or marginally) predicted NSSI in the social support (PSS) and NSSI in a psychiatric sample of adolescents, and final multivariate model, they were included in the analysis as we also explored potential mediators of this relationship. Specifically, covariates. and in a novel way, we analyzed the mediating role of borderline per­ The effect of perceived social support from family (MSPSS-Family) sonality traits. The most relevant results of this study suggest that low on NSSI was fully mediated via borderline traits (BPQ). The boot­ PSS from family and peers was indirectly related to NSSI through strapped estimate for the average causal mediation effect (ACME) was borderline personality traits, even when controlling for gender and -.004 (95% CI: -.006 to -.001). Thus, the indirect effect of MSPSS-Family current psychological distress. This finding suggests that adolescents on NSSI via BPQ was statistically significant, even when controlling for with low PSS may be especially vulnerable for developing NSSI due to current psychological distress (DASS-21) and gender. The total effect, elevated BPD traits. that is, the effect of MSPSS-Family on NSSI without controlling for the The prevalence of NSSI in our sample (58%) was similar to previous effect of the mediator, was estimated to be -.008 (95% CI: -.011 to -.001) studies with clinical samples (Kaess et al., 2013). Importantly, our and was also statistically significant. However, the average direct effect findings suggest that the presence of NSSI is associated with high clinical (ADE), that is, the effect of MSPSS-Family on NSSI after controlling for severity and suicide risk, as well as with low interpersonal, academic, the effect of the mediator, was -.005 (95% CI: -.008 to 0.00) and did not and daily functioning, regardless of the diagnosis. This finding is in line remain significant, indicating a full mediation (Fig. 2A). with previous research suggesting that NSSI may be an unspecific pre­ Similarly, the effect of perceived social support from friends (MSPSS- cursor of psychopathology and suicide (Groschwitz et al., 2015; Mes­ Friends) on NSSI was fully mediated via borderline traits (BPQ). The zaros et al., 2017). Furthermore, adolescents engaging in NSSI reported bootstrapped estimate for the average causal mediation effect (ACME) lower PSS from family and friends and exhibited a stronger preference to was -.006 (95% CI: -.000 to -.001), indicating that the indirect effect was interact virtually (e.g., connect with peers over the internet that in significant, even when controlling for current psychological distress person), compared to adolescents without NSSI. These results are (DASS-21) and gender. The total effect was not statistically significant, congruent with recent work indicating a relationship between PSS and -.002 (95% CI: -.008 to .01), and neither was the average direct effect risk of engaging in NSSI in clinical samples (Baiden et al., 2017; Miller (ADE), 004 (95% CI: -.003 to 0.01), indicating a full mediation (see et al., 2015). Moreover, consistent with our findings, previous studies Fig. 2B). have suggested that individuals with low PSS tend to use online social To evaluate the specificity of borderline traits as a mediator in the networks to compensate for the lack of social support experienced in the relationship between perceived social support and NSSI, we conducted ‘real world’ (Tang et al., 2021). two additional mediation analyses exploring current psychological Previous research shows that lower social support from family and distress (DASS-21) as a mediator (controlling for borderline traits and peers plays a crucial role in the onset and maintenance of NSSI (Brown gender). The bootstrapped estimate for the average causal mediation and Witt, 2019; Rissanen et al., 2021; Victor et al., 2019). The present effect was not significant either for the mediation analysis with MSPSS- study extends this evidence, providing data in a well-characterized Family as the independent variable (ACME = -.0005 [95% CI: -.001 to clinical sample of adolescents. Specifically, in our sample, adolescents .00]) or for the mediation analysis with MSPSS-Friends as the inde­ with NSSI perceived less support from family and friends than their pendent variable (ACME = .001 [95% CI: -.0001 to .00]). Therefore, counterparts without NSSI did, suggesting that PSS from very close re­ perceived social support from family and friends only exerts an indirect lationships (i.e., family and friends but not from significant others) is a effect on NSSI through borderline traits, but not through current psy­ factor involved in NSSI. On the one hand, this finding may suggest an chological distress. Since none of the emotion dysregulation dimensions important role of interpersonal stress in triggering NSSI. In the context of remained significant in the multivariate