The association of personality disorders with the
prospective 7-year course of anxiety disorders
E. B. Ansell1*, A. Pinto2, M. O. Edelen3, J. C. Markowitz2, C. A. Sanislow4, S. Yen5, M. Zanarini6,
A. E. Skodol7, M. T. Shea5, L. C. Morey8, J. G. Gunderson6, T. H. McGlashan1and C. M. Grilo1
1Department of Psychiatry, Yale University School of Medicine, New Haven, CT, USA
2Department of Psychiatry, Columbia University College of Physicians & Surgeons, New York State Psychiatric Institute, New York, NY, USA
3Rand Corporation, Santa Monica, CA, USA
4Department of Psychology, Wesleyan University, Middletown, CT, USA
5Department of Veterans Affairs, and Department of Psychiatry and Human Behavior, Warren Alpert Medical School of Brown University,
Providence, RI, USA
6Department of Psychiatry, Harvard Medical School and McLean Hospital, Belmont, MA, USA
7University of Arizona School of Medicine and the Sunbelt Collaborative, Tucson, AZ, USA
8Department of Psychology, Texas A&M University, College Station, TX, USA
Background. This study prospectively examined the natural clinical course of six anxiety disorders over 7 years of
follow-up in individuals with personality disorders (PDs) and/or major depressive disorder. Rates of remission,
relapse, new episode onset and chronicity of anxiety disorders were examined for specific associations with PDs.
Method. Participants were 499 patients with anxiety disorders in the Collaborative Longitudinal Personality
Disorders Study, who were assessed with structured interviews for psychiatric disorders at yearly intervals
throughout 7 years of follow-up. These data were used to determine probabilities of changes in disorder status for
social phobia (SP), generalized anxiety disorder (GAD), obsessive-compulsive disorder (OCD), post-traumatic stress
disorder (PTSD), panic disorder and panic disorder with agoraphobia.
Results. Estimated remission rates for anxiety disorders in this study group ranged from 73% to 94%. For those
patients who remitted from an anxiety disorder, relapse rates ranged from 34% to 67%. Rates for new episode onsets
of anxiety disorders ranged from 3% to 17%. Specific PDs demonstrated associations with remission, relapse, new
episode onsets and chronicity of anxiety disorders. Associations were identified between schizotypal PD with course
of SP, PTSD and GAD; avoidant PD with course of SP and OCD; obsessive-compulsive PD with course of GAD,
OCD, and agoraphobia; and borderline PD with course of OCD, GAD and panic with agoraphobia.
Conclusions. Findings suggest that specific PD diagnoses have negative prognostic significance for the course of
anxiety disorders underscoring the importance of assessing and considering PD diagnoses in patients with anxiety
Received 2 March 2010; Revised 29 July 2010; Accepted 5 August 2010; First published online 14 September 2010
Key words: Chronicity, anxiety disorders, personality disorders, relapse, remission.
Anxiety disorders are prevalent diagnoses and are as-
sociated with substantial life impairments (Roy-Byrne
& Cowley, 1994; Bruce et al. 2005; Grant et al. 2005;
Boden et al. 2007; Weisberg, 2009; Yates, 2009).
Prospective studies on the course of anxiety disorders
suggest a chronic burden, with low rates of recovery
that appear to be worsened by certain co-morbid
psychiatric conditions (Bowen et al. 2000; Yonkers et al.
2003; Bruce et al. 2005). Factors that moderate the
course of anxietydisorders, therebyreducing the prob-
ability of remission or increasing the likelihood for
relapse, are an important focus for research. One such
factor is the presence of a co-morbid personality dis-
Research has identified increased clinical severity
and decreased psychosocial functioning in patients
with anxiety disorder when PDs are co-morbid
(Skodol et al. 1995; Ozkan & Altindag, 2005). In com-
munity mental health settings in four urban com-
munities, PDs were associated with anxiety disorders
and contributed independently to global functioning
* Address for correspondence: E. B. Ansell, Ph.D., Yale University
School of Medicine, 2 Church Street South, Suite 209, New Haven,
CT 06519, USA.
Psychological Medicine (2011), 41, 1019–1028.
f Cambridge University Press 2010
impairment (Newton-Howes et al. 2010). However,
research has not consistently identified negative out-
comes in anxiety disorders co-morbid with PDs.
Although a prior study reported that co-occurring
PDs moderated poorer outcome for panic disorder
(Roy-Byrne & Cowley, 1994), a recent study found
greater improvements in psychodynamic treatment
of patients with panic disorder and cluster C PD co-
morbidity (Milrod et al. 2007). In addition, it is unclear
which specific PD diagnoses are associated with the
prospective course of specific anxiety disorders.
Few studies have examined the association between
PDs and anxiety course prospectively. Individuals
who met diagnostic criteria for at least one PD by early
adulthood had an elevated risk for anxiety disorder
onset by middle adulthood, even when controlling
for a history of Axis I disorders (Johnson et al. 2006).
In a 6-year follow-up of 284 Norwegian outpatients
using DSM-III-R diagnoses, borderline PD (BPD) pre-
dicted any anxiety disorder at follow-up, obsessive-
compulsive PD (OCPD) predicted panic disorder at
follow-up and avoidant PD (AVPD) predicted social
phobia (SP) at follow-up (Alnæs & Torgersen, 1999).
