European Psychiatry 28 (2013) 476–482

Available online at

www.sciencedirect.com

Original article

Childhood adversity in association with personality disorder dimensions:

New findings in an old debate
M.P. Hengartner *, V. Ajdacic-Gross, S. Rodgers, M. Müller, W. Rössler
Department of General and Social Psychiatry, Psychiatric University Hospital of Zurich, PO Box 1930, CH-8021 Zurich, Switzerland

A R T I C L E I N F O A B S T R A C T

Article history: Background: Various studies have reported a positive relationship between child maltreatment and
Received 23 October 2012 personality disorders (PDs). However, few studies included all DSM-IV PDs and even fewer adjusted for
Received in revised form 23 January 2013 other forms of childhood adversity, e.g. bullying or family problems.
Accepted 18 April 2013
Method: We analyzed questionnaires completed by 512 participants of the ZInEP epidemiology survey, a
Available online 5 July 2013
comprehensive psychiatric survey of the general population in Zurich, Switzerland. Associations
between childhood adversity and PDs were analyzed bivariately via simple regression analyses and
Keywords:
multivariately via multiple path analysis.
Child maltreatment
Abuse
Results: The bivariate analyses revealed that all PD dimensions were significantly related to various
Neglect forms of family and school problems as well as child abuse. In contrast, according to the multivariate
Bullying analysis only school problems and emotional abuse were associated with various PDs. Poverty was
Personality disorder uniquely associated with schizotypal PD, conflicts with parents with obsessive-compulsive PD, physical
Childhood adversity abuse with antisocial PD, and physical neglect with narcissistic PD. Sexual abuse was statistically
significantly associated with schizotypal and borderline PD, but corresponding effect sizes were small.
Conclusion: Childhood adversity has a serious impact on PDs. Bullying and violence in schools and
emotional abuse appear to be more salient markers of general personality pathology than other forms of
childhood adversity. Associations with sexual abuse were negligible when adjusted for other forms of
adversity.
ß 2013 Elsevier Masson SAS. All rights reserved.

1. Introduction tions such as parental divorce or poverty [15]. There is empirical

evidence that traumatic child abuse and (re-) victimization
Child maltreatment is a major public health and social-welfare particularly occur in specific populations from low social classes
problem. According to a comprehensive study in high-income with multiple individual, familial, and socioeconomic problems
countries about 4 to 16% of children are physically abused, [28,39,48]. Independently of child maltreatment the same
approximately 10% are neglected or emotionally abused, and 5 to burdened populations may already be at high-risk for PD
10% of girls and up to 5% of boys are sexually abused [22]. Adverse development because of factors such as poverty, low education
childhood experiences are supposed to be crucial with regard to level, parental separation, foster-care, or drug abuse. Furthermore,
the development and onset of personality disorders (PDs) childhood adversities are highly clustered and most individuals
[9,29,34]. report exposure to multiple events [32]. Thus, social conditions
In various longitudinal and cross-sectional community sam- may influence the probability of both child abuse and psycho-
ples, it has been found that child abuse or neglect was positively pathological symptoms [39].
related to PDs [1,20,21,26,29,36]. The influence of child abuse and Among other things it has been suggested that early separation
neglect on PD development has additionally been replicated in from the mother predicts elevations in PD symptoms or even may
several clinical samples [9,13,23,33,50]. cause PDs [3,7,17]. There is also evidence for an association
However, with regard to PD development it has been suggested between PD symptoms and adverse school climate [31]. Coid [15]
that the framework of adverse childhood experiences be expanded reported several exposures to childhood adversities like parental
beyond traumatic child abuse to include more long-term condi- loss, adoption, poverty, and foster-care to be predictors of different
PD diagnoses. He concluded that the social and individual
background may substantially predict PD development. However,
* Corresponding author. Tel.: +41 44 296 75 87; fax: +41 44 296 74 49. it is mostly unknown whether different PD dimensions are
E-mail address: [email protected] (M.P. Hengartner). associated with distinct or even specific forms of childhood

0924-9338/$ – see front matter ß 2013 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.eurpsy.2013.04.004
 M.P. Hengartner et al. / European Psychiatry 28 (2013) 476–482 477

