Table 1 Included relevant articles examining associations between self-injury and externalizing pathology

Aglan et al. (2008) [11]

UK

prospective; baseline: adolescents took part in a family intervention; follow up: six years later; self-poisoners

BL: M = 14.5; SD = 1.1; FU: M = 20.8; SD = 1.1

BL: 158; FU: 126

DSP

No

part of the interviewing process, question for DSH

The main risk factor for repeated DSH is MDD. However, adulthood adversity (DSH is included) is influenced by other risk factors such as: child sexual abuse and adolescent hopelessness and conduct disorder.

Conduct disorder is a risk factor in their model to DSH repetition in adulthood. Correlation coefficient: 0.17.

Bacskai et al. (2012) [18]

Hungary

cross sectional, drug dependent patients from outpatient drug centers with and without ADHD

M = 27; SD = 6.3

210

self-injury

Yes

part of the interviewing process, question for self-injury

The highest aggressive tendencies were shown by the ADHD group. ADHD patients had higher levels of depression, suicidality and self-injury.

Physical self-injury is significantly more prevalent among substance dependent patients with ADHD than without ADHD (X² = 25.2; p < 0.0001).

Carli et al. (2010) [19]

Italy

cross sectional, incarcerated volunteers

N.A.

1265

self-mutilation

Yes

part of the interviewing process, question for self-mutilation

Psychoticism, extraversion, aggression, hostility and resilience capacity were more frequent among high impulsive subjects. They engaged in self-mutilation more frequently, substance use disorder was also more frequent in this group. There was no difference between groups in suicidality.

Prevalence of self-mutilation is significantly higher in high-impulsive group than in low-impulsive group (X² = 9.27; p = 0.001), but there is no difference in the level of suicidality.

Cerutti et al. (2011) [20]

Italy

cross-sectional, normal population

13–22; M = 16.47; SD = 1.7

234

DSH

Yes

questionnaire for DSH; DSHI

They found a correlation between all the psychopathologies examined: CD, ODD, BPD, dissociative symptoms. They also found a correlation between DSH and stressful life events (psychological and sexual abuse, natural disasters and serious accidents, the loss of someone important, and the witnessing of family violence or a serious accident).

There was a strong correlation between DSH and CD/ODD (correlation coeffitients: ODD: 0,39; CD: 0.41; p < 0.01). There were no differences in the levels of correlation between DSH and CD and DSH and ODD.

Chou et al. (2014) [12]

Taiwan

retrospective, Taiwanian National Healthcare Insurance Database; patients with ADHD, matched controls; 3 years interval

<18; ADHD: M = 8.5; SD = 3.0; HC: M = 8.4; SD = 3.0

3685; 36,850

DSP

No

based on clinical data

50% of the NSSI group engaged in this behaviour once or twice; 36% four times or more. No self-harm group: lowest levels of risk factors, highest levels of protective factors. Adolescents in the NSSI group had fewer depressive symptoms, lower suicidal ideation, greater self-esteem and parental support than the suicide attempt group.

DSP incidence was significantly higher in the ADHD group than in comparsion group. Adjusted hazard ratio: 4.65 (95% CI: 2.41–8.94); p < 0.05.

Crowell et al. (2005) [13]

USA

cross-sectional, parasuicidal adolescents and age matched controls

14–18; M = 15.3; SD = 1.1

46 (23 + 23)

parasuicide

No

interview for parasuicide; LPC

Adolescents with a history of parasuicide showed reduced RSA at baseline, greater RSA reactivity during negative mood induction. There were no differences between EDR and PEP. These results support theories that emphasize the importance of emotional dysregulation and impulsivity in parasuicidal teenage girls.

There were significantly more externalization symptoms in the parasuicidal than in the control group. According to parent’s report: F = 39.3; according to self report: F = 26.6; according to teacher’s report: F = 17.2; p < 0,001.