model, they were not further low PSS, there may be a higher frequency of interpersonal stressors, which may enhance negative affect and promote NSSI as an emotion regulation strategy (Nock, 2010). On the other hand, the relationship between low PSS and NSSI may also indicate that adolescents who engage in NSSI may feel rejected or stigmatized by family and friends due to their self-harm behavior (Piccirillo et al., 2020), and may therefore feel less supported by their central social network. We conducted a multivariate analysis to better understand the rela­ tionship between PSS and NSSI, which revealed that borderline traits and psychological distress significantly predicted NSSI, while PSS did not. This finding suggests that when tested simultaneously, clinical status is a better predictor of NSSI than PSS. Importantly, the NSSI predictor with the highest odds ratio in the multivariate model was borderline traits. This is consistent with a longitudinal study reporting that BPD features (along with NSSI severity) was the best predictor of continued engagement in NSSI one year later (Glenn and Klonsky, 2011), as well as with prior research showing that BPD-symptoms are Fig. 1. ROC curve of the final logistic regression model, AUC. associated with an increased risk of engaging in NSSI (Reichl and Kaess, 209 I. Mendez et al. Journal of Affective Disorders 302 (2022) 204–213 Fig. 2. Borderline traits fully mediate the relationship between perceived social support from: (A) family (MSPSS-Family) and (B) Friends (MPSS-Friends), and the probability of engaging in non-suicidal self-injury (controlling for gender and current psychological distress). BPQ = Borderline Personality Questionnaire. MPSS = Multidimensional Scale of Perceived Social Support. DASS-21 = Depression, Anxiety and Stress Scale. * p < .05; ** p < .01; *** p < .001. 2021; Stead et al., 2019). In addition to borderline traits, current psy­ depressive symptoms), current psychological distress did not signifi­ chological distress also appeared as a significant predictor of NSSI in our cantly mediate this relationship when controlling for borderline traits. sample. Given that this predictor includes symptoms of anxiety and This result shows the relevance of long-lasting traits versus state depression in the last two weeks, our results are consistent with prior symptoms in the relationship between PSS and NSSI. This finding may findings in both adolescents (Giletta et al., 2012) and young adults be inconsistent with prior studies in non-clinical samples that highlight (Kiekens et al., 2020) suggesting that higher levels of depressive the mediating role of depressive symptoms (Madjar et al., 2021; Rissa­ symptoms in the past weeks predict NSSI (Fox et al., 2015). In contrast nen et al., 2021). However, those studies did not control for borderline with our hypothesis, emotion dysregulation did not remain as a signif­ symptomatology, possibly because the latter tends to be more common icant predictor of NSSI in the multivariate model. in clinical rather than non-clinical adolescent samples (Hutsebaut and Finally, we conducted a mediation analysis to explore if clinical Aleva, 2021; Reichl and Kaess, 2021). status (i.e., BPD traits and current psychological distress) intermediates The current study highlights the role of interpersonal factors in between PSS and the odds of engaging in NSSI. This analysis revealed adolescent NSSI, adding to existing research in clinical samples that has that borderline traits fully explained the relationship between PSS (from thus far focused on comorbid psychopathology, stressful life events and family and friends) and NSSI, even when controlling for current psy­ neglected social factors (Greenfield et al., 2015; Skabeikyte and Bar­ chological distress (i.e., symptoms of depression, anxiety, and stress). kauskiene, 2021; Strandholm et al., 2017). Despite promising findings, This would mean that, low PSS is associated with NSSI through higher the present study is not without limitations. First, due to our levels of BPD traits. This finding is consistent with a recent study sug­ cross-sectional design, we cannot establish temporal precedence in the gesting that high-perceived support from parents predicted fewer BPD relationships proposed in the mediation analyses (e.g., whether PSS symptoms in a psychiatric sample of adolescents, while invalidating predicts BPD traits or vice versa). Even though we tested a plausible, behaviors from parents were associated with greater BPD symptoms in alternative model with BPD traits as the predictor and PSS as the daily life (Vanwoerden et al., 2021). Congruently, recent research shows mediator and found that the hypothesized indirect effect there did not that adolescents with BPD perceive their caregivers to be less supportive materialize, our cross-sectional design does