Content uploaded by Mario Speranza
Author content
All content in this area was uploaded by Mario Speranza on Aug 15, 2018
Content may be subject to copyright.
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
Alexithymia in adolescents with borderline personality disorder
Gwenolé Loas
a,
⁎, Mario Speranza
b
, Alexandra Pham-Scottez
c
, Fernando Perez-Diaz
d
, Maurice Corcos
e
a
Antenne de Psychiatrie, CHU d'Amiens, Université de Picardie, Amiens, France
b
Service de psychiatrie infanto-juvénile, Centre Hospitalier de Versailles, Versailles, France. INSERM U669, PSIGIAM (Paris Sud Innovation Group In Adolescent Mental Health),
Université Paris-Sud et Université Paris Descartes, UMR-S0669, Paris, France
c
CMME, Hôpital Sainte Anne, Paris, France. INSERM U669, PSIGIAM (Paris Sud Innovation Group In Adolescent Mental Health), Université Paris-Sud et Université Paris Descartes,
UMR-S0669, Paris, France
d
Centre Emotion, CNRS - UPSR 3246, Hôpital de la Salpêtrière, Paris, France
e
Département de Psychiatrie infanto juvénile de l'Institut Mutualiste Montsouris, Paris, France. INSERM U669, PSIGIAM (Paris Sud Innovation Group In Adolescent Mental Health),
Université Paris-Sud et Université Paris Descartes, UMR-S0669, Paris, France
abstractarticle info
Article history:
Received 26 July 2011
Received in revised form 9 November 2011
Accepted 15 November 2011
Keywords:
Adolescents
Borderline personality disorders
Alexithymia
Depression
Anxiety-trait
Objective: The aim of this study was to explore the relationship between alexithymia and borderline person-
ality disorder (BPD) in adolescents.
Methods: The study investigated a sample of 59 consulting or inpatient adolescents with a well-established
diagnosis of BPD (SIDP-IV) and a control sample of healthy adolescents individually matched by gender,
age and socio-economic status. Alexithymia, depression and trait-anxiety were rated using the Twenty-
item Toronto Alexithymia Scale (TAS-20), the Beck Depression Inventory (BDI-II) and the trait-anxiety sub-
scale from the State-Trait Anxiety Inventory (STAI-T), respectively. A confirmatory factorial analysis (CFA)
was performed to test the fit of the three-factor structure of the TAS-20 in the adolescent sample
(N=140). BPD and control groups were compared on alexithymic scores using ANCOVA analyses controlling
for the potential confounding effects of depression and anxiety.
Results: The ratio of the chi-square to its degrees of freedom, the goodness-of-fit index, the adjusted
goodness-of-fit index and Steiger's root-mean-square error of approximation had satisfactory values of
1.54; 0.87; 0.83 and 0.058, respectively. The two ANCOVA demonstrated no significant difference for TAS-
20 scores. BPD subjects were more alexithymic than healthy subjects but this difference was mainly
explained by the levels of depression or anxiety.
Limitations: Since BPD subjects have high comorbidity with depression or anxiety, longitudinal studies exam-
ining the absolute and relative stability of TAS-20 scores are necessary to determine whether alexithymia
constitutes a state or a trait in BPD.
Conclusions: BPD adolescents are characterized by alexithymia, probably of a secondary or state-dependent
nature.
© 2011 Elsevier Inc. All rights reserved.
Introduction
Borderline personality disorder (BPD) is an impairing mental disor-
der characterized by a pervasive pattern of instability in affect regula-
tion, impulse control, interpersonal relationships, and self-image. BPD
is currently considered a public health problem, as it is associated with
severe psychosocial impairmentsand high mortality rates due to suicide
[1]. BPD concerns 1–2% of the general population [2],withhigherprev-
alence in adolescents compared to adults. Detecting BPD in adolescence
would allow early intervention before maladaptive behaviors become
fixed and refractory to biological or psychotherapeutic treatments.
Among the various theoretical efforts to understand BPD, great in-
terest has been raised by the proposals made by Fonagy [3,4] who
suggested conceptualizing BPD as a disorder of "mentalization" and
proposed a specific mentalization-oriented psychodynamic approach
to the disorder [5]. Mentalization refers to the process used by
humans to make sense of the social world by imagining the mental
states that explain their own and others' behaviors in social interac-
tion. Mentalization-based treatments have been empirically validated
by randomized, controlled trials as more effective than non-specific
psychiatric treatments [6,7]. Among the different conceptual overlaps
of the construct of mentalization, Choi-Kain and Gunderson [4] have
underlined the notion of affect consciousness (and its absence),
which resembles the concept of alexithymia. Alexithymia, literally
“no words for emotions”, is used to describe individuals presenting
difficulty in identifying their feelings, difficulty in communicating
their feelings, an absence of daydreaming, and an externally-
Journal of Psychosomatic Research 72 (2012) 147–152
⁎Corresponding author at: Antenne de Psychiatrie, Hôpital Nord, CHU d'Amiens,
Université de Picardie F-80000 Amiens, France. Tel.: +33 322 668 290; fax: +33 322
954 115.
E-mail address: loas.gwenole@chu-amiens.f (G. Loas).
