1. health and fitness
2. disorders

European Psychiatry 19 (2004) 123–130
www.elsevier.com/locate/eurpsy
Original article
Mental disorders in a forensic sample of sexual offenders
Anja Leue a,*, Bernd Borchard b, Jürgen Hoyer c
a
Central Institute of Mental Health, Mannheim, J5, 68159 Mannheim, Germany
b
State Forensic Hospital of Moringen, Moringen, Germany
c
Dresden University of Technology, Dresden, Germany
Received 28 August 2002; accepted 26 August 2003
Available online 20 April 2004
Abstract
Objective. – The present study examined the prevalence of DSM IV axis I disorders and DSM IV personality disorders among sexual
offenders in Forensic State Hospitals in Germany.
Method. – Current and lifetime prevalence rates of mental disorders were investigated based on clinical structured interviews among sexual
offenders (n = 55). Additionally, subgroups were analyzed on the basis of diagnostic research criteria, with 30 sexual offenders classified as
paraphiliacs and 25 sexual offenders as having an impulse control disorder (without paraphilia).
Results. – Anxiety disorders, mood disorders, and substance use disorders were common among sexual offenders, as were cluster B and
cluster C personality disorders. While social phobia was most common among paraphilic sexual offenders, major depression was most
prevalent in impulse control disordered sexual offenders.
Conclusion. – The results replicate recent findings of high psychiatric morbidity in sexual offenders placed in forensic facilities.
Furthermore, differential patterns of co-morbid mental disorders were found in paraphiliacs and impulse control disordered sexual offenders.
With regard to an effective therapy and relapse prevention co-morbid mental disorders should be a greater focus in the assessment of subgroups
of sexual offenders.
© 2004 Elsevier SAS. All rights reserved.
Keywords: Sexual offenders; Paraphilia; Impulse control disorder; Axis I disorder; Personality disorder
1. Introduction
Several studies have documented a relationship between
sexual delinquency and mental disorders [5,10,13,20,32],
mostly on the basis of questionnaires. Using this methodology, sexual offenders have described themselves as socially
anxious and depressed, or have reported regular alcohol
abuse. However, these studies are constrained because it is
unclear whether participants fulfill diagnostic criteria, e.g. of
the DSM IV [34], since questionnaires allow a screening
assessment but no assured diagnosis of mental disorders.
Recent studies focused explicitly on DSM III R or DSM IV
diagnoses in sexual offenders by using structured diagnostic
interviews. These studies found various axis I and personality disorders in sexual offenders (cf. [14,15,26,33]). In Raymond et al. [33] 93% of the 45 examined pedophilic sexual
offenders met the diagnostic criteria for at least one co* Corresponding author.
E-mail address: [email protected] (A. Leue).
© 2004 Elsevier SAS. All rights reserved.
doi:10.1016/j.eurpsy.2003.08.001
morbid axis I disorder. In the study of McElroy et al. [26], 35
(97%) of the 36 paraphilic and non-paraphilic sexual offenders suffered from at least one further axis I disorder;
28 sexual offenders (78%) showed three or more co-morbid
axis I disorders. Mood, anxiety and eating disorders were
significantly more common in paraphilic than in nonparaphilic sexual offenders. Social phobia was most prevalent among anxiety disorders, varying between 19% and 31%
in paraphilic sexual offenders. In the non-paraphilic sample,
the prevalence rate was 13% (cf. [26,33]). Hoyer et al. [14]
found increased social interaction anxiety in paraphilic
sexual offenders. These results are in line with studies, which
discuss increased social phobia as one of the accompanying
conditions of sexual delinquency [15,17,30,36,37,39].
Moreover, substance use disorders also emerged as highly
prevalent according to McElroy et al. [26] and Raymond et
al. [33]. Leygraf [21] pointed out that violent sex offenses
have been frequently committed under psychoactive substances (e.g. alcohol or drugs). Allnutt et al. [1] found a high
co-morbidity of alcoholism in paraphilic sexual offenders.
124
A. Leue et al. / European Psychiatry 19 (2004) 123–130
Estimates of prevalence rates for alcoholism varied between
51% and 67% in paraphilic and non-paraphilic sexual offenders [26,33]. It has been assumed that alcohol could be
relevant in sexual offenses because of minimizing anxiety
and increasing aggressive impulses as well as inhibiting
adequate cognitive perception or processing, and inducing
positive mood [1].
Prevalence rates of personality disorders in delinquent
groups—including sexual offender groups—have been estimated as high—between one third [21,36] and two thirds
[2,3]. Several studies revealed high prevalence rates of antisocial, borderline, schizotypal, and narcissistic personality
disorder in sexual offenders. These personality disorders
have been discussed as accompanying factors for disinhibited behavior in sexual delinquency [18,21,36]. However,
personality disorders such as avoidant, passive–aggressive,
dependent, and obsessive–compulsive personality disorder
have also been diagnosed in sexual offenders [5,26,33].
The studies cited above suggest an interrelation between
mental disorders and sexual offenders (i.e. paraphilia). Moreover, various mental disorders were found in sexual offenders without paraphilia (see [26]). Wulfert et al. [41] described
a subgroup of sexual offenders that is not prone to sexual
deviant fantasies like paraphiliacs but exhibits sexually deviant aggressive behavior as a—potentially recurring—
reaction to a situational loss of reinforcement, e.g. because of
unemployment, interpersonal conflicts, or frustrations.