not allow us to discard this and more invalidating than did controls without BPD, suggesting that alternative model, and the matter will ultimately have to be investigated their perceptions of caregiving behaviors are vital in the context of in longitudinal studies. Second, our data were collected through enduring borderline psychopathology (Bennett et al., 2019). Research in self-report questionnaires, which may entail retrospective memory non-clinical adult samples also indicates that BPD features are associ­ biases. Third, our measure of PSS is quite general and fails to capture ated with fewer supportive relationships and less satisfaction with social whether low PSS refers to childhood adversity or a difficult childhood support (Beeney et al., 2018; Lazarus et al., 2016; Zielinski and Veilleux, temperament. Fourth, although participants were thoroughly assessed, 2014). Importantly, while BPD features are associated with low PSS, no specific interview was used to diagnose BPD. Finally, although it was they are also closely associated with NSSI. For example, the number of not the aim of the present study, future research needs to address po­ BPD criteria met significantly predicted repeated (vs. single) NSSI his­ tential differences between patients with NSSI who meet the full criteria tory in an outpatient clinical sample of adolescents (Muehlenkamp et al., for NSSI disorder and patients with NSSI who do not meet the full 2011). criteria Findings from these studies may help determine the clinical While borderline traits mediated the relationship between PSS and utility of NSSI as an independent diagnostic category (Hooley et al., NSSI when controlling for current psychological distress (e.g., 2020). 210 I. Mendez et al. Journal of Affective Disorders 302 (2022) 204–213 Current findings highlight the mediating role of borderline traits in the online version, at doi:10.1016/j.jad.2022.01.065. the relationship between PSS and NSSI. Results suggest that adolescents with NSSI constitute a more severe clinical group due to more pro­ References nounced BPD-traits and less PSS. Importantly, current results are consistent with previous evidence suggesting a role of perceived inval­ Arturo, B.L., Antoni, S., Raquel, A., 2005. Psycometric properties of the spanish version of depression,anxiety and stress scales (DASS). Psicothema. idation in borderline pathology (Zanarini et al., 1997). Findings could Ayodeji, E., Green, J., Roberts, C., Trainor, G., Rothwell, J., Woodham, A., Wood, A., have several implications for prevention and treatment of NSSI in ado­ 2015. The influence of personality disorder on outcome in adolescent self-harm. Br. lescents. On the one hand, our results emphasize the importance of J. Psychiatry. https://doi.org/10.1192/bjp.bp.113.138941. Baiden, P., Stewart, S.L., Fallon, B., 2017. The role of adverse childhood experiences as evaluating BPD-traits in clinical settings, especially in those adolescents determinants of non-suicidal self-injury among children and adolescents referred to reporting low PSS. On the other hand, findings support the use of spe­ community and inpatient mental health settings. Child Abus. Negl. 69, 163–176. cific interventions aimed at improving BPD symptoms and social support https://doi.org/10.1016/j.chiabu.2017.04.011. Beeney, J.E., Hallquist, M.N., Clifton, A.D., Lazarus, S.A., Pilkonis, P.A., 2018. Social in adolescent patients with NSSI and their caregivers (e.g., the Dialec­ disadvantage and borderline personality disorder: A study of social networks. tical Behavioral Therapy; Boritz et al., 2021; McCauley et al., 2018). Personal. Disord. Theory Res. Treat. https://doi.org/10.1037/per0000234. Indirectly, our findings could also be useful in the context of the current Bennett, C., Melvin, G.A., Quek, J., Saeedi, N., Gordon, M.S., Newman, L.K., 2019. Perceived invalidation in adolescent borderline personality disorder: an pandemic, where compliance with social distancing measures may lead investigation of parallel reports of caregiver responses to negative emotions. Child to lower PSS and increased risk of NSSI, especially among adolescents Psychiatry Hum. Dev. https://doi.org/10.1007/s10578-018-0833-5. with elevated BPD traits. Future research should also consider repli­ Beranuy, M., Oberst, U., Carbonell, X., Chamarro, A., 2009. Problematic Internet and cating our findings in non-clinical samples of adolescents with mobile phone use and clinical symptoms in college students: the role of emotional intelligence. Comput. Hum. Behav. https://doi.org/10.1016/j.chb.2009.03.001. sub-clinical BPD. Blakemore, S.J., Mills, K.L., 2014. Is adolescence a sensitive period for sociocultural processing? Annu. Rev. Psychol. https://doi.org/10.1146/annurev-psych-010213- Funding 115202. Boritz, T.Z., Sheikhan, N.Y., Hawke, L.D., McMain, S.F., Henderson, J., 2021. Evaluating