Wesleyan University
From the SelectedWorks of Charles A. Sanislow, Ph.D.
$%
##! '%'!"'&&#'
*'( '$%&#" '+&#%%&
,""&%#!'# #%')
#"'(" %&#" '+&#%%&
'(+
+"'''
%
*"#"&#"
%#" #'"
%+"%"'
) ' ''$&*#%&$%&&#!% &&"& #*
BATTLE ET AL.CHILDHOOD MALTREATMENT
CHILDHOOD MALTREATMENT ASSOCIATED WITH
ADULT PERSONALITY DISORDERS: FINDINGS
FROM THE COLLABORATIVE LONGITUDINAL
PERSONALITY DISORDERS STUDY
Cynthia L. Battle, PhD, M. Tracie Shea, PhD,
Dawn M. Johnson, PhD, Shirley Yen, PhD, Caron Zlotnick, PhD,
Mary C. Zanarini, EdD, Charles A. Sanislow, PhD,
Andrew E. Skodol, MD, John G. Gunderson, MD,
Carlos M. Grilo, PhD, Thomas H. McGlashan, MD,
and Leslie C. Morey, PhD
Adverse childhood experiences such as abuse and neglect are frequently
implicated in the development of personality disorders (PDs); however,
research on the childhood histories of most PD groups remains limited. In
this multisite investigation, we assessed self-reported history of abuse
and neglect experiences among 600 patients diagnosed with either a PD
(borderline, schizotypal, avoidant, or obsessive-compulsive) or major de-
pressive disorder without PD. Results indicate that rates of childhood
maltreatment among individuals with PDs are generally high (73% re-
porting abuse; 82% reporting neglect). As expected, borderline PD was
more consistently associated with childhood abuse and neglect than
other PD diagnoses. However, even when controlling for the effect of bor-
derline PD, other PD diagnoses were associated with specific types of
maltreatment.
Personality disorders (PDs) are conceptualized as enduring, charac
-
ter-based patterns of pathology that first appear during adolescence or early
adulthood (American Psychiatric Association, 1994). Theorists from diverse
camps (Millon & Davis, 1996; Pretzer & Beck, 1996) frequently point to
childhood experiences as being central to the development of PDs. In partic
-
ular, adverse events such as abuse, neglect, prolonged separations, and
early losses have been identified as risk factors with etiological significance
(Johnson et al., 1999; Laporte & Guttman, 1996). In spite of the putative sa
-
193
Journal of Personality Disorders, 18(2), 193-211, 2004
© 2004 The Guilford Press
From the Collaborative Longitudinal PDs Study (CLPS): Brown University Medical School
(C.L.B., M.T.S., D.M.J., S.Y., C.Z.); Harvard University Medical school (M.C.Z., J.G.G.); Yale
University School of Medicine (C.A.S., C.M.G., T.H.M.); Columbia University College of Physi
-
cians and Surgeons (A.E.S.); and Texas A&M University (L.C.M.).
The CLPS is an ongoing, longitudinal, multisite study of PDs supported by National Institutes of
Mental Health grants (R10 MH 50837, 50838, 50839, 50840, and 50850).
Address correspondence to Cynthia L. Battle, Butler Hospital Psychosocial Research Program,
345 Blackstone Blvd., Providence, RI 02906; E-mail: Cynthia_Battle@brown.edu.
lience of adverse childhood experiences in the development of PDs, relatively
little empirical research has investigated the role of childhood maltreatment
in the majority of PDs (Paris, 2000; Torgersen & Alnaes, 1992). A better un
-
derstanding of the childhood events linked with specific forms of personality
pathology may contribute to more effective treatments or early interventions
to help prevent these disorders.
