Need for speed: Evaluating slopes of OCD recovery in behavior therapy
enhanced with D-cycloserine
Gregory S. Chassona, Ulrike Buhlmannb, David F. Tolinc, Sowmya R. Raoa, Hannah E. Reesea,
Theresa Rowleya, Kaitlyn S. Welsha, Sabine Wilhelma,*
aMassachusetts General Hospital/Harvard Medical School, USA
bHumboldt Universitaet zu Berlin, Germany
cInstitute of Living/Yale University School of Medicine, USA
a r t i c l e i n f o
Received 9 September 2009
Received in revised form
9 March 2010
Accepted 10 March 2010
Exposure and response prevention
a b s t r a c t
Evidence suggests that the antibiotic D-cycloserine (DCS) enhances the treatment effects of exposure and
response prevention (ERP) for ObsessiveeCompulsive Disorder (OCD). Further, evidence suggests that
the effects of DCS diminish partway through treatment, but it is unclear to what extent. In an effort to
evaluate these issues, the current study re-analyzes data from a 10-session randomized controlled trial of
ERP þDCS versus ERPþplacebo in a sample of 22 adults with OCD. We analyzed repeated-measures
mixed models with random slopes and intercepts across different intervals: sessions 1e10, 1e5, and
6e10. The results indicate that the course of ERP was 2.3 times faster over the full 10 sessions for the DCS
compared to the placebo group, and nearly six times quicker in the first half of ERP. Further interpre-
tation of the results suggests that DCS does not amplify the effects of ERP, but instead initiates treatment
effects sooner in treatment. In addition, DCS does not necessarily lose its effect over repeated use, but
instead may exhaust its maximum utility after effectively jump-starting ERP. Ultimately, DCS may
provide a means for curtailing treatment costs, decreasing treatment dropout and refusal rates, and
enhancing access to care.
? 2010 Elsevier Ltd. All rights reserved.
Exposure and response prevention (ERP) is considered the gold
standard psychological intervention for ObsessiveeCompulsive
Disorder (OCD). ERP is thought to work by facilitating fear extinc-
tion (i.e., the behavioral unlearning of acquired fears) through
systematic and prolonged exposure to anxiety-provoking stimuli
and the simultaneous prevention of fear-reducing physical and
mental actions (i.e., rituals or avoidances). Approximately 10e20
sessions is associated with symptom reduction in at least 85% of
patients, with 55% reporting substantial improvement (Jenike,
2004). However, these success rates can be misleading because
they do not account for the high rates of patient refusal or dropout.
Indeed, of those patients who initially seek behavior therapy for
OCD, 25% of patients decline participation in ERP and a similar
percentage of patients drop out once initiated (Jenike, 2004;
Schruers, Koning, Luermans, Haack, & Griez, 2005). Patient refusal
and dropout occurs for a variety of reasonsdfor example, ERP can
be time consuming, costly, and anxiety provoking.
Consistent with this effort to decrease ERP refusal and dropout,
recent research has focused on the use of an antibiotic, D-cyclo-
serine (DCS;Seromycin), as an adjunctive pharmacological
approach. When paired with exposure therapy, DCS purportedly
facilitates fear extinction and enhances associative learning by
acting as a partial agonist at N-methyl-D-aspartate receptors in the
brain (Rothbaum, 2008). A recent meta-analysis by Norberg,
Krystal, and Tolin (2008) quantitatively suggests that DCS
enhances fear extinction in animal models, as well as in human
anxious populations, including those with specific phobia (Ressler
et al., 2004) and social anxiety disorder (Hofmann et al., 2006).
Three studies of DCS and ERP for OCD have been published to
date: Kushner et al. (2007), Storch et al. (2007), and Wilhelm
et al. (2008). While both the Kushner et al. (2007) and the
Storch et al.(2007) studies
randomized control trial of ERP þDCS versus ERP þplacebo, the
results were conflicting. For Kushner et al.’s study, participants
were administered 125 mg of DCS or placebo two hours prior to
each of 10 twice-weekly sessions. Session-by-session treatment
* Correspondence to: Sabine Wilhelm, Massachusetts General Hospital, OCD and
Related Disorders Program, 185 Simches Research Bldg., Floor 2, Boston, MA 02114,
USA. Tel.: þ1 (617) 726 6766; fax: þ1 (617) 643 3080.