Wulfert et al. [41] emphasized that the underlying symptoms
correspond to an impulse control disorder, since in these
sexual offenders arousal increases prior to the sex offense
and decreases with committing the offense. To describe
sexual offenders with such a non-paraphilic, but recurring
pattern of sexual offending, Hoyer et al. [13] suggested the
use of the DSM IV category of impulse control disorder not
otherwise specified (n.o.s., DSM IV 312.30) (see below).
The present study refers to the investigation of sexual
offenders in State Forensic Hospitals—not of sexual offenders in prison. These sexual offenders have been psychiatrically assessed prior to the trial. According to §§ 20, 21 of the
German penal law it has to be examined whether a severe
mental disorder (e.g. personality disorder, schizophrenia,
paraphilia, intellectual disability) has markedly reduced persons ability of insight and/or their ability to control their
actions. If a severe mental disorder has been diagnosed as a
factor that influenced the commitment of the sex offense,
placement in a forensic facility is ordered according to §
63 penal law (placement for therapeutic purposes for an
unlimited time period) or § 64 penal law (placement for the
treatment of substance use disorders for 2 years).
Besides the diagnoses of severe mental disorders which
could have influenced the commitment of a sex offense
current and lifetime mental disorders were rarely systematically investigated in sexual offenders. Especially for differential therapeutic interventions in subgroups of sexual offenders and effective relapse prevention it is important to
know whether diagnostic subgroups of sexual offenders,
paraphiliacs and impulse control disordered persons, differ
with regard to specific mental disorders. Thus, the aim of the
present study is to provide more detailed descriptions of
mental disorders among sexual offenders placed in State
Forensic Hospitals using structured diagnostic interviews.
This method will gather more detailed information about
relevant subgroups of sexual offenders. Beside an offenserelated classification (such as rape versus child molestation)
a disorder-related classification between paraphilia and impulse control disorder according to the DSM IV criteria has
also been used. In contrast to the study of Raymond et al. [33]
who investigated pedophilic sexual offenders only—as one
subgroup of paraphiliacs with hands-on-sex-offenses—this
study includes pedophilic (DSM IV 302.2) sexual offenders
as well as sexual sadists (DSM IV 302.84). Moreover, it was
considered whether sexual offenders without paraphilia can
be classified as impulse control disordered n.o.s. (DSM IV
312.30). This is in contrast to McElroy et al. [26] who did not
further diagnostically specify their group of non-paraphilic
sexual offenders.
Thus, the present study focuses on three major issues: (1)
which axis I and personality disorders can be diagnosed in
sexual offenders in State Forensic Hospitals in general, and
more specifically, in paraphilic and impulse control disordered sexual offenders? (2) Do paraphilic and impulse control disordered sexual offenders differ in axis I disorders
only, as in McElroy et al. [26], or also in personality disorders? (3) In order to explore the extent, to which co-morbid
diagnoses are relevant for effective therapeutic treatment and
relapse prevention in sexual offenders, we investigated the
temporal relation between mental disorders and sex offenses.
2. Method
2.1. Sample
Participants were recruited from the State Forensic Hospitals in Arnsdorf, Saxony, and Moringen, Lower Saxony,
Germany. All participants were placed in the forensic facilities for therapeutic purposes according to § 63 of the German
penal law. In both above-mentioned hospitals 141 sexual
offenders were incarcerated at the time of the study. Sexual
offenders with a history of psychosis (n = 12), with analphabetism (n = 2), with intellectual (n = 36) or neurological
disabilities (n = 11) were excluded. Moreover, 21 (17%)
sexual offenders declined to participate in this study, and four
patients could not be examined because of extramural therapeutic interventions. The remaining 55 patients signed the
written consent form and participated voluntarily in this
study. All participants were male. Participants were selected
using the diagnostic criteria of the DSM IV for the hands-onparaphilias—pedophilia and sexual sadism—and impulse
control disorder n.o.s. (DSM IV 312.30).
File information was used to classify the sexual offenders
into the above-mentioned categories. The psychiatric ex-
A. Leue et al. / European Psychiatry 19 (2004) 123–130
125
Table 1
Sociodemographic variables in sexual offenders (n = 55)
Paraphilia (n = 30)
Time of incarceration (month)
Age (years)
Schoolgrade
No school grade
Lowest level school leaving certificate (“Hauptschule”)
Middle/upper level school leaving certificate (“Realschule, Abitur”)
Professional grade
Profession certificate
No profession certificate
Marital status
Single
Married
Divorced
a
M
48.6
36.5
S.D.
35.3
9.8
t-test a /U-test
Impulse control
disorder (n = 25)
M
S.D.
57.5
39.4
30.7
8.9
T(53) = 0.88, P = 0.38
T(53) = 2.29, P < 0.05
n
n
6
17
7
6
17
2
}
15
15
11
14
}
22
1
7
21
–
4
}
U = 317.5, P = 0.26
U = 352.5, P = 0.66
U = 337.5, P = 0.37
Time of incarceration and age are normally distributed (Kolmogorov–Smirnov-test: P > 0.75).
pert’s opinions and the verdicts were used to obtain information pertaining to paraphilia or impulse control disorder
n.o.s. Since for many sexual offenders it is very painful to
explicate their sexual fantasies and details of their sexual
offenses, file information was preferred in order to prevent
socially desirable responding and cognitive distortions.
