Embed Size (px)
Citation preview
REVIEW
Non-suicidal self-injury
Paul Wilkinson
Published online: 1 December 2012
� Springer-Verlag Berlin Heidelberg 2012
Abstract More than 20 % of adolescents may self-injure.
Often there is no suicidal intent; rather the intent is to
reduce distressing affect, inflict self-punishment and/or
signal personal distress to important others. Non-suicidal
self-injury (NSSI) is both deliberate and contains no desire
to die and therefore aetiology is likely to be at least partly
different to suicidal behaviour per se. Interestingly, NSSI is
associated with subsequent suicide attempts suggesting that
these behaviours and their related psychology may lie on
the same risk trajectory. NSSI does not, however, appear in
DSM-IV or ICD-10 as a disorder; and does not constitute a
component of any current anxious or depressive syndrome.
This lack of nosological recognition coupled with clear
psychopathological importance is to be recognised in the
5th edition of the Diagnostic and Statistical Manual of
Mental Disorders, with NSSI being classified as a syn-
drome in its own right. We agree that this is appropriate
and is likely to have several positive consequences
including improving communication between professionals
and patients, informing treatment and management deci-
sions and increasing research into the nature, course and
outcome of NSSI. We agree that while suicidal and non-
suicidal self-harm are often seen together, they are not the
same behaviour and that it is both valid and useful to
separate them.
Keywords Non-suicidal self-injury � Adolescence �Suicide � Depression � Borderline personality disorder
Abbreviations
NSSI Non-suicidal self-injury
DSM-5 5th edition of the Diagnostic and Statistical
Manual of Mental Disorders
Introduction
A significant minority of adolescents deliberately injure
themselves. A review from the last decade suggested that
7–14 % of adolescents deliberately injure themselves at
least once [1]. Prevalence is significantly higher in females
than males. A more recent cross-national study suggested a
higher 1-year prevalence rate of 24 % [2]. This suggests
the possibility that prevalence is rising (although different
methodologies may also influence these results).
The most commonly reported behaviour is self-cutting.
In some cases, self-harm is with the explicit intention of
trying to end their life (often by self-poisoning in Europe).
For the majority, however, there is no suicidal intent. The
most common reason for non-suicidal self-injury (NSSI) is
to relieve intense distressing affect (e.g. sadness, guilt,
flashbacks, depersonalization) by the use of sharp physical
pain, which can distract the sufferer from their unbearable
feelings [3]. Other reasons include: self-punishment, which
the adolescent sees as ‘deserved’; to gain attention, so that
other people can see their distress; to make other people
feel guilty and change their behaviour; to fit in socially
with peers who self-injure. While NSSI often results in
short-term relief from distress [3], it frequently leads to
longer-term negative consequences. Emotionally, it can
evoke more complex feelings of guilt and shame towards
the self [3]. Socially, it can lead to teasing from peers and
shock from parents who may then become over-protective.
P. Wilkinson (&)
Developmental Psychiatry Section, University of Cambridge,
Douglas House, 18b Trumpington Road,
Cambridge CB2 8AH, UK
e-mail: [email protected]
123
Eur Child Adolesc Psychiatry (2013) 22 (Suppl 1):S75–S79
DOI 10.1007/s00787-012-0365-7
Physically, it can lead to infection and scarring. Despite
awareness of the negative consequences when in a calm
mood, it can be difficult to resist the strong urge to self-
harm when angry and/or depressive feelings are intense,
especially when self-injury has been rewarded in the past
by reduction of distress. It is a problem, and adolescents
often want help in stopping it.
But is NSSI enough of a problem to warrant a diagnostic
label of its own? NSSI has been proposed as a new diag-
nostic category within DSM-5 [4]. This has been praised as
a positive move by some. However, others are critical,
particularly of the fact that ‘suicidal’ and ‘non-suicidal
self-injury’ have been separated, despite significant overlap
between the two behaviours. This review will first discuss
whether it is appropriate to include NSSI as a separate
DSM-5 psychiatric diagnosis. A discussion will follow
whether it is appropriate to separate ‘suicidal’ from ‘non-
suicidal self-injury’.
