Adolescent Self-Injury: Evaluation, Referral, and Treatment

Michigan State University

Dr Smith is assistant professor in the department of psychiatry at Michigan State University, East Lansing. He reports no conflicts of interest concerning the subject matter of this article.

ABSTRACT: The prevalence of nonsuicidal self-injury (NSSI) in adolescents may be increasing, and NSSI may be a risk factor for suicide. NSSI has a number of psycho­logical functions and may be contagious and/or addictive. Ask adolescent patients directly about self-injury and assess their general appearance for signs of scarring. Approach patients with a history of NSSI in a nonjudgmental manner and inquire about suicidality. Understanding the motivation behind the behavior—eg, conflict with family or peers or trauma, especially sexual abuse—helps guide the treatment plan. Asking questions using the Stages of Change model can help patients change behavior. Dialectical behavioral therapy, when available, can be considered the standard therapy for patients with NSSI who do not respond to initial interventions. Referral to a psychiatrist may be needed for complicated psychiatric
comorbidities or for pharmacotherapy.

In working with adolescents, pediatricians often draw on their own developmental experiences to help guide their patients; however, nonsuicidal self-injury (NSSI) is not likely to be a personal experience that pediatricians can draw on. In fact, today’s youth face new challenges, including increasing exposure to the practice of self-injury. We can best help adolescents by being aware of this problem, asking directly about self-injury, understanding basic therapeutic approaches, determining when to refer to outside providers for psychotherapy and pharmacotherapy, and having knowledge of more specific evidence-based psychotherapeutic and psychopharmacological interventions that can be used to treat NSSI.

In this article, NSSI, which is used interchangeably with self-injury, is defined as deliberate, self-directed tissue injury inflicted without conscious intent to kill oneself.1 In most studies, skin cutting is the most common form, followed by burning and self-hitting or banging; the forearms, wrists, and thighs are common locations of self-injury.2,3 Other forms of self-injury include scratching and interfering with wound healing. Klonsky4 reports that many individuals who engage in NSSI practice more than one method.

The prevalence of self-injury in adolescents and young adults, including the nonclinical population, may be increasing.4-7 One out of 3 self-injurers reports an onset of self-injurious behavior in childhood, with a peak incidence in mid- to late-adolescence.8 Relatively late development of brain circuits (pruning and myelination) involved with emotion, judgment, and inhibitory control may explain the heightened propensity of adolescents to act impulsively and ignore the negative consequences of their behavior.9 Studies have found that approximately 15% of high school adolescents and 17% of college students engage in self-injury, with estimates as high as 40% to 60% for adolescent inpatients.5,6,10 In one study, researchers found that 1 out of 5 students attending an Ivy League school endorsed these maladaptive behaviors, and 1 in 10 respondents were repeat self-injurers.5

Other psychiatric conditions often predispose patients to self-injurious behavior. Borderline personality disorder (BPD) and other personality disorders, anxiety, eating disorders, and substance abuse are all risk factors for NSSI.11-14 Additional risk factors include adolescence to college age, maltreatment, and family turmoil.5,12-14 A recent study noted that identification with Goth subculture is also a risk factor.15

Although most studies report females as being more likely to engage in NSSI, a number of studies have found no significant differences between females and males.2-6 There are several potential reasons for this disconnect. First, a majority of the earlier studies were completed with patients who had BPD, a disorder that is more common in females. Second, self-hitting and banging, which may not have been included in analyses of NSSI, may be more common among males; females are more likely to cut themselves. Third, a significant increase in male NSSI may have made a strong contribution to the overall increasing prevalence of NSSI in adolescents and young adults.5,6

Possible reasons for increased prevalence. An increase in some risk factors (eg, child abuse, substance abuse, family turmoil, anxiety) may explain the possible increase in incidence of NSSI in the adolescent population. Likewise, the phenomenon of contagion may be affecting rates of NSSI. “Behavioral contagion” is an increased tendency to engage in a behavior when socially related people also engage in that behavior.16 Contemporary routes for transmission include direct contact with self-injurers (often friends or family), direct or indirect exposure over the Internet, exposure via the media, and role modeling through the cult of celebrity, including adolescent antiheros.17,18

In addition to risk factors and modes of transmission, there are variables that motivate and reinforce the behavior of each individual. Nock and Prinstein19 hypothesized that NSSI may be automatically reinforced (eg, emotion regulation) or socially reinforced (eg, avoidance/escape, attention/communication of distress). The primary function of most NSSI appears to be reinforced by the individual, involving an affect-regulation function (reduction of negative affect).4 There is also strong support for a self-punishment function.4 Additional functions with more modest evidence include antidissociation, interpersonal influence, antisuicide, sensation seeking, and interpersonal boundaries functions.4
What should you do for your adolescent patients? Ask directly about self-injury, “Have you ever cut or done other damage to your skin on purpose?”5 In my experience, most adolescents will be forthcoming about a number of risky behaviors, including NSSI, when directly questioned without their parents present. If the adolescent seems particularly guarded under direct questioning, you may consider having the patient fill out a written self-report scale on NSSI.19-21 You should also assess the general appearance of the adolescent, noting any obvious scarring on visible skin and unusual dress, such as a Goth look or wearing long-sleeve shirts in hot weather.