However, these diagnoses were based on one follow-
up evaluation and did not examine remissions and
relapse of anxiety disorders. Only one study has ex-
amined the effect of PD co-morbidity on the prospec-
tive, naturalistic course of anxiety disorders. A 5-year
follow-up of 514 participants in the Harvard/Brown
DSM-III-R PDs decreased rates of remission for gener-
alized anxiety disorder (GAD) and SP, but not panic
disorder (Massion et al. 2002). Specifically, dependent
and AVPDs decreased remission rates for GAD, and
AVPD decreased remission rates for SP. However, this
study had several limitations. Only three anxiety dis-
orders were assessed and the rates of specific PDs
in the overall sample were low enough to preclude
analysis of some PDs as course predictors [i.e. <1%
schizotypal PD (STPD), 3–7% BPD, 5–9% OCPD]. The
analysis examined remissions but did not look at re-
lapse, chronicity or new episodes of anxiety disorders.
Although the findings suggest that the presence of
a PD impairs anxiety disorder outcome, the lack of
power for specific PDs leaves questions unanswered.
In particular, given prior findings in HARP that co-
morbid PDs are associated with course of anxiety
disorders, it is important to determine whether the
presence of any specific PD diagnoses, relative to
other PD diagnoses, are associated with the course of
anxiety disorders in a clinical group with high rates of
The Collaborative Longitudinal Personality Dis-
orders Study (CLPS) is a prospective, naturalistic
study designed to assess the course and outcome of
patients with PDs and a comparison group of patients
with current major depressive disorder (MDD) but no
PD (Gunderson et al. 2000; McGlashan et al. 2000;
Skodol et al. 2005). Patients were not excluded based
on presence of an anxiety disorder. Annual assess-
ments were made for six anxiety disorders: SP; GAD;
OCD; post-traumatic stress disorders (PTSD); panic
disorder with and without agoraphobia. Importantly,
status of each anxiety disorder was followed through-
out the study, allowing for remission and relapse
analysis. Previously published CLPS diagnostic co-
occurrence at baseline or lifetime frequencies (22.9%
for SP, 21.4% for GAD, 15.6% for OCD, 29.6%
for PTSD and 26.1% for panic disorder) indicate that
many individuals with PDs also have anxiety dis-
orders (McGlashan et al. 2000).
The purpose of this study was to examine: (1) rates
of remission, relapse and new onsets of anxiety dis-
orders in a study group with high rates of PD diag-
noses; (2) the prospective effects of a set of predictors;
four PD diagnoses, MDD, gender, age and number
of Axis I disorders, on the probability of remission,
relapse and new onsets of anxiety disorders; (3) the
prospective effects of the set of predictors on the
chronicity of the anxiety disorders over 7 years of fol-
low-up. This study represents the only naturalistic
prospective assessment of DSM-IV PDs on remission,
relapse, new episode onsets and chronicity of multiple
anxiety disorders. It also extends prior findings on
DSM-III-R PDs by examining the course of anxiety
disorders in a study group with high rates of PD di-
Participants were drawn from the CLPS, a multi-site,
prospective naturalistic longitudinal study. Recruit-
ment sought a diverse, clinically representative study
group from in- and out-patient clinical programs af-
filiated with four recruitment sites (Brown, Columbia,
Harvard, and Yale). CLPS enrolled 668 participants
aged 18–45 years with at least one of four PDs (STPD,
BPD, AVPD and OCPD), or with current MDD with-
out any PD. Details of the CLPS methods and study
group characteristics, including specific co-occurrence
patterns among Axis I and Axis II diagnoses, have
been reported (Gunderson et al. 2000; McGlashan et al.
2000; Skodol et al. 2005). The co-occurrence patterns
echo those reported for other clinical samples (Becker
et al. 2000), increasing confidence in the generaliz-
ability of this study group to other clinical groups.
The current report includes 499 participants who
met criteria for current anxiety disorder at baseline,
1020E. B. Ansell et al.
or who had new episode onsets during the 7 years
of follow-up and who had at least 12 months of
follow-up data available. Mean age was 32.5 (S.D.=8.1)
years. Of the participants, 65% were female and
35% male, 76% were Caucasian and 24% minorities
(13% African American, 8% Hispanic American and
2% ‘other’). Baseline anxiety disorder diagnoses in
this study group were 29% SP, 31% GAD, 21% OCD,
33% PTSD, 13% panic disorder, 17% panic disorder
with agoraphobia and 3% agoraphobia without panic
disorder. There were 58%, 25%, 8% and 4% who had
one, two, three and four anxiety disorder diagnosis at
baseline, respectively. Altogether, 48% had a baseline
diagnosis of MDD, 26% had no other Axis I diagnosis
at baseline in addition to the anxiety disorder and
46%, 21%, and 6% had one, two or three other diag-
noses, respectively. PD diagnosis rates at baseline
were 17% STPD, 56% AVPD, 43% OCPD and 40%
BPD; 12% had no PD diagnosis.
All participants provided written informed consent
following a full description of study procedures.