E-mail address: firstname.lastname@example.org (S. Wilhelm).
Contents lists available at ScienceDirect
Behaviour Research and Therapy
journal homepage: www.elsevier.com/locate/brat
0005-7967/$ e see front matter ? 2010 Elsevier Ltd. All rights reserved.
Behaviour Research and Therapy 48 (2010) 675e679
outcome was measured using the Subjective Units of Distress
Scale. The study found that the ERP þDCS group demonstrated
quicker reduction in symptom severity through the first four
sessions, but the ERP þplacebo group caught up during the
remaining sessions, suggesting that DCS facilitates fear extinction
for only a limited set of initial ERP sessions.
Storch and colleagues alternatively administered 250 mg of DCS
or placebo four hours prior to each of 12 weekly ERP sessions. The
investigators utilized the gold standard measure of OCD sever-
itydthe Yale-Brown ObsessiveeCompulsive
Goodman et al., 1989a, 1989b)dto measure treatment outcome at
every other session. In contrast to the Kushner et al. study, the
Storch et al. investigation found no statistically significant group
differences at post-treatment or two-month follow-up, and they
also found no differences in the rate of symptom reduction across
sessions. However, Storch and colleagues did find evidence for
a trend in which DCS resulted in greater treatment gains compared
to placebo from sessions 4 through 6, whereas the placebo group
paradoxically demonstrated a trend for greater improvement
compared to the DCS group from sessions 8 to 10.
Wilhelm et al. (2008) conducted a study with the same group
design but with 10 twice-weekly ERP sessions. The investigation
also addressed some of the methodological concerns that may have
100 mg, and they used the clinician-rated Y-BOCS as the standard-
ized measure of OCD symptom severity at pretreatment, post-
treatment, and 1-month follow-up. Despite non-significant group
differences at pretreatment, the DCS group reported significantly
lower OCD symptom severity halfwaythrough treatment relative to
the placebo group. The group differences were non-significant at
posttreatment and one-month follow-up, which was consistent
with Kushner et al.’s (2007) finding that the ERPþplacebo group
caught up to the ERPþDCS group during treatment.
For the Kushner et al. (2007) and Wilhelm et al. (2008) studies,
the results indicate significant group differences roughly midway
through treatment, indicating that DCS initially leads to quicker
reduction in symptom severity. In addition, the Storch et al. (2007)
investigation found a corroborating trend suggesting that DCS
midpoint of treatment. Thus, this line of research has elicited some
unanswered questions: how much quicker does DCS affect OCD
recovery compared to placebo, and to what extent does the rate of
recovery differ between the two halves of treatment? While the
Storch et al. (2007) trial assessed symptom severity at every other
session, they had little reason to evaluate the speed of recovery (i.e.,
slopes of symptom reduction acrosssessions)since theyfoundnon-
significant group differences. Kushner and colleagues provided
some data to evaluate these questions in a graph illustrating OCD
severity reduction per group across the first four sessions. The DCS
group appears to have recovered approximately twice as fast in that
time frame. However, these results were based on the Subjective
Unit of Distress Scale rather than a standardized measure of
symptom severity, such as the Y-BOCS. Kushner and colleagues also
did not provide information pertaining to severity reduction after
the fourth session, presumably because they only found significant
group differences at that session. Furthermore, Wilhelm and
colleagues did not report session-by-session OCD severity reduc-
tions that would allow for calculations of recovery speed. Thus, the
differential rate of reduction in OCD symptoms with ERPþDCS
relative to ERPþplacebo remains unknown. Additionally, differ-
halves of treatment.