Should NSSI be included within DSM-5?
Up until the proposed DSM-5 diagnosis, no psychopatho-
logical significance was given to this behaviour in younger
people, despite research findings that the condition is pre-
valent, impairing and under-recognised in adolescent
groups [1]. Furthermore, when present, there is nowhere in
the current classification of axis I emotional disorders in
young people to record this behaviour as a symptom. While
NSSI is a symptom of borderline personality disorder [5],
this is generally seen as an inappropriate diagnosis for
children and younger adolescents, whose personalities are
still developing. This implies by default that mental health
professionals do not need to consider this behaviour as
abnormal or even non-normative. The combination of the
feeling of intense negative distress (often characterised by
anger and/or sadness) and self-harm, together with their
adverse consequences, is both necessarily within the clin-
ical domain and of sufficient importance to warrant a
separate diagnostic category in young people.
The new proposed category of NSSI will be helpful for
several reasons: to improve communication; to reduce
problems from the lack of diagnostic specificity for NSSI;
to improve provision of treatment for adolescents who
engage in NSSI and improve research on aetiology, treat-
ment and outcome. These will be detailed below.
Improved communication
Concurrent as well as sequential co-morbidity of psychi-
atric diagnoses is common in mentally unwell and
behaviourally disturbed children and adolescents. It greatly
aids communication when all psychiatric diagnoses are
listed as part of inter-professional communication, espe-
cially if a multi-axial format is used [5]. It also helps cli-
nicians to think of all the problems that need addressing.
Elevating NSSI to a full diagnostic category will increase
the visibility of this abnormal behaviour, ensuring that it is
recorded within a clinical formulation, thereby giving
greater clarity to the range of problems suffered by an
adolescent. Of course, this will only be useful if the pres-
ence of NSSI has implication in terms of treatment and
prognosis. This is indeed the case as shall be argued below.
Reduced problems from lack of diagnostic
specificity for NSSI
The only DSM-IV psychiatric diagnosis which includes
NSSI as a criterion is borderline personality disorder
(301.83) [5]. This can lead some clinicians to make an
automatic assumption that a patient who engages in NSSI
may have this condition (BPD) which is not appropriate for
multiple reasons. Firstly, many clinicians believe that a
diagnosis of personality disorder should not be given to a
child or younger adolescent, whose personality is still
developing [6]. Secondly, NSSI is often present in patients
with psychiatric disorders, including depression, post-
traumatic stress disorder, other anxiety disorders, conduct
disorder, and substance misuse disorders [7, 8]. Thirdly,
NSSI may be present and may be a problem in a patient
who does not meet diagnostic criteria for a mental illness or
personality disorder. It is clearly crucial to make the correct
diagnosis, so that appropriate treatment is given. Assigning
a specific diagnostic category to NSSI makes it more likely
that clinicians will consider treatment and management in
more flexible ways than may occur if the assumption is
made that NSSI is simply indexing a borderline personality
disorder. Fourthly, an assumption of BPD may lead to a
bias in clinician perception of an adolescent with mental
health difficulties. We know little about borderline per-
sonality disorder components from a developmental con-
text and there is a risk of translating down to the adolescent
years information about treatment and management gen-
erated from an older and putatively different clinical pop-
ulation. Indeed, one of the reasons for including NSSI in
DSM-5 was to stop individuals who self-injure being
inappropriately diagnosed with borderline personality
disorder.
Improved provision of treatment
NSSI may be present with a psychiatric disorder, but for
many who engage in this behaviour, there is no associated
set of signs and symptoms to warrant a psychiatric diag-
nosis. In the absence of a mood or behavioural disorder (or
personality disorder in older adolescents), there is currently
S76 Eur Child Adolesc Psychiatry (2013) 22 (Suppl 1):S75–S79
123
no place for recording this behaviour. Nevertheless, ado-
lescents who engage in NSSI are deserving of treatment:
both to reduce the distress and impairment from their
negative emotional state that leads them to self-injure, and
to reduce the harms from the NSSI. In health care systems
where funding may be dependent on diagnostic labels for
payment, distressed self-harming adolescents will not
receive treatment unless they are labeled with an ‘incor-
rect’ diagnosis for their presenting complaint. It is not
known if this is currently occurring. The addition of NSSI
to DSM-5 will improve the chances of assessment, treat-
ment and management being given in the absence of other
emotional conditions.