For diagnosing paraphilia or impulse control disorder
n.o.s. on the basis of the file information, it was necessary to
specify the definitions of the DSM IV criteria [13] further:
Pedophilia (DSM IV 302.2) was diagnosed if (A) for at least
6 months repeated intense sexual deviant fantasies or behaviors occurred, and if the hands-on-sex-offense was realized
with a child of 13 years or younger. (B) These sexual deviant
fantasies or behaviors have caused clinically relevant impairment in social and professional areas. (C) The sexual offender is at least 16 years old and at least 5 years older than
the child or the children who was/were sexually abused.
Deviant sexual fantasies were either reported or there were
repetitive features throughout the offenses with an obviously
pedophilic theme, and no sex offenses that do not confirm to
these criteria.
Sexual sadism (DSM IV 302.84) was diagnosed if (A) for
at least 6 months repeated intense sexual deviant fantasies or
behaviors occurred which included at least one hands-onsex-offense in that the sexual offender is sexually aroused by
the physical or mental pain and humiliation of the victim. (B)
(C) These sexual deviant fantasies or behaviors have caused
clinically relevant impairment in social and professional areas. Deviant sexual fantasies were either reported or there
were repetitive features throughout the offenses with an obviously sadistic theme, and no sex offenses that do not confirm
to these criteria.
Impulse control disorder n.o.s. (DSM IV 312.30) was
diagnosed if (A) there was at least one hands-on-sex-offense
with a reported loss of impulse control before and during the
offense. (B) No sexual deviant fantasies were registered
before or during the sexual offense. (C) Loss of impulse
control occurred also in non-sexual areas. (D) There are
indications for increased sexual tension and arousal but no
indication of deviant sexual fantasies.
All participants could be classified to one of the suggested
diagnostic categories. Thirty of the examined sexual offenders were classified as paraphiliacs, and in 25 sexual offenders
an impulse control disorder n.o.s. was diagnosed. Classification of participants was performed by the first author. Another independent rating was performed by the second author. The interrater-reliability was j = 0.85. Within the
paraphilic sample, 18 sexual offenders were pedophiles
(DSM IV 302.2) and 10 were sexual sadists (DSM IV
302.84). The group did not differ with regard to time of
incarceration (Table 1). As also pointed out in Table 1,
paraphilic sexual offenders were older than impulse control
disordered participants. There were no significant correlations between age and diagnosed mental disorders (Kendall’s
Tau <0.22, n.s.; in all analyses). Almost one third of the
55 investigated sexual offenders did not graduate from secondary school, one half of the 55 sexual offenders had no
profession and two third of the paraphiliacs and the impulse
control disordered n.o.s. sexual offenders were single (see
Table 1). Analyzing the offense-related criteria (rape versus
child molestation) with regard to the disorder-related criteria,
it was shown that of the 21 child molesters 19 were
paraphilacs and two were impulse control disordered sexual
offenders. Rape was committed by 11 sexual offenders being
classified as paraphiliacs and by 23 classified as impulse
control disordered.
2.2. Interviews
Structured clinical interviews were used for diagnostic
assessment. The diagnosis of further axis I disorders other
than paraphilia and impulse control disorder n.o.s. was based
126
A. Leue et al. / European Psychiatry 19 (2004) 123–130
on the “Diagnostische Interview für psychische Störungen—
Kurzversion” (Mini-DIPS) (Diagnostic Interview for Psychiatric Disorders—short version) [23,25]. The Mini-DIPS is a
structured interview for point and lifetime prevalence, based
on the original DIPS [24,35] and the DSM IV criteria. The
following mental disorders can be diagnosed using the MiniDIPS: anxiety disorders, mood disorders except cyclothymia, eating disorders (bulimia and anorexia), hypochondriasis, somatization disorder, conversion disorder, pain
disorder, substance abuse, and dependence. For investigating
the reliability of the Mini-DIPS, 100 patients were examined
with the DIPS and the Mini-DIPS by two independent clinical psychologists. The interrater-reliability of both instruments was j = 0.80, and Yule’s Y was between 0.84 and 1.0
[24]. Both instruments—DIPS and Mini-DIPS—proved to
be reliable for the diagnostics of mental disorders [25].
Personality disorders were diagnosed using the German
Version of the “Structured Clinical Interview for DSM IV
Axis II” by First et al. [6]—the “Strukturierte Klinische
Interview für DSM IV Achse II” (SKID II) [40]. The SKID II
is a two-stage method consisting of a questionnaire for
screening personality disorders and a structured interview for
diagnosing personality disorders on the basis of the DSM IV
criteria. Cluster A (e.g. schizotypal personality disorder),
cluster B (e.g. antisocial personality disorder), and cluster C
personality disorders (e.g. avoidant personality disorder) can
be diagnosed using SKID II. Moderate interrater-reliability
of the SKID II was identified in a sample of 91 patients of a
psychosomatic clinic (j = 0.55, Yule’s Y = 0.72–1.00) [7].
The reliabilities reported in the interview manuals are
setting-specific. Thus, it is not clear whether similar reliabilities could be obtained for both interviews outside a forensic
setting, or in another similar facility. The diagnosis of axis I
and personality disorders was performed by the first author.
Current and lifetime diagnoses were screened separately.
Current diagnoses were assigned when the diagnostic criteria
of a mental disorder were met at the time of study. If the
diagnostic criteria of a mental disorder were met both at the
time of study and at an earlier point in time, lifetime diagnoses were assigned (cf. [16]).