In an effort to determine how much more quickly the DCS group
experienced areduction in OCDseverity,the current studyfollowed
up on the published findings from Wilhelm et al. (2008) using the
same dataset. We evaluated slopes of symptom reduction that were
derived from a session-by-session assessment of severity using
a self-report version of the Y-BOCS. Using linear mixed-effects
random slopes and intercepts modeling, three different sets of
slopes were compared across the group receiving ERP þDCS versus
the group receiving ERPþplacebo: slopes from sessions 1 through
10, sessions 1 through 5, and sessions 6 through 10.
According to Kushner et al. (2007), Wilhelm et al. (2008), and
Norberg et al. (2008), DCS appears to facilitate ERP for OCD
primarily in the first set of sessions. Thus, we hypothesized that the
speed of recovery would be significantly faster for the ERPþDCS
group relative to ERPþplacebo in the first half of treatment only.
Evaluating specific speeds of recovery may help uncover the
underlying mechanisms of DCS. In addition, the success of ERP is
hindered in large part by the high percentages of patient refusal
and dropout. Thus, the current study may provide critical infor-
approachdinformation that potentially can be communicated to
ambivalent patients in an effort to make ERP more consumer-
The sample consisted of 22 adult patients with a primary
diagnosis of OCD who were recruited from the community. The
study was conducted at two sites: Massachusetts General Hospital
(MGH) in Boston, or the Institute of Living in Hartford, Connecticut.
Initially, 33 patients signed consent to participate in the study;
three did not meet inclusion criteria, and one refused treatment
from the outset. Thus, 29 participants initiated the treatment.
Seven individuals dropped out before the midpoint of treatment
(four in the DCS group and three in the placebo group). Of the 22
patients included in the current study, 10 received ERPþDCS and
12 patients received ERPþplacebo. Sample demographics are
provided in Table 1.
According to Wilhelm et al. (2008), after randomization the two
treatment groups did not differ significantly with respect to age,
pretreatment OCD and depression severity, and proportion of
patients receiving augmentative psychiatric medication. More
specific descriptive statistics about the sample, including rates of
comorbidity, can be found in the original publication (Wilhelm
et al., 2008).
Written consent was obtained after each participant was fully
informed about the procedures of the study. Participants received
an assessment at pretreatment, after the fifth session, posttreat-
ment, and one month after treatment termination. At pretreat-
ment, participants received a diagnostic interview from a master’s
or doctoral-level clinician, and also met with study physicians to
determine if the use of DCS was contraindicated. Patients returned
for one psycho-education and treatment planning session prior to
Age M (SD)a
aMean (standard deviation).
G.S. Chasson et al. / Behaviour Research and Therapy 48 (2010) 675e679
the initiation of 10 twice-weekly, 60-min sessions of ERP, following
procedures from Kozak and Foa (1997). Advanced trainees under
the supervision of licensed clinical psychologists administered the
ERP protocol. One hour before each session, participants received
either a 100 mg pill of DCS or placebo. Medication was adminis-
tered in a double-blind fashion. Participants also completed a self-
report version of the Yale-Brown ObsessiveeCompulsive Scale
(Baer, Brown-Beasley, Sorce, & Henriques, 1993; Steketee, Frost, &
Bogart, 1996) before each session. All research procedures were
conducted in accordance with the policies set forth by each insti-
tution’s internal review board. For more details about the study
procedure, please see Wilhelm et al. (2008).
Structured clinical interview for DSM-IV axis I disorders, patient
edition (First, Spitzer, Gibbon, & Williams, 1995)
The SCID is a widely used semi-structured, clinician-adminis-
tered interview designed to establish psychiatric diagnoses based
on criteria from the Diagnostic and Statistical Manual of Mental
Disorders, Fourth Edition (APA, 2000).