Crowell et al. (2012) [21]

USA

cross-sectional with 2 phases: visit1: self-reported questionnaires; visit2: interview, physiological measurement; 3 groups: self-injuring, depressed, control females

13–17; SII group: M = 16.3; SD = 1.0; depressed group: M = 15.4; SD = 1.4; control: M = 16.1; SD = 1.3

75 (25 + 25 + 25)

self-inflicted injury (SII)

Yes

interview for SII; L-SASI same as LPC

The SII group had higher scores on both externalizing psychopathology and emotion dysregulation, and had more attenuated EDR than depressed adolescents. Self-injuring adolescents also had higher scores on borderline pathology. These findings show that there are differences between self-injuring and depressed adolescent girls in aetiology and developmental course.

There were significantly more externalization symptoms in the self-injury group than in the control group or in the depressed group. According to parent’s report: F = 23.7; according to self report: F = 34.5; p < 0.001.

Darke and Torok (2013) [22]

Australia

cross-sectional, IDU patients

21–62; M = 37.1; SD = 7.9

300

non-suicidal self-harm

Yes

part of the interviewing process, questions for non-suicidal self-harm

Childhood physical abuse was reported in 74.3%. Independent correlates of non-suicidal self-harm were: female gender, avoided home due to conflict, more extensive polydrug use. Independent correlates of attempted suicide were: severe childhood physical abuse, frequent abuse, avoided home due to conflict, female gender, a positive screen for CD, more extensive polydrug use.

There were no significantly more CD symptoms among drug abuser patients with non-suicidal self-harm than without non-suicidal self-harm. Odds ratio: 1.74 (95% CI: 0.97–3.14); non-significant.

Evren et al. (2014) [23]

Turkey

cross-sectional, normal population

M = 16.7; SD = 6.4

4938

self-harm

Yes

part of a longer questionnaire, questions for self-harm

ADHD symptom scores were higher in females and among adolescents who used tobacco, alcohol and drugs, and had self-harming behaviour and suicidal thoughts in their lifetime. ADHD symptom score was correlated with depression, anxiety, anger and sensation seeking scores, in ANCOVA ADHD symptom score correlated with depression, anxiety, anger, sensation seeking. Alcohol use and suicidal thoughts predicted the severity of ADHD.

Sscores on the questionnaire measuring ADHD symptoms were significantly higher in the self-harming population than in the non-self-harming population (t = −15.38; p < 0.001).

Feingold et al. (2014) [24]

Israel

cross-sectional, general clinical sample, inpatients

M = 15.8; SD = 4.3

238

self-injury

Yes

based on clinical data

Alcohol-abuse patients had suicide attempts and self-injurious acts more frequently than non-alcohol-abuse patients. Attention-deficit disruptive behaviour disorders, criminal activity and drug use were more common in the abusers group. Median length of stay in hospital was longer in the non-abuser group.

There was significant difference in the prevalence of self-injurious behaviour among alcohol abusers and non-abusers (X² = 7.61; p < 0.05). Externalizing pathology was also more common in the alcohol abuser group (X² = 6.29; p < 0.05).

Fulwiler et al. (1997) [25]

USA

cross sectional, incarcerated population

Self-mutilators: M = 30; SD = 7.2; suicide attempters: M = 34; SD = 7.3

31

self-mutilation

Yes

based on clinical data

Suicide attempters frequently had adult affective disorder. Self-mutilators more often had a history of childhood hyperactivity and a mixed dysthymia and anxiety that began in childhood or early adolescence.

Childhood ADHD is associated with self-mutilation among adult prisoners (X² = 15.5; p = 0.00008).

Garcia-Nieto et al. (2014) [26]

Spain

cross-sectional, clinical sample, inpatients, who had self-injurious behaviours

M = 43.3; SD = 10.3

150

suicide attempt, suicide gesture (NSSI + suicide attempt to communicate with others)

Yes

interview for suicide gestures; SITIBI

Those who engaged in suicidal gestures are more likely to have a personality disorder, especially from Cluster B (histrionic and antisocial). Narcissistic personality disorder was a risk factor for suicide attempt, and borderline personality disorder was a risk factor for both. High impulsiveness was associated with suicide. Suicide attempts and suicide gestures are two distinct phenomena.