0022-3999/$ –see front matter © 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jpsychores.2011.11.006
Contents lists available at SciVerse ScienceDirect
Journal of Psychosomatic Research
Author's personal copy
oriented way of thinking. If BPD is a disorder of mentalization, then
high levels of alexithymia should be observed in BPD subjects. More-
over, if the level of reflective functioning is conceptualized as a core
feature of BPD and as state-dependent [5], alexithymia could be a
primary feature of BPD and not only secondary to other symptoms
(e.g. anxiety or depression). Moreover, from a therapeutic point of
view, this perspective suggests several hypotheses. First, as alexithy-
mic features have been shown to be less sensitive to psychotherapy
[8], high levels of alexithymia could be associated with lower
responses to psychotherapy in BPD subjects. Second, the psychother-
apeutic approaches that have been found effective in alexithymia
could also be effective in BPD, and vice-versa. For example, Levy and
colleagues [9] assessed changes in attachment organization and re-
flective function as putative mechanisms of change over 1 year in a
3-year psychotherapy treatment for patients with BPD. BPD patients
were randomized to transference-focused psychotherapy (TFP),
dialectical behavioral therapy, or psychodynamic supportive psycho-
therapy. Although the three treatments produced significant positive
changes in several domains of psychopathology and functioning
(with TFP showing improvements in more domains than the other
two treatments: see Clarkin et al., [10]), an increase in patients' narra-
tive coherence and reflective function was only observed for
transference-focused psychotherapy.
Despite the potential interest of investigating the relationships be-
tween alexithymia and BPD, few studies have been conducted on this
topic, and they have reported contrasting results. To our knowledge,
only five studies have been published: three reported significant associ-
ations between BPD and alexithymia and two reported non-significant
associations. Berenbaum [11] investigated the relationship between
alexithymia and personality disorder in 60 adults receiving outpatient
psychotherapy. Alexithymia was rated using the well-validated Toronto
Alexithymia Scale (TAS) and personality disorders were assessed using
the Personality Diagnostic Questionnaire Revised (PDQ-R) following
DSM-III-R criteria. Significant and positive correlations were found be-
tween BPD diagnoses and two of the subscales in the TAS, namely diffi-
culty in identifying emotions and difficulty in communicating emotions.
Zlotnick et al. [12], in a sample of 252 treatment-seeking patients
reported that the 34 subjects who met the DSM-IV criteria for BPD
(using the SIDP), had significantly higher scores on the Toronto Alex-
ithymia Scale than the 218 subjects without BPD. Modestin et al. [13] ex-
plored the relationship between alexithymia and BPD in 223 medical
students or nursing personnel using the TAS-20 and the Borderline Pa-
thology Questionnaire for DSM-IV. Using the cutoff scores of ≥60 and
≤52 to categorize participants as alexithymic or non-alexithymic, they
found a prevalence of BPD of 62% and 21%, respectively, with a signifi-
cant difference between the two groups. Bach et al. [14] reported the re-
sults of a stepwise regression analysis exploring the relationships
between the TAS score (dependent variable) and the specificDSM-III-
R personality disorder dimension scores (predictors) assessed with the
PDQ-R. The sample included 182 female psychiatric outpatients. No sig-
nificantassociation was found between alexithymia and the BPD dimen-
sion score. Recently, Nicolo et al. [15] explored the relationships
betweenalexithymia andpersonality disorders in a sample of 388 adults
and adolescents requiring psychiatric treatment in an outpatient clinic.
Alexithymia was assessed with the TAS-20 and personality disorders
were rated using the SCID-II for DSM-IV. Using the validated cutoff
scores of the TAS-20 (≥60: alexithymia; 53–59: intermediate alexithy-
mia; ≤52: non-alexithymia), three groups were defined (94 alexithy-
mics, 81 intermediate alexithymics and 213 non-alexithymics). The
three groups were then compared in terms of the number of criteria
they met for clusters A, B and C on the SCID-II. No group differences
were reported for cluster B criteria. However, a weak but significant cor-
relation (r =0.22, p b0.0001) was observed between the number of BPD
traits and the TAS-20 subscale rating difficulty in identifying feelings.
These contrasting results could be partly explained by differences in
sampling and statistical methods, but also by differences in the
comorbidity profiles associated with the groups under investigation. It
has been clearly established that alexithymia and BPD are bothassociat-
ed with high levels of anxiety and depression [1,16]. A recent study [17]
in a non-clinical sample of late adolescents, found that alexithymic sub-
jects, as assessed using the TAS-20, were significantly more anxious on
the State-Trait Anxiety Inventory (STAI) than non-alexithymic subjects.
Thus, anxiety and depressive levels should be controlled for when ex-
amining the relationship between BPD and alexithymia.
Thus, although the relationship between BPD and alexithymia is
still controversial, alexithymia, as a mentalization deficit, could be
considered as a main symptom of BPD subjects and a potential major
target for the therapeutic approach to BPD. Since the relationship be-
tween alexithymia and BPD has never been studied in adolescents,
this study was designed to explore this relationship. We tested the hy-
pothesis that BPD subjects would be more alexithymic than non-BPD
subjects, independently from the impact of anxiety and depression.
Methods
Participants
The study sample was drawn from a European research project in-
vestigating the phenomenology of BPD in adolescence (the European
Research Network on Borderline Personality Disorder, EURNET BPD;
see [18] for a full description of the study methodology). Briefly, the
research network was composed of five specialist psychiatric centers
for adolescents and young adults in France, Belgium, and Switzerland.
