Journal of Affective Disorders 195 (2016) 105–118

Contents lists available at ScienceDirect

Journal of Affective Disorders

journal homepage: www.elsevier.com/locate/jad

Research paper

The prevalence and predictors of bipolar and borderline personality

disorders comorbidity: Systematic review and meta-analysis
M. Fornaro a,n, L. Orsolini b,c,d, S. Marini e, D. De Berardis f, G. Perna g, A. Valchera d,
L. Ganança a,h, M. Solmi i,j, N. Veronese k, B. Stubbs l,m
a
New York State Psychiatric Institute, Columbia University, NY, USA
b
School of Life and Medical Sciences, University of Hertfordshire, Hatfield, Herts, UK
c
Department of Psychiatry and Neuropsychology, Maastricht University, Maastricht, Netherlands
d
Hermanas Hospitalarias – Villa San Giuseppe, Ascoli Piceno, Italy
e
Department of Neuroscience & Imaging, “G. D’Annunzio” University, Chieti, Italy
f
National Health Service, Department of Mental Health, Psychiatric Service of Diagnosis and Treatment, Hospital “G. Mazzini”, ASL 4, Teramo, Italy
g
Hermanas Hospitalarias-Villa San Benedetto Menni Hospital, Department of Clinical Neurosciences, FoRiPsi, Italy
h
Departamento de Psiquiatria e Saúde Mental, Faculdade de Medicina, Universidade de Lisboa, PT
i
Department of Neuroscience, University of Padova, Padua, Italy
j
National Health Care System, Padua Local Unit ULSS 17, Italy
k
Department of Medicine, DIMED, Geriatrics Section, University of Padua, Italy
l
Physiotherapy Department, South London and Maudsley NHS Foundation Trust, London SE5 8AZ, UK
m
Health Service and Population Research Department, Institute of Psychiatry, King's College London, De Crespigny Park, London SE5 8AF, UK

art ic l e i nf o a b s t r a c t

Article history: Introduction: Data about the prevalence of borderline personality (BPD) and bipolar (BD) disorders co-
Received 18 November 2015 morbidity are scarce and the boundaries remain controversial. We conducted a systematic review and
Received in revised form meta-analysis investigating the prevalence of BPD in BD and BD in people with BPD.
4 January 2016
Methods: Two independent authors searched MEDLINE, Embase, PsycINFO and the Cochrane Library
Accepted 24 January 2016
Available online 5 February 2016
from inception till November 4, 2015. Articles reporting the prevalence of BPD and BD were included. A
random effects meta-analysis and meta-regression were conducted.
Keywords: Results: Overall, 42 papers were included: 28 considering BPD in BD and 14 considering BD in BPD. The
Bipolar disorder (BD) trim and fill adjusted analysis demonstrated the prevalence of BPD among 5273 people with BD
Borderline personality disorder (BPD)
(39.94 7 11.78 years, 44% males) was 21.6% (95% CI 17.0–27.1). Higher comorbid BPD in BD were noted in
Comorbidity
BD II participants (37.7%, 95% CI 21.9–56.6, studies ¼ 6) and North American studies (26.2%, 95% CI 18.7–
Prevalence
Predictors 35.3, studies ¼11). Meta regression established that a higher percentage of males and higher mean age
Systematic review significantly (p o0.05) predicted a lower prevalence of comorbid BPD in BD participants. The trim and fill
Meta-analysis adjusted prevalence of BD among 1814 people with BPD (32.22 7 7.35 years, 21.5% male) was 18.5% (95%
CI 12.7–26.1).
Limitations: Paucity of longitudinal/control group studies and accurate treatment records.
Conclusions: BPD-BD comorbidity is common, with approximately one in five people experiencing a
comorbid diagnosis. Based on current diagnostic constructs, and a critical interpretation of results, both
qualitative and quantitative syntheses of the evidence prompt out the relevance of differences rather
similarities between BD and BPD.
& 2016 Elsevier B.V. All rights reserved.

1. Introduction

n
Correspondence to: Columbia University, New York State Psychiatric Institute, Discriminating between borderline personality (BPD) and bi-
1051 Riverside Drive, Unit 42, Room 2729, NYC ZIP 10032, United States.
polar (BD) disorder is difficult (Barroilhet et al., 2013; Ghaemi
E-mail addresses: mf3000@cumc.columbia.edu,
dott.fornaro@gmail.com (M. Fornaro), laura.orsolini@hotmail.it (L. Orsolini), et al., 2014), as well as crucial in the clinical and critical evaluation
sfnmarini@gmail.com (S. Marini), dodebera@aliceposta.it (D. De Berardis), of the comorbidity rates between the twos (Zimmerman and
pernagp@gmail.com (G. Perna), a.valchera@ospedaliere.it (A. Valchera),
Morgan, 2013).
lg2733@cumc.columbia.edu (L. Ganança), marco.solmi83@gmail.com (M. Solmi),
ilmannato@gmail.com (N. Veronese), brendon.stubbs@kcl.ac.uk (B. Stubbs). The odds of confusing BPD with BD are particularly high for