Yale-Brown obsessiveecompulsive scale-self-report (Baer et al.,
1993; Steketee et al., 1996)
The Y-BOCS-SR was derived from the clinician-administered
version (Goodman et al., 1989a, 1989b), which is considered the
gold standard assessment of OCD severity (Grabill et al., 2008). The
instrument consists of 10 items measuring the severity of obses-
sions and compulsions using a five-point scale with higher scores
reflecting more severity. The Y-BOCS-SR has demonstrated excel-
lent reliability and validity (Steketee et al., 1996).
We fit a linear mixed effects model with random intercept and
slopes. We categorized the time into two groups: sessions 1e5 (i.e.,
first half) and 6e10 (i.e., second half). We fit a model that incor-
porated treatment group (class variable), subject level random
intercept, and a random group-by-time interaction term and used
contrasts to test the following: (i) whether the slopes between the
two treatment groups were significantly different over the 10
sessions; (ii) whether the slopes between the two treatment groups
were significantly different in the first half, and then separately for
the second half, and (iii) whether the difference in the slopes
between the two treatment groups in the first half was significantly
different from the difference in the slopes between the two treat-
ment groups in the second half. For each test, we then computed
a ratio of slopes (i.e., slope for the DCS group divided by the slope
for the non-DCS group) to provide an estimate of how much more
quickly the DCS group exhibited a reduction in OCD severity.
We did not adjust for any covariates since the characteristics of
the patients were not significantly different between the two
groups (Wilhelm et al., 2008). Of the 220 observations, only 18 were
missing. We conducted sensitivity analysis by (i) excluding the
observations, and (ii) imputing a single value based on recom-
mendations from Engels and Diehr (2003). The results were very
similar in both analyses, so we opted to utilize the non-imputed
A mixed-effects model with random slopes and intercepts
revealed a statistically significant group-by-time interaction for the
full 10 sessions of treatment, F(4,137)¼8.71, p<.0001. The ratio of
slopes suggests that the ERPþDCS group recovered 2.3 times faster
than the ERPþplacebo group. Fig. 1 illustrates average OCD
severity scores at each session across the full ten sessions of
Within the mixed-effects model, the slopes of the two groups
also were compared through the fifth session only. The results
indicate a significant group-by-time interaction for sessions one
through five, t(4, 137)¼2.02, p<.05. The ratio of slopes indicates
that ERPþDCS group recovered 5.98 times faster than the
ERPþplacebo group in the first five sessions. The slopes of the
two groups were also compared in the second half of treatment.
The results revealed a non-significant group-by-time interaction,
t(4, 137)¼0.66, p¼ .15. Fig. 2 demonstrates group differences in
slopes in the first and second half of treatment. Unstandardized
slopes and the ratio of slopes for each half of treatment are pre-
sented in Table 2.
We conducted a contrast as part of the mixed-effects model
comparing the first half between-group difference in slopes versus
the second half between-group difference in slopes. The result was
non-significant, t(4, 137)¼1.07, p¼.29, suggesting that the differ-
ence in slopes between the two groups did not significantly differ
between the two halves of treatment.
We carried out additional exploratory analyses to further clarify
our findings. We first tested the within-group differences in slopes
between the two halves to clarify if each group’s slope of recovery
changed over the full course of treatment. The ERPþDCS group did
not reveal a statistically significant difference in slope between the
two halves, t(4, 137)¼?.61, p¼ .54. Similarly, the ERPþplacebo
group did not demonstrate a statistically significant change in slope
from the first to second half of treatment, t(4, 137)¼.93,
We then evaluated whether the slopes of each groupin each half
significantly differed from zero. The ERPþDCS group’s slope was
significantly different from zero in the first half, t(4, 137)¼?3.20,
p<.01, and the second half, t(4,137)¼?3.13, p<.01. However, the
ERPþplacebo group’s slope was non-significantly different from
zero in the first half, t(4, 137)¼?0.62, p¼.54, but significantly
different in the second half, t(4,137)¼?2.48, p<.05. This suggests
that ERPþplacebo treatment was relatively inert in the first half of
Lastly, we computed one final contrast to help elucidate
whether or not 1) DCS amplifies treatment effects, 2) DCS initiates
standard ERP effects sooner in treatment, or 3) both. We evaluated
whether the speed of recovery rate for the ERPþDCS group at its
quickest (i.e., first half slope) was significantly different from the
speed of recovery for the ERPþplacebo group at its quickest (i.e.,
second half slope). The results indicated a non-significant differ-
ence, t(4,137)¼?1.26, p¼.21, suggesting that DCS does not boosts
the effects of ERP, but instead seems to initiate the effects of ERP
sooner in treatment.