Between January and December 2007, all inpatient and outpatient
adolescents (aged 15 to 19) meeting DSM-IV criteria for BPD on the
Structured Interview for DSM-IV Personality (SIDP-IV) were included
[19]. Individuals with schizophrenia and any potentially life-
threatening, chronic and/or serious medical illness were excluded.
The final study population comprised 85 BPD adolescents (11 boys,
13%, and 74 girls, 87%). The mean age was 16.3 years (SD =1.4).
67% (N=57) were inpatients. The control sample included 85
healthy adolescents individually matched for gender, age and socio-
economic status recruited by announcement in schools. Control sub-
jects were excluded if they had a history of or ongoing psychiatric
follow-up, and if they were positive for a DSM-IV diagnosis of person-
ality disorder.
This study was approved by our local institutional review board
and all data were rendered anonymous. Subjects were provided
with comprehensive information on the study objectives and proce-
dures. Written informed consent was obtained from the adolescents
and at least one of their parents in each case.
Assessments
All subjects completed a research protocol (consisting of a diag-
nostic evaluation of Axis I and Axis II disorders) and a self-
administered questionnaire to collect socio-demographic and psy-
chopathological data. Axis II disorders were investigated using the
French version of SIDP-IV which is known to have good psychometric
properties in adolescents and young adults [20]. Borderline severity
for each of the 9 criteria was coded as absent (0), subliminal (1), pre-
sent (2) and severe (3). Borderline severity scores thus varied from
0 to 27. Axis I disorders were assessed with the Schedule for Affective
Disorders and Schizophrenia for School-Age Children (K-SADS). Diag-
nostic interviews were conducted by a team of 5 clinical psycholo-
gists and psychiatrists experienced in research and the assessment
and/or treatment of DSM-IV Axis I/II disorders in adolescents. To en-
sure high levels of reliability, the research team participated in sever-
al training sessions, including commented scoring of videotaped
interviews. The inter-rater reliability for SIDP-IV was calculated
from independent ratings of ten videotaped interviews. The Kappa
coefficient for agreement on the presence or absence of a BPD was
148 G. Loas et al. / Journal of Psychosomatic Research 72 (2012) 147–152
Author's personal copy
very high (0.84) and the values for the presence/absence of other per-
sonality disorders ranged from 0.54 to 1.
For the purposes of the present study, three questionnaires evalu-
ating alexithymia, depression and anxiety were used.
Alexithymia was rated using the Twenty-item Toronto Alexithy-
mia Scale (TAS-20, [21]). The TAS-20 comprises twenty items evalu-
ating 1) difficulty in identifying emotions, 2) difficulty in describing
emotions and 3) externally-oriented way of thinking. The French ver-
sion has satisfactory validity and reliability [22]. A cutoff score equal
to or greater than 56 is used to detect alexithymia [23]. This cut-off
score has been established for the French population using the meth-
odology employed by Taylor and Parker for their initial validation of
the TAS [24] and the TAS-20 [16] which consisted of testing several
values of the TAS-20 scores to identify the value with the best positive
predictive power in detecting alexithymia as assessed using a stan-
dard reference instrument (the Beth Israël Questionnaire).
Depression was rated using the second version of the Beck De-
pression Inventory (BDI-II, [25]). The BDI-II is the 1996 revision of
the classic BDI. The French version of the BDI-II has satisfactory psy-
chometric properties [26]. The State-Trait Anxiety Inventory (STAI,
[27]) is a self-report assessment device that includes separate mea-
sures for state and trait anxiety. The French version has been validat-
ed in adults and adolescents and has good validity and reliability [28].
For the present study we chose to use the trait subscale from the STAI
to better control for a more general and long-standing type of anxiety,
which can be considered as a stronger confounder in the relationship
between alexithymia and BPD than the temporary condition of
“state”anxiety.
Statistical analysis
Firstly, the validity and reliability of the TAS-20 were investigated
in the entire sample using confirmatory factorial analysis (CFA) and
Cronbach's alpha coefficient. CFA was used to test the three-factor
solution of the TAS-20. To assess the goodness-of-fit of the model
we used the criteria already employed in the previous validation
studies [29,30]: the goodness-of-fit index (GFI>0.85), the adjusted
goodness-of-fit index (AGFI>0.80), Steiger's root-mean-square
error of approximation (RMSEAb0.08), the ratio of chi-square to its
degrees of freedom (chi-square/df ratio) with a value b5 and prefer-
ably b2. Secondly, the two groups were compared on the TAS-20 total
score and sub-scale scores using Student's t tests. Thirdly, rates of
alexithymia according to cut-off scores (alexithymia: TAS-20 ≥56;
non-alexithymia: TAS-20 b56) that were calculated for the French
version of the scale [23] were compared between groups using
chi-square analyses. Fourthly, successive analyses of covariance
(ANCOVA) were conducted using the BDI-II score or STAI trait-
anxiety subscales as covariables, the TAS-20 score as the dependent
variable and the groups (BPD versus controls) as independent vari-
ables. Statistical analyses were performed with the SAS software.
Significance was set at a p value of b0.05.