2.3. Statistical procedures
The results of the diagnosed axis I and personality disorders are presented as prevalence rates in percentages but also
in natural frequencies, since reporting percentages only may
lead to statistical and inferential misinterpretations [11]. Statistical analyses were conducted with SPSS 9.0 [38]. v2-tests
were performed to test group differences according to the
diagnosed mental disorders. Logistic regression analyses
were conducted to investigate whether the offense-related or
the disorder-related classification best accounted for the
variation in the most prevalent mental disorders. According
to Norman and Streiner [28] the question of how many cases
are needed for predictions based on logistic regression analysis is not yet answered. Therefore, Norman and Streiner [28]
refer to the statistical standard that is in general followed for
regression analysis. Ten subjects per predictor variable are
needed for performing logistic regression analysis, which
was followed here.
3. Results
Participants displayed high rates of lifetime axis I and
personality disorders in total group as well as in both subgroups. All impulse control disordered sexual offenders
(n = 25) and 93% (n = 28) of the paraphiliacs met the DSM
IV criteria for one or more lifetime axis I or personality
disorder. Additionally, 70% (n = 21) of the paraphilic sexual
offenders and 48% (n = 12) of the impulse control disordered
sexual offenders met criteria for at least one axis I disorder
and for at least one personality disorder. Ninety-three percent
of the examined sexual offenders (28 paraphiliacs, 23 impulse control disordered n.o.s.) suffered from at least one
axis I or personality disorder prior to the commitment of the
sex offense.
Anxiety disorders were common in this sample (Table 2):
73% of the paraphiliacs (n = 22) and 64% (n = 16) of the
impulse control disordered sexual offenders met the criteria
for at least one anxiety disorder. The difference between the
subgroups was not significant for the lifetime prevalence of
any anxiety disorder (v(1)2 = 0.56, P = 0.46, n.s.). The prevalence rate of anxiety disorders (69%) for all investigated
sexual offenders in this study was almost twice to three times
higher as in out-patient sexual offenders (n = 120) with a
prevalence rate for anxiety disorders of 39% [15] or in offenders (n = 202) placed in prison of 17% [4]. Compared to
the population with a lifetime prevalence of 1–14% for any
anxiety disorder [34] the lifetime prevalence in the sexual
offender group was approximately five times higher.
Social phobia, simple phobia, and post-traumatic stress
disorder were the most common anxiety disorders in this
sample. The prevalence rate of social phobia was in the
present sample almost twice to eight times higher (38%) than
in out-patient sexual offenders (12%) [15] and offenders in
prison (5%) [4]. Paraphilic sexual offenders met the diagnostic criteria for lifetime social phobia significantly more often
than impulse control disordered sexual offenders
(v(1)2 = 6.42, P = 0.01, odds ratio: 3.6). The prevalence rate of
social phobia was comparable high as in McElroy et al. [26]
with 31%. The subgroups did not significantly differ for
simple phobia (v(1)2 = 0.31, P = 0.58, n.s.) and post-traumatic
stress disorder (v(1)2 = 0.52, P = 0.47, n.s.). Social phobia and
simple phobia were present prior to the sexual offense
whereas post-traumatic stress disorder developed prior to but
also after sex offenses. Current axis I diagnoses have been
found for social phobia, simple phobia and post-traumatic
stress disorder only (Table 3). With the exception of one
impulse control disordered sexual offender suffering from
major depression at the time of study—the examined sexual
offenders fulfilled the diagnostic criteria for all the other axis
A. Leue et al. / European Psychiatry 19 (2004) 123–130
127
Table 2
Lifetime prevalence of axis I disorders
Anxiety disorders
Any anxiety disorder
Panic disorder
Agoraphobia
Social phobia**
Simple phobia
Post-traumatic stress disorder
Generalized anxiety disorder
Psychoactive substance use disorders
Any substance use disorder
Alcoholism
Drug use
Mood disorders
Any mood disorder
Major depression **
Dysthymia
Lifetime diagnoses
Total sample (n = 55)
n
%
Paraphilia (n = 30)
n
%
Impulse control disorder (n = 25)
n
%
38
3
1
21
22
15
5
69
6
2
38
40
27
9
22
2
1
16
13
7
3
73
7
3
53
43
23
10
16
1
0
5
9
8
3
64
4
0
20
36
32
12
31
31
9
56
56
16
17
17
7
57
57
23
14
14
2
56
56
8
31
29
3
56
53
6
11
9
2
30
30
7
20
20
1
80
80
4
v2-test: ** P < 0.01.
Table 3
Current diagnoses of axis I disorders
Anxiety disorders
Panic disorder
Social phobia
Simple phobia
Post-traumatic stress disorder
Current diagnoses
Total sample Paraphilia
(n = 55)
(n = 30)
n
1
9
14
4
%
2
16
25
7
n
1
7
9
1
%
3
23
30
3
Impulse
control
disorder
(n = 25)
n
%
0
0
2
8
5
20
3
12
I disorders at an earlier time but not at the time of the study.
Whether the investigated sexual offenders suffered from substance use disorders at the time of study, could not be reliably
diagnosed since during incarceration in a forensic state hospital the consumption of alcohol and drugs is not officially
allowed. Thus, the diagnoses of substance use disorders are
described for the lifetime. Approximately half of the examined sexual offenders met current diagnoses for social phobia
and simple phobia.