Each collaborating site’s Institutional Review Board
approved the study protocol. Interviewers were ex-
perienced research-clinicians with masters or doctoral
degrees in mental health disciplines who underwent
extensive standardized training to achieve reliability
vised interviewers within and across sites to maintain
reliability and prevent temporal drift.
At baseline, interviewers administered the Structured
Clinical Interview for DSM-IV Axis I Disorders –
Patient Version (SCID-I/P; First et al. 1996) to assess
Axis I psychiatric disorders and the Diagnostic Inter-
view for DSM-IV Personality Disorders (DIPD-IV;
Zanarini et al. 1996) to assess all PDs. Participants were
re-interviewed at 6 and 12 months and yearly there-
after for 7 years following baseline. The course of
anxiety disorders was assessed using the Longitudinal
Interval Follow-up Evaluation (LIFE; Keller et al.
1987). To maximize the reliability of the course of
Axis I disorders, these follow-up interviews were not
blind and were conducted by the interviewer from the
previous interval whenever possible.
The SCID-I/P (First et al. 1996), a diagnostic interview
to assess current and lifetime psychiatric disorders,
was administered at baseline. k coefficients for inter-
rater reliability for disorders ranged from 0.57 to 1.0;
k coefficients for specific anxiety disorders were as
follows: 0.65 (for panic disorder), 0.63 (for SP), 0.57
(for OCD), 0.63 (for GAD) and 0.88 (for PTSD)
(Zanarini et al. 2000).
The DIPD-IV (Zanarini et al. 1996), a semi-struc-
tured diagnostic interview to assess DSM-IV Axis II
PDs, was given at baseline. Each PD criterion is as-
sessed with one or more questions rated on a 3-point
scale (0=not present; 1=present but of uncertain
clinical significance; 2=present and clinically signifi-
cant). The DIPD-IV requires that criteria be pervasive
for at least 2 years and characteristic of most of the
person’s adult life. Inter-rater reliability (based on 84
pairs of raters independently rating 27 videotaped
assessments) k coefficients for PD diagnoses ranged
from 0.58 to 1.0 (Zanarini et al. 2000). Test–retest
reliability k coefficients (based on two direct inter-
views of 52 participants, performed 7–10 days apart
with the second interview blind to the first interview)
ranged from 0.69 (BPD) to 0.74 (OCPD).
The LIFE (Keller et al. 1987) is a semi-structured in-
terview rating system for assessing the longitudinal
course of mental disorders. The LIFE has served as the
primary measure for major longitudinal studies of
psychopathology (Yonkers et al. 2003; Bruce et al. 2005;
Grilo et al. 2005, 2007) and has good-to-excellent re-
liability (Warshaw et al. 2001). The LIFE developers
and official training staff at the Brown site trained and
certified interviewers across sites and provided on-
going training and consultation for the interview and
ratings. These methods have been shown to maintain
long-term reliability and prevent drift (Warshaw et al.
As in the National Institute of Mental Health –
Collaborative Depression Study (NIMH-CDS) and
HARP studies (Keller et al. 1982a,b; Bruce et al. 2005),
the LIFE measured the weekly presence and severity
of psychopathology. The severity of psychopathology
is quantified by weekly ‘psychiatric status ratings’
(PSRs) for each Axis I disorder present. For the current
study a 3-point scale was used: PSR=1, no symptoms;
PSR=2, moderate symptoms, but less than full diag-
nostic criteria; PSR=3, symptoms meeting full diag-
nostic criteria. Remission from anxiety disorders was
defined as eight consecutive weeks with PSR ratings
<2 (minimal or no symptoms). Relapse was defined
as four consecutive weeks with PSR ratings of 3.
These definitions parallel those used in other major
longitudinal studies of psychopathology (Keller et al.
1982a,b; Bruce et al. 2005). Relapse of GAD was ex-
amined using a four consecutive week definition and a
DSM-IV 26 consecutive week (6 month) diagnostic
definition. For individuals with a history of meeting
full criteria and time duration for GAD, emergence
of >4 weeks of clinically significant symptoms after
PD and course of anxiety disorders 1021
remission may represent a meaningful and distressing
syndromal recurrence (Kessler et al. 2005; Ruscio et al.
2005; Angst et al. 2006). Therefore, both definitions
provide relevant insight into the course of GAD.
Statistical analyses were conducted with SAS ver-
sion 9.1.3 (SAS Institute Inc., USA). For each anxiety
disorder, we used standard survival analysis methods
to obtain rates of remission, relapse and new episode
onsets on weekly PSR ratings over the 7 years of
follow-up. Anxiety disorders examined in the analyses
were SP, GAD, OCD, PTSD and panic disorder with
and without agoraphobia. Although assessed, agora-
phobia without panic disorder was excluded from
remission and relapse analyses due to small sample
size at baseline (n=12). Cox (1972) proportional
hazards regression analyses were used to examine
possible predictors of remission, relapse and new
episodes (individuals who did not report a current
anxiety disorder episode at baseline). In these analy-
ses, we simultaneously considered the following
variables: age; gender; total number of Axis I dis-
orders; presence of MDD, STPD, AVPD, OCPD and
BPD diagnoses at baseline. Analyses considered only
the first relapse in instances with multiple relapses.