Antisocial personality disorder is a risk factor for suicidal gestures. Odds ratio: 2.28 (CI 95%: 1.12–4.64); p < 0,05.

Gatta et al. (2016) [27]

Italy

case-control study, neuropsychiatric inpatients and controls

12–17; NSSI group: M = 15.0; SD = 1.4; Control group: M = 15.4; SD = 1.2

33 in NSSI group +88 in control group

NSSI

Yes

based on clinical data

Scores on the questionnaires used (measuring internalizing, externalizing and other psychological problems) were significantly higher among those who engaged in NSSI than in the control group. Habitual self-injurers were more impulsive and alexithymic than occasional self-injurers, but they sought for help more frequently.

Cases had significantly higher scores on externalizing problems in Child Behaviour Checklist (CBCL; Z = 6.42; p < 0.05) and in Youth Self-Report (YSR; Z = 4.57; p < 0.05), than controls.

Guendelman et al. (2016) [28]

USA

prospective, girls with ADHD, W1, W2 - after 5 years, W3 - after 10 years

W1: M = 9.6, SD = 1.7; W2: M = 14.3; W3: 19.710-year follow-up

140 girls with ADHD +88 in comparison group

NSSI

Yes

interview for NSSI; SIQ

Maltreated participants among ADHD diagnosed girls and comparsion group were significantly more impaired with respect to suicidal attempts, internalizing symptomatology eating disorder symptomatology, and well-being than nonmaltreated participants.

There were no signifficant differences in engaging in NSSI between maltreated and non-maltreated participants among ADHD diagnosed girls and comparsion group. Odds ratio: 1.19 (CI 95%: 0.78–1.80); non significant

Guertin et al. (2001) [29]

USA

cross-sectional, clinical sample, inpatients after suicide attempt

12–18; M = 15.1; SD = 1.5

95

self-mutilation

Yes

questionnaire for self-mutilation; FASM

In the self-mutilative (and suicide attempt too) group, diagnoses of ODD, MDD, dysthymia were more common, they also had higher scores on hopelessness, loneliness, anger, risk taking and reckless behaviour and alcohol use scales.

ODD was significantly more common in the self-mutilative + suicide attempt group, than in the suicide attempt only group (X² = 3.91; p < 0.05).

Hinshaw et al. (2012) [30]

USA

prospective, girls with ADHD, matched controls, BL, % years FU, 10 years FU

BL: 6–12; 10 years FU: 17–24; M = 19.6

BL: 221; 10 years FU: 216

self-injury

Yes

questionnaire for self-injury; SIQ

Participants with childhood ADHD had higher rates of ADHD and comorbidity, more serious impairment, suicide attempts and self-injury. There were no differences in eating pathology, substance use and driving behaviour. There were significant differences between inattentive and combined type only in suicide attempts and self-injury.

Self-injury is significantly more common among girls diagnosed with combined ADHD in childhood. Odds ratio: 4.5; p < 0.05.

Hurtig et al. (2012) [31]

Finland

birth cohort (9479), subsample of that

16–18

104 + 169

DSH

Yes

part of the interviewing process questions for DSH

ADHD group had both more suicidal ideation and more DSH than the non-ADHD group. Other factors in suicidal behaviour: female gender, childhood emotional and behavioural problems, concurrent depression and anxiety, and especially in DSH behavioural disorder, substance use disorder and strain in family relations.

ADHD prevalence was significantly higher in DSH group than in no-DSH group (69% vs. 32%, P < 0.001).