Fig. 1. Mean Y-BOCS-SR scores per session per group.
G.S. Chasson et al. / Behaviour Research and Therapy 48 (2010) 675e679
In the current study, we examined the rate at which OCD
improves when ERP is paired with DCS. Our results suggest that
a course of ERP is 2.3 times faster than a course of ERP without DCS.
By the eighth treatment session, the ERPþDCS group scored below
a YBOCS-SR clinical cutoff of 16 (Steketee et al., 1996), whereas the
ERPþplacebo group was still above the cutoff by the end of the
treatment trial. Since the trial only lasted 10 sessions, it is unclear
how many additional sessions would have been necessary for the
ERPþplacebo group to reach this clinical cutoff. Evidence from
previous studies suggests that ERP often requires more than eight
sessions to reduce OCD severity to such a level. For example, one
study of ERP (without DCS or concomitant pharmacotherapy)
found that participants still remained above the clinical cutoff on
the clinician-administered version of the YBOCS after 16 sessions
(Van Oppen et al., 1995). Indeed, even 16 sessions might be insuf-
ficient to reach sub-clinical levels. Compared to the current study,
some ERP trials have involved approximately four times as much
treatment when the total number of treatment hours is considered
(Foa et al., 2005).
Using a clinical cutoff of 16 on the YBOCS-SR, it is possible to
speculate about potential financial implications of administering
DCS. A rudimentary cost comparisondconservatively omitting
adjustments for the time value of money (e.g., potential interest
gains), fewer missed work days, decreased patient disability
income, lessened ancillary health care services, etc.dsuggests non-
negligible costs savings per patient. Using standard fees fromone of
the current institutions (MGH),100 mg of DCS costs approximately
$2.85 per capsule, a behavior therapy session costs $335.00, and
a psychopharmacology evaluation (for obtaining a DCS prescrip-
tion) costs $475.00. To reduce symptom severity below the stan-
dard clinical cutoff, the average patient receiving DCS would spend
approximately $3,178 for eight sessions of ERP, a psychopharma-
cology evaluation, and eight DCS doses. However, a patient who
receives ERP alone would spend approximately $5,360 for 16
sessions of ERP (the number of sessions was based on the afore-
mentioned clinical trial by Van Oppen et al., 1995), and likely still
not score below the clinical cutoff. Utilizing DCS could potentially
save each patient over $2100. In addition, as previously stated,
often ERP requires more than 16 sessions, further suggesting that
this is a conservative estimate of cost savings. Therefore, patients,
insurance companies, and employers could benefit from these cost
savings projected over entire populations of individuals receiving
ERP for OCD.
Reducing the required number of behavior therapy sessions
could result in other benefits to patients, treatment providers, and
society. For example, treatment refusal and dropout rates might
decrease if treatment requires fewer sessions. Exposure and
response prevention treatment can be anxiety provoking. Thus, if
patients experience success (i.e., symptom reduction) earlier in
treatment and within fewer sessions, they may be more likely to
adhere to a treatment and to complete it, as opposed to aborting it
prematurely. Furthermore, patients may be more likely to initiate
and complete a treatment that requires fewer sessions, as it is less
costly in terms of time and money. In supportof the notionthat DCS
may protect against ERP dropout, evidence from Kushner et al.