Fifty-six percent of the sexual offenders (17 paraphiliacs,
14 impulse control disordered) met the DSM IV criteria for a
lifetime substance use disorder (alcohol or drugs, Table 2).
The prevalence rate of substance use disorders in this sample
was comparably high as in male offenders in prison with 70%
[4] and out-patient sexual offenders with 41% [15].
Paraphilic and impulse control disordered sexual offenders
did not significantly differ with regard to any substance use
disorder (v(1)2 = 0.002, P = 0.96, n.s.). Approximately half of
the paraphiliacs and half of the impulse control disordered
sexual offenders were intoxicated with alcohol at the time of
their offenses and met the diagnostic criteria for alcohol
dependence. Compared to the diagnosis of alcohol and drug
dependence, alcohol abuse and drug abuse were of low fre-
quency in this sample (alcohol abuse: three paraphiliacs, one
impulse control disordered sexual offender; drug abuse: one
paraphiliac). The prevalence rate of drug dependence was
less common compared to alcohol dependence. In contrast to
the lifetime prevalence of alcohol dependence in the population, which is between 5% and 8% [34], sexual offenders
were eight times more likely to have an alcohol dependence.
Mood disorders had a very high lifetime prevalence in
impulse control disordered sexual offenders (Table 2): 80%
(n = 20) of the impulse control disordered met the diagnostic
criteria for any mood disorder, whereas just 30% (n = 11) of
the paraphiliacs met the lifetime criteria for any mood disorder (v(1)2 = 10.41, P < 0.01, odds ratio: 6.0). In impulse
control disordered sexual offenders major depression was the
most prevalent axis I disorder and major depression
(v(1)2 = 13.68, P < 0.01) was twice as often in impulse control
disordered than in paraphilic sexual offenders. Whereas in
paraphilic sexual offenders the lifetime prevalence of major
depression was 30% and thereby slightly higher than in the
male population with 10–25% [34], major depression in
impulse control disordered sexual offenders was with 80%
considerably higher (approximately three to six times). The
lifetime prevalence of mood disorders for the impulse control
disordered sexual offenders in this study was comparable to
that of out-patient sexual offenders with 72% [15] but approximately four times higher than the lifetime prevalence
for offenders in prison (22%) [4]. Sixty percent (n = 15) of
the impulse control disordered sexual offenders and 30%
(n = 9) of the paraphiliacs developed a major depression
during incarceration after sexual offending. Eating disorders
(bulimia and anorexia), hypochondriasis, somatization, conversion, and pain disorders did not occur in this sample.
Cluster B and cluster C personality disorders were also
highly prevalent in the present sample (Table 4). Forty-seven
percent (n = 14) of the paraphiliacs and 40% (n = 10) of the
128
A. Leue et al. / European Psychiatry 19 (2004) 123–130
Table 4
Lifetime prevalence of personality disorders
Lifetime diagnoses
Total sample
Paraphilia
(n = 55)
(n = 30)
Any cluster B or C
personality disorder
Cluster B
Antisocial
Borderline
Histrionic
Narcissistic
Total
Cluster C
Avoidant
Dependent
Obsessive-compulsive
Total
n
24
%
44
n
14
%
74
Impulse control
disorder
(n = 25)
n
%
10
40
19
8
0
6
33
35
15
0
11
60
11
5
0
3
19
37
17
0
10
63
8
3
0
3
14
32
12
0
12
56
13
2
6
21
24
4
11
38
9
1
3
13
30
3
10
43
4
1
3
8
16
4
12
32
was performed. The stepwise logistic regression analysis was
performed first for the offense-related classification, and
69% of the examined sexual offenders were correctly classified to the categories of child molester and rapist. Major
depression was the only significant predictor (Wald(1) = 8.00,
P < 0.01).
When conducting the stepwise logistic regression analysis
for the disorder-related classification again major depression
was entered first into the equation (Wald(1) = 8.99, P < 0.01)
and 71% of the sexual offenders were correctly classified to
the categories of paraphilia and impulse control disorder
n.o.s. Additionally, avoidant personality disorder was a significant predictor, and was entered second in the equation
(major depression: Wald(1) = 10.75, P < 0.001; avoidant
personality disorder: Wald(1) = 4.25, P < 0.05). Altogether
76% of the sexual offenders were correctly classified as
paraphilic or impulse control disordered.
4. Discussion
impulse control disordered sexual offenders met the criteria
for at least one cluster B personality disorder. Both subgroups did not significantly differ in the prevalence of cluster
B personality disorders. Regarding cluster C personality disorders both groups significantly differed: approximately
twice as many paraphiliacs (n = 12) as impulse control
disordered sexual offenders (n = 5) suffered from at least one
cluster C personality disorder (v(1)2 = 6.94, P < 0.05) at the
time of the study.
Antisocial (n = 19), avoidant (n = 13), and borderline
personality disorder (n = 8) were the most common personality disorders in the present sample. Thus, the lifetime
prevalence rates of these personality disorders were about
12 times higher in the examined sexual offenders compared
to the population with a lifetime prevalence of 1–3% [34].
The prevalence rates of these three personality disorders are
also in accordance with the prevalence rates in paraphilic
sexual offenders reported by McElroy et al. [26]. Paraphilic
and impulse control disordered sexual offenders did not
significantly differ in antisocial (v(1)2 = 0.13, P = 0.72, n.s.)
and avoidant personality disorder (v(1)2 = 1.48, P = 0.22,
n.s.). A statistical analysis for borderline personality disorder
could not be performed because of too few cases. Cluster A
personality disorders (e.g. schizotypal personality disorder)
did not occur. This could be due to the excluding criteria
described above. All diagnosed personality disorders have
been developed prior to the sex offenses.