Ilomaki et al. (2007) [32]

Finland

cross-sectional, CD population: alcohol dependents and non-dependents

12–17; alcohol dependent girls: M = 16.4; boys: M = 16.3; non-dependent girls M = 15.3; boys: M = 15.1

141

self-mutilation

Yes

part of the interviewing process questions for DSH

40.7% of CD girls and 29.3% of CD boys suffered from alcohol-dependence as well. Life-threatening suicide risk is higher among dependent girls and boys. Self-mutilation risk is also elevated in the group of dependent girls and in the group of dependent boys as well.

Prevalence of repeated self-mutilation is high among CD patients and higher among girls. The prevalence rate depends on gender and alcohol dependence of participants, the lowest was among non-alcohol-dependent boys, and the highest was among alcohol-dependent girls. Odds ratio among girls: 3.9 (CI 95%: 1.1–13.8); Odds ratio among boys: 5.3 (CI 95%: 1.1–26.5); p < 0.05.

Izutsu et al. (2006) [33]

Japan

cross-sectional, normal population

boys: M = 14.2; SD = 0.7; girls: M = 14.2; SD = 0.7

486

DSH

Yes

own questionnaire for DSH

Lifetime prevalence of self-cutting: 8.0% among boys; 9.3% among girls; lifetime prevalence of self-hitting: 27.7% among boys; 12.2% among girls. Lifetime prevalence of tobacco use: 33.1% among boys, 14.3% among girls; lifetime prevalence of alcohol use: 74.1% among boys; 63.4% among girls. ADHD scores are significantly higher in all groups with these problems (self-cutting, self-hitting, tobacco use, alcohol use) than those who do not have problems in these areas.

ADHD scores are significantly higher in all groups with these problems (self-cutting, self-hitting, tobacco use, alcohol use) compared to those who do not have problems in these areas (t = 2.55–5.56; p < 0.05).

Jacobson et al. (2008) [34]

USA

retrospective study, clinical sample, outpatients

12–19; M = 15.1; SD = 1.7

227

DSH, NSSI, SA

Yes

interview for DSH; LPC

The most frequent method of DSH was cutting, followed by overdose, burning, and strangling self. 52% did not engage in self-injury, 13% engaged in NSSI, 16% engaged in SA, 17% engaged in both SA and NSSI. Major depressive disorder and/or posttraumatic stress disorder were more common in SA group, than in NSSI group. Borderline personality features were more frequent among those who engaged in any kind of DSH. The suicidal ideation levels of those in the NSSI group were similar to those in the No DSH group.

There were non-significant differences between the DSH group and the non-DSH group regarding disruptive behaviour disorders (X² = 2.64; p = 0.45), and significant differences in borderline-features (X² = 28.69; p < 0,001)

Jenkins et al. (2015) [35]

USA

cross-sectional, specific psychopatology group, IED patients, perosnality disorder, HC

18–81; M = 35.1; SD = 10.3

1097

NSSI

Yes

questionnaires for NSSI; DSHI

Individuals with personality disorders, and particularly with comorbid IED, are at increased risk for self-injurious behaviours. Traits of aggression, impulsivity and affect lability has an effect on the relationship between personality disorders, IED and self-injurious behaviours, especially NSSI.

Individuals with personality disorders, particularly with comorbid IED are at increased risk for self-injurious behaviours (IED - control: X² = 9.89; personality disorder - control: X² = 40.85; IED and personality disorder vs. control: X² = 94.03; p < 0.05). Traits of aggression, impulsivity and affect lability has an effect on the relationship between personality disorders, IED and self-injurious behaviours, especially NSSI.

Keenan et al. (2014) [36]

USA

prospective normal population

13–14 years old girls

2180

NSSI

Yes

part of the interviewing process questions for NSSI

6.0% of the sample engaged in NSSI according to either child or parent report. In the ‘aggression’ domain, initial levels of conduct problems and self-control were significantly associated with NSSI. The initial levels of relational aggression were not associated with NSSI, but the increases in these levels over time were. In the ‘depression’ domain, initial levels of depression and a decrease in assertiveness were associated with later risk for NSSI. In the ‘environmental stressor’ domain, NSSI was associated with initial levels of peer victimization and negative life events. The most common pathway to NSSI was aggression combined with depression and environmental stressor domains, and the least common was environmental stressors only.

