A Randomized Controlled Trial of D-Cycloserine
Enhancement of Exposure Therapy for Social
Adam J. Guastella, Rick Richardson, Peter F. Lovibond, Ronald M. Rapee, Jonathan E. Gaston,
Philip Mitchell, and Mark R. Dadds
Background: Pilot research has suggested that D-cycloserine (DCS) enhances treatment outcomes for anxiety disorders when employed as
an adjunct to exposure therapy (ET). The aim of this study was to determine whether 50 mg of DCS enhances ET for social anxiety disorder
(SAD) according to a comprehensive set of symptom and life impairment measures.
Methods: In a randomized double-blind placebo-controlled trial, we administered 50 mg of DCS or placebo in combination with ET to 56
participants who met primary diagnosis for SAD.
Results: Participants administered DCS reported greater improvement on measures of symptom severity, dysfunctional cognitions, and
life-impairment from SAD in comparison with placebo-treated participants. Effect sizes were mostly in the medium range. Results also
indicated that the amount of adaptive learning about one’s ability to give speeches in front of an audience interacted with DCS to enhance
Conclusions: This study shows that the administration of DCS before ET enhances treatment outcomes for SAD. Results also provide the
first preliminary evidence to suggest that DCS moderates the relationship between a reduction in negative appraisals about one’s speech
performance and improvement in overall SAD symptoms.
Key Words: Cognitive-behavior therapy, D-cycloserine, social anx-
iety, treatment efficacy
when administered immediately before or even shortly after
extinction training (1). It has been suggested that DCS strength-
ens extinction memories so they might be more easily retrieved
during subsequent exposures to fear-relevant cues. On the basis
of these pre-clinical studies, it was suggested that DCS might
enhance the extinction of fear in humans when combined with
exposure-based therapy (ET) procedures (2).
Research shows that DCS has no effect on symptoms of
anxiety when administered chronically to patients over weeks
and months (3). However, when administered acutely in combi-
nation with exposure-based procedures, pilot studies suggest
that DCS enhances treatment outcome. In a first study (4), 27
patients diagnosed with height phobia (acrophobia) were as-
signed to three conditions: placebo, 50 mg DCS, or 500 mg DCS;
and all received two sessions of virtual reality ET. At 1-week and
3-months post-treatment, participants assigned to the DCS con-
dition, regardless of dose, experienced less fear as indicated by
self-reported fear levels in a virtual reality environment, acropho-
bic beliefs, and an increased number of self-exposures to heights
in real-world environments. In a second study (5), 27 patients
diagnosed with social anxiety disorder (SAD) were randomly
n nonhuman animals, Pavlovian conditioning studies have
shown that the partial N-methyl d-aspartate (NMDA) agonist
D-cycloserine (DCS) facilitates the extinction of learned fear
assigned to receive DCS (50 mg) or placebo 1 hour before each
of four public speaking tasks. Speeches were recorded and then
viewed. At the conclusion of therapy and at 1-month follow-up,
participants who received DCS showed larger improvements
relative to participants who received a placebo, as measured by
clinician-rated and self-report ratings of symptom severity. In a
third study (6), 32 patients diagnosed with obsessive-compulsive
disorder were assigned to receive DCS (125 mg) or placebo 2
hours before each of 10 exposure and ritual prevention therapy
sessions. Results indicated that DCS facilitated anxiety reduction
of obsession-related distress after four sessions of ET. The DCS
effects were, however, weaker and nonsignificant over the
following six therapy sessions. The authors concluded that DCS
might be best applied as an adjunct to brief therapy treatment
Our research group has published two studies investigating
DCS effects on fear extinction in non-clinical samples (7,8). In the
first study (7), non-clinical students (n ? 100) who reported
heightened spider fear were given DCS (50 mg or 500 mg) or
placebo in combination with a single-session of ET. Across two
experiments, results showed that DCS did not enhance treatment
outcomes as indicated by behavioral, physiological, or self-report
fear measures taken in the presence of spiders at 3.5 weeks after
treatment. In the second study (8), 238 participants were given
either DCS (50 or 500 mg) or placebo 2–3 hours before extinction
training with a differential shock conditioning paradigm (9).