http://dx.doi.org/10.1016/j.jad.2016.01.040
0165-0327/& 2016 Elsevier B.V. All rights reserved.
106 M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118
severe bipolar cases (Ghaemi and Barroilhet, 2015), essentially due
to differential emphasis placed on similarities rather than differ-
ences between the twos (Agius et al., 2012; Ghaemi et al., 2014;
Vieta and Suppes, 2008; Zimmerman and Morgan, 2013).
A bio-psychosocial approach promoting an explanatory psy-
chological effect of a biological (cyclothymic) temperament in the
understanding of the controversies surrounding BD-BPD nosolo-
gical dilemmas has not been unanimously accepted, tough being
ontologically and clinically suggestive (Khalili, 2014).
Moreover, BPD and BD share substantial overlap in the noso-
logical validator of mood lability, especially for currently de-
pressed BD Type-II (BD-II) cases (Henry et al., 2001), “soft bipolar”
atypical forms of depressions sharing a common cyclothymic
temperament diathesis (Perugi et al., 2011, 2003) and “ultra-rapid”
(Mackinnon and Pies, 2006), “stably instable” bipolar cases
(Akiskal, 1994). Mood lability is a Diagnostic and Statistical Manual
for Mental Disorders, Fifth Edition (DSM-5) (APA, 2013) criterion
for BPD, but not BD, though being common also in this latter
(Goodwin and Jamison, 2007). DSM-defined atypical or manic
features are infrequent in BPD compared to BD, though high rates
of mixed features have been documented by large-sampled cross-
sectional studies on major depressive episode patients (either
DSM-defined “unipolar” or “bipolar” cases) with both BD and BPD
according to permissive definitions (Allen et al., 2012; Perugi et al.,
2013, 2015; Young et al., 2012).
Similarly, the symptom of impulsivity is closely allied to mood
lability, and it is often seen as manifesting as sexual impulsivity in
both BPD and BD, although it can also be physical, aggressive, fi-
nancial (Ghaemi et al., 2014) or binge eating-related (Nagata et al.,
2013; Perugi and Akiskal, 2002b). The affective lability of BPD vs.
BD-II/cyclothymic patients nonetheless shows differential fre-
quency and intensity patterns using both self-report and clinician-
administered measures (Reich et al., 2012; Swann et al., 2013). BD
and BPD differ notably on a number of diagnostic validators,
especially the course of illness of past sexual abuse (Bayes et al.,
2015; Briere and Elliott, 2003; Conus et al., 2010; Fossati et al.,
1999; Maniglio, 2014) and history of para-suicidal self-harm (Joyce
et al., 2010; Nock and Kessler, 2006). Genetic validators, treatment
response, and neurobiological differences are also consistent be-
tween the twos (Ghaemi et al., 2014).
Unsurprisingly, stating the controversy surrounding the re-
lationship between BD and BPD, the most studied question con-
cerns their actual diagnostic concordance, not only for in-
dependent samples but also for comorbid BD-BPD cases (Zim-
merman and Morgan, 2013). Across studies, approximately 10% of
BPD patients had BD-I, an additional 10% had BD-II. Likewise, ap-
proximately 20% of BD-II patients were diagnosed with BPD
(Zimmerman and Morgan, 2013). While the comorbidity rates are
substantial, each disorder is nonetheless usually diagnosed in the
absence of the other across the studies, whereas studies directly
comparing patients with BPD to BD cases (or BPD to BD) found
significant differences over a broad-range of validators, actually
challenging the notion of BPD being part of the broad bipolar
spectrum (Zimmerman and Morgan, 2013). On the other side,
while the “pragmatic approach” of the DSM-IV (and DSM-5) aims
at reducing the rates of over-diagnosis, the “strict” validity of some
diagnostic categories as BD and BPD has been questioned (Stein
et al., 2010), meaning that it cannot be granted that BD and BPD
necessarily represent clear-cut distinct diagnostic entities. In this
view, the absence of any DSM guidance soliciting the assessment
of BPD in BD or the opposite could ultimately lead to under-
estimation of comorbidity rates between the twos.
To our knowledge, no meta-analysis has specifically in-
vestigated BPD and BD comorbidity and predictors. Adopting a
meta-analytic approach would therefore provide the advantage of
pooling data from numerous studies in a logical manner towards a
more accurate effect size which is closer to the true prevalence
than when individual studies are considered separately (Ioannidis,
2009).
In contrast to previous meta-analytic reports assessing a broad
range of mood disorders comorbid with varying personality dis-
orders (Friborg et al., 2014), the present systematic review and
meta-analysis, first of its kind to best of our knowledge, rather
focuses on the prevalence and predictors of comorbid BPD2BD in
adults.
2. Materials and methods
The present meta-analysis was conducted according to the
Meta-analysis Of Observational Studies in Epidemiology (MOOSE)
guidelines (Stroup et al., 2000), and the Preferred Reporting Items
for Systematic Reviews and Meta-Analyses (PRISMA) guidelines
(Liberati et al., 2009).
2.1. Information sources and search strategy
Two independent authors searched MEDLINE, Scopus, Embase,
PsycINFO and the Cochrane Library. The search strategy combined
free text terms and exploded MESH headings for the topics of bi-
polar disorder and borderline personality disorder as following:
(((((((Bipolar disorder) OR BD) OR Bipolar) OR Manic depressive
disorder) OR Manic depressive) OR Manic)) AND (((Borderline
personality disorder) OR Borderline) OR BPD). This latter MEDLINE
strategy was then adapted for use in the other databases (Ap-
pendix A). Studies published in English through November 4, 2015
were included. We further assessed the reference listing of re-
trieved relevant articles for potential inclusion of additional
contributes.
2.2. Inclusion criteria
2.2.1. Study population and study design
We considered studies that included comorbid cases of BD and
BPD providing accurate diagnostic definitions based on either the
DSM or the International Classification of Diseases ICD (any edition
or text revision). Accounted BD populations at study included ei-
ther BD-I, BD-II and/or BD-Not Otherwise Specified (BD-NOS)
cases. Participants of both sexes, 18 years of age or older were
considered.
Both population-based and hospital-based studies were in-
cluded. Among hospital-based studies, inpatients, day-hospital
and outpatient subjects were included; emergency care records
excluded as considered non-representative. All experimental and
observational study designs were included apart from case re-
ports, opinion articles/letters to the Editor or conference pro-
ceedings or reviews.
2.2.2. Outcome measures
Primary outcomes were (i) lifetime prevalence of comorbid
BPD in BD patients and (ii) lifetime prevalence of comorbid BD in
BPD patients.
2.2.3. Study selection and data extraction
Identified studies were independently reviewed for eligibility
by three authors (MF, LO, SM) in a two-step based process; a first
screening was performed based on title and abstract while full
texts were retrieved for the second screening. Disagreements by
reviewers were resolved by consensus at both stages. Data were
extracted by two authors (BS, MF) and supervised by six additional
authors (AV, LG, GP, DDB, MS and NV) using a purpose built data
extraction spreadsheet. The data extraction spreadsheet was
 M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118
piloted on 10 randomly selected papers and modified accordingly.
Both auto- and hand-searches for “type-I” (“duplicates among/
across different databases”) and “type-II” (“duplicate publications
in different Journals/issues”) (Qi et al., 2013) were performed
based on Thompson Endnote X7s for Microsoft Windowss.
Specifically, the recorded variables included the followings:
author(s), year publication, study design, sample size, eventual
follow-up or control group, socio-demographic status, concurrent
psychotherapy (any type) or history of drug and/or physical
treatment, outcome measures, conclusions, limitations, quality
score, and quality differentiation. Whenever documented and re-
liably defined and/or most likely ascertainable by included original
sources, accounted moderators focused on course of illness and
pivotal variables reported in the literature with regard to the BD
and BPD controversial boundaries, namely mood lability, im-
pulsivity, history of childhood sexual abuse and/or lifetime phy-
sical self-harm.
2.2.4. Quality assessment
Two authors (LO and SM) independently assessed the quality of
selected studies using the checklist developed by Downs and Black
both for randomized and non- randomized studies (Downs and
Black, 1998), as reported in Tables 1 and 2 for those studies in-
cluded in our systematic review. Eventual disagreements by re-
viewers were resolved by consensus.
2.2.5. Meta-analysis
We pooled individual study data using DerSimonian-Laird
proportion method (DerSimonian and Laird, 1986) with Compre-
hensive Meta-Analysiss software (version 3). Due to anticipated
heterogeneity, a random effects meta-analysis was employed.
First, we calculated the prevalence of BPD comorbidity in BD
participants. Second, we calculated the prevalence of BD partici-
pants in people with BPD. For both meta-analyses, when three or
more studies were available, we conducted sub group meta-ana-
lyses investigating BPD-BD comorbidity according to geographical
region, study setting (inpatient versus outpatient/community),
and BD classification (BD-I versus BD-II and mixed). We also an-
ticipated on conducting separate pooled prevalence of BD ac-
cording to gender and classification of BD itself. We assessed
publication bias with the visual inspection of a funnel plot (Hig-
gins and Green, 2011) and the Begg (Begg and Mazumdar, 1994)
and Egger (Egger et al., 1997) tests. In addition, for the main pre-
valence analysis we conducted a trim and fill adjusted analysis to
remove the most extreme small studies from the positive side of
the funnel plot, re-computing the effect size at each iteration, until
the funnel plot is symmetric about the (new) effect size (Duval and
Tweedie, 2000). Finally, we conducted several meta-regression
analyses (if N Z3) to investigate potential moderators (age, per-
centage males and study population, as well as the pivotal differ-
ential features expected to differentiate between BD and BPD)
with Comprehensive Meta-Analysiss (version 3).
3. Results
One thousand four hundred twenty-four potential contributes
were identified from searching the selected databases and listing
references of relevant articles. After removing duplicates, 453 ar-
ticles were retrieved. Studies were screened and selected on the
basis of pre-specified inclusion and exclusion criteria (Fig. 1).
Forty-two articles were ultimately included in the systematic re-
view: 28 articles about comorbid BPD in BD and 14 articles about
comorbid BD in BPD. Out the included 42 studies, only 3 (Alnaes
and Torgersen, 1988; Comtois et al., 1999; Zimmerman and Mattia,
1999) concurrently assessed the prevalence rates and clinical
107
features associated to either of BD in BPD cases or the opposite
(please refer to Tables 1 and 2 respectively for additional details
and referencing).
3.1. Comorbid BPD in BD
3.1.1. Study characteristics
The characteristics of included studies are reported in Table 1.
Eight studies were cross-sectional studies, 7 were case-control,
7 prospective cohort studies, 4 cohort studies and 2 clinical trials.
Two (7.14%) studies were population-based while 26 (92.86%)
were hospital-based ones. A total of 5273 BD patients were re-
presented among the 28 included studies. The mean age of BD
patients was 39.947 11.78 years and 44% were males. As detailed
in Table 1, twelve studies were carried in Europe, 11 in North
America, 3 in South America and 2 (Perugi et al., 2013, 2015) across
Europe, Asia and Africa.
3.1.2. Quality of results
The average score of the studies assessing BPD comorbidity in
BD (n¼ 28) was 23.43 out of 31 (range¼ 11–25). Please refer to
Table 1 for details about the given studies.
3.1.3. Meta-analysis of the Pooled prevalence of BPD in BD patients
The pooled prevalence of BPD comorbidity among 5273 people
with BD over 28 studies was 18.6% (95% CI 14.44–23.71, I2 ¼91%)
(Fig. 2). Whilst the Egger test did not indicate any publication bias
( 0.88, p ¼0.33) the trim and fil analysis adjusted five studies and
calculated a new prevalence of BPD comorbidity at 21.6% (95% CI
17.0–27.1).
3.2. Sub group meta-analyses
All sub group analyses results investigating BPD and BD co-
morbidity are presented in Table 3 but will briefly be explored.
3.2.1. BPD prevalence according to the BD type
BPD prevalence among three studies conducted in BD-I pa-
tients was 12.5% (95% CI 6.97–21.27, n¼422), and BPD prevalence
was 27.3% in BD-II participants (95% CI 16.6–41.3%, n ¼377). The
prevalence of BPD among studies conducted in multi-diagnostic
subsets of BD patients (BD-I, BD-II, BD-NOS) was 18.5% (95% ci
13.1–25.3%, n ¼4474). The between group difference in prevalence
was not significant (p ¼0.1). Whilst none of the pooled prevalence
demonstrated any evidence of publication bias (see Table 2 for
Egger test and p values), the prevalence of BPD comorbidity in-
creased in each BD type after the adjustment for publication.
Notably, the prevalence of BPD comorbidity was 37.7% (95% CI
21.9–56.6) in BD-II participants.
3.2.2. Geographical variations in the prevalence of BPD
The prevalence of BPD was 16.5% (95% CI 10.91–24.29, studies
12, n ¼880) in Europe and 22.1% (95% CI 14.54–32.04, studies 11,
n¼2059) in North America (see Table 3). The adjustment for
publication bias in the trim and fill analysis resulted in the pre-
valence of BPD increasing to 26.2% (95% CI 18.7–35.3) in North
America BD participant studies.
3.2.3. BPD prevalence according to the study setting
No between group differences were observed in the prevalence
of BPD in BD according to study setting (see Table 3), although the
highest prevalence was observed in outpatient settings (20.8%,
95% CI 15.2–27.8, studies 20, n ¼3067).
3.2.4. Predictors of the prevalence of BPD in BD patients
Meta regression analysis demonstrated that a higher
 108
Table 1
Studies assessing BPD comorbidity in BD samples according to our inclusion/exclusion criteria for systematic review.