the effectiveness of the family connections program for caregivers of youth with This research was supported by the Catalan government with a PERIS mental health challenges, part I: a quantitative analysis. Heal. Expect. https://doi. fellowship, reference number SLT006/17/00159. CS is supported by the org/10.1111/hex.13205. Briones-Buixassa, L., Alí, Í., Schmidt, C., Nicolaou, S., Pascual, J.C., Soler, J., Vega, D., Chilean government, specifically by the Becas-Chile fellowship program 2021. Predicting non-suicidal self-injury in young adults with and without of the National Agency for Research and Development (ANID), reference borderline personality disorder: a multilevel approach combining ecological number 2018-72190624. SN is supported by a felowship from "La Caixa" momentary assessment and self-report measures. Psychiatr. Q. 1–20. https://doi. org/10.1007/s11126-020-09875-7. Foundation (ID 100010434; code: LCF/BQ/DR21/11880013). Brown, R.C., Witt, A., 2019. Social factors associated with non-suicidal self-injury (NSSI). Child Adolesc. Psychiatry Ment. Health. https://doi.org/10.1186/s13034-019-0284- CRediT authorship contribution statement 1. Brunner, R., Kaess, M., Parzer, P., Fischer, G., Carli, V., Hoven, C.W., Wasserman, C., Sarchiapone, M., Resch, F., Apter, A., Balazs, J., Barzilay, S., Bobes, J., Corcoran, P., Iria Mendez: Visualization, Project administration, Investigation, Cosmanm, D., Haring, C., Iosuec, M., Kahn, J.P., Keeley, H., Meszaros, G., Nemes, B., Writing – original draft, Writing – review & editing. Anna Sintes: Podlogar, T., Postuvan, V., Saiz, P.A., Sisask, M., Tubiana, A., Varnik, A., Wasserman, D., 2014. Life-time prevalence and psychosocial correlates of adolescent Visualization, Project administration, Investigation, Writing – original direct self-injurious behavior: a comparative study of findings in 11 European draft, Writing – review & editing. Juan Carlos Pascual: Visualization, countries. J. Child Psychol. Psychiatry Allied Discip. https://doi.org/10.1111/ Project administration, Investigation, Writing – original draft, Writing – jcpp.12166. review & editing. Joaquim Puntí: Visualization, Project administration, Calderón, C., Ferrando, P.J., Lorenzo-Seva, U., Gómez-Sánchez, D., Fernández- Montes, A., Palacín-Lois, M., Antoñanzas-Basa, M., Rogado, J., Manzano- Investigation, Writing – original draft, Writing – review & editing. Anaís Fernández, A., Ferreira, E., Asensio-Martínez, E., Jiménez-Fonseca, P., 2021. Lara: Visualization, Project administration, Investigation, Writing – Multidimensional scale of perceived social support (MSPSS) in cancer patients: original draft, Writing – review & editing. Laia Briones-Buixassa: psychometric properties and measurement invariance. Psicothema. https://doi.org/ 10.7334/psicothema2020.263. Project administration, Investigation, Writing – original draft, Writing – Canty-Mitchell, J., Zimet, G.D., 2000. Psychometric properties of the multidimensional review & editing. Stella Nicolaou: Writing – review & editing, Formal scale of perceived social support in urban adolescents. Am. J. Commun. Psychol. analysis, Writing – original draft. Carlos Schmidt: Writing – review & https://doi.org/10.1023/A:1005109522457. Chanen, A.M., Jovev, M., Djaja, D., McDougall, E., Yuen, H.P., Rawlings, D., Jackson, H. editing, Formal analysis, Writing – original draft. Soledad Romero: J., 2008. Screening for borderline personality disorder in outpatient youth. J. Pers. Visualization, Writing – original draft, Writing – review & editing. Disord. https://doi.org/10.1521/pedi.2008.22.4.353. Marta Fernández: Visualization, Writing – original draft, Writing – Chapman, A.L., Gratz, K.L., Brown, M.Z., 2006. Solving the puzzle of deliberate self- harm: the experiential avoidance model. Behav. Res. Ther. 44, 371–394. https://doi. review & editing. Cristina Carmona i Farrés: Project administration, org/10.1016/j.brat.2005.03.005. Project administration, Investigation, Writing – original draft, Writing – Cipriano, A., Cella, S., Cotrufo, P., 2017. Nonsuicidal self-injury: a systematic review. review & editing. Joaquim Soler: Visualization, Writing – original Front. Psychol. https://doi.org/10.3389/fpsyg.2017.01946. Claes, L., Luyckx, K., Baetens, I., Van de Ven, M., Witteman, C., 2015. Bullying and draft, Writing – review & editing. Pilar Santamarina-Perez: Visuali­ victimization, depressive mood, and non-suicidal self-injury in adolescents: the zation, Writing – original draft, Writing – review & editing. Daniel moderating role of parental support. J. Child Fam. Stud. https://doi.org/10.1007/ Vega: Visualization, Writing – original draft, Project administration, s10826-015-0138-2. Investigation, Writing – review & editing, Formal analysis. Claes, L., Vandereycken, W., Vertommen, H., 2003. Eating-disordered patients with and without self-injurious