Over three experiments, DCS was shown to have no clear
influence on fear-extinction as assessed by skin conductance and
self-reported shock expectancy measures. To explain discrepan-
cies between clinical and non-clinical findings, some researchers
(10) have suggested that DCS effects might be more difficult to
show in non-clinical studies, because of potential floor effects
from extinction procedures and/or the use of only a single
administration of DCS in this population.
Although there is promising evidence that DCS enhances
exposure-based treatments for clinical anxiety disorders, clinical
From the School of Psychology (AJG, RR, PFV, MRD); School of Psychiatry
(RMR, JEG), Macquarie University, Sydney, Australia.
Address reprint requests to Adam J. Guastella, Ph.D., School of Psychology,
University of New South Wales, Sydney, NSW 2052, Australia; E-mail:
BIOL PSYCHIATRY 2008;63:544–549
© 2008 Society of Biological Psychiatry
studies have been limited by small sample size. To have strong
evidence that DCS augments anxiety treatments, there is a need
for replication with significantly larger sample sizes. Further-
more, there is a need for process measures to understand how
DCS facilitates ET. The aim of this study was to conduct the first
major replication of DCS-enhanced treatment of SAD for a large
sample of community patients. Social anxiety disorder has an
estimated lifetime prevalence of 12.1% (11). Patients with this
disorder suffer significant impairment in functioning character-
ized by social fear, avoidance, dysfunctional cognitions, and
life-impairment (12,13). We also aimed to extend previous
studies by tracking weekly changes in symptoms to identify the
point in treatment at which DCS enhances ET. This information
might be used by researchers to plan the number of treatment
sessions needed to show DCS effects on outcome. Finally, we
wanted to test the effect of DCS on patients learning to reduce
negative self-appraisals of speech performance. If DCS works by
enhancing adaptive learning during exposure, then reductions in
anxiety and avoidance should be associated with a reduction in
negative self-appraisals about how one appears and communi-
cates in front of an audience during an in-session speech
exposure task. We predicted that patients receiving DCS would
display larger reductions in negative self-appraisals across treat-
ment sessions and that this learning would be more strongly
associated with SAD symptom reduction than in those patients
given a placebo.
Methods and Materials
Following the procedures of previous research (5), we re-
cruited participants recruited from the community if they met
DSM-IV diagnosis for SAD with the Anxiety Disorder Interview
Schedule for Adults (ADIS-IV ) and also reported fear of
public speaking on self-report measures. All participants were
recruited through the University of New South Wales (UNSW)
Psychology Clinic between May and October 2006 (all proce-
dures were approved by the UNSW Human Research Ethics
Committee [#04145]; this trial was registered on the Australian
Clinical Trials Registry ). All assessment and
treatment sessions were conducted by therapists who were
registered or provisionally registered clinical psychologists. All
therapists were supervised and trained by a senior clinical
A total of 110 participants self-referred to the UNSW Psychol-
ogy Clinic from advertisements in local media. Participants who
seemed eligible after an initial phone screen were then assessed
for SAD with the ADIS-IV clinical interview. Exclusion criteria
included: a primary diagnosis of major depression, bipolar
disorder, psychotic disorders (such as schizophrenia), severe
kidney disease, epilepsy, reported pregnancy, current substance
dependence, and current participation in any other psychologi-
cal therapy. Non-psychiatric exclusion criteria were determined
by self-report. Participants were asked to refrain from caffeine,
nicotine, and alcohol on days that they received treatment. Of the
110 participants, 80 adults were eligible for treatment and were
offered participation; 65 accepted (signed informed consent and
completed initial assessments). Trial sample size was based on
previous studies that report moderate–large effects from DCS in
combination with ET (4 – 6). After the first treatment session
(before drug assignment), 9 participants failed to return for the
second and third treatment session and were not included in the
analysis. The remaining 56 participants were randomly assigned
to DCS (n ? 28) or placebo (n ? 28) at the start of the second
therapy session. After drug assignment, 6 participants (1 ? DCS;
5 ? placebo) failed to attend at least three group exposure
sessions between session 2 and 5 and dropped out of treatment.