Reference Study design Country Study population Sample size of comorbid BD-BPD cases Diagnosis assessment Quality

Population-based studies: BPD comorbidity in BD samples

McDermid et al. Longitudinal population- Canada Non-institutionalized civilian population (multi-diag- 360 Out of 1172 BD-I (9%) DSM-IV 25/31
(2015) based study nostics sample) of 50 United States (1172 BD-I; 428 BD- 101 Out of 428 BD-II (24%). total BD cases of
II). Total BD cases, n¼ 1600. comorbid BPD, n ¼461 (28.8%).
Barbato and Hafner Population study Australia 42 BD-I patients in remission coming from the Noar- 6 (14.29%) DSM-IV; IPDE; BSI; patients’ 11/31
(1998) lunga Community Health Service questionnaire

Hospital-based studies: BPD comorbidity in BD samples

Joyce et al. (2004) Pharmacological clinical New Zealand Post-hoc analysis of two studies: 195 currently de- 6 Out of 19 (31.6%) and 13 out 22 (59%). DSM-IV 26/31
trial pressed patients (of whom, 19 BD-II) and 135 bulimic Total ¼ 19 out of 41 (46.3%)
females (of whom, 22¼ BD-II). Total subset n ¼41 de-
pressed BD-II cases)
Vieta et al. (2001) Cohort study Spain 129 Consecutive remitted BD-I outpatients 8 Out of 129 (6.2%) DSM-III; SCID-I; SCID-II; YMRS; HAM- 26/31

M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118

D; SADS-C
Benazzi (2002) Cross-sectional study Italy 78 BD-II patients 9 Out of 78 BD-II (11.5%) DSM-IV 25/31
Perugi et al. (2013) Cross-sectional study Europe, Asia, 5632 hospital-based OR community psychiatry-based 256 Out of 1761 BD (14.5%) DSM-IV; MINI; GAF; HCL-32 24/31
North Africa patients with current major depressive episode
(BD ¼1761, no clear-cut distinction between BD-I and
BD-II)
Comtois et al. (1999) Case-control study USA 278 outpatients with Axis I Disorders (BD, n ¼ 34) 8 Out of 34 BD cases (23.5%) DSM-III-R; SCID-I; SCID-II; HAM-D; 24/31
BDI; STAI
Preston et al.(2004) Cohort study USA 35 BD-I patients who participated in two multi-center 14 Out of 35 (40%) DSM-IV; SCID-I, SCID-II; WURS; MRS; 24/31
studies of lamotrigine, were retrospectively evaluated HAM-D; GAF
for the incidence of BPD
Vieta et al. (2000) Case-control study Spain 40 BD-II outpatients recruited from primary care psy- 5 Out of 40 (12.5%) DSM-III; SCID-II; SADS-C 24/31
chiatric setting
Perugi et al. (2015) Cross-sectional study Europe, Asia, 2811 Hospital-based or community psychiatry-based BD-I subset with comorbid BPD, n¼ 42 (9% of DSM-IV; DSM-5; GAF 23/31
North Africa patients with a current major depressive episode. DSM- the BD total) and BD-II, n ¼17 (3.7%)
IV-define BD cases, n ¼ 464. Of whom, BD-I¼ 288; BD-
II ¼176.
Wilson et al. (2007) Case-control study USA 173 Outpatients admitted to a multisite project on mood 15 Out of 30 (50%) DSM-IV; BDI; HAM-D; BIS-11; B-DHI 23/31
disorders and suicidal behavior (143 MDD; 30 BD-II)
Garno et al. (2005) Cross-sectional study USA 100 BD patients (73 BDI-I; 27 BD-II), of whom, 95 out- 17 Out of 100 (17%) DSM-IV-TR; SCID-I; SCID-II; HAM-D; 22/31
patients; 5 inpatients. CTQ; YMRS; schedule for affective
disorders and schizophrenia
Rossi et al. (2001) Prospective study Italy 188 Patients consecutively admitted to a psychiatric re- 21 Out of 71 (29.6%) DSM-III; SCID-P; SCID-II-PQ; HAM-D 22/31
search unit for a major depressive index episode (71 BD;
117 MDD)
George et al. (2003) Psychotherapy clinical USA 52 Remitted BD-I patients 2 Out of 52 (3.8%) DSM-III; PDE; SADS-C 22/31
trial
Dunayevich et al. Prospective study USA 59 Manic or mixed inpatients followed during a 12- 3 Out of 59 (5.4%) DSM-III-R; YMRS; HAM-D 21/31
(2000) month course of illness
Gasperini et al. Cohort study Italy 213 Lithium treated outpatients (100 MDD and 113 BD, 3 Out of 54 (5.5%) DSM-III; HAM-D; MRS; SIDP 21/31
(1993) unspecified ratio of BD-I or BD-II cases, of whom only 54
were assessed for the presence of BPD)
Loftus (2006) Cohort study USA 51 Remitted BD-I patients, of whom, 4 inpatients; 47 10 Out of 51 (19.6%) DSM-IV-TR; SCID.I; SCID-II; HAM-D; 21/31
outpatients CARS-Mania; MSIF
Perugi et al. (2011) Prospective study Italy 107 Consecutive outpatients with major depressive 12 Out of 25 BD-II cases (48%), none among the DSM-IV-TR and modified criteria for 21/31
episode with atypical features. BD-II, n ¼ 25 and BD- BD-I subset. overall, the subset of “bipolar the diagnosis of BD; SCID-I; SCID-II;
I¼ 46. Note: the diagnostic codes for BD encompassed spectrum” cases with comorbid BPD included 39 HAM-D; ADDS; SID; HSCL-90
non-DSM-IV-TR defined cases too. Overall, “bipolar patients (47%)
spectrum” cases comprised 83 patients.
Zimmerman and Cross-sectional study USA 409 Outpatients recruited by Hospital Department Of 5 Out 8 BD-I (62.5%); 5 out of 15 BD-II (33.3%) DSM-IV; SCID-I; SCID-II 21/31
Mattia (1999) Psychiatry BD cases, n ¼ 28, of whom, BD-I ¼8, BD-II ¼15 and 2 out 5 (40%) BD-NOS
 and 5 additional BD-NOS.
Pica et al. (1990) Case-control study Australia 26 recent-onset BD patients (unspecified ratio of sub- 3 Out of 26 (11.5%) DSM-III; SIDP; SCID-P; RPMIP; BDI; 19/31
types) compared with 35 recent-onset schizophrenic BRMS; SAPS; SANS
patients
Benazzi (2008) Prospective study Italy 138 Consecutive remitted BD-II outpatients from private 63 Out of 138 (45.9%) DSM-IV; SCID-I; SCID-II; GAF 19/31
practice
Carpiniello et al. Cross-sectional study Italy 57 Outpatients from university community Mental 18 Out of 57 (31.6%) DSM-IV-TR; BIS; AQ; CGI; GAF; 19/31
(2011) Health Center: 29 BDI-I and 28 BD-II MCMI; MMPI
Brieger et al. (2003) Case-control study Germany 177 Patients: 117 MDD and 60 BD (of whom, n ¼ 44 BD-I 4 Out of 60 BDs (6.7%) DSM-III-R; FFM-NEO; BR-MAS; CDRS; 17/31
manic or mixed cases) reviewed for any DSM-IV-defined MWT; MMSE
personality disorder at time of hospital admission
Peselow et al. (1995) Prospective study USA 66 Outpatients who had a lifetime diagnosis of BD and 11 Out of 47 (23.4%) DSM-III; SCID-I; SCID-II; SADS-C; 16/31
met minimum Research Diagnostic Criteria for hypo- IMPS; CGI
mania (47 BD who successfully recovered from the hy-
pomanic episode)
Ucok et al. (1998) Case-control study Turkey 90 BD-I outpatients compared with 58 control subjects 9 Out of 90 (10%) DSM-III; SCID-I; SCID-II 13/31
Benazzi (2000) Case-control study Italy 113 Patients: 63 unipolar MDE (MDD and dysthymic 6 Out of 50 BDs (12%) DSM-IV; SCID-I; SCID-II; MADRS; GAF 9/31
disorders) and 50 BD-II MDE recruited in a private
practice

M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118

Carpenter et al. Cross-sectional study USA 23 BD-I patients None. DSM-III-R; PDE; BPRS; SAS–SR; GAS 9/31
(1995)
Alnaes and Torger- Cross-sectional study Norway 289 Consecutive outpatients with Axis I assessed for None. DSM-III 9/31
sen (1988) Axis-II comorbidity. BD total, n ¼30, of whom n¼ 19 BD-I
cases.

BD: bipolar disorder; BD-I: bipolar disorder type I; BD-II: bipolar disorder type II; BPD ¼Borderline Personality Disorder; MDE: major depressive episode; MDD: major depressive disorder; NOS: not otherwise specified; DSM:
Diagnostic and Statistical Manual of Mental Disorders; IPDE: International Personality Disorder Examination; BSI: Brief Symptom Inventory; FFM-NEO: Five Factor Inventory-NEO; BR-MAS: Bech-Rafaelsen Mania Scale; CDRS:
Cornell Dysthymia Rating Scale; MWT: Multiple Choice Vocabulary Test; MMSE: Mini Mental State Examination; PDE: Personality Disorder Examination; BPRS: Brief Psychiatric Rating Scales; SAS-SR: Social Adjustment Scale-Self
Report Version; GAS: Global Assessment Scale; BIS-11: Barratt Impulsiveness Scale; AQ: Aggressiveness Questionnaire; MCMI: Millon Clinical Multiaxial Inventory-III; MMPI-II: Multiphasic Personality Inventory-II; YMRS: Young
Mania Rating Scale; CTQ: Childhood Trauma Questionnaire; MRS: Mania Rating Scale; PDE: Personality Disorder Examination; SADS-C: Schedule for Affective Disorders and Schizophrenia-Change Version; CARS Mania: Clinician-
Administered Rating Scale for Mania; MSIF: Multidimensional Scale for Independent Functioning; SID: Semi-structured Interview for Depression; ADDS: Atypical Depression Diagnostic Scale; HSCL-90: Hopkins Symptoms Check
List; IMPS: Inpatient Multidimensional Psychopathology Scale; SCID-P: Structured Clinical Interview for DSM-III-R; RPMIP: Royal Park Multi-diagnostic Instrument for Psychosis; BDI: Beck Depression Inventory; BRMS: Bech-
Rafaelsen Mania Scale; SAPS: Scale for the Assessment of Positive Symptoms; SANS: Scale for the Assessment of Negative Symptoms; WURS: Wender Utah Rating Scale; GAF: Global Assessment of Functioning; B-DHI: Buss-Durkee
Hostility Inventory; MINI: Mini International Neuropsychiatric Interview; HCL-32: Hypomania Checklist for self-assessment.

109
 110
Table 2
Studies assessing BD comorbidity in BPD samples according to our inclusion/exclusion criteria for systematic review.