behaviours: a comparison of psychopathological features. Eur. Eat. Disord. Rev. 11, 379–396. https://doi.org/10.1002/erv.510. Declaration of Competing Interest Colón-Soto, M., Díaz, V., Soto, O., Santana, C., 2005. MINI KID Mini International Neuropsychiatric Interview. Tampa, Florida. Crowell, S.E., Beauchaine, T.P., Linehan, M.M., 2009. A biosocial developmental model All authors declare that they have no conflicts of interest. of borderline personality: elaborating and extending Linehan’s theory. Psychol. Bull. 135, 495–510. https://doi.org/10.1037/a0015616. Acknowledgements Daukantaitė, D., Lundh, L.G., Wångby-Lundh, M., Claréus, B., Bjärehed, J., Zhou, Y., Liljedahl, S.I., 2020. What happens to young adults who have engaged in self- injurious behavior as adolescents? A 10-year follow-up. Eur. Child Adolesc. We thank Roger Borràs at the Fundació Clínic per a la Recerca Bio­ Psychiatry. https://doi.org/10.1007/s00787-020-01533-4. mèdica for his contribution in the statistical analysis. Fruzzetti, A.E., Shenk, C., Hoffman, P.D, 2005. Family interaction and the development of borderline personality disorder: a transactional model. Dev. Psychopathol. 17, 1007–1030. Supplementary materials Fonseca-Pedrero, E., Paino, M., Lemos-Giráldez, S., Sierra-Baigrie, S., González, M.P.G. P., Bobes, J., Mûniz, J., 2011. Borderline personality traits in nonclinical young adults. J. Pers. Disord. https://doi.org/10.1521/pedi.2011.25.4.542. Supplementary material associated with this article can be found, in 211 I. Mendez et al. Journal of Affective Disorders 302 (2022) 204–213 Fox, K.R., Franklin, J.C., Ribeiro, J.D., Kleiman, E.M., Bentley, K.H., Nock, M.K., 2015. Madjar, N., Daka, D., Zalsman, G., Shoval, G., 2021. Depression symptoms as a mediator Meta-analysis of risk factors for nonsuicidal self-injury. Clin. Psychol. Rev. https:// between social support, non-suicidal self-injury, and suicidal ideation among Arab doi.org/10.1016/j.cpr.2015.09.002. adolescents in Israel. Sch. Psychol. Int. https://doi.org/10.1177/ Ghinea, D., Koenig, J., Parzer, P., Brunner, R., Carli, V., Hoven, C.W., Sarchiapone, M., 0143034321998741. Wasserman, D., Resch, F., Kaess, M., 2019. Longitudinal development of risk-taking Marino, C., Gini, G., Vieno, A., Spada, M.M., 2018. The associations between problematic and self-injurious behavior in association with late adolescent borderline personality Facebook use, psychological distress and well-being among adolescents and young disorder symptoms. Psychiatry Res. 273, 127–133. https://doi.org/10.1016/j. adults: A systematic review and meta-analysis. J. Affect. Disord. https://doi.org/ psychres.2019.01.010. 10.1016/j.jad.2017.10.007. Giletta, M., Scholte, R.H.J., Engels, R.C.M.E., Ciairano, S., Prinstein, M.J., 2012. McCauley, E., Berk, M.S., Asarnow, J.R., Adrian, M., Cohen, J., Korslund, K., Avina, C., Adolescent non-suicidal self-injury: A cross-national study of community samples Hughes, J., Harned, M., Gallop, R., Linehan, M.M., 2018. Efficacy of dialectical from Italy, the Netherlands and the United States. Psychiatry Res. https://doi.org/ behavior therapy for adolescents at high risk for suicide a randomized clinical trial. 10.1016/j.psychres.2012.02.009. JAMA Psychiatry. https://doi.org/10.1001/jamapsychiatry.2018.1109. Glenn, C.R., Klonsky, E.D., 2011. Prospective prediction of nonsuicidal self-injury: a 1- Mészáros, G., Győri, D., Horváth, L.O., Szentiványi, D., Balázs, J., 2020. Nonsuicidal self- year longitudinal study in young adults. Behav. Ther. https://doi.org/10.1016/j. injury: its associations with pathological internet use and psychopathology among beth.2011.04.005. adolescents. Front. Psychiatry 11. https://doi.org/10.3389/fpsyt.2020.00814. Greenfield, B., Henry, M., Lis, E., Slatkoff, J., Guilé, J.M., Dougherty, G., Zhang, X., Meszaros, G., Horvath, L.O., Balazs, J., 2017. Self-injury and externalizing pathology: a Raz, A., Eugene Arnold, L., Daniel, L., Mishara, B.L., Koenekoop, R.K., de Castro, F., systematic literature review. BMC Psychiatry 17, 160. https://doi.org/10.1186/ 2015. Correlates, stability and predictors of borderline personality disorder among s12888-017-1326-y. previously suicidal youth. Eur. Child Adolesc. Psychiatry. https://doi.org/10.1007/ Miller, A.B., Esposito-Smythers, C., Leichtweis, R.N., 2015. Role of social support in s00787-014-0589-9. adolescent suicidal ideation and suicide attempts. J. Adolesc. Heal. 56, 286–292. Groschwitz, R.C., Plener, P.L., Kaess, M., Schumacher, T., Stoehr, R., Boege, I., 2015. The https://doi.org/10.1016/j.jadohealth.2014.10.265. situation of former adolescent self-injurers as young adults: a follow-up study. BMC Muehlenkamp, J., Brausch, A., Quigley, K., Whitlock, J., 2013. Interpersonal features and Psychiatry. https://doi.org/10.1186/s12888-015-0555-1. functions of nonsuicidal self-injury. Suicide Life Threat. Behav. 