Be nonjudgemental. After NSSI is disclosed, it is important to proceed with the interview and approach the patient in a nonjudgmental manner. Families (and sometimes clinicians) may overreact when they learn of self-injury. In many cases, NSSI may be a form of brief experimentation causing superficial tissue injury as part of the separation-individuation process; rarely does the self-mutilation involve suicidal intent.4,22 I would be reluctant to recommend psychiatric hospitalization with NSSI as the chief concern in the absence of other signs or symptoms of severe disturbance. However, even though NSSI usually does not constitute an acute emergency involving imminent danger, it does increase overall suicide risk, and clinicians should inquire about suicidality.23,24 In one study, adolescents with NSSI who had a history of suicide attempt were likely to have more depressive symptoms, higher suicidal ideation, and lower self-esteem and parental support than the group with NSSI with no history of suicide attempt.25

Determine functions of NSSI. Following disclosure of NSSI, the clinician should attempt to ascertain why the patient is engaging in such behaviors and the function(s) of NSSI (Table 1). Because NSSI is nonspecific, like a fever, it is necessary to understand the motivation behind the behavior to develop a treatment plan. Similarly, it is important to inquire about the emotions that precede and follow the NSSI. Ask about the common precipitants to NSSI, especially conflict with family or peers. Also inquire about past or present trauma, especially sexual abuse.

Evaluate social context and reinforcement. It may be helpful to ask, “How did you learn to hurt yourself?” This examines both the functions of NSSI and the possibility of contagion. Contact with peers who also engage in NSSI may positively reinforce this maladaptive behavior. Knowledge of active behavioral contagion suggests the need for changes in the adolescent’s social environment, which may include limiting unsupervised access to certain friends or the Internet.

Initiate thinking about change. While the clinician will likely view NSSI as a serious issue, the adolescent may not see it as a problem. It is extremely important to ask, “Do you think this is a problem?” Discovering where the adolescent is on the Stages of Change (Transtheoretical) model will inform the therapeutic approach that will most likely be successful.26 The Stages of Change model is a general approach to helping patients change behavior.26 It recognizes that changes in behavior may involve a patient moving gradually from being uninterested in change (precontemplation) to considering change (contemplation) to deciding and preparing to make a change.27

Examine benefits and barriers to change. Asking questions using the Stages of Change model as a framework may allow a patient to begin thinking about change as well as the benefits and barriers to change (see Table 1). In addition, it is useful to help patients identify alternative ways to reach goals associated with this behavior, such as feeling strong, connected, in control, and independent.27
Support the desire for feeling in control. Clarifying questions may be very powerful, such as, “Who is in control of your life—the cutting or you?” It is also wise to highlight the other aspects of NSSI that may be distasteful to the patient, including scarring, secrecy, shame, wound infections, doctor visits, stigma, the potential for addiction to the behavior, and accidental death.

Maintain confidentiality. Although direct communication between the adolescent and the parent is most often ideal, the issue of NSSI may remain confidential between the patient and clinician at the clinician’s discretion unless there are immediate safety risks.

Schedule follow-up. After the initial evaluation, schedule a follow-up session for further monitoring and discussion of the NSSI. In my experience, the majority of NSSI cases improve following disclosure and the patient-centered approaches described. NSSI also may be reduced through the treatment of comorbid mental health conditions.

If self-injury continues despite attempts with the interventions already described, if the patient has significant comorbid psychiatric issues, or if the self-injurious behavior is severe, then consider referral to an outside provider. Comorbid psychiatric issues that complicate matters include substance abuse (especially if engaging in NSSI while intoxicated), psychosis, suicidality, homicidality, borderline personality disorder, and a history of psychological trauma. Severe family discord may necessitate a referral for family therapy. Concern should be increased when the patient has no motivation to cease the behavior, identifies himself or herself as a “cutter,” self-harms on a daily basis, or engages in NSSI that causes significant scarring or requires medical interventions, such as sutures or surgical repair.