(2007) demonstrated that patients in the ERPþDCS group
completed treatment at a significantly higher rate compared to the
ERPþplacebo group. While the current studycould not statistically
investigate meaningful differences in dropout or refusal rates
between groups because of the small sample size, future research
with larger samples could examine the extent to which dropout
rates improve. Similarly, another benefit of shortening the time to
OCD recovery is that it may free up treatment providers to see more
patients. This is especially important in light of recent allusions to
thelackof available and appropriatetreatment resourcesfor OCD in
the community (Pollard, 2008).
While the overall evidence supports the notion that DCS
shortens the time to recovery over the full course of treatment, the
findings were less clear regarding the benefits of DCS versus
placebo in each half of ERP. On one hand, there is evidence for
differences in group slopes across halves: a significant group-by-
time interaction in the first half but not the second, and the
ERPþplacebo was relatively inert in the first half but active in the
second half. On the other hand, there is contradictory evidence that
group slopes remain unchanged from the first to second half: the
within-group differences across halves for the ERP þDCS and
ERPþplacebo groups were non-significant, plus the difference
between group slopes in the first half was not statistically different
from the difference in group slopes in the second half.
In evaluating the full landscape of results, however, one inter-
pretation suggests that DCS quickens the course of ERP because it
may trigger the effects of ERP sooner in treatment than if ERP were
administered alone (based on the significant group-by-time inter-
action in the first half), but it does not boost the effects of ERP per se
(based on the exploratory contrast that the ERPþDCS group slope,
slope at its steepest). As a result of DCS jump-starting the effects of
ERP sooner, and relative to the initial inertness of the ERPþplacebo
group, there was a near six-fold advantage to receiving DCS in the
Evidence also seems to suggest that DCS does not stop working
after the first half (based on the non-significant within-group
difference for the ERPþDCS group across halves), but rather, it may
become redundant (based on the non-significant group-by-time
interaction in the second half), possibly because it may have
alreadyexhausted its maximum utility. The pattern of results might
also be explained by an ERP ceiling effect; it is possible that DCS
facilitates the effects of treatment sooner, but perhaps ERP has an
upper limit on its effect size.
The findings of the current study may have implications for
future discussions and research on DCS for OCD. For example, it
might be misleading to state that DCS amplifies the effects of ERP.
Fig. 2. Slopes of recovery per group per half of treatment.
Slopes and ratios of slopes.
Ratio Sessions 1e10
G.S. Chasson et al. / Behaviour Research and Therapy 48 (2010) 675e679
Once the effects of ERP are initiated and maximized, the speed of
recovery in a given moment seems to be the same for each group.
Thus, it may be more appropriate to specify that DCS speeds up
the course of ERP. In addition, understanding and investigating the
biological mechanisms of DCS might make more sense from the
framework of exhausting maximum utility rather than losing effec-
tiveness over time. Also, since OCD is a heterogeneous disorder, it
could be important to evaluate if the effects of DCS differ with
respect tosymptom presentation. Along this vein, the current study
did investigate different patterns of results based on obsessions
versus compulsions using the YBOCS-SR subscale scores. The
findings for both subscales were similar to the findings for the total
score, with two exceptions. First, the DCS þERP group recovered
significantly faster in the first half of treatment compared to the
DCSþplacebo group with respect to the obsessions subscale
scores, but group differences on the compulsions subscale only
approached statistical significance (p¼.07). Second, for the
obsessions subscale only, the comparison of the first half difference
between group slopes versus the second half difference between
group slopes was statistically significant. Based on these two
exceptions, it is possible that any differential effect of DCS across
ERP halves is more pronounced for obsessions rather than
compulsions, but further research is needed to clarify this
It appears that the treatment effects for the ERPþplacebo group
groupbeing non-significantlydifferentfromzero inthefirst,butnot
second, half). It is unclear why the effects of ERPþplacebo lagged,
but the relative lack of effect for the ERPþplacebo in the first half
support the idea that DCS could help patients circumvent a smaller,
or even temporarily inert, ERP effect in the early sessions of
Not all of our results of the study fit neatly into the above
interpretation. Indeed, it is still unclear why the comparison of the
first half difference between group slopes versus the second half
difference between group slopes was non-significant, or why the
within-group effect for the ERP þplacebo group was non-signifi-
cant. The small sample size, and hence limited statistical power,
might explain the inconsistent results. Evaluating these research
questions with a larger sample might clarify the findings further.