43, 67–80. https:// Haber, M.G., Cohen, J.L., Lucas, T., Baltes, B.B., 2007. The relationship between self- doi.org/10.1111/j.1943-278X.2012.00128.x. reported received and perceived social support: a meta-analytic review. Am. J. Muehlenkamp, J.J., Ertelt, T.W., Miller, A.L., Claes, L., 2011. Borderline personality Commun. Psychol. 39, 133–144. https://doi.org/10.1007/s10464-007-9100-9. symptoms differentiate non-suicidal and suicidal self-injury in ethnically diverse Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E., 2010. Multivariate data analysis. adolescent outpatients. J. Child Psychol. Psychiatry Allied Discip. https://doi.org/ Vectors. 10.1111/j.1469-7610.2010.02305.x. Hankin, B.L., Abela, J.R.Z., 2011. Nonsuicidal self-injury in adolescence: prospective Nemati, H., Sahebihagh, M.H., Mahmoodi, M., Ghiasi, A., Ebrahimi, H., Atri, S.B., rates and risk factors in a 2½ year longitudinal study. Psychiatry Res. 186, 65–70. Mohammadpoorasl, A., 2020. Non-suicidal self-injury and its relationship with https://doi.org/10.1016/j.psychres.2010.07.056. family psychological function and perceived social support among Iranian high Hooley, J.M., Fox, K.R., Boccagno, C., 2020. Nonsuicidal self-injury: diagnostic school students. J. Res. Health Sci. 20 https://doi.org/10.34172/jrhs.2020.04. challenges and current perspectives. Neuropsychiatr. Dis. Treat. https://doi.org/ Nock, M.K., 2010. Self-injury. Annu. Rev. Clin. Psychol. 6, 339–363. https://doi.org/ 10.2147/NDT.S198806. 10.1146/annurev.clinpsy.121208.131258. Homan, K.J., Sim, L.A., Fargo, J.D., Twohig, M.P., 2017. Five-year prospective Nock, M.K., Joiner, T.E., Gordon, K.H., Lloyd-Richardson, E., Prinstein, M.J., 2006. Non- investigation of self-harm/suicide-related behaviors in the development of suicidal self-injury among adolescents: diagnostic correlates and relation to suicide borderline personality disorder. Personal Disord. Theory Res. Treat. https://doi.org/ attempts. Psychiatry Res. 144, 65–72. https://doi.org/10.1016/j. 10.1037/per0000169. psychres.2006.05.010. Hutsebaut, J., Aleva, A., 2021. The identification of a risk profile for young people with Orben, A., Tomova, L., Blakemore, S.J., 2020. The effects of social deprivation on borderline personality pathology: a review of recent literature. Curr. Opin. Psychol. adolescent development and mental health. Lancet Child Adolesc. Heal. https://doi. https://doi.org/10.1016/j.copsyc.2020.06.004. org/10.1016/S2352-4642(20)30186-3. Jacobson, C.M., Muehlenkamp, J.J., Miller, A.L., Turner, J.B., 2008. Psychiatric Perez, J., Venta, A., Garnaat, S., Sharp, C., 2012. The difficulties in emotion regulation impairment among adolescents engaging in different types of deliberate self-harm. scale: Factor structure and association with nonsuicidal self-injury in adolescent J. Clin. Child Adolesc. Psychol. https://doi.org/10.1080/15374410801955771. inpatients. J. Psychopathol. Behav. Assess. 34, 393–404. https://doi.org/10.1007/ Kaess, M., Brunner, R., Chanen, A., 2014. Borderline personality disorder in adolescence. s10862-012-9292-7. Pediatrics. https://doi.org/10.1542/peds.2013-3677. Piccirillo, M.L., Burke, T.A., Moore-Berg, S.L., Alloy, L.B., Heimberg, R.G., 2020. Self- Kaess, M., Parzer, P., Mattern, M., Plener, P.L., Bifulco, A., Resch, F., Brunner, R., 2013. stigma toward nonsuicidal self-injury: an examination of implicit and explicit Adverse childhood experiences and their impact on frequency, severity, and the attitudes. Suicide Life Threat. Behav. 10.1111/sltb.12640. individual function of nonsuicidal self-injury in youth. Psychiatry Res. https://doi. Plener, P.L., 2021. COVID-19 and nonsuicidal self-injury: the pandemic’s influence on an org/10.1016/j.psychres.2012.10.012. adolescent epidemic. Am. J. Public Health. https://doi.org/10.2105/ Karaer, Y., Akdemir, D., 2019. Parenting styles, perceived social support and emotion AJPH.2020.306037. regulation in adolescents with internet addiction. Compr. Psychiatry 92, 22–27. Plener, P.L., Schumacher, T.S., Munz, L.M., Groschwitz, R.C., 2015. The longitudinal https://doi.org/10.1016/j.comppsych.2019.03.003. course of non-suicidal self-injury and deliberate self-harm: A systematic review of Kiekens, G., Hasking, P., Nock, M.K., Boyes, M., Kirtley, O., Bruffaerts, R., Myin- the literature. Borderl. Personal. Disord. Emot. Dysregulation. https://doi.org/ Germeys, I., Claes, L., 2020. Fluctuations in affective states and self-efficacy to resist 10.1186/s40479-014-0024-3. non-suicidal self-injury as real-time predictors of non-suicidal self-injurious thoughts R Core Team, 2021. R A Language and Environment for Statistical Computing.(R and behaviors. Front. Psychiatry. https://doi.org/10.3389/fpsyt.2020.00214. Foundation for Statistical Computing. Vienna, Austria. URL. https://www.R-project. Kleiman, E.M., Liu, R.T., 2013. Social support as a protective factor in suicide: Findings org. from two nationally representative samples. J. Affect. Disord. 