adversity beyond child maltreatment. Furthermore, as Samuels First, a total of 9829 Swiss males and females aged 20 to 41
stated in a recent review [41], there is still a need for years at the onset of the survey and representative of the general
epidemiological studies of PD risk factors in the general popula- population of the canton of Zurich, Switzerland, were screened by
tion. computer-assisted telephone interview (CATI) using the Symptom
Thus, the major aim of the present study was to explore Checklist-27 (SCL-27) [27]. All participants were randomly
potential risk factors for different PD dimensions that have drawn selected through the residents’ registration offices of all munici-
little research attention so far. In addition to child maltreatment, palities of the canton of Zurich. Residents without Swiss
we deliberately assessed a broad range of non-traumatic forms of nationality were excluded from the study. The CATI was conducted
adverse childhood experiences by examining environmental risk by Growth for Knowledge (GfK), a major market and field research
factors such as poverty, family and school problems, which are institute, in accordance with instructions from the research team.
very common in Western societies, but often disregarded in PD The overall response rate was 53.6%. Reasons for non-response
research. were no response, only telephone responder, incorrect telephone
number, communication impossible, unavailability during the
2. Materials and methods study period, or refusal by a third person or the target person itself.
In cases where potential subjects were available, the response rate
2.1. Study design and sampling was 73.9%.
Second, 1500 subjects were randomly selected from the initial
The study was conducted within the scope of the epidemiology screening sample for subsequent face-to-face interviews (response
survey of the ‘‘Zurich Programme for the Sustainable Development rate: 65.2%). We applied a stratified sampling procedure including
of Mental Health Services’’ (ZInEP; in German: Zürcher Impul- 60% high-scorers (scoring above the 75th percentile of the global
sprogramm zur nachhaltigen Entwicklung der Psychiatrie), a severity index of the SCL-27) and 40% low-scorers (scoring below
research and health care programme involving several psychiatric the 75th percentile of the global severity index). The basic
research divisions and mental health services from the canton of sampling design was adapted from the longitudinal Zurich cohort-
Zurich, Switzerland. The epidemiology survey is one of the six study [4] and was chosen to enrich the sample with subjects at
ZInEP subprojects and consists of four components: high-risk of mental disorders. Such a two-phase procedure with
initial screening and subsequent interview with a stratified
 a short telephone screening; subsample is fairly common in epidemiological surveys [19].
 a comprehensive semi-structured face-to-face interview fol- Face-to-face interviews were conducted by experienced and
lowed by self-report questionnaires; extensively trained clinical psychologists. The interviews took
 tests in the sociophysiological laboratory; place either at the participants’ home or at the Psychiatric
 a longitudinal survey (Fig. 1). University Hospital in Zurich. All participants who completed
the semi-structured interview were required to complete addi-
Telephone screening and semi-structured interviews started in tional questionnaires. For this purpose, the sample was divided
August 2010, the tests at the sociophysiological laboratory in into subsamples focusing either on psychosis (n = 820) or on PDs
February 2011, and the longitudinal survey in April 2011. The (n = 680), respectively. Out of a total of 680 subjects in the PD
screening ended in May 2012 and all other components in subsample, 168 (24.7%) refused to return or to complete all
September 2012. questionnaires required for the present study, resulting in a
reduced final sample size of n = 512.
The ZInEP epidemiology survey was approved by the Zurich
State Ethical Committee (KEK) to fulfil all legal and data privacy
protection requirements and is in strict accordance with the
declaration of Helsinki of the World Medical Association. All
participants gave their written informed consent.

2.2. Instruments and measures

To assess dimensional PD scores we used the Assessment of

DSM-IV Personality Disorders Questionnaire (ADP-IV) [42]. The
ADP-IV design allows a dimensional trait-score and a categorical
PD diagnosis for each of the DSM-IV PDs. The ADP-IV is a paper-
pencil self-report instrument consisting of 94 items, which
represent the 80 criteria of the 10 DSM-IV PDs and the 14 research
criteria of the depressive and the passive-aggressive PD. Each trait-
question is rated on a 7-point Likert scale, ranging from ‘‘totally
disagree’’ to ‘‘totally agree’’. The dimensional score of a given PD is
computed by adding all scores of its respective items and by
dividing this value by the number of items. For the present study
we used a German translation by Doering et al. [18]. Internal
consistency of the ADP-IV dimensional PD scales is good for the
original Dutch version [44] and for the German adaptation [18]
(median Cronbach’s a = 0.77 and 0.76, respectively). Test-retest
reliability and concurrent validity of the dimensional ADP-IV trait-
scores is also satisfactory [18,44]. Most importantly, the ADP-IV
showed good concordance with the SCID-II interview [43] and may
Fig. 1. The sampling procedure of the epidemiology survey of Zürcher be considered as an economic and valid alternative to semi-
Impulsprogramm zur nachhaltigen Entwicklung der Psychiatrie (ZInEP). structured interviews.
478 M.P. Hengartner et al. / European Psychiatry 28 (2013) 476–482