Reference Study design Country Study population Sample size of comorbid BD-BPD cases Diagnosis Quality
assessment

Hospital-based studies: BD comorbidity in BPD samples

M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118

Comtois et al. (1999) Cross-sectional study USA 278 Patients (38 BPD, 108 other personality 8 (No clear-cut stratification of BD types) SCID; DSM-III-R 24/31
disorders, 132 no personality disorders)
Zimmerman and Mattia Cross-sectional study USA 409 Outpatients recruited by Hospital de- 5 Out of 8 BD-I total (62.5% OR 8.47% of BPD sample); 5 out of 15 BD-II total DSM-IV; SCID-I; 21/31
(1999) partment of psychiatry BPD cases, n¼ 59. (33.3% OR 8.47% of BPD sample) and 12 out 28 BD-NOS (40% OR 20.34% of SCID-II
BPD sample)
Zanarini et al. (1998) Cross-sectional study USA 504 Patients (379 BPD, 125 other personality 36 BD-II out of 38 total (9.5% of BPD sample) SCID; DSM-III-R 25/31
disorders)
Skodol et al. (1999) Cross-sectional study USA 571 Patients (240 BPD) of whose 60 BD (45 32 Out of 240 (13.3%): 22 BD-I (9.17%); 10 BD-II (4.17%) SCID; DSM-IV 26/31
BD-I; 15 BD-II)
Akiskal et al. (1985) Prospective longitudinal USA 100 BPD patients 17 Out of 100 (17%) DSM-III 23/31
follow-up
Deltito et al. (2001) Cross-sectional study USA 16 BPD patients 10 Out of 16 (62.5%) SCID; DSM-III 21/31
Hudziak et al. (1996) Cross-sectional study USA, Italy 87 BPD patients 14 Out of 87 (16.09%) DSM-III-R 22/31
Pope et al. (1983) Prospective longitudinal USA 39 Patients (33 BPD) 3 Out of (9.09%) DSM-III-R 24/31
follow-up
Zanarini et al. (2004) Prospective longitudinal USA 362 Patients (290 BPD, 72 other personality 15 Out of 290 (5.17%) SCID; DSM-III-R 24/31
follow-up disorders)
Gunderson et al. (2006) Prospective longitudinal USA 629 Patients (196 BPD, 433 other personality 38 Out of 196 (19.39%): 23 BD-I (11.73%) and 15 BD-II (7.65%) DSM-IV 24/31
follow-up disorders)
Gunderson et al. (2014) Prospective longitudinal USA 223 BPD patients (161 MDD, 34 BD-I, 28 BD- 62 Out of 223 (31.63%): 34 BD-I (15.25%); 28 BD-II (12.6%) SCID; DSM-IV 24/31
follow-up II)
Links et al. (1988) Prospective longitudinal Canada 88 BPD patients None DSM-III 10/31
follow-up
Alnaes and Torgersen Cross-sectional study Norway 289 Consecutive outpatients with Axis I as- None DSM-III 9/31
(1988) sessed for Axis-II comorbidity. BPD ¼44.
Prasad et al. (1990) Cross-sectional study USA 21 BPD patients 5 Out of 21 (23.8%): 4 BD-I (16.7%); 1 BD-II (4.8%) DSM-III 9/31

BD: bipolar disorder; BD-I: bipolar disorder type I; BD-II: bipolar disorder type II; BPD ¼Borderline Personality Disorder; MDE: major depressive episode; MDD: major depressive disorder; NOS: not otherwise specified; DSM:
Diagnostic and Statistical Manual of Mental Disorders (Third [DSM-III], Third-Revised [III-R], Fourth [DSM-IV], Fourth, Text-Revised [DSM-IV-TR] or Fifth [DSM-5] editions). SCID: Structured Clinical Interview for Mental Disorders.
 M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118 111

Records identified through

Identification
database searching Additional records identified
(N=1424, of which 718 though through other sources (n=8
PUBMED; Scopus=277; edited books or manual search)
Embase=276; PsycINFO=147;
Cochrane library=6)

Records after duplicates removed

(n=453)
Screening

Thirty-three
records
Potential relevant
excluded
citations (n=85)
after screening
of abstract and
title
Eligibility

Ten articles were excluded, as

they did not fulfill the
Full-text articles assessed inclusion and exclusion
for eligibility criteria. These included eight
(n=52) articles that analyzed the
prevalence of BPD in BD
without specify the rates for
BPD or BD, and two articles
Included

described studies performed

in specific populations less
likely to be representative of
the average BD-BPD patient.
Studies included in
qualitative synthesis
(n=42)

Fig. 1. Adapted PRISMA 2009 flow diagram.

percentage of males predicted a lower prevalence of comorbid BPD
in BD participants (β ¼ 0.0581, 95% CI 0.0848 to  0.0315,
p o0.001, studies ¼20) and explained over half of the between
study heterogeneity (R2 ¼0.58). A higher mean age also predicted
a lower prevalence of BPD comorbidity in BD participants
(β ¼  0.0782, 95% CI  0.1569 to  0.0006, p ¼0.04, R2 ¼0.15,
studies ¼20). The year of study publication was not related to the
prevalence of BPD across the studies (β ¼  0.0007, 95% CI  0.0017
to 0.0003, p ¼ 0.19, R2 ¼0.07, studies ¼28).
3.3. Comorbid BD in BPD
3.3.1. Study characteristics
The characteristics of included studies are reported in Table 2.
Eight studies were cross-sectional studies and 6 prospective co-
hort studies. All studies were hospital-based ones. A total of 1814
BPD patients were represented among the 14 included studies. The
mean age of BPD patients was 32.22 77.35 years and 21.5% were
males. As detailed in Table 2, one study was conducted in Europe,
12 in North America, and 1 (Gunderson et al., 2014) across Europe
and North America.
112 M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118

Study name Statistics for each study Event rate and 95% CI

Event Lower Upper

rate limit limit

Wilson et al 2007 0.500 0.328 0.672

Perugi et al. 2011 0.470 0.365 0.577
Joyce et al. 2004 0.463 0.319 0.615
Benazzi 2008 0.457 0.375 0.540
Zimmermann and Mattia 1999 0.429 0.262 0.613
Preston et al. 2004 0.400 0.253 0.567
Carpiniello et al 2011 0.316 0.209 0.447
Rossi et al 2001 0.296 0.201 0.411
McDermind et al 2015 0.288 0.266 0.311
Comtois et al 1999 0.235 0.122 0.405
Peselov et al. 1995 0.234 0.135 0.375
Loftus and Jaeger 2006 0.196 0.109 0.327
Garno et al 2005 0.170 0.108 0.257
Perugi et al 2013a 0.145 0.130 0.163
Barbato and Hafner 1998 0.143 0.066 0.283
Vieta et al 2000 0.125 0.053 0.267
Benazzi 2000 0.120 0.055 0.242
Pica et al. 1990 0.115 0.038 0.303
Benazzi 2002 0.115 0.061 0.207
Ucok et al 1998 0.100 0.053 0.181
Perugi et al 2015 0.091 0.068 0.120
Brieger P. et al., 2003 0.067 0.025 0.165
Vieta et al 2001 0.062 0.031 0.119
Gasperini et al 1993 0.056 0.018 0.159
Dunayevich et al 2000 0.051 0.016 0.146
George et al. 2003 0.038 0.010 0.141
Carpenter et al 1995 0.021 0.001 0.259
Alnaes and Torgersen 1988 0.016 0.001 0.211
0.186 0.144 0.237
-1.00 -0.50 0.00 0.50 1.00

Prevalence BPD
Fig. 2. Pooled prevalence of BPD in BD participants across all studies of BPD in BD patients.

Table 3
Sub-group analyses results investigating BPD comorbidity in BD.

Analysis Number of study Meta-analysis Heterogeneity Publication bias

estimates
Prevalence BPD 95% CI Between group p I2 Egger test (p Trim and fill (95% CI) [adjusted
value value) studies]

Main analysis 28 18.6% 14.44 23.71 91%  0.88, p ¼ 0.33 21.6% (17.0–27.1) (Alnaes and
Torgersen, 1988)
BD type 0.10
BD I 7 12.46% 6.97 21.27 80  2.7, p¼ 0.23 13.1% (6.7–24.1) (Agius et al.,
2012)
BD II 6 27.26% 16.66 41.28 89  5.1, p ¼0.12 37.7% (21.9–56.6) (Akiskal, 1994)
BD mixed 15 18.49% 13.12 25.42 90  0.95, p ¼ 0.51 20.6% (15.1–27.5) (Akiskal, 1994)

Geographical region 0.49

Europe 12 16.55% 10.91 24.29 90  0.39, p ¼ 0.08 16.8% (9.9–27.2) (Agius et al.,
2012)
North America 11 22.07% 14.54 32.04 78  0.23, p ¼ 0.35 26.2% (18.7–35.3) (Akiskal, 1994)
Oceania 3 22.84% 10.01 44.06 89  7.7, p¼ 0.28 Unchanged
Various 2 11.56% 4.44 26.89 88 N/A N/A

Setting 0.49
Inpatient 4 11.83% 5.22 24.65 82  5.21, p¼ 0.68 Unchanged
Outpatient 20 20.79% 15.16 27.83 90  2.1, p ¼0.21 22.1% (16.1–29.5) (Akiskal, 1994)
Mixed 2 18.21% 6.65 41.05 0 N/A N/A

Key: N/A¼ not applicable, BD¼ bipolar disorder, BPD¼ borderline personality disorder.

3.3.2. Quality of results 3.3.3. Meta-analysis of the pooled prevalence of BD comorbidity in

The average score of the studies assessing BD comorbidity in BPD patients
BPD (n¼ 14) was 27.57 out of 31 (range¼ 9–24). Please refer to The pooled prevalence of BD among 1814 people with BPD was
Table 2 for additional details. 16.58% (95% CI 11.38–23.53, I2 ¼87%, Studies 14) (Fig. 3). Whilst no
 M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118 113

Study name Event rate and 95% CI

Event Lower Upper
rate limit limit
Delito J. Et al. 2001 0.625 0.377 0.821
Zimmerman MC. and Mattia JI. 1999 0.373 0.260 0.502
Gunderson JG. Et al. 2014 0.278 0.223 0.340
Prasad RB et al. 1990 0.238 0.103 0.460
Comtois KA. Et al. 1999 0.211 0.109 0.368
Gunderson JG. Et al. 2006 0.194 0.144 0.255
Akiskal HS. Et al. 1985 0.170 0.108 0.257
Hudziak JJ. Et al. 1996 0.161 0.098 0.254
Skodol AE. Et al. 1999 0.133 0.096 0.182
Zanarini MC et al. 1998 0.095 0.069 0.129
Pope HG. Et al. 1983 0.091 0.030 0.247
Zanarini MC et al. 2004 0.052 0.031 0.084
Alnaes R. et al. 1988 0.011 0.001 0.154
Links PS. And al. 1988 0.006 0.000 0.083
0.166 0.114 0.235
-1.00 -0.50 0.00 0.50 1.00
Prevalence BD
Fig. 3. Pooled prevalence of BD in BPD participants across all studies of BPD in BD patients.

publication bias was evident (Egger¼  0.85, p¼ 0.61), the trim
and fill analysis adjusted for publication bias demonstrated the
prevalence of BD in BPD was slightly higher at 18.5% (95% CI 12.7–
26.1).
3.4. Sub group meta-analyses
Full details of all sub group analyses are presented in Table 4
but will briefly be summarized.
and BD II (12.65%, 95% CI 4.79–29.47), although the between group
differences were not significantly different (p ¼0.24).
3.4.2. Geographical variations in the prevalence of BD
The prevalence of BD across 12 studies in North America was
17.50% (95% CI 11.70–25.36) which slightly increased after the
adjustment for publication bias (18.8%, 95% CI 12.5–27.1). Only one
study was conducted in Europe and another study recruited par-
ticipants from North America and Europe (see Table 4).