In the present study, sexual offenders suffered from a
variety of axis I and personality disorders. Thus, the predictive value of the most prevalent mental disorders (e.g. social
phobia, simple phobia, post-traumatic stress disorder, major
depression, alcoholism, avoidant, and antisocial personality
disorder) was compared for the disorder-related categories
(i.e. paraphilia and impulse control disorder) and for the
offense-related categories (i.e. child molestation and rape). A
stepwise logistic regression analysis (Likelihood-RatioTest), in that predictors are sequentially included to analysis,
The present study shows that sexual offenders with handson-sex-offenses suffer from a variety of axis I disorders and
personality disorders. The prevalence of mental disorders in
this sample exceeded lifetime prevalence rates in the population and partly in out-patient sexual offenders [15] and offenders in prison [4]. Anxiety, mood, and substance use
disorders were, in accordance with McElroy et al. [26], most
prevalent in this sample. In contrast to McElroy et al. [26],
anxiety disorders were of comparable prevalence in
paraphilic as well as in impulse control disordered sexual
offenders. In accordance with results reported by McElroy et
al. [26] and Raymond et al. [33] social phobia was the most
common anxiety disorder in paraphilic sexual offenders followed by simple phobia. An important difference compared
to the results of McElroy et al. [26] was that post-traumatic
stress disorder was not only diagnosed in paraphilic but also
in impulse control disordered sexual offenders. In contrast to
Allnutt et al. [1] who diagnosed alcohol dependence more
often in paraphilic sexual offenders than in rapists, alcohol
dependence had almost the same prevalence in both subgroups in our study. In contrast to the study of Kunst et al.
[19] with impulse control disordered sexual offenders being
twice as often intoxicated with alcohol as paraphilic sexual
offenders during their sexual offense, approximately half of
the paraphiliacs and of the impulse control disordered sexual
offenders in the present study became sexually delinquent
with intoxication of alcohol and met the diagnostic criteria of
alcohol dependence. Whereas in paraphiliacs social phobia
was most prevalent, major depression was the most prevalent
axis I disorder in impulse control disordered sexual offenders
in this sample. Especially, our results for major depression
are in contrast to McElroy et al. [26], where paraphilic and
non-paraphilic sexual offenders did not significantly differ
with regard to mood disorders.
Moreover, considering time of onset of the diagnosed
mental disorders it has been noted that—with the exception
A. Leue et al. / European Psychiatry 19 (2004) 123–130
of post-traumatic stress disorder and major depression—the
diagnosed mental disorders have been developed prior to the
sexual offenses. Major depression developed in impulse control disordered sexual offenders during incarceration twice as
often as in paraphiliacs. One possible post-hoc explanation is
that this could be due to the aggressive impulses demonstrated by impulse control disordered sexual offenders during their sexual offenses which they probably become aware
of during incarceration and treatment. Major depression in
paraphilic sexual offenders during incarceration might be
due to feeling impaired by incarceration and social interaction problems with other patients. According to these assumptions, major depression in paraphiliacs might possibly
be a co-morbid condition of social phobia rather than an
offense-related consequence as in impulse control disordered
sexual offenders.
In the present study cluster B and cluster C personality
disorders were also highly prevalent. Paraphilic sexual offenders met the diagnostic criteria for any cluster C personality disorder significantly more often than impulse control
disordered sexual offenders did. In accordance with results
reported by Moran [27] in offenders, and especially, in sexual
offenders [26] antisocial personality disorder was the most
prevalent personality disorder, followed by avoidant and borderline personality disorder (cf. also [3]). As in McElroy et
al. [26], the investigated subgroups in this study also did not
significantly differ in the prevalence rates of specific personality disorders. As shown above, there are fewer current
diagnoses than lifetime diagnoses. This could be due to
several aspects: (1) Therapeutic interventions have been effective so that the impairment of the mental disorders has
decreased. (2) There was a spontaneous remission of symptoms. (3) A spontaneous remission of symptoms could be
also due to the placement. The placement in a forensic hospital could be—inspite of the fact of incarceration—possibly
also a protective living condition that eliminates conflict
situations that have led prior to the placement to specific
symptoms of a mental disorder (e.g. social phobia).
The present study is limited by three aspects: (1) Except
for paraphilia and impulse control disorder n.o.s., there was
no independent diagnosis by a second clinician of other axis
I disorders or the personality disorders. Thus, the determination of interrater-reliability was only possible for the sexual
offender disorder-related categories. (2) Statistical analyses
could not be conducted for some mental disorders because of
small cell size. Based on the results of this study, further
studies in sexual offenders should be conducted for testing
effect size and to replicate the existing results. On the basis of
the non-significant results shown in the present study inferences should be drawn with caution. Some non-significant
results in the present study may have been due to low statistical power, which was affected by the small sample size (see
[8]). Additional studies with larger sample sizes will allow
for analyses with increased statistical power and for more
detailed examinations whether subgroups of sexual offenders
differ significantly in further mental disorders—beyond so-
129
cial phobia and major depression. (3) The prevalence rates of
the diagnosed mental disorders in this sample were compared
not only to the population, but also to in-patient sexual
offenders, out-patient sexual offenders and offenders in
prison [4,15,26]. Due to the features of the forensic hospital
sample, such as criminal history or socio-economic status,
one cannot draw conclusions to the prevalence rates of the
population exclusively. The results of the present study are
limited to sexual offenders with hands-on-sex-offenses in
State Forensic Hospitals. Conclusions about the prevalence
rates of mental disorders in other sexual offender groups
(e.g. not incarcerated sexual offenders or sexual offenders in
prison) should be also done with caution.