Higher initial levels of conduct problems (odds ratio: 1.08 (CI 95%: 1.01–1.15); p < 0.05), lower initial levels of self-control (0.92 (CI 95%: 0.87–0.98); p < 0.05) and increases in relational aggression over time (1.36 (CI 95%: 1.10–1.70); p < 0.05) were associated with NSSI.

Kirkcaldy et al. (2006) [37]

Germany

cross-sectional, clinical sample, inpatients

3–24; M = 13.4; SD = 3.4

3649

SIB

Yes

based on clinical data, regularly used: questionnaire for SIB, German Inventory

SIB and suicidal behaviours correlated, but aggression (or disruptive behaviour) was not correlated with suicide or SIB. Age and family disharmony were risk factors for suicide, but there was no association with disruptive behaviour and SIB. Intelligence and age were significant predictors of aggression among females.

There was no correlation between SIB and disruptive behaviour.

Lam et al. (2005) [14]

Australia

cross-sectional, patients owing injuries

Range: 5–15

18.729

self-harm

No

based on clinical data

Participants from the age group 11–15 years-old were more likely to be involved in intra- and interpersonal violence than those from the age group 5–10-years-old. Female gender showed a stronger association with suicide or self-harm, while socioeconomically advantaged background was related to interpersonal violence. Suicide attempts and engagement in self-inflicted injuries were associated with staying in hospital for longer time. 3 cases resulted in death.

Compared to other causes of injury, patients whose cause of hospital admission was suicide were more likely to be diagnosed with ADD, while patients with self-harm were also more likely to have comorbid ADD as compared to those with other types of injuries (odds ratio in univariate model: 3.76 (CI 95%: 1.73–8.15); odds ratio in multivariate model: 6.27 (CI 95%: 2.76–14.26)).

McCloskey et al. (2008) [38]

USA

cross-sectional, adults with intermittent explosive disorder

M = 36.10; SD = 9.32

376

SIB

Yes

own interview for SIB

16% reported self-aggression, 12.5% suicide attempts, 7.4% non-lethal self-injurious behaviours. MDD, drug dependence and Axis-I comorbidities all predicted SA or SIB in IED. Women were at increased risk for self-aggressive behaviour overall, but it was not significant for either SA or SIB.

16% of IED patients reported self-aggressive behaviour, with 12.5% attempting suicide and 7.4% engaging in SIB. This shows a relationship between IED and self-aggression.

Meza et al. (2016) [39]

USA

prospective, patients (girls) with ADHD and control group

W1, M = 9.6, range6–12); 5 years follow-up (W2, M = 14.2, range 11–18);10-year follow-up (W3, M = 19.6, range 17–24)

228 with (n = 140) and without (n = 88) childhood ADHD

NSSI

Yes

interview for NSSI; SIQ

Wave 1 commission errors (indicator of response inhibition) predicted Wave 3 suicide ideation with marginal significance and significantly predicted suicide attempts and NSSI severity. Social preference was a significant partial mediator of the Wave 1 commission errors and suicide ideation: indirect effect. Wave 2 peer victimization was a significant partial mediator in the Wave 1 commissions and Wave 3 NSSI severity.

There were a significant differences in NSSI severity between ADHD group and comparsion group (Cohen’s d:0.60; p < 0.001). Childhood response inhibition predicted NSSI in young adulthood in longitudinal study (B = 0,16; p < 0,05; R2 = 0.03).

Nock et al. (2006) [40]

USA

cross-sectional, case series, inpatient unit patients who engaged in NSSI

12–17 (M = 14,7)

89

NSSI, SIB

Yes

questionnaire for self-mutilation; FASM

87.6% met at least one Axis I criterion, internalization 51.7%, externalization 62.9%, substance use 59.6%; 67.3% of female patients met Axis II criteria (51.7% BPD was the most common). 70% reported at least one suicide attempt, 55% reported more than one suicide attempt.