NSSI has prognostic implications
NSSI in itself has negative emotional, physical and social
consequences. However, it is also associated with the more
worrying behaviour of suicide attempts. Cross-sectional
studies have demonstrated that adolescents who engage in
NSSI are more likely to make suicide attempts than those
who do not engage in NSSI [9]. Longitudinal research has
demonstrated that a history of NSSI before the index sui-
cide attempt is more likely in adolescent in-patients who
have attempted suicide than those who have suicidal ide-
ation but not suicide attempts [10]. Data from one of our
own studies (a secondary analysis of the ADAPT data)
demonstrate that the strongest predictor of a suicide
attempt over a 6-month follow-up in depressed adolescents
receiving treatment under randomized controlled trial
conditions is NSSI, but not suicide attempts, at baseline
[11]. This significant association holds when controlling
for other potential confounders at entry, including suici-
dality, gender, severity of depression and current family
dysfunction. This finding has since been replicated in a
similar population [12]. A large follow-up study of people
(of all ages), who presented to hospital with self-harm,
demonstrated that completed suicide is predicted by both
NSSI and suicide attempt as the index presentation, with no
difference in the likelihood of future suicide between
baseline suicidal and non-suicidal self-harm [13]. This
longitudinal research suggests that NSSI is a risk factor for
future suicide attempts. As only a minority of people who
have suicidal ideation attempt suicide, it may be that the
presence of NSSI is a marker for crossing the boundary
between ideas and suicidal acts. From the neuropsycho-
logical perspective it may be that it represents a behav-
ioural outcome of impaired behavioural inhibition at the
time of an acute rise in negative emotional tone (fear,
anger, irritability or sadness), increasing impulsivity for
and decreasing regulatory factors against adverse self-
harming behaviour [7]. Once NSSI is established it is
possible that recurrence habituates people to the pain from
self-harm suggesting that they are insensitive to punishing
outcomes and/or more directed to the acute rewards [7].
Most studies to date have not been large enough to deter-
mine predictors for completed suicide, but suicide attempts
are the strongest risk factor for future suicide in adolescents
[14]. Therefore, NSSI is likely to be a predictor of com-
pleted suicide adding to the evidence that the behaviour is
clinically important.
The association of NSSI with future suicide risk means
that adolescents who engage in NSSI without any apparent
emotional or behavioural disorder should be properly
assessed for suicide risk. Furthermore adolescents with
other known psychiatric risks factors for suicide, such as
existing clinical depression, substance misuse and conduct
disorder, should be assessed for a current and lifetime
history of NSSI [14]. Finally, NSSI may well have prog-
nostic significance for the development of borderline per-
sonality disorder in young adult life and earlier recognition
and treatment may reduce the risk of developing this per-
vasive and disabling condition.
Specific treatment for NSSI
The presence of short- and long-term adverse consequences
of NSSI means that treatment needs to be considered in its
own right and not just if the adolescent meets criteria for a
mental illness or behavioural disorder. Successful treat-
ment of existing psychiatric disorders may mean that NSSI
stops, but this does not always happen. Assigning NSSI as
a separate disorder is likely to increase the chance that
specific treatment will be offered. Once this occurs there is
a greater chance than hitherto that effective treatments will
be developed. Sadly, most research that has investigated
strategies to reduce self-injury (whether separated into
suicidal and non-suicidal self-injury, or whether these two
categories were combined) have failed to demonstrate
significant efficacy or effectiveness more than control
treatments. The only treatment demonstrated to reduce
self-harm in non-selected groups at any age in a recent
NICE review was flupentixol (a small trial of a drug with
significant side-effects) [15, 16].