Child maltreatment was assessed with the Childhood Trauma values of AFI for a good model fit are an CFI > 0.95, an
Questionnaire (CTQ) [11]. The CTQ is a popular retrospective RMSEA < 0.06, and an SRMR < 0.08 [45]. Estimation of effect sizes
measure of child abuse and neglect. The short-form [12] consists of relied on Cohen’s f2, a formula for the interpretation of variance
28 items divided into a control-scale named denial and the four estimates [14]. The path analysis was conducted with Mplus
domains emotional abuse, emotional neglect, sexual abuse, Version 7 for Macintosh [37].
physical abuse and physical neglect. The items are rated on a 5-
point Likert scale ranging from ‘‘never true’’ to ‘‘very often true’’. 3. Results
For the present study we applied a German adaptation [49].
Internal consistency of the different domains is high and test-retest Descriptive statistics are reported in Table 1. The bivariate
reliability is also good. The CTQ showed good convergent and associations between childhood adversity and PD dimensions are
discriminant validity and can be considered to be a sensitive and shown in Table 2. Every PD dimension was associated with various
valid screening questionnaire for child maltreatment [10,12,49]. childhood adversity variables. Poverty, conflicts with parents,
Beside child maltreatment, childhood adversity was assessed bullying victimization, emotional abuse and neglect as well as
with another self-report questionnaire that was designed by the physical abuse and neglect were significantly related to all 10 PDs.
authors following the schemes of the Zurich Study [4]. The Parents separated or divorced was associated exclusively with
questionnaire was divided into two sections, the first labelled schizotypal PD, whereas parental substance abuse was uniquely
school problems and the second familial problems. Every question related to borderline and histrionic PD. Sexual abuse was
referred to childhood and adolescence and was answered by the significantly related to all PD dimensions except for schizoid,
standard response options ‘‘yes’’ or ‘‘no’’. Based on these data, we histrionic, and narcissistic PD. A continuous total adversity score
defined six dichotomous outcome variables (absent vs. present) was computed by adding the values of all childhood adversity
according to the following algorithms: variables. This total score was significantly associated with all PDs,
which indicates a dose-response relationship. Additional analyses
 bullying victimisation in school: if the respondent endorsed revealed that all significant associations corresponded to at least
either ‘‘were you often physically assaulted in school?’’, ‘‘were small effect sizes (all Pearson’s r > 0.090). Those associations were
you frequently insulted in school?’’, or ‘‘were you regularly thus all substantial and not merely a statistical artefact based on
excluded and ignored in school?’’; the large sample size.
 conduct problems in school: if either ‘‘did you fight frequently in The model fit of the path analysis was excellent. The x2 was
school?’’, ‘‘did your teacher frequently punish you?’’ or ‘‘did you 22.662 (df = 21); P = 0.362 (cut-off: P > 0.05), the CFI was 0.999
deliberately miss or skip school class a lot?’’; (cut-off: > 0.95), the RMSEA was 0.013 (cut-off: < 0.06), and the
 parents’ separation or divorce: if the item ‘‘my parents separated SRMR was 0.016 (cut-off: < 0.08). The path diagram is shown on
or divorced.’’ was endorsed; Fig. 2 and the results are reported in Table 3. There were three
 parents’ poverty: if the question ‘‘sometimes we had hardly outstanding multivariate predictors of PD trait-scores, namely,
enough money to live’’ was endorsed; bullying victimization and conduct problems in school, and
 conflicts with parents: if the respondent endorsed either ‘‘I was emotional abuse. The latter was particularly strongly associated
frequently punished by my parents.’’ or ‘‘my relationship with with schizotypal, borderline, histrionic, and avoidant PD (all
my parents was a constant up and down’’; b > 0.200). Parents separated/divorced and parental substance
 parent’s substance abuse: if the question ‘‘did your parents drink abuse were not related to any PD dimension in the path analysis.
far too much alcohol?’’ or ‘‘did your parents use illicit drugs?’’ Poverty was uniquely related to schizotypal PD, conflicts with
was endorsed. parents to obsessive-compulsive PD, physical abuse to antisocial

2.3. Statistical analysis Table 1

Descriptive statistics.

First, we analyzed the bivariate associations between every Categorical measures n %

predictor variable and each dimensional PD trait-score by applying Female sex 284 55.5
a series of generalized linear regression models (GLM). All Parents separated/divorced 100 19.5
dependent variables (i.e. PD dimensions) were right skewed; Poverty 54 10.5
therefore we fitted models with gamma distribution and log-link Conflicts with parents 198 38.7
Parental substance abuse 79 15.4
function. A robust estimator was used to reduce the effects of
Victim of bullying 276 53.9
outliers and influential observations. Results were reported with Conduct problems 156 30.5
unstandardized regression coefficients (b) and standard errors
(SE). The GLM were performed with SPSS version 20 for Macintosh. Continuous measures Mean (SD) Range
Second, we examined multivariate associations by fitting a
Age 29.61 (6.74) 20–41
multiple path analysis, where predictors were adjusted for each Emotional abuse 8.11 (3.86) 5–25
other and covariance between dependent variables was accounted Emotional neglect 9.58 (4.53) 5–25
for. Path analysis is a helpful tool used to minimize the unwanted Physical abuse 5.79 (2.06) 5–24
Physical neglect 6.40 (2.11) 5–17
effects of several interrelated predictor variables (multicollinear-
Sexual abuse 5.80 (2.60) 5–25
ity) and correlated dependent variables (endogeneity). Each PD Paranoid PD 2.27 (0.94) 1–7
dimension was regressed on its bivariately significant predictors. Schizoid PD 2.17 (0.85) 1–7
Path coefficients were indicated with standardized regression Schizotypal PD 2.12 (0.90) 1–7
coefficients (b) and standard errors (SE). To evaluate the goodness Antisocial PD 1.65 (0.70) 1–7
Borderline PD 2.30 (1.05) 1–7
of model fit we considered the x2-test of model fit and the
Histrionic PD 2.26 (0.89) 1–7
following approximate fit indices (AFI): the comparative fit index Narcissistic PD 2.24 (0.80) 1–7
(CFI), the root mean square error of approximation (RMSEA), and Avoidant PD 2.44 (1.15) 1–7
the standardized root mean square residual (SRMR). According to Dependent PD 2.19 (0.89) 1–7
the x2-test a good fitting model should provide an insignificant Obsessive-compulsive PD 2.77 (0.92) 1–7

result (i.e. above the 0.05 threshold) [8]. Recommended cut-off PD: personality disorder.
Table 2
Results of a series of generalized linear models: bivariate predictors of personality disorder dimensions.