The ?2analysis showed the difference between the two groups
in drop-out rates after drug assignment approached significance,
(p ? .08). No drop-outs occurred over the 1-month follow-up
assessment period. The progress of participants is shown in
Supplements 1 and 2.
Of the 56 participants (mean age ? 35.48, SD ? 11.35, range ?
18–60) recruited, 57% were male, 76.8% Caucasian, 66.1% single,
and 53.5% had a tertiary degree. The other major ethnic origin
was Asian (10.7%). The t tests and ?2analysis indicated that there
were no differences across drug condition on any sample
characteristics such as gender, age, baseline clinical ratings,
ethnicity, education, or antidepressant use. All participants met
criteria for SAD, and 30.34% of the participants (n ? 17) were
given an additional secondary diagnosis. Of these, 8 participants
were diagnosed with an additional anxiety disorder, 8 were
diagnosed with an additional secondary mood disorder, and 1
was diagnosed with an additional anxiety and mood disorder.
Only 21% (n ? 12) of participants were taking medication, and
they were evenly distributed across DCS and placebo groups.
They were stabilized for a period of at least 6 weeks on
antidepressant drugs (including sertraline hydrochloride [n ? 2],
mirtazapine [n ? 2], and paroxetine hydrochloride [n ? 1]) (n ?
5), immune-suppressant drugs (n ?2), appetite suppressant
drugs (n ? 2), blood pressure medication (n ? 2), or herbal
preparations (n ? 1).
D-cycloserine is approved as an antibiotic for treatment of
tuberculosis by the Therapeutic Goods Administration (15) and
the Food and Drug Administration (United States). Peak blood
levels are reported within 2–8 hours after dosing, and half-life is
estimated at 10 hours (16). The compounding chemist purchased
DCS powder directly from Eli-Lilly (Indianapolis, Indiana) to
make 50-mg DCS capsules, along with identical placebo. We
used 50 mg of DCS in combination with ET, because previous
research had shown it to be an effective dose (4,5) and differ-
ences in treatment efficacy have not been reported between 50
mg and 500 mg (4). A random allocation sequence was gener-
ated by numbering containers with the medication. This random-
ization sequence was developed by the compounding chemist
before the trial and concealed from all individuals involved in
patient care, evaluation, or supervision until follow-up assess-
ments were completed.
All participants received a five-session group ET protocol
based on that used by Hofmann et al. (5). Treatment sessions
were scheduled 1 week apart. Participants were provided with
an exposure-based model of treatment for SAD in the first
treatment session (60 min). At the start of session 2, participants
randomly selected one coded bottle that contained the DCS or
placebo capsules that were to be taken before each ET session.
In sessions 2–5, participants received one blinded-study pill,
waited 1 hour, and then began exposure-based group therapy.
Sessions 2–5 (90 min each) involved giving increasingly difficult
speeches about topics chosen by the therapists in front of the
other group members and a video camera. All participants gave a
speech, and recorded speeches were played back to the group. At
the beginning and conclusion of each exposure-based session,
A.J. Guastella et al.
BIOL PSYCHIATRY 2008;63:544–549 545
home-practice exposure strategies were reviewed and encouraged
(e.g., giving speeches in front of a mirror). Homework was consid-
ered part of treatment. Participants were asked to refrain from
alternative treatment until the final assessment session.
conducted within the month preceding the initial treatment
session and 1 month after the completion of the last treatment
session. The DSM-IV diagnoses were determined with the
ADIS-IV (14). The ADIS-IV contains diagnostic questions about
each anxiety disorder and other diagnostic categories that are
important for differential diagnosis (e.g., affective disorders and
substance abuse/dependence). Upon completion of diagnostic
assessments, clinicians completed a Global Assessment of Func-
tioning (GAF) on a scale of 0–100, in accordance with Axis-V of
the diagnostic axis for DSM-IV (17).