Dialectical behaviour therapy (DBT) and mentalization-
based therapy (MBT) have been demonstrated to be more
effective than treatment as usual at reducing suicidal and
non-suicidal self-injury in adults with the specific diagnosis
of borderline personality disorder [17, 18]. However, such
lengthy and intense treatment may not be effective or
acceptable (to patients or treatment funders) in patients
who self-harm, but do not have a personality disorder.
Treatment evidence for NSSI is even more limited in
adolescents. There have been no randomized controlled
trials of DBT in adolescents and the non-randomized pilot
studies that have taken place have not demonstrated any
Eur Child Adolesc Psychiatry (2013) 22 (Suppl 1):S75–S79 S77
123
difference between DBT and treatment as usual [19]. A
recent RCT of MBT in 80 adolescents with depression and
self-harm demonstrated a reduction in self-harm and
depression, although whether results would generalise to
self-harming adolescents without depression is unknown
[20]. Despite encouraging findings from an earlier study, a
specific group intervention (using CBT and DBT tech-
niques) led to higher self-harm rates than treatment as usual
in a study of self-injuring adolescents [21]. Two random-
ized controlled trials have demonstrated that adding cog-
nitive-behavioural therapy (CBT) to selective serotonin
reuptake inhibitor (SSRI) antidepressants leads to a sig-
nificant additional reduction in depressive symptoms. In
contrast neither led to a significant additional reduction in
self-injury, whether suicidal and non-suicidal self-injury
were classified together [22] or separately [23].
The lack of evidence for effective treatments for NSSI in
young people demonstrates the need to develop treatments
that are more specific than assuming indirect response via
treating the primary psychiatric condition; and subsequently
test these in large, adequately-powered studies.
Improved research
Although there is an abundance of descriptive research on
adolescent self-injury, most fail to distinguish suicidal from
non-suicidal behaviours. As I argue below, these behav-
iours are different. It is important that future research
distinguishes the two forms of self-injury. This will provide
more information on the prevalences of these behaviours,
thus revealing the public health implications and treatment
needs. A more detailed study of mechanisms of self-injury
and comparing these with suicide will ensue, including the
neural systems that may underpin each or both. Such
mechanistic studies would provide more information than
that is currently present on the aetiologies of the two
behaviours and is likely to inform the design of specific
treatments. Most importantly, it should lead to treatment
studies testing specific interventions for NSSI in the pres-
ence and absence of other psychopathologies.
We believe that the inclusion of NSSI in DSM-5 will
encourage research specifically on NSSI, rather than
including it within the broader category of self-injury.
Should non-suicidal and suicidal self-injury be
separated?
It is certainly true that suicidal and non-suicidal self-injury
are overlapping behaviours, and that NSSI is a strong risk
factor for suicide attempts [11]. There are common aetio-
logical factors, which increase the risk of both behaviours:
patients with depression often experience intensely
distressing affect; this may lead to either NSSI (to relieve
distressing affect) or suicide attempts (when death is seen
as the best way out); the combination of impulsivity and a
lack of alternative coping strategies may lead to a mentally
ill young person to cross the threshold between thinking of
self-injury and carrying out the act; a childhood history of
abuse can lead to both suicidality and NSSI [3]. Neuro-
psychologically, patient groups with either suicidal or non-
suicidal self-injury have some similar problems with
poorer response inhibition and more risky decision making
than controls [24–27].