The results of the present study provide a significant
indication for the relevance of a detailed structured diagnostics in mentally disordered sexual offenders. In addition to
offense-related aspects, psychopathological features should
be acknowledged in sexual offender research more intensively in future. According to Hanson and Bussière [9] more
empirical research in sexual offenders should be initiated on
the relevance of “psychological features to sexual recidivism” (p. 349). The disorder-related perspective in sexual
offender research may advance theoretical and diagnostic
knowledge about mentally disordered sexual offenders who
especially show a high risk for recidivism. Furthermore, it
could help to develop more effective and more specific treatment programs for sexual offenders with different patterns of
mental disorders [12,14]. However, structured treatment programs have little or no focus on co-morbid mental disorders
[22,29,31] so far. Thus, co-morbidity of mental disorders in
sexual offenders is probably rather indirectly treated by combining therapeutic interventions (e.g. social competence
training, self-esteem building, improving coping effectiveness) in single or group therapy of sexual offenders. The
differentiation of therapeutic programs for subgroups of
sexual offenders could lead to more direct treatment of comorbid mental disorders.
Acknowledgements
We are grateful to directors, staff and participants in the
State Forensic Hospitals at Arnsdorf (Saxony) and Moringen
(Lower Saxony), Germany, for making this study possible.
We also wish to thank Heike Kunst (Kassel, Germany) for
advises regarding classificatory specifications, André Beauducel (Mannheim, Germany) for discussing statistical aspects, Steven Feelgood (Berlin, Germany) and Renee Goodwin (NewYork) for helpful comments on an earlier version of
this paper.
References
[1]
Allnutt SH, Bradford JMW, Greenberg DM, Curry S. Co-morbidity of
alcoholism and the paraphilias. J Forensic Sci 1996;41:234–9.
130
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
A. Leue et al. / European Psychiatry 19 (2004) 123–130
Berger P, Berner W, Bolterauer J, Guitierrez K, Berger K. Sadistic
personality disorder in sexual offenders: relationship to antisocial
personality disorder and sexual sadism. J Pers Disord 1999;13:175–
86.
Berner W, Berger P, Guitierrez K, Jordan B, Berger K. The role of
personality disorders in the treatment of sexual offenders. J Offender
Rehabil 1992;18:25–37.
Brink JH, Doherty D, Boer A. Mental disorder in federal offenders: a
Canadian prevalence study. Int Law Psychiatr 2001;24:339–56.
Chantry K, Craig RJ. Psychological screening of sexually violent
offenders with the MCMI. J Clin Psychol 1994;50:430–5.
First MB, Gibbon M, Spitzer RL, Williams JBW, Benjamin L. Structured clinical interview for DSM-IV axis II personality disorders.
Washington, DC: American Psychiatric Press; 1997.
Fydrich T, Schmitz B, Hennch C, Bodem M. Zuverlässigkeit und
Gültigkeit diagnostischer Verfahren zur Erfassung von Persönlichkeitsstörungen. [Reliability and validity of instruments for the diagnostics of personality disorders]. In: Fydrich T, Limbacher K, editors.
Persönlichkeitsstörungen: Diagnostik und Psychotherapie [Personality disorders: diagnostics and psychotherapy]. Weinheim: Beltz/PVU;
1996. p. 91–113.
Gigerenzer G. The superego, the ego, and the id in statistical reasoning. In: Keren G, Lewis C, editors. A handbook for data analysis in the
behavioral sciences: methodological issues. New Jersey: LEA; 1993.
p. 311–39.
Hanson RK, Bussière M. Predicting relapse: a meta-analysis of sexual
offender recidivism studies. J Consult Clin Psychol 1998;66:348–62.
Hayashino DS, Wurtele SK, Klebe KJ. Child molester—an examination of cognitive factors. J Interpersonal Violence 1995;10:106–16.
Hoffrage U, Lindsey S, Hertwig R, Gigerenzer G. Communicating
statistical information. Science 2000;290:2261–2.
Hoyer J, Borchard B, Kunst H. Diagnostik und störungsspezifische
Therapie bei Sexualdelinquenten mit psychischen Störungen. [Diagnostics and disorder-related therapy in mentally disordered sexual
offenders]. Verhaltenstherapie 2000;10:7–15.
Hoyer J, Kunst H, Borchard B, Stangier U. Paraphile versus Impulskontrollgestörte Sexualstraftäter: Eine psychologisch valide Differenzierung? [Paraphilic versus impulse control disordered sexual
delinquents]. Z Klin Psychol 1999;28:37–44.
Hoyer J, Kunst H, Schmidt A. Social phobia in sexual offenders. J
Nerv Ment Dis 2001;189:463–70.
Kafka MP, Hennen J. A DSM-IV axis I comorbidity study of males
(n = 120) with paraphilias and paraphilia-related disorders. Sexual
Abuse 2002;14:349–66.