Self-report measures were obtained immediately before the
two clinical assessment interviews and immediately after the
completion of the last treatment session (treatment session 5).
Following previous research (5), we used two self-report mea-
sures to assess SAD symptoms: the Social Phobia and Anxiety
Inventory (SPAI ), and the Liebowitz Social Anxiety Scale
(LSAS ). In addition, we employed other measures to assess
additional dimensions of SAD. Firstly, the Brief Fear of Negative
Evaluation Scale (BFNE ) assesses how characteristic certain
dysfunctional thoughts relating to negative evaluation are to an
individual (e.g., “I am afraid that others will not approve of me”).
Responses are measured on a 5-point Likert scale of how
characteristic the statement is (1 ? “not at all,” and 5 ?
“extremely”). Secondly, the Life Interference Scale (LIS )
provides a measure of the impact of an individual’s social fears
on various components of their life, including work, family life,
and leisure activities (e.g., “My social anxiety interferes with my
home/family life”; “My social anxiety makes my day-to-day living
unpleasant”). Scores are assessed on a 9-point likert scale (0 ?
“not at all,” and 8 ? “extremely”) with the total score ranging
from 0 (no interference) to 48 (extreme interference). Past
analysis from members of our team shows the scale demonstrates
excellent internal consistency (? ? .90), and the total score
correlates with other life-impairment measures (21).
Process Measures. We administered the Credibility/Expect-
ancy Questionnaire (CEQ ) to measure expectancy and cred-
ibility of treatment. Participants were asked to rate the perceived
logic of treatment on a 9-point scale (1 ? “not at all logical,” 9 ?
“very logical”) after the first treatment session. Before each
session and before drug administration, participants completed a
weekly version of LSAS (19). Participants also completed the
Speech Performance Questionnaire (SPQ ) immediately after
each in-session speech exposure task (Sessions 2–5). The SPQ is
thought to tap into the common negative self-appraisals dis-
played by SAD patients when performing in front of a group
(e.g., “Seemed to tremble or shake”; “Seemed nervous”; “Made a
good impression” ) on a 5-point scale (1 ? “not at all,” 5 ?
“very much”). A higher total score (0–68) indicated a more
positive view of one’s performance. The scale demonstrates
good internal consistency (? ? .79–.86) and inter-rater reliability
(22,24,25) and differentiates between patients with SAD, treated
SAD patients, and non-clinical participants (23).
Clinical assessment interviews were
Data were entered by a research assistant blind to drug
assignment and analyzed with the SPSS statistical software pack-
age (SPSS V14; Chicago, Illinois). Last observation carried for-
ward was used to replace missing data. We employed intention-
to-treat analysis for all participants assigned to receive DCS or
placebo throughout. For outcome analysis, a Drug (DCS, pla-
cebo) ? Time (pretreatment, post-session 5 treatment, 1-month
follow-up) repeated-measures multivariate analysis of variance
(MANOVA) was conducted on total scores from the SPAI, LSAS,
BFNE, and LIS. We also conducted a Drug (DCS, placebo) ?
Time (pretreatment, 1-month follow-up) repeated-measures
analysis of variance (ANOVA) on the clinician-rated GAF scores.
We computed effect sizes (Cohen’s d ) by dividing the
difference between the mean change of the DCS group and the
mean change of the placebo group by the pooled SD. To
determine the number of treatment sessions required to show
differences between DCS and placebo groups, we ran a Drug
(DCS, placebo) ? Treatment Session (Sessions 1–5) repeated-
measures ANOVA on total scores of the LSAS obtained before
each therapy session. In addition, we conducted a Drug (DCS,
placebo) ? Exposure Treatment Session (Sessions 2–5) repeated-
measures ANOVA to determine whether there were differences
between drug groups on the SPQ immediately after all four
speech tasks. Finally, we tested the relationship between change
in negative appraisals of the speech task and change in fear and
avoidance in the real world and investigated whether DCS
moderated this relationship. Moderation of these two variables, A
(Change on SPQ) and C (LSAS Session 5), by a categorical
variable B (Drug Group), was tested with regression in which A,
B, and the product term AB were used to predict C (27,28). If AB
is a significant predictor over and above A and B, then B is
moderating A’s effect on C.