Results
Of the 170 participants included in the study, 26 BPD and 4 control adolescents did
not fully complete the three questionnaires and were excluded from the analyses. No
significant differences were observed in terms of socio-demographic variables be-
tween subjects with complete and incomplete data. The Borderline severity score did
not differ between the two groups, with values of 17.3 (SD = 4.1) and 18.4
(SD=3.41) respectively for subjects with complete and incomplete data. Subjects in-
cluded in the study did not differ from the others in terms of the current prevalence
of major depressive disorders (35.6% vs 38.5%, p = ns). The final sample included 59
BPD adolescents and 81 control subjects. No significant differences were observed be-
tween BPD and control subjects for any socio-demographic variables. The most fre-
quent co-morbid diagnosis was major depressive disorder which was found in 35.6%
of the BPD sample and was absent in the control sample.
Results concerning the fit of the three-factor solution of the TAS-20 for the current
data were satisfactory. The value of the chi-square ratio, GFI, AGFI and RMSEA were
1.54, 0.87, 0.83, and 0.058 respectively. The factor loadings of the items are given in
Table 1. All items were significant except for two (#16, #20) belonging to the
Externally-Oriented Thinking factor. The value of Cronbach's alpha was 0.80 for the
TAS-20 and 0.84, 0.77 and 0.53 for the three subscales (Difficulty Identifying Feelings,
Difficulty Describing Feelings, Externally-Oriented Thinking) respectively. The preva-
lence of alexithymia in the borderline and control groups was 66.1% (N = 39) and
25.9% (N=21) respectively, with a significant difference between the two groups
(χ
2
=22.5, df= 1, pb0.001). The BPD subjects had significantly higher TAS-20 scores
than the control subjects with values of 59.6 (SD = 10.4) and 49.1 (SD = 9.8) respec-
tively (t (138)=6.1 (pb0.001))(Table 2).
Two ANCOVA were performed to control for the levels of anxiety and depression.
The dependent variable was the TAS-20, the covariables were either the STAI-trait or
the BDI-II scores and the independent variable was the group (borderline versus con-
trols). No significant group effect, but a significant effect of the covariable was ob-
served for each ANCOVA. Using the BDI-II score as the covariable, the group effect
was not significant (F (1,137) =0.07, p =0.79), although the covariable effect was sig-
nificant (F (1,137) =41.6, pb0.001). Using the STAI–trait score as the covariable, the
group effect was not significant (F (1,137) = 0.02, p =0.88), although the covariable
effect was significant (F (1,137) =38.8, pb0.001).
Discussion
The present study confirms the overall satisfactory psychometric
properties of the TAS-20 in adolescents, but points to poor reliability
for the Externally-Oriented Thinking factor, as has already been
highlighted in the literature. Five CFA have been reported demon-
strating the fit of the three-factor solution for the TAS-20 in adoles-
cents and/or children [31–35]. The value of Cronbach's alpha
coefficient for the TAS-20 total score was similar to those reported
by Säkkinen et al. [32], Parker et al. [34] and Loas et al. [35] (range:
0.64–0.82). In addition, the value of Cronbach's alpha coefficient for
the Externally-Oriented Thinking factor was low (0.53), confirming
the low values derived from the five CFA studies [31–35]. Parker et
al. [34] studied the psychometric properties of the TAS-20 in subjects
in their early teens (13–14 years), mid-teens (15–16 years) and late
teens (17–18 years). The values of Cronbach's alpha coefficients
were 0.68, 0.75 and 0.82, respectively, for the TAS-20 total score
and 0.49, 0.52 and 0.68, respectively, for the EOT factor. The authors
concluded that age has a negative impact on the quality of measure-
ment, with poorer reliability the younger the age. The poor reliability
of the Externally-Oriented Thinking factor observed in several
Table 1
Parameters estimated from the results of the CFA of the TAS-20 for the sample of ado-
lescents (N=140).
Item DIF DDF EOT
Difficulty identifying feelings (DIF)
1 0.72
3 0.51
6 0.62
7 0.53
9 0.79
13 0.76
14 0.66
Difficulty describing feelings (DDF)
2 0.81
4 0.69
11 0.68
12 0.55
17 0.41
Externally-oriented thinking (EOT)
5 0.38
8 0.49
10 0.60
15 0.27
16 0.04⁎
18 0.26
19 0.62
20 0.19⁎
⁎p>0.05.
149G. Loas et al. / Journal of Psychosomatic Research 72 (2012) 147–152
Author's personal copy
adolescent samples therefore strongly argues in favor of using the
TAS-20 total score rather than its subscores.
This study reports, for the first time, a high prevalence of alexithy-
mia in BPD adolescents (66.1%) compared to non-BPD healthy con-
trols (25.9%). A similar association between alexithymia and BPD
has already been reported by Modestin and colleagues using a differ-
ent study design [13]. These authors found higher levels of BPD (62%)
in medical students or nursing personnel with alexithymia compared
to subjects without alexithymia (21%). Although the association be-
tween alexithymia and BPD seems to be confirmed by these results,
it must be emphasized that in both of these studies subjects pre-
sented a high prevalence of concurrent mood disorders which could
partially explain the association between alexithymia and BPD. In
our study, 35.6% of BPD subjects presented a major depressive disor-
der versus no subjects in the control group. Similarly, in the study by
Modestin et al. [13], among the alexithymic subjects, 46% presented a
depressive disorder and 62% had BPD, whereas the rates were signif-
icantly much lower in the non-alexithymic group (16% and 21% re-
spectively for depressive disorder and BPD). The potential influence
of depression on the relationship between alexithymia and BPD was
confirmed by the results of the ANCOVA analyses which showed
that the relationship between these two variables was mainly
explained by the associated symptoms of depression and/or anxiety.