Personality disorder dimensions

PAR SZ ST AS BDL HIS NAR AV DEP OC

Parents separated/divorced 0.040 (0.05) 0.029 (0.04) 0.102 (0.05) 0.040 (0.05) 0.077 (0.05) 0.033 (0.04) 0.028 (0.04) –0.005 (0.05) –0.035 (0.05) 0.021 (0.04)
Poverty 0.235 (0.06) 0.153 (0.05) 0.327 (0.05) 0.142 (0.06) 0.272 (0.06) 0.163 (0.06) 0.142 (0.05) 0.225 (0.06) 0.152 (0.06) 0.120 (0.04)
Conflicts with parents 0.274 (0.04) 0.147 (0.03) 0.271 (0.04) 0.209 (0.04) 0.287 (0.04) 0.184 (0.03) 0.198 (0.03) 0.198 (0.04) 0.145 (0.04) 0.158 (0.03)
Parental substance abuse 0.090 (0.05) –0.011 (0.05) 0.060 (0.05) 0.091 (0.05) 0.130 (0.05) 0.094 (0.05) 0.069 (0.05) 0.029 (0.06) 0.035 (0.05) 0.023 (0.04)
Victim of bullying 0.180 (0.04) 0.183 (0.03) 0.272 (0.04) 0.189 (0.04) 0.237 (0.04) 0.117 (0.03) 0.140 (0.03) 0.236 (0.04) 0.179 (0.04) 0.130 (0.03)
Conduct problems 0.216 (0.04) 0.171 (0.04) 0.235 (0.04) 0.361 (0.04) 0.221 (0.04) 0.158 (0.04) 0.181 (0.03) 0.130 (0.04) 0.067 (0.04) 0.075 (0.03)
Emotional abuse 0.156 (0.02) 0.099 (0.02) 0.182 (0.02) 0.131 (0.02) 0.200 (0.02) 0.118 (0.02) 0.098 (0.02) 0.159 (0.02) 0.121 (0.01) 0.081 (0.01)

M.P. Hengartner et al. / European Psychiatry 28 (2013) 476–482

Emotional neglect 0.158 (0.02) 0.101 (0.02) 0.173 (0.02) 0.126 (0.02) 0.187 (0.02) 0.105 (0.02) 0.101 (0.02) 0.161 (0.02) 0.113 (0.02) 0.074 (0.01)
Physical abuse 0.102 (0.01) 0.055 (0.01) 0.116 (0.01) 0.114 (0.02) 0.127 (0.02) 0.079 (0.01) 0.074 (0.01) 0.088 (0.02) 0.083 (0.01) 0.051 (0.01)
Physical neglect 0.113 (0.02) 0.066 (0.02) 0.129 (0.02) 0.102 (0.02) 0.141 (0.02) 0.085 (0.02) 0.095 (0.02) 0.102 (0.02) 0.077 (0.02) 0.054 (0.01)
Sexual abuse 0.048 (0.02) 0.026 (0.02) 0.081 (0.02) 0.058 (0.02) 0.087 (0.02) 0.031 (0.02) 0.012 (0.01) 0.051 (0.02) 0.051 (0.02) 0.031 (0.01)

Total adversity score 0.169 (0.02) 0.104 (0.02) 0.198 (0.02) 0.156 (0.02) 0.211 (0.02) 0.120 (0.02) 0.116 (0.02) 0.155 (0.02) 0.118 (0.02) 0.085 (0.01)

Statistically significant associations are indicated in bold (P < 0.05). Continuous predictors were standardized using z-transformation. PAR: paranoid; SZ: schizoid; ST: schizotypal; AS: antisocial; BDL: borderline; HIS: histrionic;
NAR: narcissistic; AV: avoidant; DEP: dependent; OC: obsessive-compulsive.

Table 3
Results of a path analysis: multivariate predictors of personality disorder dimensions.

Personality disorder dimensions

PAR SZ ST AS BDL HIS NAR AV DEP OC

Parents separated/divorced – – 0.042 (0.03) – – – – – – –

Poverty 0.043 (0.04) 0.043 (0.05) 0.118 (0.04) 0.028 (0.04) 0.075 (0.04) 0.077 (0.05) 0.013 (0.05) 0.042 (0.05) 0.038 (0.05) 0.033 (0.05)
Conflicts with parents 0.085 (0.05) 0.046 (0.05) 0.044 (0.05) 0.037 (0.05) 0.031 (0.05) 0.059 (0.05) 0.098 (0.05) –0.026 (0.05) –0.014 (0.05) 0.104 (0.05)
Parental substance abuse – – – – 0.022 (0.03) 0.030 (0.03) – – – –
Victim of bullying 0.088 (0.04) 0.164 (0.05) 0.171 (0.04) 0.081 (0.04) 0.105 (0.04) 0.019 (0.05) 0.090 (0.05) 0.159 (0.04) 0.142 (0.05) 0.114 (0.05)
Conduct problems 0.098 (0.04) 0.106 (0.04) 0.101 (0.04) 0.335 (0.04) 0.071 (0.03) 0.125 (0.04) 0.143 (0.04) 0.011 (0.03) – 0.028 (0.04)
Emotional abuse 0.167 (0.07) 0.056 (0.08) 0.238 (0.07) 0.054 (0.07) 0.285 (0.07) 0.233 (0.07) 0.062 (0.07) 0.201 (0.07) 0.193 (0.07) 0.119 (0.08)
Emotional neglect 0.149 (0.07) 0.135 (0.08) 0.060 (0.07) 0.053 (0.07) 0.064 (0.07) –0.039 (0.08) 0.033 (0.08) 0.164 (0.07) 0.060 (0.08) 0.029 (0.08)
Physical abuse 0.034 (0.05) –0.061 (0.05) –0.017 (0.05) 0.110 (0.05) 0.001 (0.05) 0.005 (0.05) 0.002 (0.05) –0.033 (0.05) 0.011 (0.06) –0.066 (0.06)
Physical neglect 0.054 (0.05) –0.005 (0.06) 0.047 (0.05) -0.001 (0.05) 0.060 (0.05) 0.039 (0.06) 0.128 (0.06) 0.005 (0.05) 0.003 (0.06) 0.014 (0.06)
Sexual abuse –0.037 (0.04) – 0.069 (0.03) 0.017 (0.04) 0.076 (0.03) – – –0.001 (0.04) 0.046 (0.04) 0.068 (0.04)