However, they are not the same behaviours. Firstly, the
intent of both behaviours is often different. The aim of
NSSI is often intense pain, which will lead to the reduction
of distressing effect; some patients actually do this to
reduce the suicidal thoughts that a strong part of them does
not want to have. People usually carry out suicide attempts
to try to end their life, often as painlessly as possible. It is
true that some people are unsure of their intent. But many
are clear that there is no suicidal intent in their behaviour
and that they definitely do not want to end their lives; such
individuals can be accurately classified. Secondly, the
degree of planning often varies: NSSI is often rapid with
little planning, particularly if it is a highly repeated and
habitual behaviour. Suicide attempts are more likely to be
preceded by careful planning, particularly if suicidal intent
is high. Thirdly, there are some differences in aetiological
factors. In our own ADAPT study, anxiety disorders were
only associated with NSSI, while poor family function was
only associated with suicide attempts [11]. Fourthly, the
two have different associations with other psychiatric dis-
orders: adolescents who only attempt suicide are more
likely to have a concurrent clinical disorder of depression
or PTSD than those who only engage in NSSI; those who
only engage in NSSI are more likely to have features of
borderline personality disorder [8]. Fifthly, the statistical
association between the two types of self-harm is not total.
For example, in the ADAPT study of depressed adoles-
cents, those with NSSI in the month pre-baseline had a
55 % risk of a suicide attempt over 28 weeks of follow-up.
This was substantially elevated over the 18 % risk in those
with no NSSI in the month pre-baseline. However, this was
less than the 69 % risk of NSSI over follow-up in those
with baseline NSSI. And it also means that nearly half of
those with baseline NSSI did not subsequently make a
suicide attempt. It may therefore be better to treat these as
overlapping behaviours, not as the same behaviour. This
mirrors other findings of high co-morbidity in mental
health between distinct disorders, such as depressive and
anxiety disorders, and attention-deficit/hyperactivity dis-
order and conduct disorder [28].
One clear argument against DSM-5 defining ‘non-sui-
cidal self-injury’ is that by its very name, people (including
S78 Eur Child Adolesc Psychiatry (2013) 22 (Suppl 1):S75–S79
123
emergency department staff) will see this behaviour as not
important as it is ‘only non-suicidal’; they may therefore
not think it so important to treat these patients (Naveen
Kapur, personal communication). While this is possible,
the problem here could be better seen as ignorant staff, not
a bad diagnosis. In fact, the opposite could be argued: the
presence of the DSM-5 diagnosis of NSSI may lead to
people being made more aware of it and its strong asso-
ciation with suicidal behaviour, hence leading patients
presenting with it to be better assessed and treated. The key
good education of our professional colleagues.
Conclusions
Non-suicidal self-injury is common in adolescents. It is
associated with intense negative thoughts and feelings, and
negative emotional, physical and social consequences. The
behaviour may occur in isolation or in association with one
of a number of specific psychiatric syndromes. The inclu-
sion of NSSI within DSM-5 should lead to better research
into, and treatment of, this important and pathological
behaviour.
Conflict of interest The corresponding author states that there are
no conflicts of interest. This article is part of the supplement ‘‘The
Future of Child and Adolescent Psychiatry and Psychology: The
Impact of DSM 5 and of Guidelines for Assessment and Treatment’’.
This supplement was not sponsored by outside commercial interests.
References
1. Hawton K, James A (2005) Suicide and deliberate self harm in
young people. BMJ 330(7496):891–894
2. Giletta M et al (2012) Adolescent non-suicidal self-injury: a
cross-national study of community samples from Italy, the
Netherlands and the United States. Psychiatry Res 197:66–72
3. Briere J, Gil E (1998) Self-mutilation in clinical and general
population samples: prevalence, correlates, and functions. Am J
Orthopsychiatry 68(4):609–620
4. APA (2010) Non-Suicidal Self Injury. In: APA DSM-5 Devel-
opment. http://www.dsm5.org/ProposedRevisions/Pages/propose
drevision.aspx?rid=443#. Accessed 25 May 2012
5. APA, American Psychiatric Association (2000) Diagnostic and
statistical manual of mental disorders: text revision, 4th edn.
American Psychiatric Association, Washington, DC
6. WHO (1992) The ICD-10 classification of mental and behav-
ioural disorders: clinical descriptions and diagnostic guidelines.
World Health Organization, Geneva
7. Nock MK et al (2006) Non-suicidal self-injury among adoles-
cents: diagnostic correlates and relation to suicide attempts.