Participant Beliefs About Treatment
Participants assigned to both drug conditions reported no
significant side effects. The t tests showed there was no differ-
ence between drug conditions at any time point in regard to
guessing what treatment they were receiving [Largest t(54) ? .88,
p ? .38]. On average, 36.5% of DCS-assigned participants and
39.5% of placebo-assigned participants believed they had re-
ceived DCS. Participants were asked about both their expecta-
tions and credibility for the treatment approach after the educa-
tional session (session 1). Participants’ ratings on these scales
were moderate–high and not significantly different between the
two groups [Largest t(54) ? .46, p ? .65].
Table 1 shows the means and SDs of the four self-report
outcome measures (SPAI, LSAS, BFNE, LIS) at pretreatment, post-
treatment, and 1-month follow-up in the two groups. A Drug (DCS,
placebo) ? Time (Pre, Post, 1-month follow-up) repeated-measures
MANOVA with all total scores from self-report outcome measures
revealed a main effect of Time [F(8,47) ? 14.77, p ? .001].
Examination of 95% confidence intervals for all measures showed
that there was a reduction in self-reported SAD symptoms from
pretreatment to immediately after treatment and this difference
with pretreatment scores was maintained at follow-up. There
was no main effect for Drug, [F ? 1.0] on these measures but
there was a significant Drug ? Time interaction [F(8,47) ? 3.75,
p ? .002]. Examination of 95% confidence intervals for all
measures, except the SPAI, showed that DCS-treated participants
showed a greater reduction in SAD symptoms than placebo-
treated participants from pretreatment to immediately after treat-
546 BIOL PSYCHIATRY 2008;63:544–549
A.J. Guastella et al.
ment and this effect was maintained at follow-up (Table 1).
When the analysis was restricted to those participants who
completed treatment (DCS ? 27, placebo ? 23), the Drug ?
Time interaction remained significant and in the same direction
[F (8,41) ? 3.07, p ? .008]. Effect sizes on all measures (.42–.70)
except the SPAI (.22–.26) were in the moderate range (Figure
1)—analysis on SPAI sub-scales show effects sizes indicating
DCS improvement in comparison with placebo [Social Phobia
Subscale: Post ES ? .42; follow-up ES ? .45; Agoraphobia
Subscale: Post ES ? .48; follow-up ES ? .52]; the SPAI total or
difference score reflects the subtraction of the Agoraphobia
sub-scale score from the Social Phobia sub-scale score; this
subtraction lowered the overall effect size. Table 1 also shows
the means and SDs for GAF ratings made from clinical interviews
conducted pretreatment and at 1-month follow-up. A Drug (DCS,
placebo) ? Time (Pre, 1-month follow-up) repeated-measures
ANOVA comparing pretreatment with follow-up clinician-rated
GAF scores revealed a main effect for Time [F(1,54) ? 31.42, p ?
.001], no main effect for Drug [F(1,54) ? 1.79, p ? .19], and a
significant Drug ? Time interaction [F(1,54) ? 4.78, p ? .03].
Treatment improved clinician ratings of general life functioning,
and these improvements were greater in the DCS-treated group
in comparison with the placebo-treated group. The effect size
was in the moderate range (Figure 1).
Weekly Symptom Tracking
A Drug (DCS, placebo) ? Treatment Session (Sessions 1–5)
repeated-measures ANOVA on LSAS total symptom scores indi-
cated a main effect of Treatment Session [F(4,51) ? 8.73, p ?
.001]. As expected, anxiety symptoms scores decreased over the
five treatment sessions. There was no main effect of Drug
[F(1,54) ? 1.45, p ? .23], but there was a significant Drug ?