Statistically significant associations are indicated in bold (P < 0.05). PAR: paranoid; SZ: schizoid; ST: schizotypal; AS: antisocial; BDL: borderline; HIS: histrionic; NAR: narcissistic; AV: avoidant; DEP: dependent; OC: obsessive-
compulsive.

479
480 M.P. Hengartner et al. / European Psychiatry 28 (2013) 476–482

Fig. 2. Results of a path analysis: multivariate predictors of personality disorder dimensions. Only significant path coefficients are depicted (P < 0.05).

PD, and physical neglect to narcissistic PD. Sexual abuse was 4. Discussion
statistically significantly associated with schizotypal and border-
line PD, but the corresponding effects sizes were small (both To the best of our knowledge, this is the first study to examine
b < 0.080). Considerably strong associations were found between all 10 DSM-IV PDs in association with various forms of childhood
antisocial PD and conduct problems (b = 0.335) as well as between adversity, including familial problems, school problems, and child
borderline PD and emotional abuse (b = 0.285). maltreatment in a general population-based community sample.
Childhood adversity accounted overall for a substantial Overall, the findings show that childhood adversity is highly
proportion of variance explained in all 10 PD dimensions (Table associated with PD symptomatology, explaining a large propor-
4). Estimates of variance explained ranged from 9.0% for obsessive- tion of total variance. Furthermore, adding the scores of the
compulsive PD (corresponding to f2 = 0.10) to 27.8% for borderline various childhood adversity variables is associated with higher
PD (f2 = 0.39) and 29.3% for schizotypal PD (f2 = 0.41). According to scores on every PD dimension, suggesting that childhood
Cohen’s f2 those variance estimates represent medium to large adversity impacts PDs in a dose-response relationship. In
effect sizes (with f2 = 0.02 denoting a small effect, f2 = 0.15 a bivariate analyses poverty, conflicts with parents, being a victim
medium effect and f2 = 0.35 a large effect). of bullying in school, emotional abuse and neglect as well as
physical abuse and neglect were significantly related to all 10 PD
dimensions. In contrast to the bivariate analyses, the multivariate
Table 4 associations via path analysis indicated that only specific forms of
Proportion of variance explained (R2) in personality disorder dimensions.
childhood adversity were associated with various PD dimensions.
PD dimension Estimate (SE) Significance Multivariate analyses allow more detailed conclusions and enable
Paranoid 0.230 (0.034) P < 0.001 examination of the associations of a given form of childhood
Schizoid 0.113 (0.027) P < 0.001 adversity when adjusted for others, which has rarely been done to
Schizotypal 0.293 (0.035) P < 0.001 date. Bullying victimization and conduct problems in school as
Antisocial 0.232 (0.033) P < 0.001
well as emotional abuse were by far the strongest multivariate
Borderline 0.278 (0.035) P < 0.001
Histrionic 0.132 (0.028) P < 0.001
predictors, yielding significant associations with most PD
Narcissistic 0.135 (0.029) P < 0.001 dimensions. Relatively strong associations were found between
Avoidant 0.166 (0.031) P < 0.001 antisocial PD and conduct problems as well as between borderline
Dependent 0.121 (0.028) P < 0.001 PD and emotional abuse. Finally, sexual abuse showed no practical
Obsessive-compulsive 0.090 (0.025) P < 0.001
significance.
 M.P. Hengartner et al. / European Psychiatry 28 (2013) 476–482 481