This last result suggests considering alexithymia in BPD mainly as a
secondary phenomenon resulting from the concurrent depressive
and anxiety symptoms so frequently observed in BPD (see review in
[1]), and not as a specific feature of the disorder. As mood disorders
are extremely frequent in BPD, to the point of being considered as
part of the disorder [36], studying the relationship between alexithy-
mia and depression in BPD subjects with no depressive symptoms can
be thought to be an artificial situation. Moreover, the primary or sec-
ondary nature of alexithymia in medical or psychiatric disorders has
been widely debated in the literature. As has been convincingly
shown by Luminet and colleagues in depressive patients [37], al-
though alexithymia scores may be influenced by the severity of de-
pression, this dimension shows relative stability over time, thus
supporting the view of this construct as a stable personality trait rath-
er than a state-dependent phenomenon. Unfortunately, no studies
have directly investigated the absolute and relative stability of alex-
ithymia in BPD patients, notwithstanding the fact that this could
help disentangle the relationships between alexithymia and anxi-
ety/depression in BPD individuals. Among the studies that have indi-
rectly approached this point, it is worthwhile quoting the studies by
Grabe and colleagues [38] and Honkalampi and colleagues [39].
Grabe et al. [38] administered the TAS-20, the Temperament and
Character Inventory (TCI) and the Symptoms Check list (SCl-90-R)
to 254 psychiatric patients to explore how far alexithymia could pre-
dict a broad range of psychiatric symptoms. High scores on the TAS-
20 subscale measuring difficulties in identifying feelings emerged as
a major predictor of current psychopathology for all SCL-90-R sub-
scales and in particular for the depression and anxiety subscales. For
the depressive subscale, a multiple regression analysis showed that
the relationship with TAS-20 subscale difficulties identifying feelings
was independent from high levels of novelty seeking and low levels of
self-directedness on the TCI which were also significantly associated
with the depressive subscale. Moreover, several studies [40,41] have
reported high levels of novelty seeking and low levels of self-
directedness in BPD subjects compared to controls. This result could
be explained by depression, as the level of depression was not includ-
ed as covariate in the statistical analyses. Thus, the relationship be-
tween alexithymia, depression, high levels of novelty seeking and
low levels of self-directedness could be independent from BPD.
Honkalampi and colleagues [39] examined the stability of alex-
ithymia over a 6-month period in a sample of 121 outpatients with
major depressive disorders and with or without a cluster C personal-
ity disorder (assessed by the SCID-II). The results of the regression an-
alyses showed that the stability of alexithymia was greater among
depressive patients with a comorbid cluster C personality disorder
than among depressive patients without a comorbid cluster C person-
ality disorder, for which the stability was more related to the severity
of depression. Although these results concern only cluster C personal-
ity disorders, they indirectly suggest that the presence of a comorbid
personality disorder increases the stability of alexithymia features in
patients with major depressive disorders.
Notwithstanding the interest of these studies, however, it is not pos-
sible, in the absence of longitudinal data, to clearly establish any causal
relationships between BPD, alexithymia and anxiety or depression.
In fact, several alternative pathways could be suggested to explain
the relationships between these variables. For example, anxiety and
depression could act as mediators between alexithymia and BPD
symptoms: due to their cognitive limitations in emotional regulation,
subjects with high levels of alexithymia could be easily overwhelmed
in stressful situations [42] with marked symptoms of anxiety and de-
pression which they might try to manage by resorting to maladaptive
behaviors, such as the self-stimulatory behaviors observed in BPD pa-
tients. Conversely, anxiety/depression and alexithymia in BPD could
reinforce one another in a circular manner: BPD symptoms could di-
rectly increase alexithymic features making these subjects highly vul-
nerable to anxiety and depression. BPD could also indirectly increase
levels of alexithymia by increasing levels of anxiety and depression
which, in turn, can further affect recognition and regulation of emo-
tions [43].
Further longitudinal studies are needed to establish the primary
nature of alexithymia in BPD individuals and, in line with the menta-
lization deficit of BPD [3], to support the hypothesis that alexithymia
could correspond to a low level of reflective functioning, and could be
a core symptom of BPD. If confirmed by further studies, the specific
association between alexithymia and BPD could imply several impor-
tant consequences.
From a prognostic point of view, alexithymia could prove to be a
key feature to explore in BPD individuals. To our knowledge, only
one study has explored this hypothesis in adult BPD patients. Karaklic
et al. [44] reported an 18-month follow-up study on 75 adult BPD pa-
tients compared to 40 subjects presenting other personality disorders
(OPD). At the 18-month follow-up, 57 BPD patients and 22 OPD were
re-evaluated: 45% of the BPD and 50% of the OPD were in remission.
Using logistic regressions, the authors found that low levels of alex-
ithymia and good global functioning at baseline were independent
predictors of good outcome. Unfortunately, the authors did not
include a measure of depression or anxiety to adjust the effects of
alexithymia.