3.4.1. Prevalence of BD according to different BD type 3.4.3. Prevalence of BD according to setting of the study
BD mixed was evident in 19.89% (95% CI 12.23–30.67) of people Nine studies reported the prevalence of BD in BPD participants
with BPD, which was higher than BD I (15.30%, 95% CI 6.47–32.06) in outpatient settings with the pooled prevalence at 19.34% (95% CI

Table 4
Sub-group analyses results investigating BD comorbidity in BPD.

Analysis Number of study Meta-analysis Heterogeneity Publication bias

estimates
Prevalence BD 95% CI Between group p I2 Egger test (p Trim and fill (95% CI) [ad-
value value) justed studies]

Main analysis 14 16.58% 11.38 23.53 87  0.85, p¼ 0.61 18.5% (12.7–26.1) (Akiskal,
1994)
BD type 0.24
BD I 3 15.30% 6.47 32.06 0  1.67, p ¼ 17.9% (11.9–25.7) (Agius et al.,
2012)
BD II 2 12.65% 4.79 29.47 77 N/A N/A
BD mixed 8 19.89% 12.23 30.67 91  0.69, P ¼0.80 Unchanged

Geographical region 0.19

Europe 1 1.11% 4.87 20.56 0 N/A N/A
North Americaþ Europe 1 16.09% 3.91 47.46 0 N/A N/A
North America 12 17.50% 11.70 25.36 89  0.37, p ¼0.85 18.8% (12.5–27.1) (Agius et al.,
2012)

Setting 0.65
Inpatient 1 9.50% 2.23 32.61 0 N/A N/A
Outpatient 9 19.34% 11.96 29.72 88  0.53, p¼ 0.86 Unchanged
Mixed 2 14.60% 5.36 34.08 0 N/A N/A
Unclear 2 10.46% 2.40 35.76 80 N/A N/A

Key: N/A ¼not applicable, BD¼ bipolar disorder, BPD¼ borderline personality disorder, BD ¼bipolar disorder.
114 M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118
11.96–29.72). No significant differences were observed in other
settings (see Table 4).
3.4.4. Predictors of the prevalence of BD in BPD patients
Meta regression analysis demonstrated mean age moderated a
higher prevalence of BD in BPD participants (β ¼0.292, 95% CI
0.2019–0.382, p o0.001, R2 ¼1.00) although only restricted to four
studies. The year the study was published (β ¼0.0394, 95% CI
0.0178 to 0.0966, p ¼0.17, R2 ¼ 0.05, studies ¼14) and the per-
centage of males (β ¼  0.0136, 95% CI  0.083 to 0.0558, p¼ 0.77,
R2 ¼0.01, studies¼ 5).
4. Discussion
The current systematic review established that comorbid BPD
among people with BD and BD among people with BPD is com-
mon. Specifically, in the first meta-analysis of its kind we found
that after the adjustment for potential outliers the prevalence of
BPD among 5273 people with BD was 21.6% (95% CI 17.0–27.1). We
established that 18.5% of people with BPD have a comorbid BD
diagnosis. These prevalence rates are broadly concordant with a
previous narrative systematic-review which found that approxi-
mately 10% of patients with BPD had BD-I and another 10% had
BD-II (roughly, a total of about 20% of patients with BPD were
diagnosed with comorbid BD) (Zimmerman and Morgan, 2013).
Likewise, this is also concordant with the previous qualitative
synthesis documenting approximately 20% of BD-II patients diag-
nosed with comorbid BPD, though only 10% of BD-I patients were
diagnosed with comorbid BPD (Zimmerman and Morgan, 2013).
Not only did we extend the literature by conducting the first
formal meta-analysis, but also we were able to investigate po-
tential moderators and subgroup differences in BD and BPD co-
morbidity. For instance, after the adjustment for publication bias
we found higher rates of BPD in BD-II participants (37.7%, 95% CI
21.9–56.6) and North America studies (26.2%, 95% CI 18.7–35.3).
Meta-regression analysis demonstrated that a higher percen-
tage of males and increasing age predicted a lower comorbid
prevalence of BPD in people with BD (i.e. more common in females
and older adults). This latter trend is also in line with previous
evidence pointing out higher rates of female patients, and younger
mean age of onset, in those comorbid cases of either BD in BPD or
BPD in BD vs. BD samples without comorbid BPD (Barbato and
Hafner, 1998; McDermid et al., 2015; Sajatovic et al., 2006), despite
the only meta-analytic study assessing the prevalence rates of
personality disorders in mood disorder samples failed to provide
sufficient information on the matter with respect to the BD2BPD
comorbidity to allow reliable comparisons (Friborg et al., 2014).
4.1. Critical evaluation of the synthesis and implications for the
clinical practice and future lines of research
“Comorbidity” is a term originally introduced to refer to the
coexistence of two essentially independent and distinct disorders
(Amerio et al., 2015; Fountoulakis, 2014; Maj, 2005).
According to this early concept, there exists an “index” or
“primary” disorder and a comorbid separate second disorder
which potentially affects the selection of treatment and the
prognosis of the index one (Feinstein, 1970).
In Feinstein's formulation, the implication was that a com-
pletely different and independent disease occurred at the same
time as another disease. These two diseases co-occurred, more
often than not, randomly (Amerio et al., 2015). On the contrary, the
DSM explicitly produces overlapping clinical criteria for many di-
agnoses, allowing comorbidity in quite a different sense than in
the medical meaning of the term as co-occurrence of independent
diseases (Maj, 2005). Using DSM definition, it is unclear whether
concomitant diagnoses actually reflect the presence of distinct
clinical entities or refer to multiple manifestations of a single
clinical entity (Maj, 2005).
Indeed, this is a critical issue also with respect to the BD2BPD
comorbidity, which is still source of vivid debate and opposite
views about the actual overlap (Perugi et al., 2011) or distinction
(Ghaemi and Barroilhet, 2015; Ghaemi et al., 2014) existing be-
tween the twos, especially in case of Type-II bipolar depression
(Bayes et al., 2015).
Moreover, the DSM fails to provide any rule soliciting clinicians
or researchers to explore the possibility of a comorbid personality
disorder in BD, nor conversely to include or exclude BD as an ex-
planation for the emotional instability (and often overt mood
swings) seen in BPD patients (Henry et al., 2012). The distribution
of the affective-emotional lability continuum features across BD
(especially BD-II depression with atypical features and/or cy-
clothymic temperament) and BPD is not bimodal, as this is the
case of anxious-sensitivity and impulsivity – though presenting
with peculiar neuropsychological differences in BPD vs. BD (Akis-
kal, 2004; Perugi and Akiskal, 2002a) and, indeed, at least a partial
overlap or continuum between the two conditions seem to exist
(Paris et al., 2007).
This is a crucial issue to bear in mind in the interpretation of
the quantitative synthesis provided by the present meta-analytic
study, as this may actually be on the lower side. it may represent a
relevant issue for the clinical practice including general primary
care setting (Stubbs et al., 2016). as well hinder the search for
reliable and valid endophenotype, and/or genotypic vulnerability
markers for BD2BPD interface (Siever et al., 2002), as already
postulated for BD underestimation in some cases of MDD (Fornaro
et al., 2011). Among other consequences, underestimation of BD
comorbidity rates in BPD cases would lead to delayed compre-
hensive diagnosis and delayed delivery of otherwise appropriate
multi-disciplinary management. This is also compelling, con-
sidering that real world data indicate that a considerably high
number of BPD patients receive psychotropic medications, often as
part of polypharmacy regimens, including antidepressant mono-
therapy, despite the absence of any therapeutic guideline endor-
sing such practice (Bridler et al., 2015; Fornaro et al., 2012). Con-
versely, polypharmacy rates would possibly inflate even among
those BD patients whose comorbid BDP is actually underestimated
(Fornaro et al., 2016). Therefore, underestimation of BPD-BD in-
terface would lead to increased risk of exposure towards improper
or harmful medications too.
Furthermore, comparisons of BPD vs. BD samples (or the op-
posite) are almost invariably done by following a cross-sectional
rather than a longitudinal approach recommended by Kraepelin in
the assessment of manic-depressive illness (Angst and Sellaro,