Our bivariate associations between child maltreatment and PD climate caused increases in PD symptoms. Early detection and
dimensions are in line with the literature [1,9,21,26,29], indicating prevention of school problems are therefore extremely important,
that emotional and physical abuse and neglect are highly related to since schools are key social contexts for shaping development and
PDs. The examination of sexual abuse in persons with antisocial or behaviours [31]. Furthermore, PDs may originate very early, that is,
borderline PD has a long tradition in PD research [51]. Many authors during late childhood or early adolescence [30], an age when most
argue that sexual abuse is closely and/or specifically related to one of persons mainly spend their time at school. In a recent review
these PDs [9,13,29,38]. In our data sexual abuse was associated with article on bullying victimization and mental health problems, the
all PD dimensions except for schizoid, histrionic, and narcissistic PD, authors concluded that bullying contributes – independently of
although this finding applied only to the bivariate analyses. When other factors – to the development and persistence of long-lasting
controlled for other forms of childhood adversity, sexual abuse was severe mental health problems [5]. Nevertheless, further research,
uniquely associated with schizotypal and borderline PD. However, it especially of a longitudinal nature, is needed to examine the
is important to note that although these associations were interplay between school problems and early personality dysfunc-
statistically significant, they proved to be relatively weak. The tion in order to disentangle antecedents and consequences of PDs
magnitude of the corresponding effect size was not of practical and to understand the aetiopathological process underlying this
significance. As emphasized by other authors [39], adjustment for relationship.
the social context in which abuse occurs may drastically change the The results of this study need to be interpreted in the context of
associations between child maltreatment and mental health out- the following limitations: First, because of the cross-sectional
comes. Rind and Tromovitch [39] state in their meta-analytic review design the assessment of childhood adversity was necessarily
that associations with psychological maladjustment disappear or retrospective. Thus we were not able to determine a clear temporal
become negligible in most community studies that adjusted sexual order and accordingly cannot draw causal conclusions from our
abuse for confounders such as emotional abuse or familial data. Second, a serious limitation of self-report instruments may be
environment. Such findings illustrate the need for multivariate a recall bias, suggesting that subjects tend to under-report or deny
analyses of population-based samples that account for various forms traumatic events. Nevertheless, the reliability of self-reports of
of childhood adversity, including familial background and the social victimization is satisfactory in both subjects with mental disorders
environment. [25] and subjects without mental disorders [24]. It has also been
Our path analysis revealed that especially bullying victimization, found that official state registry records and self-reports of child
conduct problems and emotional abuse were related to various PD abuse show good correspondence [29]. Furthermore, only a few
dimensions. It thus appears that those forms of childhood adversity studies [29,34] have used official records from a public authority
represent covariates of general personality dysfunction rather than registry to analyse associations between child abuse and PDs; most
of specific PD dimensions. Compared to emotional abuse, physical have used retrospective self-report questionnaires. Thus this is a
abuse and sexual abuse were much less related to PDs. This finding is general limitation of research on childhood adversity. Third, our
consistent with previous studies, which found that in persons with measures of familial environment relied on simple dichotomous
PDs emotional abuse is much more prevalent than physical or sexual items (i.e. present vs. absent) without consideration of severity and
abuse [9,13,40]. The estimates of variance explained were out- frequency and PD measures were assessed by self-report ques-
standingly high in borderline and schizotypal PD; thus it appears tionnaire. Validity and reliability of those measures may thus be
that childhood adversity plays a crucial role in those disorders. restricted.
Conduct problems in school were considerably strongly related to
antisocial PD, which is obvious given that conduct problems in
5. Conclusion
childhood are a prerequisite criterion of antisocial PD [2]. Unlike
other studies [15,32] poverty was uniquely related to schizotypal PD
In this study we outline the importance of child maltreatment
when adjusted for other childhood adversity variables. However, it
and certain forms of non-traumatic childhood adversity in
has previously been shown that if parental care factors are
association with PDs that have been disregarded for a long time.
statistically controlled for, the effect of poverty on child and
In this respect we suggest that schools in particular may be a
adolescent developmental problems is considerably reduced
crucial environment in which PD symptoms become apparent or
[16,35]. Thus it appears that inclusion of various forms of childhood
accentuated and which may also contribute to PD development.
adversity as covariates counterbalanced the effects of poverty in the
Bullying and violence in school as well as emotional abuse may
present study.
play a major role in the aetiopathology of PDs. By contrast, sexual
In particular we want to emphasize the considerable impact of
abuse appears to be of negligible practical significance when
school problems on PD symptomatology. All 10 PD dimensions in
adjusted for other forms of adversity. We suggest that prevention
the path analysis were significantly predicted by childhood
and intervention programmes that are nested within the context of
adversity in the context of school (i.e. bullying victimization
family and school might be further explored for their appropriate-
and/or conduct problems). Negative experiences in the school
ness for PD symptomatology. Furthermore, longitudinal studies
environment thus appear to be universal markers of general
that focus on the aetiopathology of PDs are needed.
personality dysfunction, since they are associated with all PD
dimensions. Because our cross-sectional design does not allow for
causal conclusions we deliberately use the term marker, but recent Disclosure of interest
prospective studies suggest that being a victim of constant bullying
may constitute a causal risk factor for severe subsequent mental The authors declare that they have no conflicts of interest
health problems such as antisocial PD and anxiety disorders [47], concerning this article.
general internalizing problems [6] or psychotic symptoms [46]. To
our knowledge there is only one prospective study that has focused Acknowledgement
on the relationship between school environment and PDs [31].
Adjusting for several covariates such as child maltreatment, The ZInEP project was supported by a private donation. The
socioeconomic status and Axis I disorders, the authors found that donor had no further role in the experimental design, the
a positive and supportive school climate resulted in subsequent PD collection, analysis, and interpretation of data, the writing of this
symptom decreases, whereas a negative and conflictual school report, or the decision to submit this paper for publication.
482 M.P. Hengartner et al. / European Psychiatry 28 (2013) 476–482