Psychotherapy for self-injury. A referral for dialectical behavioral therapy (DBT) may be an appropriate first-line treatment for more complicated NSSI. DBT is perhaps the best-studied psychotherapeutic intervention for NSSI. DBT is a variation of cognitive-behavioral therapy (CBT) that also includes mindfulness training. Through individual and group skills training, patients learn emotional regulation, how to cope with negative affect, and problem-solving techniques. DBT has demonstrated direct and sustained effects for individuals with BPD and self-injury in at least 7 well-controlled trials with different patient populations, including adolescent

However, traditional DBT may not be unique in its ability to treat NSSI. In randomized controlled trials, a program that uses CBT techniques called “manual-assisted cognitive treatment” demonstrated efficacy for reducing both the severity and frequency of deliberate self-injury in patients with BPD.31

There is also some evidence for therapeutic interventions that do not rely on the principles of CBT. In a recent long-term trial comparing DBT with nonbehavioral community treatment, NSSI was reduced equally by both protocols.29

Pharmacotherapy for self-injury. Consider referral to a psychiatrist for pharmacotherapy when DBT is not available or when other barriers to receiving psychotherapy exist. An adolescent with NSSI who has significant psychiatric comorbidities or who has not responded to psychotherapy may respond to more aggressive pharmacotherapy. There is evidence, albeit limited, for pharmacotherapy for NSSI. It should be noted that there are no medications currently labeled for the treatment of self-injury or BPD. Medications are used to target the neurotransmitter systems most likely to contribute to self-injury. Opioid, serotonin, and dopamine receptors are commonly implicated.

The opioid system is implicated because of higher pain thresholds and stress-induced analgesia in BPD, likely mediated by increased activity of the dorsal prefrontal cortex with deactivation of the anterior cingulate and amygdala.32 NSSI may also involve addictive qualities, including postcessation withdrawal dysphoria.14 Open-label studies of the opioid antagonist naltrexone have shown reduced self-injury.33,34

Decreased serotonin levels have been linked to impulsive, aggressive, and suicidal behaviors.35 Self-injury is usually an impulsive act, with half of cutters thinking of the behavior less than an hour before committing the act.22 Fluoxetine is the serotonergic medication with the most evidence for use in NSSI, with at least 2 controlled trials showing reductions in impulsive aggression.36,37 In an open-label trial, venlafaxine was also beneficial.38

Self-injury may also present in syndromes that involve the dopamine system, such as Lesch-Nyhan syndrome and Tourette syndrome.39 Self-biting behaviors may be elicited by stimulants that primarily work through a dopaminergic mechanism of action. The use of dopamine antagonists has been studied for NSSI. Although there is some evidence for the efficacy of all of the atypical antipsychotics for reducing aggression and/or impulsivity, olanzapine has had the best results in the most rigorous trial designs.37,40-44

There is a paucity of literature regarding combination therapy. Surprisingly, adding fluoxetine to DBT was not found to give any additional benefit.45 A trial of olanzapine plus DBT reduced the frequency of impulsivity/aggressive behavior more than placebo plus DBT, although self-injurious behavior did not decrease significantly.46

In addition to serotonergic, dopaminergic, and opioid medications there are other drug categories being studied for NSSI. There are promising case reports for glutamate-decreasing agents, including riluzole, N-acetylcysteine, and topiramate.47 Similarly, the glutamate-modulating agent lamotrigine was found to decrease behavioral dyscontrol, impulsivity, and anger in open-label trials, retrospective trials, and controlled trials.48-50 There have been controlled studies of lithium, divalproex, carbamazepine, and omega-3 fatty acids, which have all demonstrated decreases in impulsive aggression, but not NSSI specifically, in individuals with BPD.51-54 There was also an open-label study of oxcarbazepine that demonstrated statistically significant decreases in impulsivity, affective instability, and outbursts of anger in small sample populations with BPD.55 Clonidine, an a2-agonist, decreased the urge to commit self-injurious behavior in an open study.56

Table 2 provides a proposed algorithm for the treatment of NSSI. Medications should be considered the second line of treatment in most cases. Before pharmacological treatment is initiated for NSSI, preexisting medications for other mental health concerns should be optimized. In the absence of an existing psychotropic regimen or clear indications for pharmacotherapy of comorbid mental health conditions, fluoxetine, currently the best-studied antidepressant for NSSI, is a good first option. If fluoxetine is poorly tolerated or of insufficient benefit, consider switching to olanzapine, which appears to be the most effective atypical antipsychotic for NSSI.

The suggestion of first initiating a trial of an SSRI, as opposed to an atypical antipsychotic, has more to do with side-effect profile than potential efficacy. In fact, a meta-analysis of randomized controlled trials of pharmacotherapy against core traits of BPD found that antidepressants and mood stabilizers did not produce significant benefits against impulsivity and aggression or suicidality (although they were effective against affective instability and anger).57 On the other hand, antipsychotics, as a class, had a positive effect in reducing impulsivity and aggression. For severe and/or refractory cases, consider other SSRIs, serotonin-norepinephrine reuptake inhibitors, other atypical antipsychotics, naltrexone, clonidine, mood stabilizers, or glutamate-modulating agents.