2000).
These issues are likewise critical in the evaluation of the results
provided in the present study. This is with a special emphasis
toward the systematic lack of information about the course of BD
(and BPD). In fact, none of the studies included in our systematic
review and meta-analysis provide satisfactory information about a
number of otherwise clinically meaningful potential moderators.
Namely, the pooled studies did not provide information about
atypical or sub-threshold mixed features of current bipolar de-
pressive episodes (whenever this was the syndromal case), despite
their clinical relevance for a number of course predictors and
features, including overall treatment adherence (Fornaro et al.,
2013; Fornaro et al., 2014). Neither, they systematically docu-
mented the eventual rapid-cycling course of BD in their subsets.
Similarly, the pooled samples involving BPD patients failed to
provide enough quantitative information about the otherwise
clinically suggestive distinctive features of BPD vs. BD (namely,
 M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118
affective instability, rather than mood instability, or different
psychometric profiles of impulsivity, or history of childhood
trauma or self-injury behavior).
In addition, both the BD and BPD subsets documented in the
studies included in the present systematic review and meta-ana-
lysis almost invariably failed to provide information about current
or lifetime treatment response towards pharmacological and non-
pharmacological treatments, suicidal behavior, substance abuse, or
presence of (hypo-)manic features. Similarly, no sufficient data
were provided across the included studies with regard to quanti-
tative measure of predominant affective temperament, which a
special emphasis toward cyclothymia, which has nonetheless
proposed to be the underpinning between BD (especially BD-II
depression) and the “mood/affective” (actually “emotional”) core
of BPD (Perugi et al., 2011). Though, it must be remarked that BPD
is a complex construct encompassing varying psychopathological
dimensions other than the sole “mood/affective” instability or
“impulsivity” shared (yet characterized by differential psycho-
pathological and psychometric profiles) between BPD and BD, to
encompass also other clinical variables including, but not limited
to, childhood trauma, dissociative experience, and self-injury/
para-suicidal behavior (Ghaemi and Barroilhet, 2015). This is of
particular relevance considering that the aforementioned variables
should represent the translation to BPD of the classical constructs
originally developed by Robins and Guze (Robins and Guze, 1970)
for psychiatric nosological research (namely, “symptoms”, “ge-
netics”, “course”, “treatment response”) (Ghaemi et al., 2014).
Ultimately, these intrinsic limitations of the present study
hinder the actual appreciation of additional clinically sound and
neurobiological (Chen et al., 2010; Pally, 2002) features associated
to comorbid BPD and BD, thus warranting future studies to sys-
tematically assess these issues.
4.2. Study limitations
As afore mentioned, the main limitations of the present study
are essentially related to the DSM “pragmatic approach” towards
the diagnosis of BD and BPD and the elusive definition of “mental
disorder” as a whole nosological entity by either the DSM-IV or the
DSM-5 (Stein et al., 2010), and the study design and analysis
strategy of the included studies, as documented in the quality
assessment scale used.
Most studies are observational and based on retrospective as-
sessments. The use of retrospective assessment scales with low
sensitivity in discriminating self-report history of actual childhood
traumata confabulations or dissociative states may have biased
results towards an overestimation of such symptom prevalence.
Some of the studies do not include a control group. Small sample
size and enrollment of subjects mainly from BD-BPD outpatient
units may limit generalizability of these results. Potential con-
founding factors in these studies include demographic and his-
torical illness variables, which often were not appropriately stra-
tified across multivariate modeling. Moreover, in our analyses we
anticipated some extent of heterogeneity, which is nonetheless
often anticipated when conducting meta-analysis of observational
studies (Stroup et al., 2000). We therefore attempted to address
following the MOOSE guidelines by (a) stratifying our results
where possible and (b) conducting meta-regression analyses. Fi-
nally, we also encountered some evidence of publication in our
analyses but conducted Trim and fill analyses to diminish the in-
fluence of this (Duval and Tweedie, 2000).
We acknowledge that the exclusion of original studies provid-
ing side-by-side comparison of BD against BPD cases (beyond the
mere information about BD2BPD comorbidity) might have
115
hindered a direct comparison of twos in terms of similarities and
differences beyond the sole comorbid cases. Nonetheless, we
purposely followed this strategy aiming at enhancing the quality
of pooled data. In fact, most of the studies comparing BD2BPD
are qualitative rather than quantitative reports and/or based on
chart-review or post-hoc records, meaning that the subsets of the
sample(s) at study were almost invariably measured based on fully
non-comparable ratings. Specifically, with few notable recent,
rigorous exceptions testing a broad ranges of alternative defini-
tions and moderators in non-comorbid cases of either BD or BPD
(Bayes et al., 2015), BD cases are simply not systematically as-
sessed for history of childhood trauma or self-injury behavior,
otherwise relevant to BPD. Ultimately, this would have led to a
major measurement bias invariably affecting the overall validity of
quantitative comparisons between BD and BPD as critically dis-
cussed over in the text.
Nevertheless, allowing for these caveats our study is a first and
contains numerous strengths. First, the strength of the selected
studies is that the diagnosis of BD and BPD were consistently
based on the DSM criteria and were established by trained in-
vestigators using validated assessment scales mainly with inter-
rater reliability. The main strength of this review is its being sys-
tematic and its including the entire scientific evidence published
so far on the main medical databases.
5. Conclusion
In our meta-analysis, we established that (i) BD and BPD and
(ii) BPD and BD comorbidity affect approximately one in five
people. Further higher rates of BDP-BD comorbidity would none-
theless be expected in the clinical practice beyond the DSM-IV
(and DSM-5) approach, ultimately influencing the therapeutic
choices and outcomes as well as the accuracy and homogeneity of
the diagnostic samples at study for endophenotype and other
genetic investigations. Above all, BD-BPD comorbidity is common
and necessitates appropriate diagnosis and treatment in clinical
practice. Future longitudinal prospective studies are required to
better understand the diagnostic boundaries between these two
conditions. Comprehensive quantitative assessment of the over-
lapping and differential clinical moderators is also warranted in
order to better understand the actual boundaries of BPD and BD
toward the delivery of more accurate therapeutic interventions
and a better insight about the potential biomarkers and genetics
validators allowing an accurate distinction between the twos.
Contributors
MF and BS conceived the study and extensively edited the
main-text and its attachments. LO, SM, DDB, GP, AV, LG, MS and NV
assisted either in manuscript revisions, preparations of the tables
and figures or interpretation of results. Finally, all the co-authors
substantially contributed to the present piece of work before ap-
proving it for final submission.
Role of funding source
There is no funding source to disclose in conjunction with the present piece of
work.
Acknowledgments
None to state.
116 M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118