Some sort of externalizing disorder was present in 62.9% of the NSSI sample (CD: 49.4%; ODD: 44.9%). Substance use disorder was present in 59.6%, while some sort of internalizing disorder was present in 51.7%.

Preyde et al. (2012) [41]

Canada

prospective, clinical population

5–18; M = 11.57 (SD = 2.75).

210 (Self-harm data available for 169)

NSSI, self-harm

Yes

part of the interviewing process questions for self-harm; BCFPI

39% who reported self-harm at admission were less than 12 years of age. Most of the differences between self-harmers and non-self-harmers on symptom severity at intake disappeared by the time of follow up. Youth who engaged in self-harm had higher symptom severity on Attention and Impulsivity regulation, Managing Mood, Internalizing Behaviour and Total Mental Health. At discharge, symptom severity only differed on the Total Mental Health subscale, and no differences were evident at 12 to 18 months post-discharge or 36 to 40 months post-discharge.

Higher symptom severity among self-harmers compared to non-self-harmers on Attention and Impulsivity regulation and Total Mental Health was measured. Mean scores (SD; CI 95%) on Attention and Impulsivity regulation: no self-harm: 70.9 (10.2; 68.8–73.0); self harm: 75.3 (7.8; 73.0–77.5); mean scores (SD; CI 95%) on Total Mental Health: no self-harm: 77.4 (11.4; 75.1–79.8); self harm: 82.5 (11.1; 79.2–85.9); p < 0.05.

Semiz et al. (2008) [42]

Turkey

cross-sectional, male offenders in psychiatry department of a military

20–36; M = 22.7; SD = 2.9

105

SIB

Yes

own interview for SIB

92% reported SIB with 57% reporting more than 10 episodes. The onset of SIB was between 5 and 23 years with mean onset at 14.8 years, and the mean duration of the behaviour was 7.2 years. 65% had received medical treatment for their SIB. APD participants who had comorbid ADHD did report a significant increase in suicide attempts but not in SIB or criminal behaviours. The number of ADHD symptoms was significantly correlated with frequency of SIB. ADHD total score was significantly correlated with frequency of SIB and negatively correlated with age at onset of SIB.

The number of ADHD symptoms was significantly correlated with frequency of SIB (correlation coefficient: 0.32; p < 0.05). ADHD total score was significantly correlated with frequency of SIB (correlation coefficient: 0.38; p < 0.05) and negatively correlated with age of onset of SIB (correlation coefficient: −0.23; p < 0.05).

Swanson et al. (2014) [43]

USA

prospective, patients (girls) with ADHD and control group

6–12; M = 9.1

140 girls with ADHD +88 in comparison group

NSSI

Yes

interview for NSSI; SIQ

Combined type of ADHD engaged in the most severe forms of NSSI using the most methods compared to childhood-diagnosed inattentive type of ADHD and comparison groups. Both ADHD subtypes showed increased NSSI frequency compared to the comparison group but did not significantly differ from each other. Persistent ADHD showed increased NSSI compared to those with transient ADHD or the comparison group. Participants with persistent ADHD also had a higher rate of suicide attempts. Externalizing symptoms and a lab-based measure of response inhibition/impulsivity were significant partial mediators between ADHD and NSSI while internalizing symptoms during adolescence was a significant partial mediator of the ADHD-suicide attempt linkage.

There were significant differences in NSSI frequency, variety and severity between combined ADHD and comparsion groups (inattentive ADHD and control; Cohen’s d: 0.20–0.85; p < 0.05). Persistent ADHD also differed significantly from comparsion groups (transient ADHD and control) in NSSI frequency, variety and severity (Cohen’s d: 0.32–0.87; p < 0.05). Externalizing symptoms mediate between ADHD and NSSI (IE = 0.29; SE = 0.11; CI 95%: 0.10–0.51).