150, 540–545. https:// Reichenberger, J., Eibl, J.J., Pfaltz, M., Wilhelm, F.H., Voderholzer, U., Hillert, A., doi.org/10.1016/j.jad.2013.01.033. Blechert, J., 2017. Don’t praise me, don’t chase me: Emotional reactivity to positive Kleindienst, N., Hauschild, S., Liebke, L., Thome, J., Bertsch, K., Hensel, S., Lis, S., 2019. and negative social-evaluative videos in patients with borderline personality A negative bias in decoding positive social cues characterizes emotion processing in disorder. J. Pers. Disord. https://doi.org/10.1521/pedi_2016_30_238. patients with symptom-remitted borderline personality disorder. Borderl. Personal Reichl, C., Kaess, M., 2021. Self-harm in the context of borderline personality disorder. Disord. Emot. Dysregulation. Curr. Opin. Psychol. https://doi.org/10.1016/j.copsyc.2020.12.007. Klomek, A., Snir, A., Apter, A., Carli, V., Wasserman, C., Hadlaczky, G., Hoven, C.W., Reininger, B.M., Pérez, A., Aguirre Flores, M.I., Chen, Z., Rahbar, M.H., 2012. Sarchiapone, M., Balazs, J., Bobes, J., Brunner, R., Corcoran, P., Cosman, D., Perceptions of social support, empowerment and youth risk behaviors. J. Prim. Prev. Haring, C., Kahn, J.-P., Kaess, M., Postuvan, V., Sisask, M., Tubiana, A., Varnik, A., 33, 33–46. https://doi.org/10.1007/s10935-012-0260-5. Žiberna, J., Wasserman, D., 2016. Association between victimization by bullying and Rijnhart, J.J.M., Twisk, J.W.R., Eekhout, I., Heymans, M.W., 2019. Comparison of direct self injurious behavior among adolescence in Europe: a ten-country study. Eur. logistic-regression based methods for simple mediation analysis with a dichotomous Child Adolesc. Psychiatry 25, 1183–1193. https://doi.org/10.1007/s00787-016- outcome variable. BMC Med. Res. Methodol. 19 https://doi.org/10.1186/S12874- 0840-7. 018-0654-Z. Klonsky, E.D., 2007. The functions of deliberate self-injury: a review of the evidence. Rissanen, M.L., Kekkonen, V., Laukkanen, E., Hintikka, J., Tolmunen, T., 2021. Social Clin. Psychol. Rev. 27, 226–239. https://doi.org/10.1016/j.cpr.2006.08.002. support may prevent self-cutting in adolescence: A 5-year follow-up study. J. Child Klonsky, E.D., Glenn, C.R., 2009. Assessing the functions of non-suicidal self-injury: Adolesc. Psychiatr. Nurs. https://doi.org/10.1111/jcap.12341. psychometric properties of the inventory of statements about self-injury (ISAS). Sadeh, N., Londahl-Shaller, E.A., Piatigorsky, A., Fordwood, S., Stuart, B.K., McNiel, D.E., J. Psychopathol. Behav. Assess. https://doi.org/10.1007/s10862-008-9107-z. David Klonsky, E., Ozer, E.M., Yaeger, A.M., 2014. Functions of non-suicidal self- Lai, C.C.W., Ma, C.M.S., 2016. The mediating role of social support in the relationship injury in adolescents and young adults with borderline personality disorder between psychological well-being and health-risk behaviors among Chinese symptoms. Psychiatry Res. https://doi.org/10.1016/j.psychres.2014.02.018. university students. Heal. Psychol. Open 3. https://doi.org/10.1177/ Skabeikyte, G., Barkauskiene, R., 2021. A systematic review of the factors associated 2055102916678106. with the course of borderline personality disorder symptoms in adolescence. Borderl. Laye-Gindhu, A., Schonert-Reichl, K.A., 2005. Nonsuicidal self-harm among community Personal Disord. Emot. Dysregulation. https://doi.org/10.1186/s40479-021-00151- adolescents: Understanding the “whats” and “whys” of self-harm. J. Youth Adolesc. z. 34, 447–457. https://doi.org/10.1007/s10964-005-7262-z. Stead, V.E., Boylan, K., Schmidt, L.A., 2019. Longitudinal associations between non- Lazarus, S.A., Southward, M.W., Cheavens, J.S., 2016. Do borderline personality disorder suicidal self-injury and borderline personality disorder in adolescents: A literature features and rejection sensitivity predict social network outcomes over time? Pers. review. Borderl. Personal. Disord. Emot. Dysregulation. https://doi.org/10.1186/ Individ. Differ. https://doi.org/10.1016/j.paid.2016.02.032. s40479-019-0100-9. 212 I. Mendez et al. Journal of Affective Disorders 302 (2022) 204–213 Stepp, S.D., Whalen, D.J., Scott, L.N., Zalewski, M., Loeber, R., Hipwell, A.E., 2014. Wang, Q., Wang, H., Liu, X., 2020. Loneliness, non-suicidal self-injury, and friendship Reciprocal effects of parenting and borderline personality disorder symptoms in quality among Chinese left-behind adolescents: the role of parent-child cohesion. adolescent girls. Dev. Psychopathol. 26, 361–378. https://doi.org/10.1017/ J. Affect. Disord. 271, 193–200. S0954579413001041. Washburn, J.J., Potthoff, L.M., Juzwin, K.R., Styer, D.M., 2015. Assessing DSM-5 Strandholm, T., Kiviruusu, O., Karlsson, L., Pankakoski, M., Pelkonen, M., Marttunen, M., nonsuicidal self-injury disorder in a clinical sample. Psychol. Assess. 