From a therapeutic point of view, the high rates of alexithymia ob-
served in BPD subjects call for specific attention. Several studies have
shown that alexithymia negatively influences treatment outcome in
Table 2
Differences between the borderline personality disorder adolescents and their controls
on the self-administered questionnaires.
Measure Borderline personality
disorders
Controls
Mean (SD) Mean (SD)
TAS-20 total* 59.6 (10.44) 49.1 (9.83)
DIF-subscale* 22.7 (5.89) 15.5 (5.42)
DDF-subscale* 17.4 (4.43) 14.1 (4.24)
EOT-subscale 19.5 (4.83) 19.4 (4.27)
BDI-II* 28.9 (12.11) 8 (6.64)
STAI-T* 61 (12.26) 38.9 (10.27)
*Significant effect at alpha =0.05.
Acronyms: TAS-20 = Twenty-item Toronto Alexithymia Scale; DIF = Difficulty Identifying
subscale of the TAS-20; DDF = Difficulty Describing subscale of the TAS-20; EOT = Externally
Oriented thinking subscale of the TAS-20; BDI-II = Beck Depression Inventory; STAI-T =
State-Trait Anxiety Inventory-trait.
150 G. Loas et al. / Journal of Psychosomatic Research 72 (2012) 147–152
Author's personal copy
various psychiatric conditions [45]. Several studies suggest that alex-
ithymic patients are less responsive to psychodynamic psychotherapy
[45]. Results are less consistent concerning the predictive value of
alexithymia for the outcome of cognitive-behavioral therapy, but
studies are lacking [see review in 45]. A recent naturalistic study of
a short-term cognitive-behavioral group therapy (CBGT) for panic
disorder found that baseline alexithymia did not predict the outcome
of CBGT, either at post-treatment or at follow-up [46]. Recently, sev-
eral studies have tried to investigate the processes through which
alexithymia has negative effects on the outcome of psychotherapy.
Ogrodniczuk and colleagues [45] have shown, in a randomized con-
trolled trial on two forms of group therapy for complicated grief,
that while alexithymic features were associated with a less favorable
outcome, this relationship was largely mediated by therapist reac-
tions and perceptions of the patient's positive qualities, and personal
compatibility. In another similar study, Rasting and colleagues [47]
videotaped and evaluated the facial affects displayed in dyadic thera-
peutic interactions in a sample of psychosomatic patients. They ob-
served that the predominant emotional reaction among therapists
to negative affects expressed by alexithymic patients was contempt,
thus liable to produce negative interaction in therapy. While the liter-
ature suggests overall that cognitive-behavioral group therapies
could be more effective in alexithymic BPD patients than psychody-
namic psychotherapies [46], the studies on the potential impact of
alexithymic features on psychotherapists' attitudes suggests that psy-
chotherapeutic approaches that focus on therapist reactions (such as
the mentalizing approach) could be also indicated.
To our knowledge, no study up to now has investigated the effec-
tiveness of different psychotherapeutic approaches in BPD patients
taking into account their levels of alexithymia. Only one study [48]
has indirectly explored the negative role played by alexithymia in
communication within families of BPD patients, by comparing the
levels of alexithymia in parents of patients with BPD with parents of
subjects with restricting anorexia nervosa. Family members of BPD
subjects had higher levels of alexithymia compared to family mem-
bers of anorexia nervosa or nonclinical subjects. Thus, alexithymia
in parents could represent a pathogenic or maintenance factor for
their BPD offspring, and should be taken into account in therapeutic
approaches.
Several limitations of our study should be acknowledged. Firstly,
the results of this study need to be confirmed in larger samples. Sec-
ondly, the cross-sectional design of the study did not make it possible
to establish the absolute and relative stability of alexithymia and de-
pression/anxiety in BPD adolescents. As this study sample is currently
being followed longitudinally over 3 years, the stability of alexithy-
mia will be further investigated. Thirdly, we did not use a dimensional
approach to BPD although categorical and dimensional aspects of
alexithymia were taken into account. The choice of a dimensional ap-
proach to BPD could be worthwhile since the use of DSM-IV criteria
can lead to the inclusion of BPD subjects with different levels of sever-
ity. However, in the present study, BPD subjects had high values on
the borderline severity index (mean= 17.3) suggesting that our sam-
ple was composed of a homogeneous severe group of BPD adoles-
cents. Fourthly, 26 BPD adolescents did not complete the three
questionnaires and were excluded from the final analyses. It is possi-
ble that, according to their levels of depression, the association be-
tween alexithymia and BPD could have been strengthened or
weakened. However, the mean scores on the Borderline severity
index and the prevalence of major depressive disorders did not differ
between included and excluded adolescents, thus limiting the bias of
the high rates of incomplete data in our sample of BPD adolescents. It
should be noted that the compliance of BPD adolescents seems lower
compared to what is usually observed in BPD adults.
In conclusion, this study confirms the psychometric properties of
the TAS-20 in adolescents and reports high levels of alexithymia,
probably of a secondary or state-dependent nature, in BPD
adolescents compared to controls. Longitudinal studies are required
to explore the stability and the prognostic value of alexithymia, and
to better understand any causal relationship between alexithymia,
anxiety and depression in BPD adolescents.
Role of funding source
The funding source was not involved in the study design, data
collection, analysis, and interpretation, writing of the report or in
the decision to submit the paper for publication.