Psychiatry Res 144(1):65–72
8. Jacobson CM et al (2008) Psychiatric impairment among ado-
lescents engaging in different types of deliberate self-harm. J Clin
Child Adolesc Psychol 37(2):363–375
9. Brunner R et al (2007) Prevalence and psychological correlates of
occasional and repetitive deliberate self-harm in adolescents.
Arch Pediatr Adolesc Med 161(7):641–649
10. Zlotnick C et al (1997) Affect regulation and suicide attempts in
adolescent inpatients. J Am Acad Child Adolesc Psychiatry
36(6):793–798
11. Wilkinson P et al (2011) Clinical and psychosocial predictors of
suicide attempts and nonsuicidal self-injury in the adolescent
depression antidepressants and psychotherapy trial (ADAPT).
Am J Psychiatry 168(5):495–501
12. Asarnow J et al (2011) Suicide attempts and nonsuicidal self-
injury in the treatment of resistant depression in adolescents
(TORDIA) study. J Am Acad Child Adolesc Psychiatry
50(8):772–781
13. Cooper J et al (2005) Suicide after deliberate self-harm: a 4-year
cohort study. Am J Psychiatry 162(2):297–303
14. Bridge JA, Goldstein TR, Brent DA (2006) Adolescent suicide
and suicidal behavior. J Child Psychol Psychiatry 47(3–4):
372–394
15. NICE (2011) Self-harm: longer-term management. National
Collaborating Centre for Mental Health, London
16. Montgomery S et al. (1979) Maintenance therapy in repeat sui-
cidal behaviour: a placebo controlled trial. In: Proceedings of the
10th International Congress for Suicide Prevention and Crisis
Intervention. Ottawa, Canada
17. Bateman A, Fonagy P (2008) 8-year follow-up of patients treated
for borderline personality disorder: mentalization-based treatment
versus treatment as usual. Am J Psychiatry 165(5):631–638
18. Lynch TR et al (2007) Dialectical behavior therapy for borderline
personality disorder. Annu Rev Clin Psychol 3:181–205
19. Katz LY et al (2004) Feasibility of dialectical behavior therapy
for suicidal adolescent inpatients. J Am Acad Child Adolesc
Psychiatry 43(3):276–282
20. Rossouw TI, Fonaghy P (2012) Mentalization-based treatment for
self-harm in adolescents: a randomised control trial. JAACAP.
doi:10.1016/j.jaac.2012.09.018
21. Hazell PL et al (2009) Group therapy for repeated deliberate self-
harm in adolescents: failure of replication of a randomized trial.
J Am Acad Child Adolesc Psychiatry 48(6):662–670
22. March J et al (2004) Fluoxetine, cognitive-behavioral therapy,
and their combination for adolescents with depression: treatment
for Adolescents With Depression Study (TADS) randomized
controlled trial. J Am Med Assoc 292(7):807–820
23. Brent DA et al (2009) Predictors of spontaneous and systemati-
cally assessed suicidal adverse events in the treatment of SSRI-
resistant depression in adolescents (TORDIA) study. Am J Psy-
chiatry 166(4):418–426
24. Carrion VG et al (2008) Posttraumatic stress symptoms and brain
function during a response-inhibition task: an fMRI study in
youth. Depress Anxiety 25:514–526
25. Janis IB, Nock MK (2009) Are self-injurers impulsive?:results
from two behavioral laboratory studies. Psychiatry Res 169(3):
261–267
26. Jollant F et al (2010) Decreased activation of lateral orbitofrontal
cortex during risky choices under uncertainty is associated with
disadvantageous decision-making and suicidal behavior. Neuro-
image 51:1275–1281
27. Raust A et al (2007) Prefrontal cortex dysfunction in patients
with suicidal behavior. Psychol Med 37:411–419
28. Angold A, Costello EJ, Erkanli A (1999) Comorbidity. J Child
Psychol Psychiatry 40(1):57–87
Eur Child Adolesc Psychiatry (2013) 22 (Suppl 1):S75–S79 S79
123