Appendix A. MEDLINE search strategy

Set Medline

1 Bipolar disorder
2 BD
3 Bipolar
4 Manic depressive disorder
5 Manic depressive
6 Manic
7 Bipolar disorder
8 Sets 1–7 were combined with “OR”
9 Borderline personality disorder
10 Borderline
11 BPD
12 Borderline personality disorder
13 Sets 9–12 were combined with “OR”
14 Sets 8 and 13 were combined with “AND”
15 Set 14 was limited to November 4, 2015, Humans, English language, adult: 19 þ years

Words written in italic were used as MeSH headings, the others were used as free text.

References

Agius, M., Lee, J., Gardner, J., Wotherspoon, D., 2012. Bipolar II disorder and bor-
derline personality Disorder-co-morbidity OR spectrum? Psychiatr. Danub. 24
(Suppl 1), S197–S201.
Akiskal, H.S., Chen, S.E., Davis, G.C., Puzantian, V.R., Kashgarian, M., Bolinger, J.M.,
1985. Borderline: an adjective in search of a noun. J. Clin. Psychiatry 46 (2),
41–48.
Akiskal, H.S., 1994. The temperamental borders of affective disorders. Acta Psy-
chiatr. Scand. Suppl. 379, 32–37.
Akiskal, H.S., 2004. Demystifying borderline personality: critique of the concept and
unorthodox reflections on its natural kinship with the bipolar spectrum. Acta
Psychiatr. Scand. 110, 401–407.
Allen, D.M., Parry, P.I., Purssey, R., Spielmans, G.I., Jureidini, J., Rosenlicht, N.Z.,
Healy, D., Feinberg, I., 2012. BRIDGE study warrants critique. Arch. Gen. Psy-
chiatry 69 (643–644), 644–645.
Alnaes, R., Torgersen, S., 1988. The relationship between DSM-III symptom dis-
orders (axis I) and personality disorders (axis II) in An outpatient population.
Acta Psychiatr. Scand. 78, 485–492.
Amerio, A., Odone, A., Tonna, M., Stubbs, B., Ghaemi, S.N., 2015. Bipolar disorder and
its comorbidities between Feinstein and the diagnostic and statistical manual
of mental disorders. Aust. N. Z. J. Psychiatry 49, 1073.
Angst, J., Sellaro, R., 2000. Historical perspectives and natural history of bipolar
disorder. Biol. Psychiatry 48, 445–457.
APA, 2013. Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition
(DSM-5). American Psychiatric Publishing, Washington.
Barbato, N., Hafner, R.J., 1998. Comorbidity of bipolar and personality disorder.
Aust. N. Z. J. Psychiatry 32, 276–280.
Barroilhet, S., Vohringer, P.A., Ghaemi, S.N., 2013. Borderline versus bipolar: dif-
ferences matter. Acta Psychiatr. Scand. 128, 385–386.
Bayes, A.J., McClure, G., Fletcher, K., Ruiz, Y., Hadzi-Pavlovic, D., Stevenson, J.L.,
Manicavasagar, V.L., Parker, G.B., 2015. Differentiating the bipolar disorders
from borderline personality disorder. Acta Psychiatrica Scandinavica.
Begg, C.B., Mazumdar, M., 1994. Operating characteristics of a rank correlation test
for publication bias. Biometrics 50, 1088–1101.
Benazzi, F., 2000. Borderline personality disorder and bipolar II disorder in private
practice depressed outpatients. Compr. Psychiatry 41 (2), 106–110.
Benazzi, F., 2002. Borderline personality disorder comorbidity in early- and late-
onset bipolar II disorder. Can. J. Psychiatry 47, 195–196.
Benazzi, F., 2008. A relationship between bipolar II disorder and borderline per-
sonality disorder? Progr. Neuro-Psychopharmacol. Biol. Psychiatry 32,
1022–1029.
Bridler, R., Häberle, A., Müller, S.T., Cattapan, K., Grohmann, R., Toto, S., Kasper, S.,
Greil, W., 2015. Psychopharmacological treatment of 2195 in-patients with
borderline personality disorder: a comparison with other psychiatric disorders.
Eur. Neuropsychopharmacol. 25, 763–772.
Brieger, P., Ehrt, U., Marneros, A., 2003. Frequency of comorbid personality dis-
orders in bipolar and unipolar affective disorders. Compr. Psychiatry 44, 28–34.
Briere, J., Elliott, D.M., 2003. Prevalence and psychological sequelae of self-reported
childhood physical and sexual abuse in a general population sample of men and
women. Child. Abus. Negl. 27, 1205–1222.
Carpenter, D., Clark, J.F., Glick, I.D., et al., 1995. Personality pathology among mar-
ried adults with bipolar disorder. J. Affect. Disord. 34, 269–274.
Carpiniello, B., Loreal, L., Pirarba, S., et al., 2011. Impulsivity and aggressiveness in
bipolar disorder with co-morbid borderline personality disorder. Psychiatry
Res. 188, 40–44.
Chen, G., Henter, I., Manji, H., 2010. Translational research in bipolar disorder:
emerging insights from genetically based models. Mol. Psychiatry 15, 883–895.
Comtois, K.A., Cowley, D.S., Dunner, D.L., Roy-Byrne, P.P., 1999. Relationship be-
tween borderline personality disorder and axis I diagnosis in severity of de-
pression and anxiety. J. Clin. Psychiatry 60, 752–758.
Conus, P., Cotton, S., Schimmelmann, B.G., Berk, M., Daglas, R., McGorry, P.D.,
Lambert, M., 2010. Pretreatment and outcome correlates of past sexual and
physical trauma in 118 bipolar I disorder patients with a first episode of psy-
chotic mania. Bipolar Disord. 12, 244–252.
Deltito, J., Martin, L., Riefkohl, J., Austria, B., Kissilenko, A., Morse, P.C.C., 2001. Do
patients with borderline personality disorder belong to the bipolar spectrum? J.
Affect. Disord. 67, 221–228.
DerSimonian, R., Laird, N., 1986. Meta-analysis in clinical trials. Control. Clin. Trials
7, 177–188.
Downs, S.H., Black, N., 1998. The feasibility of creating a checklist for the assess-
ment of the methodological quality both of randomised and non-randomised
studies of health care interventions. J. Epidemiol. Community Health 52,
377–384.
Dunayevich, E., Sax, K.W., Keck Jr., P.E., et al., 2000. Twelve-month outcome in bi-
polar patients with and without personality disorders. J. Clin. Psychiatry, 61.
Duval, S., Tweedie, R., 2000. Trim and fill: a simple funnel-plot-based method of
testing and adjusting for publication bias in meta-analysis. Biometrics 56,
455–463.
Egger, M., Davey Smith, G., Schneider, M., Minder, C., 1997. Bias in meta-analysis
detected by a simple, graphical test. BMJ (Clin. Res. ed.) 315, 629–634.
Feinstein, A.R., 1970. The pre-therapeutic classification of co-morbidity in chronic
disease. J. Chronic Dis. 23, 455–468.
Fornaro, M., Martino, M., De Pasquale, C., Moussaoui, D., 2012. The argument of
antidepressant drugs in the treatment of bipolar depression: mixed evidence or
mixed states? Expert Opin. Pharmacother. 13 (14), 2037–2051. http://dx.doi.
org/10.1517/14656566.2012.719877, Epub 2012 Sep 4.
Fornaro, M., De Berardis, D., Iasevoli, F., Pistorio, M.L., D'Angelo, E., Mungo, S.,
Martino, M., Ventriglio, A., Cattaneo, C.I., Favaretto, E., Del Debbio, A., Romano,
A., Ciampa, G., Elassy, M., Perugi, G., De Pasquale, C.J., 2013. Treatment ad-
herence towards prescribed medications in bipolar-II acute depressed patients:
relationship with cyclothymic temperament and "therapeutic sensation seek-
ing" in response towards subjective intolerance to pain. Affect. Disord 151 (2),
596–604. http://dx.doi.org/10.1016/j.jad.2013.07.004, Epub 2013 Jul 30. Erratum
in: J Affect Disord. 2015 Jan;170:22.PMID:23906864".
Fornaro, M., De Berardis, D., Iasevoli, F., Pistorio, M.L., D′Angelo, E., Mungo, S.,
Martino, M., Ventriglio, A., Cattaneo, C.I., Favaretto, E., Del Debbio, A., Romano,
A., Ciampa, G., Elassy, M., Perugi, G., De Pasquale, C.J., 2014. Corrigendum to
 M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118
"Treatment adherence towards prescribed medications in bipolar-II acute de-
pressed patients: Relationship with cyclothymic temperament and "therapeutic
sensation seeking" in response towards subjective intolerance to pain". Affect.
Disord 170C, 22. http://dx.doi.org/10.1016/j.jad.2014.08.014 [Epub ahead of
print]. PMID: 25217760.
Fornaro, M., De Berardis, D., Koshy, A.S., Perna, G., Valchera, A., VanCampfort, D.,
Stubbs, B., 2016. Prevalence and clinical features associated with bipolar dis-
order polypharmacy: a systematic review. Neuropsychiatric Dis. Treat., in press.
Fornaro, M., Aguglia, E., Dell'Osso, L., Perugi, G., 2011. Could the underestimation of
bipolarity obstruct the search for novel antidepressant drugs? Expert Opin.
Pharmacother. 12, 2817–2831.
Fossati, A., Madeddu, F., Maffei, C., 1999. Borderline personality disorder and
childhood sexual abuse: a meta-analytic study. J. Personal. Disord. 13, 268–280.
Fountoulakis, K.N., 2014. Bipolar Disorder: An Evidence-Based Guide to Manic
Depression. Springer, Germany.
Friborg, O., Martinsen, E.W., Martinussen, M., Kaiser, S., Overgard, K.T., Rosenvinge,
J.H., 2014. Comorbidity of personality disorders in mood disorders: a meta-
analytic review of 122 studies from 1988 to 2010. J. Affect. Disord. 152–154,
1–11.
Garno, J.L., Goldberg, J.E., Ramirez, P.M., et al., 2005. Bipolar disorder with comorbid
cluster B personality disorder features: impact on suicidality. J. Clin. Psychiatry,
66.
Gasperini, M., Scherillo, P., Manfredonia, M.G., Franchini, L., Smeraldi, E., 1993. A
study of relapses in subjects with mood disorder on lithium treatment. Eur.
Neuropsychopharmacol. 3, 103–110.
George, E.L., Miklowitz, D.J., Richards, J.A., et al., 2003. The comorbidity of bipolar
disorder and axis II personality disorders: prevalence and clinical correlates.
Bipolar Disord. 5, 115–122.
Ghaemi, S.N., Barroilhet, S., 2015. Confusing borderline personality with severe
bipolar illness. Acta Psychiatr. Scand. 132, 281–282.
Ghaemi, S.N., Dalley, S., Catania, C., Barroilhet, S., 2014. Bipolar OR borderline: a
clinical overview. Acta Psychiatr. Scand. 130, 99–108.
Goodwin, F.K., Jamison, K.R., 2007. Manic-depressive Illness: Bipolar Disorders and
Recurrent Depression. Oxford University Press, United Kingdom.
Gunderson, J.G., Weinberg, I., Daversa, M.T., Kueppenbender, K.D., Zanarini, M.C.,
Shea, M.T., Skodol, A.E., Sanislow, C.A., Yen, S., Morey, L.C., 2006. Descriptive and
longitudinal observations on the relationship of borderline personality disorder
and bipolar disorder. Am. J. Psychiatry 163 (7), 1173–1178.
Gunderson, J.G., Stout, R.L., Shea, M.T., Grilo, C.M., Markowitz, J.C., Morey, L.C., Sa-
nislow, C.A., Yen, S., Zanarini, M.C., Keuroghlian, A.S., 2014. Interactions of
borderline personality disorder and mood disorders over ten years. J. Clin.
Psychiatry 75, 829–834.
Henry, C., Mitropoulou, V., New, A.S., Koenigsberg, H.W., Silverman, J., Siever, L.J.,
2001. Affective instability and impulsivity in borderline personality and bipolar
II disorders: similarities and differences. J. Psychiatr. Res. 35, 307–312.