CHILDHOOD MALTREATMENT AND BORDERLINE PD
The diagnostic group that has received the majority of research attention
with regard to adverse childhood experiences is borderline PD (BPD)
(Gunderson & Sabo, 1993; Ruegg & Frances, 1995). Although a full sum
-
mary is beyond the scope of this article, Zanarini (2000) provides a compre
-
hensive review of empirical findings and an analysis of research trends of
the past three decades. In general, research indicates that BPD patients re
-
port higher rates of both childhood abuse (Herman, Perry, & van der Kolk,
1989; Ogata et al., 1990; Soloff, Lynch & Kelly, 2002; Zanarini et al., 1989,
1997) and childhood neglect (Johnson et al., 2000; Zanarini et al., 1989,
1997) than individuals with other PDs (Zanarini et al., 1989) or other Axis I
psychiatric disorders (Ogata et al., 1990).
Although there are discrepancies across studies regarding the types of
abuse and neglect most salient in the childhood histories of BPD patients,
childhood sexual abuse has often been identified as a factor that discrimi-
nates patients with BPD from those diagnosed with other PDs or other psy-
chiatric disorders (Weaver & Clum, 1993). Studies examining multiple
adverse childhood events have typically found that sexual abuse is common
among BPD patients and that it often occurs within a context of multiple
forms of abuse and neglect from both parents (Ogata et al., 1990; Zanarini et
al., 1997, 2000a).
CHILDHOOD MALTREATMENT AND OTHER PDs
The substantial literature on the childhood histories of individuals with
BPD contrasts with the relatively limited information regarding childhood
maltreatment associated with other PD groups. Although some investiga
-
tors have included patients with a variety of PD diagnoses to serve as Axis II
controls in studies of childhood antecedents of BPD, data from these par
-
ticipants are typically collapsed into an “other PD” category. It is difficult to
discern which, if any, adverse childhood experiences are common among
individuals with specific PDs other than BPD. In existing research that has
investigated the childhood histories of other distinct PD groups, typically
three approaches have been used. First, a small number of studies have
examined the childhood maltreatment histories reported by adult psychi
-
atric patients with a nonborderline PD diagnosis (e.g., Arbel and
Stravynski’s [1991] examination of the childhood histories of individuals
diagnosed with avoidant PD; Grilo and Masheb’s [2002] investigation of
childhood maltreatment and PDs among adults with binge eating disor
-
der). Using a different approach, a few published studies assessed for per
-
sonality pathology among individuals with documented childhood
194 BATTLE ET AL.
maltreatment histories. For example, Luntz and Widom (1994) assessed
antisocial PD traits and diagnosis among adults with documented child
abuse or neglect histories and among a matched group of controls. Finally,
a third set of studies assessed for childhood maltreatment history and cur
-
rent PD features among individuals in a college-based or community sam
-
ple. An example is Gibb, Wheeler, Alloy and Abramson’s (2001) study of PD
features and self-reported sexual, physical, and emotional maltreatment
history in a sample of undergraduates selected for cognitive vulnerability
to depression. In general, these studies indicate that other PD groups, in
addition to BPD, report or have documented histories of childhood
maltreatment, and that differential patterns of abuse and neglect may
exist for specific PD groups.
RATIONALE FOR CURRENT STUDY
This emerging body of research on childhood maltreatment associated
with adult PDs has several limitations. First, as noted earlier, relatively few
investigations have targeted distinct PD groups other than BPD, and con
-
sistent findings have not emerged from these studies. Second, investiga-
tions often rely on self-report measures of PD features, rather than a more
stringent approach such as structured diagnostic interviews. Third, a
comparison group is often not included, or the comparison group is a het-
erogeneous group of Axis II controls. Fourth, when diagnostic interviews
have been used, the study often focused narrowly on one PD group (Luntz
& Widom, 1994), or a comparison between two PD groups (Zanarini et al.,
1989). Comparing findings across these studies is difficult given the vari-
ability in how pathological childhood experiences are conceptualized and
assessed.
To date, no published studies have compared the childhood maltreatment
experiences of patients from a large sample of multiple PD groups. Without
such a comparison, it is difficult to determine the specific associations of
abuse and neglect experiences among different PD groups and whether find
-
ings regarding childhood histories of BPD patients hold when multiple
groups are examined concurrently. Thus, the primary goal of the present in
-
vestigation is to expand the body of knowledge regarding adverse childhood
experiences that precede adult PDs by examining not only BPD, but also
other PD groups and a psychiatric comparison group.