Linear regression analyses examined the chronicity of
anxiety disorder episodes. Chronicity was determined
by calculating the proportion of weeks of follow-up
that were spent in episode (PSR 2 or PSR 3) for in-
dividuals with anxiety disorders at baseline and for
those with new episode onsets after baseline. The
same aforementioned variables were used in the
chronicity analyses. For all tests, two-tailed tests with
a=0.05 were considered statistically significant.
Figure 1 displays the survival curve for probability
of remission from each anxiety disorder and Fig. 2
displays the survival curve for probability of relapse of
each anxiety disorder. Results from the proportional
hazard regressions appear in Table 1 for remission,
Table 2 for relapse and Table 3 for new onsets of each
Of the 135 subjects with SP at baseline, life-
table analyses estimated that 79.7% had remitted by
year 7. Subjects with AVPD at baseline were signifi-
cantly less likely to remit from SP compared with
those without AVPD [hazard ratio (HR)=0.482,
p=0.0025]. For the 104 subjects with SP at baseline
who remitted, 43% relapsed by year 7 of follow-up.
Patients with STPD at baseline were more likely
to have a SP relapse than those without STPD
(HR=2.469, p=0.03). In total, 46 new episode onsets of
SP occurred over 7 years of follow-up, five of which
had a lifetime history of the disorder; the probability
of new episode onset was 0.10. Patients with baseline
AVPD were more likely to have a new episode onset of
SP compared with those without AVPD (HR=3.667,
chronicity of SP episodes show that the presence of
AVPD versus no AVPD at baseline had a significantly
greater effect relative to the other predictors on the
proportion of weeks spent in episode of SP (b=0.19,
0 52104 156
Weeks since intake
Probability of remission
208 260312 364
Panic with agor
Fig. 1. Remission survival curves for anxiety disorders over
7 years (364 weeks) of follow-up. GAD, generalized anxiety
disorder; OCD, obsessive-compulsive disorder; PTSD,
post-traumatic stress disorder; agor, agoraphobia.
Weeks since intake
Probability of relapse
GAD 4 week
GAD 26 week
Panic with agor
Fig. 2. Relapse survival curves for anxiety disorders
over 7 years (364 weeks) of follow-up. GAD, generalized
anxiety disorder; OCD, obsessive-compulsive disorder;
PTSD, post-traumatic stress disorder; agor, agoraphobia.
1022E. B. Ansell et al.
Of the 148 subjects with GAD at baseline, life-
table analyses estimated that 83.5% had remitted by
year 7. For the 118 subjects with GAD at baseline who
remitted, 67% relapsed using a 4-week definition for
GAD. Using a 26-week definition, 41% relapsed by
year 7. Patients with OCPD at baseline were more
likely to have a GAD relapse using a 4-week definition
compared with those without OCPD (Table 2;
HR=1.784, p=0.0328). A greater number of Axis I
diagnoses were also associated with 4-week GAD
relapse (HR=1.461, p=0.05). This did not hold for a
26-week definition of relapse. There were 77 new epi-
sode onsets (using a 26-week definition) of GAD by
year 7; two of which had a lifetime history of the dis-
order. The probability of new episode onset was 0.17.
Subjects with baseline OCPD were more likely than
those without OCPD to have a new episode onset
(HR=2.209, p=0.001). Patients diagnosed with BPD at
intake were more likely to have a new episode onset of
GAD than those without baseline BPD (HR=2.162,
p=0.001). Regression analyses examining the chroni-
city of GAD episodes show that the presence of STPD
Table 1. Risk for remission of anxiety disorders
Social phobiaGAD OCD PTSD
GAD, Generalized anxiety disorder; OCD, obsessive-compulsive disorder; PTSD, post-traumatic stress disorder; MDD,
major depressive disorder; STPD, schizotypal personality disorder; AVPD, avoidant personality disorder; OCPD, obsessive-
compulsive personality disorder; BPD, borderline personality disorder.
Table 2. Risk for relapse of anxiety disorders
Social phobiaGAD 4 weeksGAD 26 weeksOCDPTSD
GAD, Generalized anxiety disorder; OCD, obsessive-compulsive disorder; PTSD, post-traumatic stress disorder;
MDD, major depressive disorder; STPD, schizotypal personality disorder; AVPD, avoidant personality disorder;
OCPD, obsessive-compulsive personality disorder; BPD, borderline personality disorder.
PD and course of anxiety disorders1023
versus no STPD at intake had a significantly greater
effect than other predictors on the proportion of
weeks spent in episode of GAD (b=0.14, t=2.26,
Of the 91 subjects with obsessive-compulsive disorder
(OCD) at baseline, life-table analyses estimated that
73.6% remitted by year 7. Of the 64 subjects with OCD
at baseline who remitted, 36% relapsed by year 7.
without BPD to relapse in OCD (HR=3.56, p=0.014).
There were 35 new episode onsets of OCD by year 7;
four of which had a lifetime history of the disorder,
with risk for new episode onset 0.07 over 7 years.