Treatment Session interaction [F(4,51) ? 3.00, p ? .02]. As can be
seen in Figure 2, DCS-treated participants reported a greater
reduction across treatment sessions than placebo-treated partic-
ipants (results on both fear and avoidance sub-scales of the LSAS
were similar). Follow-up t tests, with a Bonferroni adjustment,
confirmed that the difference between drug conditions was
significant only at the fifth-session assessment point [t(54) ?
2.54, p ? .01].
A Drug (OT, placebo) ? Exposure Treatment Session (Ses-
sions 2–5) repeated-measures ANOVA conducted on the SPQ
indicated a main effect for Session [F(3,52) ? 10.23, p ? .001]. As
expected, participants’ appraisals of their speech performance
improved as sessions progressed. There was no main effect of
Drug [F ? 1.0] or an interaction effect between Session and Drug
[F ? 1.0]. To address our hypothesis that participants receiving
DCS would show a stronger association between appraisals of
their speech performance and subsequent anxiety reduction,
change scores were created by subtracting the score reported in
session 5 from the score reported in session 2; for the SPQ, lower
scores indicated greater perceived improvements; for the LSAS,
higher scores meant greater reductions in social anxiety. In the
DCS group, significant negative correlations were found between
change scores on the LSAS and the SPQ (r ? ?.38, p ? .04), but
this was not significant in the placebo condition (r ? ?.08, p ?
.73). This suggests that the more participants assigned to DCS
believed that their performance had improved across the treat-
Table 1. Means and SDs for Self-Report and Clinician Ratings Across
SPAI (range ? 27–161)
LSAS (range ? 12–140)
BFNE (range ? 21–60)
LIS (range ? 4–46)
GAF (range ? 40–80)
SPAI, Social Phobia Anxiety Inventory (Total Score); DCS, D-cycloserine;
LSAS, Leibowitz Social Anxiety Scale (Total Score); BFNE, Brief Fear of Nega-
Functioning (Clinician Rated).
Figure 1. Controlled effect sizes on self-report and clinician-rated outcome
Anxiety Scale; BFNE, Brief Fear of Negative Evaluation Scale; LIS, Life Inter-
ference Scale; GAF, Global Assessment of Functioning.
Figure 2. Total scores on the Liebowitz Social Anxiety Scale for each drug
group before each treatment session. DCS, D-cycloserine.
A.J. Guastella et al.
BIOL PSYCHIATRY 2008;63:544–549 547
ment sessions, the greater their reduction in general social fear
and avoidance symptoms. In contrast, there was no relationship
between these variables in the placebo group.
To test whether DCS was moderating the relation between
SPQ and the LSAS, we ran a regression (27,28) to predict the
Session 5 LSAS score with pretreatment LSAS scores in block 1,
Drug (DCS ? 1, placebo ? ?1) and change on the SPQ in block
2, and the interaction between Drug and Change on the SPQ in
block three (Table 2). The model predicted 69% of the variance
[F(4,51) ? 31.81, p ? .001]. Lower Pre-LSAS scores, being
assigned to the DCS group, and reporting improvement between
sessions 2 and 5 on the SPQ were all positively associated with
lower LSAS scores at the final treatment session. More impor-
tantly, after controlling for these factors, the interaction between
change on the SPQ and Drug was significant. This suggests DCS
moderated improvement such that the relationship between
improvement on the SPQ and anxiety reduction was stronger in
the DCS group in comparison with placebo.
The present study was a double-blind placebo-controlled trial
of DCS to augment ET with SAD and had a sample twice as large
as any previous studies. The ET reduced SAD symptoms regard-
less of whether participants received DCS or placebo. However,
the results of this trial also showed that DCS enhanced ET
treatment further, as indicated by a range of SAD symptom
measures, including social fear, avoidance, dysfunctional cogni-
tions, life impairment from SAD, and clinician ratings of the
quality of general life functioning.
Effect sizes on most outcome measures (i.e., LSAS, BFNE, LIS,
and GAF but not the SPAI [see note on SPAI in Results: Symptom
Outcome section]) were in the moderate range. In comparison
with the previous SAD study (5), our data replicates closely what
was found on the LSAS: moderate effect sizes showing greater
improvement in the DCS group compared with placebo at
post-treatment assessment, with effects maintained at 1-month
follow-up. In contrast, our treatment outcomes on the SPAI were
moderate and substantially smaller than the large effect size
reported previously (5).