We addressed the following three questions in this study. First, what is
the frequency of childhood maltreatment among individuals with four
common PDs: borderline (BPD), schizotypal (STPD), avoidant (AVPD), and
obsessive-compulsive (OCPD), and how does this compare to those with a
psychiatric comparison condition, major depressive disorder without PD
(MDD)? We hypothesized that PD patients will, as a group, report higher
rates of abuse and neglect compared with MDD patients. Second, do indi
-
viduals with BPD report more childhood maltreatment than those with
other PD diagnoses? Based on the literature, we hypothesized that pa
-
tients with BPD will report more childhood maltreatment, particularly sex
-
ual abuse, compared with patients with other PDs. Third, are particular
types of childhood maltreatment associated with other, non-BPD diagno
-
CHILDHOOD MALTREATMENT 195
ses? On a more exploratory basis, we examined whether unique associa
-
tions exist between other PD diagnoses and specific adverse childhood
experiences.
METHOD
The present study uses baseline data from the Collaborative Longitudinal
PDs Study (CLPS), which involves four U.S. research sites: Boston, New Ha
-
ven, New York, and Providence. The overall design and methodology have
been described in detail earlier (Gunderson et al., 2000; McGlashan et al.,
2000); the following is an overview of the sample, assessment procedures,
and analyses relevant to this investigation. Earlier related work from the
CLPS has examined risk factors and early manifestations of AVPD (Rettew et
al., 2002) and the relation between lifetime traumatic experiences, PTSD,
and BPD (Yen et al., 2002). The present analyses are the first to evaluate
comprehensively multiple types of childhood maltreatment reported by
patients in all four PD groups.
PARTICIPANTS
Participants were patients aged 18 through 45 years who were seeking treat-
ment at clinics affiliated with the four CLPS sites; some additional partici-
pants were recruited from advertisements targeting individuals who had
sought psychiatric treatment elsewhere. Participants were screened to ex-
clude individuals with (a) acute substance intoxication/withdrawal, (b) ac-
tive psychosis, (c) an IQ less than or equal to 85, (d) cognitive impairment,
and (e) a history of schizophrenia, schizophreniform, or schizoaffective dis-
orders. In addition, individuals had to meet inclusion criteria for one or more
of the PDs targeted in the CLPS (STPD, BPD, AVPD, OCPD), as assessed by
the Diagnostic Interview for DSM-IV PDs (DIPD-IV; Zanarini, Frankenburg,
Sickel, & Yong, 1996), or for the comparison group, major depressive disor
-
der without PD (MDD), as assessed by the Structured Clinical Interview for
DSM-IV Disorders (SCID-I/P: First, Gibbon, Spitzer, & Williams, 1996). To
promote validity of PD diagnoses, raters’ diagnoses had to be confirmed by
at least one other method of Axis II assessment, either the self-report Sched
-
ule for Adaptive and Nonadaptive Personality (SNAP; Clark, 1993) or the cli
-
nician-rated Personality Assessment Form (PAF; Shea, Glass, Pilkonis,
Watkins, & Docherty, 1987). Although participants were selected for inclu
-
sion based on having one of the PDs targeted in the CLPS, many had other
Axis II diagnoses as well; overall, the mean number of co-occurring PD diag
-
noses was 1.4 (McGlashan et al., 2000). The majority of PD patients also had
comorbid Axis I disorders, with a mean number of Axis I lifetime diagnoses of
3.4 (McGlashan et al., 2000).
The total CLPS sample consisted of 668 participants, 63% of whom were
women and 76% of whom were Caucasian. Before participation, all pa
-
tients were provided with a full explanation of study procedures and signed
a written informed consent form. Childhood experiences data were avail
-
able for 600 of 668 participants. The final sample included in the present
investigation consisted of 600 individuals, including 517 who were diag
-
196 BATTLE ET AL.