Patients with OCPD or AVPD diagnosis at intake
were more likely to have a new episode onset of OCD
than those without these baseline diagnoses (OCPD
HR=1.958, p=0.05 AVPD HR=2.039, p=0.05).
Of the 142 subjects with PTSD at baseline, 76.8% had
remitted by year 7. Subjects with STPD at baseline
were significantly less likely to remit from PTSD com-
pared with those without STPD (HR=0.564, p=0.05).
Older patients were slightly less likely to remit from
PTSD (HR=0.968, p=0.0337). Of the 102 subjects with
PTSD at baseline who remitted, 34% had relapsed by
year 7. Proportional hazard regression analyses indi-
cate that subjects with OCPD at intake were less likely
to have a PTSD relapse than those without OCPD at
intake (HR=0.332, p=0.047). There were 39 new epi-
sode onsets of PTSD by 7 years of follow-up, 12 of
which had a lifetime history of the disorder, with risk
for new episode onset of 0.09. Older patients were
slightly more likely to have a new episode onset of
PTSD than younger patients (HR=1.048, p=0.0216).
Panic disorder without agoraphobia
Of the 63 subjects with panic without agoraphobia at
baseline, 93.6% remitted by year 7. Of the 57 subjects
with panic without agoraphobia at baseline who
remitted, 35% relapsed by year 7. Having a BPD di-
agnosis at intake resulted in a HR of 3.14 for relapse of
panic without agoraphobia and approached signifi-
cance (p=0.06). There were 34 new episode onsets of
panic without agoraphobia by year 7, three of which
had a lifetime history of the disorder, with the prob-
ability risk for new episode onset 0.07. Only AVPD
approached significance in moderating weeks spent in
Panic disorder with agoraphobia
Of the 82 subjects with panic with agoraphobia diag-
noses at baseline, 73.2% remitted by year 7. Of the 56
subjects with baseline panic with agoraphobia who
remitted, 64% relapsed by year 7. Subjects with AVPD
at intake were less likely to relapse in panic with
agoraphobia than those without baseline AVPD
(HR=0.354, p=0.047), Women were more likely to
relapse (HR=3.131, p=0.0276) and the total number
of Axis I disorders was associated with an increased
Table 3. New episode onsets of anxiety disorders
Social phobia GADa
GAD, Generalized anxiety disorder; OCD, obsessive-compulsive disorder; PTSD, post-traumatic stress disorder;
MDD, major depressive disorder; STPD, schizotypal personality disorder; AVPD, avoidant personality disorder;
OCPD, obsessive-compulsive personality disorder; BPD, borderline personality disorder.
aGAD onset using a 26-week definition for diagnosis.
1024E. B. Ansell et al.
relapse risk (HR=2.116, p=0.0261). There were 16
new episode onsets of panic with agoraphobia by
year 7, with four cases having a lifetime history of the
disorder; the probability of a new episode onset was
0.03. Subjects with BPD at baseline were significantly
more likely than those without BPD to have a new
episode onset of panic with agoraphobia (HR=3.043,
Agoraphobia without panic disorder
There were 26 new episode onsets of agoraphobia
without panic during the 7-year follow-up, one of
which had a lifetime history for the disorder. The
probability of a new episode onset was 0.05. Those
with OCPD at baseline were significantly more likely
than those without OCPD to have a new episode onset
of agoraphobia without panic (HR=2.667, p=0.016).
Older age had a small but significant effect on the
proportion of weeks spent in episode (b=0.02, t=2.49,
This study provides the first examination of the
naturalistic, 7-year course of six anxiety disorders in
a clinical group with a high rate of PDs. Remission
estimates in this study group indicate that by year 7,
between 73% and 94% of individuals had anxiety
disorders remit and between 34% and 67% of those
individuals had anxiety disorders relapse. When com-
pared with findings from the HARP study (37–82%
remission and 39–56% relapse; Bruce et al. 2005), these
high rates of remission and relapse suggest that anxi-
ety disorders are remitting and relapsing more fre-
quently in a clinical group with a high rate of PD
diagnoses. Our findings highlight the complex and
variable course of anxiety disorders, characterized by
frequent remissions and relapses over time as well as
the occurrence of new onsets of anxiety disorders in
patient groups with high rates of PDs. Our findings
also highlight the negative prognostic significance of
PDs on the course of anxiety disorders. These findings
underscore the importance of considering both anxiety
disorders and PDs by researchers and clinicians who
work with either patient group.