Weekly tracking measures indicated that DCS facilitated the
reduction of social fear and avoidance gradually. Significant
differences between DCS and placebo drug groups emerged at
the fifth-session assessment point, which was after the third DCS
exposure-treatment session. In combination with the findings of
Kushner et al. (6), these results suggest that using DCS acutely as
an adjunct with three or four ET sessions provides the best
opportunity to show DCS enhancement on anxiety treatment. It
should be noted that effects of DCS might be detected at an
earlier session if a larger sample was employed. Results also
showed that participants were not able to identify whether they
had received DCS, because the number of participants who
believed they had taken DCS was low and there was no
difference between drug groups. This might suggest that partic-
ipants might be more likely to attribute therapeutic gain to the
therapy process itself rather than to any acute effect from the
Animal research has shown that DCS consolidates fear-extinc-
tion learning (1). Attempts to identify the mechanisms of how
DCS enhances loss of fear in humans have provided few insights
(7,8). This study provides the first preliminary evidence that the
amount of adaptive learning between exposure treatment ses-
sions was associated with DCS effects in humans. For participants
given DCS, a relationship was found between improvements
from session 2 to 5 on appraisals about participants’ speech
performance and reported improvements in social fear and
avoidance symptoms in the real-world. Thus, the present data
extend current knowledge of the effect of DCS by suggesting that
anxiety reductions are associated with adaptive learning that one
can perform in front of a group successfully and that this learning
might be facilitated by DCS’s action as a partial agonist at the
NMDA receptor (29,30). It is interesting to note that DCS did not
directly increase the amount of learning that took place within
each exposure session, as indicated by self-ratings of speech
performance. This finding is not inconsistent with the animal
literature that shows that DCS does not enhance within-session
extinction (1). Our data suggest that DCS facilitates the general-
ization of learning that occurs during within-session extinction
and thus is more likely to lead to broader and more durable
reductions of SAD symptoms. Finally, our findings do not
provide evidence of causality. Future research might wish to
manipulate the degree of learning that occurs in each exposure
session to better evaluate the causal role of DCS on learning and
subsequent anxiety reduction.
Overall, this trial confirms and extends previous demon-
strations that DCS enhances ET for anxiety disorders in
humans. This research adds further support to a radical new
approach to the treatment of anxiety disorders by enhancing
the adaptive learning that occurs in therapy via medication.
Research is now required across various anxiety disorders in
non-specialist community-based clinics to better determine
the potential impact of DCS on lowering the burden of disease
from anxiety disorders.
This research was supported by a project Grant from the
National Health and Medical Research Council (#350963).
We thank Dr. Stefan Hofmann for providing treatment man-
uals and advice upon which our procedures were based; Alice
Shires, the UNSW Psychology Clinic Manager; and our re-
search assistants, Subodha Willamaleera, Rochelle Cox, Alex
Howard, Scott Nash, Elizabeth Stanton, Nicola Rich, and
Stephanie Heffner. We also acknowledge Prof. Richard Bryant
for his helpful comments on this manuscript; Dr. Kevin Bird
for statistical advice; and Dr. David Sutherland, who served
as the external medical monitor.
Table 2. Degree Speech Performance and Drug Group Predicts Social
Fear and Avoidance at Session 5
Ch Speech Perf (centered)
Ch Speech Perf Drug
F(4,51) ? 31.81, p ? .001.
B, Standardized beta-weight; LSAS, Leibowitz Social Anxiety Scale; Ch
Speech Perf, Change in Speech Performance Ratings (Sessions 5–2); Drug,
ap ? .001.
bp ? .05.
548 BIOL PSYCHIATRY 2008;63:544–549
A.J. Guastella et al.
Drs. Guastella, Richardson, Lovibond, Rapee, Gaston, Mitch-
ell, and Dadds report no biomedical financial interests or
potential conflicts of interest.