References [26] Grover KE, Carpenter LL, Price LH, Gagne GG, Mello AF, Mello MF, et al. The
relationship between childhood abuse and adult personality disorder symp-
[1] Afifi TO, Mather A, Boman J, Fleisher W, Enns MW, Macmillan H, et al. toms. J Pers Disord 2007;21(4):442–7.
Childhood adversity and personality disorders: results from a nationally [27] Hardt J, Egle UT, Kappis B, Hessel A, Brahler E. Symptom Checklist SCL-27:
representative population-based study. J Psychiatr Res 2011;45(6):814–22. results of a representative German survey. Psychother Psychosom Med Psy-
[2] American Psychiatric Association. Diagnostic and Statistical Manual of Mental chol 2004;54(5):214–23.
Disorders DSM-IV-TR, 4th text revision ed., Washington, DC: American [28] Holzman CG. Counseling adult women rape survivors: Issues of race, ethnicity,
Psychiatric Association;; 2000. and class. Women & Therapy 1996;19(2):47–62.
[3] Anglin DM, Cohen PR, Chen H. Duration of early maternal separation and [29] Johnson JG, Cohen P, Brown J, Smailes EM, Bernstein DP. Childhood maltreat-
prediction of schizotypal symptoms from early adolescence to midlife. Schi- ment increases risk for personality disorders during early adulthood. Arch Gen
zophr Res 2008;103(1–3):143–50. Psychiatry 1999;56(7):600–6.
[4] Angst J, Gamma A, Neuenschwander M, Ajdacic-Gross V, Eich D, Rössler W, [30] Johnson JG, Cohen P, Kasen S, Skodol AE, Brook J. Age-related change in
et al. Prevalence of mental disorders in the Zurich cohort study: a twenty-year personality disorder symptom levels between early adolescence and adult-
prospective study. Epidemiol Psichiatr Soc 2005;14(2):68–76. hood: a community-based longitudinal investigation. Acta Psychiatr Scand
[5] Arseneault L, Bowes L, Shakoor S. Bullying victimization in youths and mental 2000;102:265–75.
health problems: ‘much ado about nothing’? Psychol Med 2010;40(5):717–29. [31] Kasen S, Cohen P, Chen H, Johnson JG, Crawford TN. School climate and
[6] Arseneault L, Milne BJ, Taylor A, Adams F, Delgado K, Caspi A, et al. Being continuity of adolescent personality disorder symptoms. J Child Psychol
bullied as an environmentally mediated contributing factor to children’s Psychiatry 2009;50(12):1504–12.
internalizing problems: a study of twins discordant for victimization. Arch [32] Kessler RC, Davis CG, Kendler KS. Childhood adversity and adult psychiatric
Pediatr Adolesc Med 2008;162(2):145–50. disorder in the US National Comorbidity Survey. Psychol Med 1997;27(5):
[7] Bandelow B, Krause J, Wedekind D, Broocks A, Hajak G, Ruther E. Early 1101–19.
traumatic life events, parental attitudes, family history, and birth risk factors [33] Laporte L, Guttman H. Traumatic childhood experiences as risk factors for
in patients with borderline personality disorder and healthy controls. Psychi- borderline and other personality disorders. J Pers Disord 1996;10(3):247–59.
atry Res 2005;134(2):169–79. [34] Luntz BK, Widom CS. Antisocial personality disorder in abused and neglected
[8] Barrett P. Structural equation modelling: adjudging model fit. Pers Individ Dif children grown up. Am J Psychiatry 1994;151(5):670–4.
2007;42:815–24. [35] McLoyd VC. Socioeconomic disadvantage and child development. Am Psychol
[9] Battle CL, Shea MT, Johnson DM, Yen S, Zlotnick C, Zanarini MC, et al. Childhood 1998;53(2):185–204.
maltreatment associated with adult personality disorders: findings from the [36] Miller PM, Lisak D. Associations between childhood abuse and personality
Collaborative Longitudinal Personality Disorders Study. J Pers Disord disorder symptoms in college males. J Interpers Violence 1999;14(6):642–56.
2004;18(2):193–211. [37] Muthén LK, Muthén BO. Mplus User’s Guide, Seventh Edition, Los Angeles, CA:
[10] Bernstein DP, Ahluvalia T, Pogge D, Handelsman L. Validity of the Childhood Muthén & Muthén; 1998-2012.
Trauma Questionnaire in an adolescent psychiatric population. J Am Acad [38] Paris J, Zweig-Frank H, Guzder J. Psychological risk factors for borderline
Child Adolesc Psychiatry 1997;36(3):340–8. personality disorder in female patients. Compr Psychiatry 1994;35(4):301–5.
[11] Bernstein DP, Fink L. Childhood Trauma Questionnaire: a retrospective self- [39] Rind B, Tromovitch P. A meta-analytic review of findings from national
report. San Antonio: Psychological Corperation;; 1998. samples on psychological correlates of child sexual abuse. J Sex Res
[12] Bernstein DP, Stein JA, Newcomb MD, Walker E, Pogge D, Ahluvalia T, et al. 1997;34(3):237–55.
Development and validation of a brief screening version of the Childhood [40] Saleptsi E, Bichescu D, Rockstroh B, Neuner F, Schauer M, Studer K, et al.