The high rates of PD diagnoses in this study group
allowed examination of specific PD associations with
the course of specific anxiety disorders while control-
ling for the effects of age, gender, number of Axis I
disorders and the other study PD diagnoses. OCPD
was associated with increased risk for new onset of
OCD, GAD and agoraphobia episodes, increased risk
of GAD relapse and decreased risk of PTSD relapse
over and above other factors. AVPD was associated
with decreased likelihood of SP remission, increased
likelihood of new SP and OCD episode onset and
greater chronicity in SP episodes. AVPD was also as-
sociated with decreased risk for relapse of panic dis-
order with agoraphobia. BPD was associated with
increased risk of new episode onsets for GAD and
panic disorder with agoraphobia, as well as an in-
creased risk of relapse of OCD. STPD was associated
with decreased remission rates in PTSD, increased
risk for relapse of SP and greater chronicity in GAD
Several of these findings replicate associations
identified in prior studies conducted over shorter
duration or without continuous follow-up. This study
replicates findings that AVPD is predominately associ-
ated with the course of SP (Massion et al. 2002) and
corroborates speculation that these two disorders
share a common dimension (Ralevski et al. 2005;
Reich, 2009). Prior research reviewed by Reich (2000)
also associated cluster A PDs, such as STPD, with SP,
although this finding has been dismissed by the
authors as an erroneous assignment of paranoid fears
to social anxiety. However, the present findings that
STPD increases risk for SP in a carefully diagnosed
clinical group suggest that either the prior findings
may have been accurate or, alternatively, that the cri-
teria for SP are constructed in a manner that makes it
difficult to ascertain the differences in pathological
mechanisms underlying the symptomatic expression.
We replicated findings associating BPD with increased
risk for new onsets of anxiety disorder episodes
(Alnæs & Torgersen, 1999), specifically panic disorder
and GAD. The BPD and STPD associations with GAD
in this study group replicate the findings from a large
epidemiological survey (Grant et al. 2009). Our finding
associating BPD with increased risk of OCD relapse is
consistent with prior findings that individuals with
a BPD diagnosis and self-mutilation exhibited signifi-
cantly greater OCD symptomatology (McKay et al.
2000). The present finding that PDs were not associ-
ated with remission from panic disorder also re-
plicates prior research (Massion et al. 2002). Contrary
to prior findings, we did not identify OCPD as a risk
factor for new onsets or chronicity of panic disorder
(Alnæs & Torgersen, 1999), although it was associated
with new onsets of OCD (Maina et al. 2008) and with
relapse of GAD (Grant et al. 2009). Consistent with
theoretical assertions in OCPD regarding control of
worry, individuals with OCPD may have increased
vulnerability for worry when they find their perfec-
tionistic motives are unattainable or their rigid stan-
dards are unmet.
These findings have several clinical and research
implications. Treatment planning for individuals with
co-occurring PDs and anxiety disorders may need to
PD and course of anxiety disorders 1025
take into account these findings when considering the
course of symptoms. In particular, individuals who
present for treatment with PD diagnoses and who
have a history of a co-occurring anxiety disorder
(BPD with OCD, or AVPD with SP) may warrant ad-
ditional treatment focus on preventing the recurrence
of symptoms associated with anxiety disorder. Further
research is needed to examine whether existing treat-
ments for anxiety disorders are equally effective in
samples with co-occurring PDs. In addition, the
associations in course between disorders may be me-
chanistically explained by personality traits, which
underlie both disorders. Future directions should
consider whether maladaptive personality traits in-
crease vulnerability for course of anxiety disorders.
Several potential limitations should be considered
when interpreting our findings. One limitation is the
absence of a comparison group with no PD for each
anxiety disorder. However, this study group did
contain a MDD with no PD comparison group, a di-
agnosis that is frequently co-morbid with anxiety dis-
orders. This MDD control group expands the potential
variance in ways that serve to strengthen the analytic
approach in lieu of a purified contrast group. In
addition, the study was a naturalistic study of treat-
ment-seeking individuals, yielding greater variability
among participants than a controlled treatment pro-
tocol and thereby increasing the generalizability of
these findings to patients with PDs. However, these
findings may not generalize to non-treatment-seeking
individuals or to individuals who are treatment seek-
ing but do not express interest in participating in re-
search. In addition, although the LIFE methodology is
well validated and reliable, it is potentially vulnerable
to bias or participant difficulties in recall of symptoms.
Weekly assessment of symptoms may result in differ-
ent associations than were identified in the current
The moderating influence of PDs on anxiety dis-
order course may represent differential stylistic ap-
proaches that use cognitive, affective and behavioral
mechanisms to differentially mitigate dispositional
anxiety, increasing risk for relapse, delaying remission
and increasing chronicity for anxiety disorders and
decreasing risk for relapse for other anxiety disorders.
For example, a diagnosis of BPD may make individ-
uals predisposed to OCD tendencies more vulnerable
to mitigating increasing anxiety resulting from affect-
ive and interpersonal instability through compulsive
or impulsive behaviors, such as self-harm. Alterna-
tively, individuals with AVPD may establish a life-
style in which they effectively avoid social, work and
recreational situations, resulting in little opportunity
to experience situations in which agoraphobic re-
sponses may occur, thus ‘protecting’ against relapse
of the panic disorder. AVPD and panic disorder with
agoraphobia may have a synergistic, or pathoplastic,
relationship such that the presence of personality trait
tendencies and anxious predispositions interact to
promote the presence of one or the other disorder.
Future research will need to examine the prospective
interactive effects of these disorders to determine the
risk, protective and/or pathoplastic mechanisms
underlying these findings. In particular, future studies
will need to examine the concurrent course of PDs and
anxiety disorders and consider whether alternative
definitions of remission, relapse and chronicity, or
whether an alternative construal of the disorders
(e.g. changes within diagnostic categories in the DSM),
affect findings. These data not only raise questions
but also inform the boundaries of these classes of dis-
order relevant to the development of the DSM-V.