Taylor et al. (2014) [44]

New-Zealand

case–control study, patients with retrospective ADHD and control group

M = 31.9; SD = 1.6

66 (35 with adult ADHD and 31 with no ADHD)

DSH

Yes

questionnaire for DSH; DSHI

Significant associations were present between ADHD symptom severity and history of self-harm behaviour, suicidal ideation and suicide attempts. These relationships were mediated by comorbidity (mood, anxiety, drug and alcohol abuse disorders) and emotion-focused coping style. The results suggest that comorbid mental health disorders and emotion-focused coping might mediate the relationship between self-injurious behaviour and ADHD.

Significant associations were present between ADHD symptom severity and history of self-harm behaviour (B = 0.52; SE = 0.24; p < 0.05). These relationships were mediated by comorbidity (B = 0.68; SE = 0.23; p < 0.01) and emotion-focused coping style (B = 0.72; SE = 0.22; p < 0.01).

Vaughn et al. (2015) [15]

USA

prospective, normal population

18–49

19.073

DSH

No

part of the interviewing process questions for self-harm

DSH was linked to ethnicity since African-American, Latinos and Asians were less likely to report DSH than Whites. DSH was associated with all forms of child maltreatment: child sexual abuse, physical abuse, child neglect and exposure to serious conflict in the home. DSH was also associated with lifetime victimization: intimate partner violence, violent victimization, and to ever having been stalked. DSH was also associated with MDD, avoidant personality disorder and schizotypal personality disorder. DSH was also related to engaging in violent behaviours, particularly robust effects were shown with “having forced someone to have sex with you against their will”.

Besides DSH being linked to substance abuse (odds ratios: 2.90–7.02), it was also associated with violent behaviours such as robbery, intimate partner violence, forced sex, cruelty to animals and use of a weapon (odds ratios: 3.32–12.73).

You et al. (2011) [45]

Hong Kong

cross-sectional, normal population

11–19; M = 14.7; SD = 1.9

6374

NSSI

Yes

part of a longer questionnaire, questions for NSSI

Repetitive self-injurers had more impulse-control and emotional problems, than episodic self-injurers. Severe self-injurers were more impulsive than mild self-injurers. The frequency and severity of NSSI are two possible dimensions that distinguish NSSI subgroups.

There was a strong link between NSSI frequency (odds ratio: 1.37 (CI 95%: 1.32–1.42); p < 0.01) and severity (1.12 (CI 95%: 1.08–1.17); p < 0.01) and impulse control problems.

Young et al. (2009) [16]

UK

cross-sectional, prisoner population

M = 30.0; SD = 8.2

198

critical incident, self-injury

No

based on institutional data

24% of prisoners met ADHD criteria in the present or in the past. 23% of them were fully symptomatic, 33% were in partial remission and 44% were in full remission. Critical incidents (verbal and physical aggression, damage to property, self-injury, and severity of aggression) were more common among the ADHD group than in the non-ADHD group, the most severe forms found among fully symptomatic ADHD group.

The mean number of critical incidents was significantly higher in the ADHD group than in the non-ADHD group (F = 5.01; p < 0.001).

Zlotnick et al. (1999) [17]

USA

cross-sectional, psychiatric outpatients

M = 40.6; SD = 14.0

256

self-mutilation

No

own questionnaire for self-mutilation

From the Axis-I disorders, substance abuse, posttraumatic stress disorder and IED showed significant association to self-mutilative behaviour, independent of BPD and antisocial personality disorder. Moreover, higher scores of dissociation were associated with self-mutilation while controlling for age, BPD, sex, education, sexual abuse and physical abuse.

Self-mutilation was associated with substance abuse (X² = 13.73; p < 0.05), IED (X² = 6.35; p < 0.05) and antisocial personality disorder (X² = 8.03; p < 0.05).