Trauma Questionnaire. Child Abuse Negl 2003;27(2):169–90. Negative and positive childhood experiences across developmental periods in
[13] Bierer LM, Yehuda R, Schmeidler J, Mitropoulou V, New AS, Silverman JM, et al. psychiatric patients with different diagnoses - an explorative study. BMC
Abuse and neglect in childhood: relationship to personality disorder diagno- Psychiatry 2004;4:40.
ses. CNS Spectr 2003;8(10):737–54. [41] Samuels J. Personality disorders: epidemiology and public health issues. Int
[14] Cohen J. Power analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum; Rev Psychiatry 2011;23(3):223–33.
1988. [42] Schotte CKW, de Doncker D. ADP-IV Questionnaire. Antwerp, Belgium: Uni-
[15] Coid J. Aetiological risk factors for personality disorders. Br J Psychiatry versity Hospital Antwerp; 1994.
1999;174:530–8. [43] Schotte CKW, de Doncker D, Dmitruk D, van Mulders I, D’Haenen H, Cosyns P.
[16] Conger RD, Ge X, Elder Jr GH, Lorenz FO, Simons RL. Economic stress, coercive The ADP-IV questionnaire: differential validity and concordance with the
family process, and developmental problems of adolescents. Child Dev semi-structured interview. J Pers Disord 2004;18(4):405–19.
1994;65(2):541–61. [44] Schotte CKW, de Doncker D, Vankerckhoven C, Vertommen H, Cosyns P. Self-
[17] Crawford TN, Cohen PR, Chen H, Anglin DM, Ehrensaft M. Early maternal report assessment of the DSM-IV personality disorders. Measurement of trait
separation and the trajectory of borderline personality disorder symptoms. and distress characteristics: the ADP-IV. Psychol Med 1998;28(5):1179–88.
Dev Psychopathol 2009;21(3):1013–30. [45] Schreiber JB, Stage FK, King J, Nora A, Barlow EA. Reporting structural equation
[18] Doering S, Renn D, Hofer S, Rumpold G, Smrekar U, Janecke N, et al. Validation modeling and confirmatory factor analysis results: a review. J Educ Res
of the German version of the Assessment of DSM-IV Personality Disorders 2006;99(6).
(ADP-IV) Questionnaire. Z Psychosom Med Psychother 2007;53(2):111–28. [46] Schreier A, Wolke D, Thomas K, Horwood J, Hollis C, Gunnell D, et al. Prospec-
[19] Dunn G, Pickles A, Tansella M, Vazquez-Barquero JL. Two-phase epidemiologi- tive study of peer victimization in childhood and psychotic symptoms in a
cal surveys in psychiatric research. Br J Psychiatry 1999;174:95–100. nonclinical population at age 12 years. Arch Gen Psychiatry 2009;66(5):527–
[20] Fergusson DM, Boden JM, Horwood LJ. Exposure to childhood sexual and 36.
physical abuse and adjustment in early adulthood. Child Abuse Negl [47] Sourander A, Jensen P, Ronning JA, Niemela S, Helenius H, Sillanmaki L, et al.
2008;32(6):607–19. What is the early adulthood outcome of boys who bully or are bullied in
[21] Gibb BE, Wheeler R, Alloy LB, Abramson LY. Emotional, physical, and sexual childhood? The Finnish ‘‘From a Boy to a Man’’ study. Pediatrics
maltreatment in childhood versus adolescence and personality dysfunction in 2007;120(2):397–404.
young adulthood. J Pers Disord 2001;15(6):505–11. [48] Widom CS, Czaja SJ, Dutton MA. Childhood victimization and lifetime revicti-
[22] Gilbert R, Widom CS, Browne K, Fergusson D, Webb E, Janson S. Burden and mization. Child Abuse Negl 2008;32(8):785–96.
consequences of child maltreatment in high-income countries. Lancet [49] Wulff H. Childhood Trauma Questionnaire. Entwicklung einer deutschspra-
2009;373(9657):68–81. chigen Version und Überprüfung bei psychiatrisch-psychotherapeutisch
[23] Golier JA, Yehuda R, Bierer LM, Mitropoulou V, New AS, Schmeidler J, et al. The behandelten Patienten [Inauguraldissertation]. Lübeck: Universität Lübeck;
relationship of borderline personality disorder to posttraumatic stress disor- 2006.
der and traumatic events. Am J Psychiatry 2003;160(11):2018–24. [50] Yen S, Shea MT, Battle CL, Johnson DM, Zlotnick C, Dolan-Sewell R, et al.
[24] Goodman LA, Corcoran C, Turner K, Yuan N, Green BL. Assessing traumatic Traumatic exposure and posttraumatic stress disorder in borderline, schizo-
event exposure: general issues and preliminary findings for the Stressful Life typal, avoidant, and obsessive-compulsive personality disorders: findings
Events Screening Questionnaire. J Trauma Stress 1998;11(3):521–42. from the collaborative longitudinal personality disorders study. J Nerv Ment
[25] Goodman LA, Thompson KM, Weinfurt K, Corl S, Acker P, Mueser KT, et al. Dis 2002;190(8):510–8.
Reliability of reports of violent victimization and posttraumatic stress disorder [51] Zanarini MC. Childhood experiences associated with the development of
among men and women with serious mental illness. J Trauma Stress borderline personality disorder. Psychiatr Clin North Am 2000;23(1):
1999;12(4):587–99. 89–101.