Trial registry: A Randomized Controlled Trial to Evaluate the
Effect of D-Cycloserine in Combination With Exposure Therapy
in the Treatment of Social Anxiety Disorder to Improve the
Severity of Social Anxiety Disorder Symptoms
Registry number: 012606000352505
Registry URL: http://www.actr.org.au/
Supplementary material cited in this article is available
1. Richardson R, Ledgerwood L, Cranney J (2004): Facilitation of fear ex-
tinction by D-cycloserine: Theoretical and clinical implications. Learn
tyric acid in fear extinction: Clinical implications for exposure therapy.
al. (2002): Pilot-controlled trial of D-cycloserine for the treatment of
post-traumatic stress disorder. Int J Neuropsychopharmacol 5: 301–307.
E, et al. (2004): Cognitive enhancers as adjuncts to psychotherapy: Use
K, et al. (2006): Augmentation of exposure therapy for social anxiety
disorder with D-Cycloserine. Arch Gen Psychiatry 63:298–304.
6. Kushner MG, Kim SW, Donahue C, Thuras P, Adson D, Kotlyar M, et al.
pulsive disorder. Biol Psychiatry 62:835–838.
randomized controlled trial of the effect of D-Cycloserine on exposure
therapy for spider fears. J Psychiatr Res 41:466–471.
ing and extinction in humans. Behav Res Ther 45:663–672.
9. Lovibond PF (1992): Tonic and phasic electrodermal measures of hu-
in the National Comorbidity Survey Replication. Arch Gen Psychiatry
nity sample: Prevalence, impact on functioning, and diagnostic classifi-
in a community sample of adolescents and young adults: Prevalence,
risk factors and co-morbidity. Psychol Med 29:309–323.
14. Brown T, Di Nardo PA, Barlow DH (1994): Anxiety Disorders Interview
Schedule Adult Version (ADIS-IV). San Antonio, Texas: Psychological Cor-
poration/Graywind Publications Incorporated.
15. Australian Government Department of Health and Ageing (2007): Aus-
tralian Registrar of Therapeutic Goods, Canberra. Available at: http://
www.tga.gov.au/. Accessed December 19, 2007.
of Therapeutics, 10th ed. New York: McGraw Hill.
Text Revision. Washington DC: American Psychiatric Association.
tory: Manual. Toronto, Ontario: Multi-Health Systems.
19. Leibowitz MR (1987): Social phobia. Mod Probl Pharmacopsychiatry 22:
21. Rapee RM, Abbott MJ, Baillie AJ, Gaston JE (2007): Treatment of social
phobia through pure self help and therapist-augmented self help. Br J
of performance in social phobics. J Abnorm Psychol 101:728–731.
23. Abbott M, Rapee R (2004): Post-event rumination and negative self-
24. Rodebaugh TL, Chambless DL (2002): The effects of video feedback on
self-perception of performance: A replication and extension. Cogn Ther
25. Rapee RM, Hayman K (1996): The effects of video feedback on the
self-evaluation of performance in socially anxious subjects. Behav Res
26. Cohen J (1988): Statistical Power Analysis for the Behavioral Sciences.
Hillsdale, New Jersey: Lawrence Erlbaum Associates.
27. Cohen J, Cohen P, West S, Aiken L (2003): Applied Multiple Regression/
Correlation Analysis for the Behavioral Sciences, 3rd ed. Hillsdale, New
Jersey: Lawrence Erlbaum Associates.
28. Aiken LS, West SG (1991): Multiple Regression: Testing and Interpreting
Interactions. Newbury Park, California: Sage.
a positive modulator of the N-methyl-D-aspartate receptor, enhances
30. Hood WF, Compton RP, Monahan JB (1989): D-cycloserine: A ligand for
the N-methyl-D-aspartate coupled glycine receptor has partial agonist
characteristics. Neurosci Lett 98:91–95.
A.J. Guastella et al.
BIOL PSYCHIATRY 2008;63:544–549 549