Conflict of interest
The other authors declare that they have no competing interests.
Acknowledgments
This research was supported by the Wyeth Foundation for Child
and Adolescent Health and the Lilly Foundation.
References
[1] Skodol AE, Gunderson JG, Pfohl B, Widiger TA, Livesley WJ, Siever LJ. The border-
line diagnosis I: psychopathology, comorbidity, and personality structure. Biol
Psychiatry 2002;51:936–50.
[2] Torgensen S, Kringlen E, Cramer V. The prevalence of personality disorders in a
community sample. Arch Gen Psychiatry 2001;58:590–6.
[3] Fonagy P, Luyten P. A developmental, mentalization-based approach to the under-
standing and treatment of borderline personality disorder. Dev Psychopathol
2009;21:1355–81.
[4] Choi-Kain LW, Gunderson JG. Mentalization: ontogeny, assessment, and applica-
tion in the treatment of borderline personality disorder. Am J Psychiatry
2008;165:1127–35.
[5] Bateman A, Fonagy P. Mentalization based treatment of borderline personality
disorder. J Pers Disord 2004;18:36–51.
[6] Bateman A, Fonagy P. 8-year follow-up of patients treated for borderline person-
ality disorder: mentalization-based treatment versus treatment as usual. Am J
Psychiatry 2008;165:631–8.
[7] Bateman A, Fonagy P. Randomized controlled trial of outpatient mentalization-
based treatment versus structured clinical management for borderline personali-
ty disorder. Am J Psychiatry 2009;166:1355–64.
[8] McCallum M, Piper WE, Ogrodniczuk JS, Joyce AS. Relationships among psycho-
logical mindedness, alexithymia and outcome in four forms of short-term psycho-
therapy. Psychol Psychother 2003;76:133–44.
[9] Levy KN, Meehan KB, Kelly KM, Reynoso JS, Weber M, Clarkin JF, et al. Change in
attachment patterns and reflective function in a randomized control trial of
transference-focused psychotherapy for borderline personality disorder. J Consult
Clin Psychol 2006 Dec;74(6):1027–40.
[10] Clarkin JF, Levy KN, Lenzenweger MF. Evaluating three treatments for borderline
personality disorder: a multiwave study. Am J Psychiatry 2007;164:922–8.
[11] Berenbaum H. Childhood abuse, alexithymia and personality disorder. J Psycho-
som Res 1996;41:585–95.
[12] Zlotnick C, Mattia JI, Zimmerman M. The relationship between posttraumatic
stress disorder, childhood trauma and alexithymia in an outpatient sample.
J Trauma Stress 2001;14:177–88.
[13] Modestin J, Furrer R, Malti T. Study on alexithymia in adult non-patients. J Psycho-
som Res 2004;56:707–9.
[14] Bach M, de Zwaan M, Ackard D, Nutzingzer DO, Mitchell JE. Alexithymia: relation-
ship to personality disorders. Compr Psychiatry 1994;35:239–43.
[15] Nicolo G, Semerari A, Lysaker PH, Dimaggio G, Conti L, D'Angerio S, Procacci M,
Popolo R, Carcione A. Alexithymia in personality disorders: correlations with
symptoms and interpersonal functioning. Psych Research 2011;190:37–42.
[16] Taylor GJ, Bagby RM, Parker JDA. Disorders of affect regulation alexithymia in
medical and psychiatric illness. Cambridge, UK: Cambridge University Press;
1997.
[17] Karukivi M, Hautala L, Kaleva O, Haapasalo-Pesu KM, Liuksila PR, Joukamaa M,
et al. Alexithymia is associated with anxiety among adolescents. J Affect Disord
2010;125:383–7.
[18] Corcos M, Pham-Scottez A, Speranza M. European Research Network on Border-
line Personality Disorder (EURNET-BPD). 57th Annual Meeting of the American
Academy of Child and Adolescent Psychiatry. New research poster session N
312. New York: 28–31 October 2010; 2010.
[19] Pfohl B, Blum N, Zimmerman M. Structured Interview for DSM-IV Personality.
Washington, DC: American Psychiatric Press; 1997.
[20] Chabrol H, Chouicha K, Montovany A, Callahan S, Duconge E, Sztulman H. Person-
ality disorders in a nonclinical sample of adolescents. Encéphale 2002;28:520–4.
[21] Bagby RM, Parker JDA, Taylor GJ. The twenty-item Toronto Alexithymia Scale-I.
Item selection and cross validation of the factor structure. J Psychosom Res
1994;38:23–32.
151G. Loas et al. / Journal of Psychosomatic Research 72 (2012) 147–152
Author's personal copy
[22] Loas G, Parker JDA, Otmani O, Verrier A, Fremaux D. Confirmatory factor analysis
of the French translation of the 20-item Toronto Alexithymia Scale. Percept Mot
Skills 1997;85:1018.
[23] Loas G, Otmani O, Fremaux D, Lecercle C, Duflot M, Delahousse J. External validity,
reliability and basic score determination of the Toronto Alexithymia Scales (TAS
and TAS-20) in a group of alcoholic patients. Encéphale 1996;22:35–40.
[24] Taylor GJ, Bagby RM, Ryan DP, Parker JD, Doody KF, Keefe P. Criterion validity of
the Toronto Alexithymia Scale. Psychosom Med 1988;50:500–9.