Declaration of Interest
Supported by NIMH grants MH 50837, 50838, 50839,
50840, 50850; MH016545; MH073708; MH080221.
This publication has been reviewed and approved
by the Publications Committee of the Collaborative
Longitudinal Personality Disorders Study.
Alnæs R, Torgersen S (1999). A 6-year follow-up study
of anxiety disorders in psychiatric outpatients:
development and continuity with personality disorders
and personality traits as predictors. Nordic Journal of
Psychiatry 53, 409–416.
Angst J, Gamma A, Joseph Bienvenu O, Eaton WW,
Ajdacic V, Eich D, Rossler W (2006). Varying temporal
criteria for generalized anxiety disorder: prevalence and
clinical characteristics in a young age cohort. Psychological
Medicine 36, 1283–1292.
Becker DF, Grilo CM, Edell WS, McGlashan TH (2000).
Comorbidity of borderline personality disorder with
other personality disorders in hospitalized adolescents
and adults. The American Journal of Psychiatry 157,
Boden JM, Fergusson DM, Horwood LJ (2007). Anxiety
disorders and suicidal behaviours in adolescence and
young adulthood: findings from a longitudinal study.
Psychological Medicine 37, 431–440.
Bowen RC, Senthilselvan A, Barale A (2000). Physical illness
as an outcome of chronic anxiety disorders. Canadian
Journal of Psychiatry 45, 459–464.
Bruce SE, Yonkers KA, Otto MW, Eisen JL, Weisberg RB,
Pagano M, Shea MT, Keller MB (2005). Influence of
psychiatric comorbidity on recovery and recurrence in
generalized anxiety disorder, social phobia, and panic
disorder: a 12-year prospective study. American Journal
of Psychiatry 162, 1179–1187.
1026 E. B. Ansell et al.
Cox DR (1972). Regression models and life tables. Journal
of the Royal Statistical Society Series B (Methodological) 34,
First MB, Spitzer RL, Gibbon M, Williams JBW
(1996). Structured Clinical Interview for DSM-IV Axis I
Disorders – Patient Version (SCID-I/P), New York State
Psychiatric Institute: New York.
Grant BF, Goldstein RB, Chou SP, Huang B, Stinson FS,
Dawson DA, Saha TD, Smith SM, Pulay AJ,
Pickering RP, Ruan WJ, Compton WM (2009).
Sociodemographic and psychopathologic predictors of
first incidence of DSM-IV substance use, mood and anxiety
disorders: results from the Wave 2 National Epidemiologic
Survey on Alcohol and Related Conditions. Molecular
Psychiatry 14, 1051–1066.
Grant BF, Hasin DS, Blanco C, Stinson FS, Chou SP,
Goldstein RB, Dawson DA, Smith S, Saha TD, Huang B
(2005). The epidemiology of social anxiety disorder in the
United States: results from the National Epidemiologic
Survey on Alcohol and Related Conditions. Journal of
Clinical Psychiatry 66, 1351–1361.
Grilo CM, Pagano ME, Skodol AE, Sanislow CA,
McGlashan TH, Gunderson JG, Stout RL (2007). Natural
course of bulimia nervosa and of eating disorder not
otherwise specified: 5-year prospective study of
remissions, relapses, and the effects of personality
disorder psychopathology. Journal of Clinical Psychiatry 68,
Grilo CM, Sanislow CA, Shea MT, Skodol AE, Stout RL,
Gunderson JG, Yen S, Bender DS, Pagano ME, Zanarini
MC, Morey LC, McGlashan TH (2005). Two-year
prospective naturalistic study of remission from major
depressive disorder as a function of personality disorder
comorbidity. Journal of Consulting and Clinical Psychology 73,
Gunderson JG, Shea MT, Skodol AE, McGlashan TH,
Morey LC, Stout RL, Zanarini MC, Grilo CM,
Oldham JM, Keller MB (2000). The Collaborative
Longitudinal Personality Disorders Study: development,
aims, design, and sample characteristics. Journal of
Personality Disorders 14, 300–315.
Johnson JG, Cohen P, Kasen S, Brook JS (2006). Personality
disorders evident by early adulthood and risk for anxiety
disorders during middle adulthood. Journal of Anxiety
Disorders 20, 408–426.
Keller MB, Lavori PW, Friedman B, Nielsen E, Endicott J,
McDonald-Scott P, Andreasen NC (1987).
The Longitudinal Interval Follow-up Evaluation.
A comprehensive method for assessing outcome in
prospective longitudinal studies. Archives of General
Psychiatry 44, 540–548.
Keller MB, Shapiro RW, Lavori PW, Wolfe N (1982a).
Recovery in major depressive disorder: analysis with the
life table and regression models. Archives of General
Psychiatry 39, 905–910.
Keller MB, Shapiro RW, Lavori PW, Wolfe N (1982b).
Relapse in major depressive disorder: analysis with the
life table. Archives of General Psychiatry 39, 911–915.
Kessler RC, Brandenburg N, Lane M, Roy-Byrne P,
Stang PD, Stein DJ, Wittchen HU (2005). Rethinking the
duration requirement for generalized anxiety disorder:
evidence from the National Comorbidity Survey
Replication. Psychological Medicine 35, 1073–1082.