Henry, C., Phillips, M., Leibenluft, E., M'bailara, K., Houenou, J., Leboyer, M., 2012.
Emotional dysfunction as a marker of bipolar disorders. Front. Biosci. (Elite Ed.)
4, 2722.
Higgins, J., Green, S., 2011. Cochrane handbook for systematic reviews of inter-
ventions, version 5.1.0. The Cochrane Collaboration.
Hudziak, J.J., Boffeli, T.J., Kriesman, J.J., Battaglia, M.M., 1996. Clinical study of the
relation of borderline personality disorder to Briquet's syndrome (hysteria),
somatization disorder, antisocial personality disorder, and substance abuse
disorders. Am. J. Psychiatry 153, 1598–1606.
Ioannidis, J.P., 2009. Integration of evidence from multiple meta-analyses: a primer
on umbrella reviews, treatment networks and multiple treatments meta-ana-
lyses. CMAJ : Can. Med. Assoc. J. ¼ J. l'Association Medicale Canadienne 181,
488–493.
Joyce, P.R., Luty, S.E., McKenzie, J.M., et al., 2004. Bipolar II disorder: personality and
outcome in two clinical samples. Aust. N. Z. J. Psychiatry 38, 433–438.
Joyce, P.R., Light, K.J., Rowe, S.L., Cloninger, C.R., Kennedy, M.A., 2010. Self-mutila-
tion and suicide attempts: relationships to bipolar disorder, borderline per-
sonality disorder, temperament and character. Aust. N. Z. J. Psychiatry 44,
250–257.
Khalili, N., 2014. A biopsychosocial approach to bipolar-borderline debate: psy-
chological effect of a biological temperament. Acta Psychiatr. Scand. 130,
158–159.
Liberati, A., Altman, D.G., Tetzlaff, J., Mulrow, C., Gotzsche, P.C., Ioannidis, J.P., Clarke,
M., Devereaux, P.J., Kleijnen, J., Moher, D., 2009. The PRISMA statement for
reporting systematic reviews and meta-analyses of studies that evaluate
healthcare interventions: explanation and elaboration. BMJ (Clin. Res. ed.) 339,
b2700.
Links, P.S., Steiner, M., Offord, D.R., Eppel, A.B., 1988. Characteristics of borderline
personality disorder: a Canadian study. Can. J. Psychiatry/La Rev. Can.
Psychiatrie.
Loftus, S.T., Jaeger, J., 2006. Psychosocial outcome in bipolar I patients with a per-
sonality disorder. J. Nerv. Ment. Dis. 194, 967–970.
Mackinnon, D.F., Pies, R., 2006. Affective instability as rapid cycling: theoretical and
clinical implications for borderline personality and bipolar spectrum disorders.
Bipolar Disord. 8, 1–14.
Maj, M., 2005. ‘Psychiatric comorbidity’: an artefact of current diagnostic systems?
Br. J. Psychiatry 186, 182–184.
Maniglio, R., 2014. Prevalence of sexual abuse among children with conduct dis-
order: a systematic review. Clin. Child Fam. Psychol. Rev. 17, 268–282.
McDermid, J., Sareen, J., El-Gabalawy, R., Pagura, J., Spiwak, R., Enns, M.W., 2015. Co-
morbidity of bipolar disorder and borderline personality disorder: findings
117
from the national Epidemiologic survey on alcohol and related conditions.
Compr. Psychiatry 58, 18–28.
Nagata, T., Yamada, H., Teo, A.R., Yoshimura, C., Kodama, Y., van Vliet, I., 2013. Using
the mood disorder questionnaire and bipolar spectrum diagnostic scale to
detect bipolar disorder and borderline personality disorder among eating dis-
order patients. BMC Psychiatry 13, 69.
Nock, M.K., Kessler, R.C., 2006. Prevalence of and risk factors for suicide attempts
versus suicide gestures: analysis of the national Comorbidity survey. J. Abnorm.
Psychol. 115, 616–623.
Pally, R., 2002. The neurobiology of borderline personality disorder: the synergy of
“nature and nurture”. J. Psychiatr. Pract. 8, 133–142.
Paris, J., Gunderson, J., Weinberg, I., 2007. The interface between borderline per-
sonality disorder and bipolar spectrum disorders. Compr. Psychiatry 48,
145–154.
Perugi, G., Akiskal, H.S., 2002a. The soft bipolar spectrum redefined: focus on the
cyclothymic, anxious-sensitive, impulse-dyscontrol, and binge-eating connec-
tion in bipolar II and related conditions. Psychiatr. Clin. North Am. 25, 713–737.
Perugi, G., Akiskal, H.S., 2002b. The soft bipolar spectrum redefined: focus on the
cyclothymic, anxious-sensitive, impulse-dyscontrol, and binge-eating connec-
tion in bipolar II and related conditions. Psychiatr. Clin. North Am. 25, 713–737.
Perugi, G., Toni, C., Travierso, M.C., Akiskal, H.S., 2003. The role of cyclothymia in
atypical depression: toward a data-based reconceptualization of the border-
line-bipolar II connection. J. Affect. Disord. 73, 87–98.
Perugi, G., Fornaro, M., Akiskal, H.S., 2011. Are atypical depression, borderline
personality disorder and bipolar II disorder overlapping manifestations of a
common cyclothymic diathesis? World Psychiatry: Off. J. World Psychiatr. As-
soc. 10, 45–51.
Perugi, G., Angst, J., Azorin, J.M., Bowden, C., Vieta, E., Young, A.H., Group, B.S., 2013.
The bipolar-borderline personality disorders connection in major depressive
patients. Acta Psychiatr. Scand. 128, 376–383.
Perugi, G., Angst, J., Azorin, J.M., Bowden, C.L., Caciagli, A., Mosolov, S., Vieta, E.,
Young, A.H., Group, B.R.-I.-M.S., 2015. Relationships between mixed features
and borderline personality disorder in 2811 patients with major depressive
episode. Acta Psychiatrica Scandinavica.
Peselow, E.D., Sanfilipo, M., Fieve, R.R., 1995. Relationship between hypomania and
personality disorders before and after successful treatment. Am. J. Psychiatry
52, 232–238.
Pica, S., E.J., Jackson, H.J., et al., 1990. Personality disorders in recent on set bipolar
disorder. Compr. Psychiatry 31, 499–510.
Pope, H.G., Jonas, J.M., Hudson, J.I., Cohen, B.M., Gunderson, J.G., 1983. The validity
of DSM-III borderline personality disorder: a phenomenologic, family history,
treatment response, and long-term follow-up study. Arch. Gen. Psychiatry 40,
23–30.
Prasad, R.B., Val, E.R., Lahmeyer, H.W., Gaviria, M., 1990. Associated diagnoses (co-
morbidity) in patients with borderline personality disorder. Psychiatric J. Univ.
Ottawa.
Preston, G.A., Marchant, B.K., Reimherr, F.W., et al., 2004. Borderline personality
disorder in patients with bipolar disorder and response to lamotrigine. J. Affect.
Disord. 79, 297–303.
Qi, X., Yang, M., Ren, W., Jia, J., Wang, J., Han, G., Fan, D., 2013. Find duplicates among
the PubMed, EMBASE, and Cochrane library databases in systematic review.
PloS One 8, e71838.
Reich, D.B., Zanarini, M.C., Fitzmaurice, G., 2012. Affective lability in bipolar disorder
and borderline personality disorder. Compr. Psychiatry 53, 230–237.
Robins, E., Guze, S.B., 1970. Establishment of diagnostic validity in psychiatric ill-
ness: its application to schizophrenia. Am. J. Psychiatry 126, 983–987.
Rossi, A., Marinangeli, M.G., Butti, G., et al., 2001. Personality disorders in bipolar
and depressive disorders. J. Affect. Disord. 65, 3–8.
Sajatovic, M., Blow, F.C., Ignacio, R.V., 2006. Psychiatric comorbidity in older adults
with bipolar disorder. Int. J. Geriatr. Psychiatry 21, 582–587.
Siever, L.J., Torgersen, S., Gunderson, J.G., Livesley, W.J., Kendler, K.S., 2002. The
borderline diagnosis III: identifying endophenotypes for genetic studies. Biol.
Psychiatry 51, 964–968.
Skodol, A.E., Stout, R.L., McGlashan, T.H., Grilo, C.M., Gunderson, J.G., Shea, M.T.,
Morey, L.C., Zanarini, M.C., Dyck, I.R., Oldham, J.M., 1999. Co‐occurrence of
mood and personality disorders: a report from the collaborative longitudinal
personality disorders study (CLPS). Depression Anxiety 10, 175–182.
Stein, D.J., Phillips, K.A., Bolton, D., Fulford, K., Sadler, J.Z., Kendler, K.S., 2010. What
is a mental/psychiatric disorder? from DSM-IV to DSM-V. Psychol. Med. 40,
1759–1765.
Stroup, D.F., et al., 2000. Meta-analysis of observational studies in epidemiology: a
proposal for reporting. JAMA 283 (15), 2008–2012.
Stroup, D.F., Berlin, J.A., Morton, S.C., Olkin, I., Williamson, G.D., Rennie, D., Moher,
D., Becker, B.J., Sipe, T.A., Thacker, S.B., 2000. Meta-analysis of observational
studies in epidemiology: a proposal for reporting. Meta-analysis Of Observa-
tional Studies in Epidemiology (MOOSE) group. JAMA 283, 2008–2012.
Stubbs, B., Vancampfort, D., Solmi, M., Veronese, N., Fornaro, M., 2016. How com-
mon is bipolar disorder in general primary care attendees? A systematic review
and meta-analysis investigating prevalence determined according to structured
clinical assessments Aust. N. Z. J. Psychiatry. pii: 0004867415623857. [Epub
ahead of print].
Swann, A.C., Lijffijt, M., Lane, S.D., Steinberg, J.L., Moeller, F.G., 2013. Antisocial
personality disorder and borderline symptoms are differentially related to
impulsivity and course of illness in bipolar disorder. J. Affect. Disord. 148,
384–390.
Ucok, A., Karaveli, D., Kundakci, T., et al., 1998. Comorbidity of personality
118 M. Fornaro et al. / Journal of Affective Disorders 195 (2016) 105–118
disorderswith bipolarmood disorders. Compr. Psychiatry 39, 72–74.
Vieta, E., Suppes, T., 2008. Bipolar II disorder: arguments for and against a distinct
diagnostic entity. Bipolar Disord. 10, 163–178.
Vieta, E., Colom, F., Martinez-Aran, A., et al., 2000. Bipolar II disorder and co-
morbidity. Compr. Psychiatry 41, 339–343.
Vieta, E., Colom, F., Corbella, B., et al., 2001. Clinical correlates of psychiatric co-
morbidity in bipolar I patients. Bipolar Disord. 3, 253–258.
Wilson, S.T., Stanley, B., Oquendo, M.A., et al., 2007. Comparing impulsiveness,
hostility, and depression in borderline personality disorder and bipolar II dis-
order. J. Clin. Psychiatry 68, 1533–1539.
Young, A.H., Angst, J., Azorin, J.M., Vieta, E., Perugi, G., Gamma, A., Bowden, C.L.,
2012. BRIDGE study warrants Critique-Reply. Arch. Gen. Psychiatry 69,
644–645.
Zanarini, M.C., Frankenburg, F.R., Dubo, E.D., Sickel, A.E., Trikha, A., Levin, A., Rey-
nolds, V., 1998. Axis I comorbidity of borderline personality disorder. Am. J.
Psychiatry 155, 1733–1739.
Zanarini, M.C., Frankenburg, F.R., Hennen, J., Reich, D.B., Silk, K.R., 2004. Axis I co-
morbidity in patients with borderline personality disorder: 6-year follow-up
and prediction of time to remission. Am. J. Psychiatry 161, 2108–2114.
Zimmerman, M., Mattia, J.I., 1999. Axis I diagnostic comorbidity and borderline
personality disorder. Compr. Psychiatry 40, 245–252.
Zimmerman, M., Morgan, T.A., 2013. Problematic boundaries in the diagnosis of
bipolar disorder: the interface with borderline personality disorder. Curr. Psy-
chiatry Rep. 15, 422.