[25] Beck AT, Steer RA, Ball R, Ranieri W. Comparison of Beck Depression Inventories
-IA and -II in psychiatric outpatients. J Pers Assess 1996;67:588–97.
[26] Centre de Psychologie Appliquée. Validation de la version française du BDI-II.
Paris: Editions du Centre de Psychologie Appliquée; 1996.
[27] Spielberger CD. Manual for the State Trait Anxiety Inventory (form Y). Palo Alto,
Ca: Consulting Psychologists Press; 1983.
[28] Centre de Psychologie Appliquée. Validation de la version française du STAI. Paris:
Editions du Centre de Psychologie Appliquée; 1983.
[29] Parker JDA, Bagby RM, Taylor GJ, Endler NS, Schmitz P. Factorial validity of the 20-
item Toronto Alexithymia Scale. Eur J Pers 1993;7:221–32.
[30] Taylor GJ, Bagby RM, Parker JDA. The 20-item Toronto Alexithymia Scale IV. Reli-
ability and factorial validity in different languages and cultures. J Psychosom Res
2003;55:277–83.
[31] Rieffe C, Oosterveld P, Terwogt MM. An alexithymia questionnaire for children:
factorial and concurrent validation results. Pers Individ Dif 2006;40:123–33.
[32] Sakkinen P, Riittakerttu KH, Ranta K, Haataja R, Joukamaa M. Psychometric prop-
erties of the 20-item Toronto Alexithymia Scale and prevalence of alexithymia in
a Finnish adolescent population. Psychosomatics 2007;48:154–61.
[33] Zimmermann G, Quartier V, Bernard M, Salamin V, Maggiori C. The 20-item
Toronto alexithymia scale: structural validity, internal consistency and preva-
lence of alexithymia in a Swiss adolescent sample. Encéphale 2007;33:941–6.
[34] Parker JD, Eastabrook JM, Keefer KV, Wood LM. Can alexithymia be assessed in ad-
olescents? Psychometric properties of the 20-item Toronto Alexithymia Scale in
younger, middle, and older adolescents. Psychol Assess 2010;22:798–808.
[35] Loas G, Dugré-Lebigre C, Fremaux D, Verrier A, Wallier J, Berthoz S, et al. The Alex-
ithymia Questionnaire for Children (AQC): French translation and validation
study in a convenience sample of 80 children. Encéphale 2010;36:302–6.
[36] Koenigsberg HW, Anwunah I, New AS, Mitropoulou V, Schopick F, Siever LJ. Rela-
tionship between depression and borderline personality disorder. Depress Anxi-
ety 1999;10:158–67.
[37] Luminet O, Bagby RM, Taylor GJ. An evaluation of the absolute and relative stabil-
ity of alexithymia in patients with major depression. Psychother Psychosom
2001;70:254–60.
[38] Grabbe HJ, Spitzer C, Freyberger HJF. Alexithymia and persona lity in relation to di-
mensions of psychopathology. Am J Psychiatry 2004;161:1299–301.
[39] Honkalampi K, Hintikka J, Antikainen R, Lehtonen J, Viinamäki H. Alexithymia in
patients with major depressive disorder and comorbid cluster C personality disor-
ders: a 6-month follow-up study. J Pers Disord 2001;15:245–54.
[40] Pukrop R. Dimensional personality profiles of borderline personality disorder in
comparison with other personality disorders and healthy controls. J Pers Disord
2002;16:135–47.
[41] Fassino S, Amianto F, Gastaldi F, Abbate-Daga G, Brambilla F, Leombruni P. Person-
ality trait interactions in parents of patients with borderline personality disorder:
a controlled study using the Temperament and Character Inventory. Psychiatry
Res 2009;165:128–36.
[42] Parker JD, Taylor GJ, Bagby RM. Alexithymia: relationship with ego defense and
coping styles. Compr Psychiatry 1998;39:91–8.
[43] Haviland MG, Hendryx MS, Shaw DG, Henry JP. Alexithymia in women and men
hospitalized for psychoactive substance dependence. Compr Psychiatry
1994;35:124–8.
[44] Karaklic D, Thuile J, Granger B, Secret I, Bungener C. Predictors of 18-month out-
come for patients with borderline personality disorder. Eur Psychiatry 2011;26:
1626.
[45] Ogrodniczuk JS, Piper WE, Joyce AS. The negative effect of alexithymia on the out-
come of group therapy for complicated grief: what role might the therapist play?
Compr Psychiatry 2005;46:206–13.
[46] Rufer M, Albrecht R, Zaum J, Schnyder U, Mueller-Pfeiffer C, Hand I, et al. Impact of
alexithymia on treatment outcome: a naturalistic study of short-term cognitive-
behavioral group therapy for panic disorder. Psychopathology 2010;43:170–9.
[47] Rasting M, Brosig B, Beutel ME. Alexithymic characteristics and patient-therapist
interaction: a video analysis of facial affect display. Psychopathology 2005;38:
105–11.
[48] Guttman H, Laporte L. Alexithymia, empathy and psychological symptoms in a
family construct. Compr Psychiatry 2002;43:448–55.
152 G. Loas et al. / Journal of Psychosomatic Research 72 (2012) 147–152