Kleinbaum DG, Kupper L, Morgenstern H. Epidemiologic
research—principles and quantitative methods. New York: VNR;
1982.
Knight RA, Prentky RA. Classifying sexual offenders—the development and corroboration of taxonomic models. In: Marshall WL,
Laws DR, Barbaree HE, editors. Handbook of sexual assault. New
York: Plenum Press; 1990. p. 23–52.
Kröber HL. Strafrechtliche Begutachtung von Persönlichkeitsstörungen. [Juridical appraisal of personality disorders]. Persönlichkeitsstörungen 1997;4:161–1117.
Kunst H, Hoyer J, Borchard B. Alkoholeinfluss bei Sexualdelikten
unter differential-diagnostischer Perspektive. [Alcohol intoxication in
sex offenders from differential-diagnostic perspective]. Sucht 2000;
46:137–41.
Lanyon RI. Theory and treatment in child molestation. J Consult Clin
Psychol 1986;54:176–82.
[21] Leygraf N. Psychisch kranke Straftäter. [Mentally disordered offenders]. Berlin: Springer Verlag; 1988.
[22] Linehan MM. Trainingsmanual zur Dialektisch-Behavioralen Therapie der Borderline-Persönlichkeitsstörung. [Manual for dialecticbehavioral therapy of the borderline personality disorder]. München:
CIP-Medien; 1996.
[23] Margraf J, Schneider S, Soeder U, Neumer S, Becker ES. Mini-DIPS,
Diagnostisches Interview bei Psychischen Störungen (Forschungsversion), Interviewleitfaden, Version 1.1, 7/96. [Diagnostic interview
for psychological disorders, research version, user manual, version
1.1 7/96]. Dresden: Technische Universität; 1996.
[24] Margraf J, Schneider S, Ehlers A. DIPS. Diagnostisches Interview bei
Psychischen Störungen [DIPS. Diagnostic interview for psychological disorders]. Berlin: Springer; 1991.
[25] Margraf J. Diagnostisches Kurz-Interview bei psychischen
Störungen—Mini-DIPS, Handbuch und Interviewleitfaden. [MiniDIPS user manual]. Berlin: Springer; 1994.
[26] McElroy SL, Soutullo CA, Taylor P, et al. Psychiatric features of
36 men convicted of sexual offenses. J Clin Psychol 1999;60:414–20.
[27] Moran P. The epidemiology of antisocial personality disorder. Soc
Psychiatry Psychiatr Epidemiol 1999;34:231–42.
[28] Norman GR, Streiner DL. Biostatistics: the bare essentials. 2nd ed.
London: BC Becker; 2000.
[29] Nuhn-Naber C, Rehder U, Wischka B. Behandlung von Sexualstraftätern mit kognitiv-behavioralen Methoden: Möglichkeiten und
Grenzen. [Treatment of sexual offenders with cognitive-behavioral
methods]. Mschrkrim 2002;4:271–80.
[30] Overholser JC, Beck S. Multimethod assessment of rapists, child
molesters, and three control groups on behavioral and psychological
measures. J Consult Clin Psychol 1986;54:682–7.
[31] Pithers WD. Process evaluation of a group therapy component
designed to enhance sex offenders empathy for sexual abuse survivors. Behav Res Ther 1994;32:565–70.
[32] Rader CM. MMPI profile types of exposer, rapists, and assaulters in a
court services population. J Consult Clin Psychol 1977;45:61–9.
[33] Raymond NC, Coleman E, Ohlerking F, Christenson GA, Miner M.
Psychiatric comorbidity in pedophilic sexual offenders. Am J Psychiatry 1999;156:786–8.
[34] Saß H, Wittchen HU, Zaudig M. Diagnostisches und Statistisches
Manual Psychischer Störungen DSM IV. [Diagnostic and statistical
manual of psychiatric disorders DSM IV]. Göttingen: Hogrefe; 1996.
[35] Schneider S, Margraf J, Spörkel H, Franzen U. Therapiebezogene
Diagnostik: Reliabilität des Diagnostischen Interviews bei psychischen Störungen (DIPS) [Therapy-related diagnostics: reliability of the
Diagnostic Interview for Psychiatric Disorders (DIPS)]. Diagnostica
1992;38:209–27.
[36] Schüler-Springorum H, Berner W, Cirullies B, et al. Sexualstraftäter
im Maßregel-vollzug. [Sexual offenders in state forensic hospitals].
Mschr Krim Strafrechtsreform 1996;79:147–209.
[37] Segal ZV, Marshall WL. Heterosexual skills in a population of rapists
and child molesters. J Consult Clin Psychol 1985;53:55–63.
[38] SPSS, Inc. SPSS for Windows. Chicago: Author; 1999 Release 9.0.
[39] Stermac LE, Segal ZV, Gillis R. Social and cultural factors in sexual
assault. In: Marshall WL, Laws DR, Barbaree HE, editors. Handbook
of Sexual Assault. New York: Plenum Press; 1990. p. 143–59.
[40] Wittchen HU, Zaudig M, Fydrich T. Strukturiertes klinisches Interview für DSM IV Achse-I und Achse-II, Handanweisung. [Structured
Clinical Interview for DSM IV axis I and axis II, Manual]. Göttingen:
Hogrefe; 1997.
[41] Wulfert E, Greenway DE, Dougher MJ. A logical functional analysis
of reinforcement-based disorders: alcoholism and pedophilia. J Consult Clin Psychol 1996;64:1140–51.