McGlashan TH, Grilo CM, Skodol AE, Gunderson JG,
Shea MT, Morey LC, Zanarini MC, Stout RL (2000). The
Collaborative Longitudinal Personality Disorders Study:
baseline Axis I/II and II/II diagnostic co-occurrence.
Acta Psychiatrica Scandinavica 102, 256–264.
McKay D, Kulchycky S, Danyko S (2000). Borderline
personality and obsessive-compulsive symptoms.
Journal of Personality Disorders 14, 57–63.
Maina G, Albert U, Salvi V, Pessina E, Bogetto F (2008).
Early-onset obsessive-compulsive disorder and personality
disorders in adulthood. Psychiatry Research 158, 217–225.
Massion AO, Dyck IR, Shea MT, Phillips KA, Warshaw
MG, Keller MB (2002). Personality disorders and time
to remission in generalized anxiety disorder, social
phobia, and panic disorder. Archives of General Psychiatry
Milrod BL, Leon AC, Barber JP, Markowitz JC, Graf E
(2007). Do comorbid personality disorders moderate
panic-focused psychotherapy? An exploratory
examination of the American Psychiatric Association
practice guideline. Journal of Clinical Psychiatry 68, 885–891.
Newton-Howes G, Tyrer P, Anagnostakis K, Cooper S,
Bowden-Jones O, Weaver T (2010). The prevalence of
personality disorder, its comorbidity with mental state
disorders, and its clinical significance in community
mental health teams. Social Psychiatry and Psychiatric
Epidemiology 45, 453–460.
Ozkan M, Altindag A (2005). Comorbid personality
disorders in subjects with panic disorder: do personality
disorders increase clinical severity? Comprehensive
Psychiatry 46, 20–26.
Ralevski E, Sanislow CA, Grilo CM, Skodol AE,
Gunderson JG, Tracie Shea M, Yen S, Bender DS,
Zanarini MC, McGlashan TH (2005). Avoidant
personality disorder and social phobia: distinct enough
to be separate disorders? Acta Psychiatrica Scandinavica 112,
Reich J (2000). The relationship of social phobia to avoidant
personality disorder: a proposal to reclassify avoidant
personality disorder based on clinical empirical findings.
European Psychiatry: Journal of the Association of European
Psychiatrists 15, 151–159.
Reich J (2009). Avoidant personality disorder and its
relationship to social phobia. Current Psychiatry Reports 11,
Roy-Byrne PP, Cowley DS (1994). Course and outcome in
panic disorder: a review of recent follow-up studies.
Anxiety 1, 151–160.
Ruscio AM, Lane M, Roy-Byrne P, Stang PE, Stein DJ,
Wittchen HU, Kessler RC (2005). Should excessive worry
be required for a diagnosis of generalized anxiety
disorder? Results from the US National Comorbidity
Survey Replication. Psychological Medicine 35, 1761–1772.
Skodol AE, Gunderson JG, Shea MT, McGlashan TH,
Morey LC, Sanislow CA, Bender DS, Grilo CM,
Zanarini MC, Yen S, Pagano ME, Stout RL (2005). The
Collaborative Longitudinal Personality Disorders Study
PD and course of anxiety disorders 1027
(CLPS): overview and implications. Journal of Personality Download full-text
Disorders 19, 487–504.
Skodol AE, Oldham JM, Hyler SE, Stein DJ, Hollander E,
Gallaher PE, Lopez AE (1995). Patterns of anxiety and
personality disorder comorbidity. Journal of Psychiatric
Research 29, 361–374.
Warshaw MG, Dyck I, Allsworth J, Stout RL, Keller MB
(2001). Maintaining reliability in a long-term psychiatric
study: an ongoing inter-rater reliability monitoring
program using the longitudinal interval follow-up
evaluation. Journal of Psychiatric Research 35, 297–305.
Weisberg RB (2009). Overview of generalized anxiety
disorder: epidemiology, presentation, and course.
Journal of Clinical Psychiatry 70 (Suppl. 2), 4–9.
Yates WR (2009). Phenomenology and epidemiology
of panic disorder. Annals of Clinical Psychiatry: Official
Journal of the American Academy of Clinical Psychiatrists 21,
Yonkers KA, Bruce SE, Dyck IR, Keller MB (2003).
Chronicity, relapse, and illness – course of panic disorder,
social phobia, and generalized anxiety disorder:
findings in men and women from 8 years of follow-up.
Depression and Anxiety 17, 173–179.
Zanarini MC, Frankenburg FR, Sickel AE, Yong L (1996).
Diagnostic Interview for DSM-IV Personality Disorders
(DIPD-IV). McLean Hospital: Belmont, MA.
Zanarini MC, Skodol AE, Bender D, Dolan R, Sanislow C,
Schaefer E, Morey LC, Grilo CM, Shea MT, McGlashan
TH, Gunderson JG (2000). The Collaborative
Longitudinal Personality Disorders Study: reliability
of axis I and II diagnoses. Journal of Personality Disorders
1028 E. B. Ansell et al.