27, 31–41. 2017. Stability and change in personality disorder symptoms in 1-year follow-up of https://doi.org/10.1037/PAS0000021. depressed adolescent outpatients. J. Nerv. Ment. Dis. https://doi.org/10.1097/ Wester, K., Trepal, H., King, K., 2018. Nonsuicidal self-injury: increased prevalence in NMD.0000000000000623. engagement. Suicide Life Threat. Behav. 48, 690–698. https://doi.org/10.1111/ Tang, W.C., Lin, M.P., You, J., Wu, J.Y.W., Chen, K.C., 2021. Prevalence and sltb.12389. psychosocial risk factors of nonsuicidal self-injury among adolescents during the Wolff, J.C., Thompson, E., Thomas, S.A., Nesi, J., Bettis, A.H., Ransford, B., Scopelliti, K., COVID-19 outbreak. Curr. Psychol. https://doi.org/10.1007/s12144-021-01931-0. Frazier, E.A., Liu, R.T., 2019. Emotion dysregulation and non-suicidal self-injury: A Taylor, P.J., Jomar, K., Dhingra, K., Forrester, R., Shahmalak, U., Dickson, J.M., 2018. systematic review and meta-analysis. Eur. Psychiatry 59, 25–36. https://doi.org/ A meta-analysis of the prevalence of different functions of non-suicidal self-injury. 10.1016/j.eurpsy.2019.03.004. J. Affect. Disord. 227, 759–769. https://doi.org/10.1016/j.jad.2017.11.073. Wolke, D., Schreier, A., Zanarini, M.C., Winsper, C., 2012. Bullied by peers in childhood Tham, S.G., Ibrahim, S., Hunt, I.M., Kapur, N., Gooding, P., 2020. Examining the and borderline personality symptoms at 11 years of age: a prospective study. J. Child mechanisms by which adverse life events affect having a history of self-harm, and Psychol. Psychiatry 53, 846–855. https://doi.org/10.1111/J.1469- the protective effect of social support. J. Affect. Disord. 263, 621–628. https://doi. 7610.2012.02542.X. org/10.1016/j.jad.2019.11.037. Yurkowski, K., Martin, J., Levesque, C., Bureau, J.F., Lafontaine, M.F., Cloutier, P., 2015. Tingley, D., Yamamoto, T., Hirose, K., Keele, L., Imai, K., 2014. Mediation: R package for Emotion dysregulation mediates the influence of relationship difficulties on non- causal mediation analysis. J. Stat. Softw. https://doi.org/10.18637/jss.v059.i05. suicidal self-injury behavior in young adults. Psychiatry Res. 228, 871–878. Vega, D., Sintes, A., Fernández, M., Puntí, J., Soler, J., Santamarina, P., Soto, À., Lara, A., Zanarini, M.C., Williams, A.A., Lewis, R.E., Reich, R.B., Vera, S.C., Marino, M.F., Méndez, I., Martínez-Giménez, R., Romero, S., Pascual, J.C., 2018. Review and Levin, A., Yong, L., Frankenburg, F.R., 1997. Reported pathological childhood update on non-suicidal self-injury: who, how and why? Actas Esp. Psiquiatr. 46, experiences associated with the development of borderline personality disorder. Am. 146–155. J. Psychiatry 154, 1101–1106. Vega, D., Torrubia, R., Soto, À., Ribas, J., Soler, J., Pascual, J.C., Rodríguez-Fornells, A., Zetterqvist, M., 2015. The DSM-5 diagnosis of nonsuicidal self-injury disorder: a review Marco-Pallarés, J., 2017. Exploring the relationship between non suicidal self-injury of the empirical literature. Child Adolesc. Psychiatry Ment. Health 9, 31. https://doi. and borderline personality traits in young adults. Psychiatry Res. https://doi.org/ org/10.1186/s13034-015-0062-7. 10.1016/j.psychres.2017.07.008. Zhang, S., Tian, Y., Sui, Y., Zhang, D., Shi, J., Wang, P., Meng, W., Si, Y., 2018. Victor, S.E., Davis, T., Klonsky, E.D., 2017. Descriptive characteristics and initial Relationships between social support, loneliness, and internet addiction in chinese psychometric properties of the non-suicidal self-injury disorder scale. Arch. Suicide postsecondary students: a longitudinal cross-lagged analysis. Front. Psychol. 9, 1707. Res. https://doi.org/10.1080/13811118.2016.1193078. https://doi.org/10.3389/fpsyg.2018.01707. Victor, S.E., Hipwell, A.E., Stepp, S.D., Scott, L.N., 2019. Parent and peer relationships as Zielinski, M.J., Veilleux, J.C., 2014. Examining the relation between borderline longitudinal predictors of adolescent non-suicidal self-injury onset. Child Adolesc. personality features and social support: the mediating role of rejection sensitivity. Psychiatry Ment. Health. https://doi.org/10.1186/s13034-018-0261-0. Personal. Individ. Differ. https://doi.org/10.1016/j.paid.2014.07.005. Victor, S.E., Klonsky, E.D., 2016. Validation of a brief version of the difficulties in Zimet, G., Powell, S., Farley, G., Werkman, S., Berkoff, K., 1990. Psychometric emotion regulation scale (DERS-18) in five samples. J. Psychopathol. Behav. Assess. characteristics of the multidimensional scale of perceived social support. J. Personal. https://doi.org/10.1007/s10862-016-9547-9. Assess. https://doi.org/10.1207/s15327752jpa5503&4_17. Wan, Y., Chen, R., Ma, S., Mcfeeters, D., Sun, Y., Hao, J., Tao Background, F., 2018. Associations of adverse childhood experiences and social support with self-injurious behaviour and suicidality in adolescents. 10.1192/bjp.2018.263. 213