The inconsistent findings might also be the result of choosing the
midpoint as the time differentiation, a selection which could be as
arbitrary as picking any other session. In the current study, the
biggest drop in OCD severity in the ERPþDCS group occurred
between the first two ERP sessions, but the ERPþplacebo group
started working most noticeablyaround the seventh session. When
utilizing the fifth session as the cutoff, this pattern of data could
make it more difficult to detect differences across halves because
within-half variability could potentially wash away effects.
However, selecting an appropriate session cutoff point is currently
difficult due to conflicting evidence. In contrast to the current
study, results from Kushner et al. (2007) indicate that the
ERPþDCS group was different than the ERPþplacebo group
between the second and fourth session. Further, results from Storch
et al. (2007) suggest a trend in favor of the ERPþDCS group for
sessions 4 through 6 only. Future research is necessary to reconcile
the conflicting results of the current study with respect to speeds of
treatment recovery across halves.
In conclusion, data from the current investigation suggests that
the course of ERP is 2.3 times faster when sessions are paired with
DCS. Therefore, it may be possible to reduce the required number of
ERP sessions by simply supplementing treatment with a cost-
effective antibiotic. If patients perceive treatment as toocostly, time
intensive, oranxiety provoking, minimizing the number of required
exposure therapy sessions could help persuade them to initiate or
resume treatment. Ultimately, individuals with OCD could benefit
from decreased treatment costs and time without compromising
treatment outcome. Furthermore, incorporating DCS into treat-
ment for OCD could have substantial public health implications as
managed care gives incentives to providers to adopt efficient
APA. (2000). Diagnostic and statistical manual of mental disorders (4th, Text Revision
ed.). Washington, DC: American Psychiatric Association.
Baer, L., Brown-Beasley, M. W., Sorce, J., & Henriques, A. I. (1993). Computer-assisted
telephone administration of a structured interview for obsessiveecompulsive
disorder. American Journal of Psychiatry, 150(11), 1737e1738.
Engels, J. M., & Diehr, P. (2003). Imputation of missing longitudinal data:
a comparison of methods. Journal of Clinical Epidemiology, 56(10), 968e976.
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1995). Structured clinical
interview for DMS-IV axis I disorders, patient edition (SCID-P version 2.0). New
York: New York State Psychiatric Institute, Biometrics Research Department.
Foa, E. B., Liebowitz, M. R., Kozak, M. J., Davies, S., Campeas, R., Franklin, M. E., et al.
(2005). Randomized, placebo-controlled trial of exposure and ritual prevention,
clomipramine, andtheir combination
veecompulsive disorder. American Journal of Psychiatry, 162, 151e161.
Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Fleischmann, R. L.,
Hill, C. L., et al. (1989a). The Yale-Brown Obsessive Compulsive Scale-I: devel-
opment, use, and reliability. Archives in General Psychiatry, 46, 1006e1011.
Goodman, W. K., Price, L. H., Rasmussen, S. A., Mazure, C., Delgado, P.,
Heninger, G. R., et al. (1989b). The Yale-Brown obsessive compulsive inventory
(Y-BOCS): validity. Archives in General Psychiatry, 46, 1012e1016.
Grabill, K., Merlo, L., Duke, D., Harford, K. L., Keeley, M. L., Geffken, G. R., et al. (2008).
Assessment of obsessiveecompulsive disorder: a review. Journal of Anxiety
Disorders, 22(1), 1e17.
Hofmann, S. G., Meuret, A. E., Smits, J. A., Simon, N. A., Pollack, M. H.,
Eisenmenger, K., et al. (2006). Augmentation of exposure therapy with
D-cycloserine for social anxiety disorder. Archives in General Psychiatry, 63,
Jenike, M. A. (2004). Clinical practice. Obsessiveecompulsive disorder. New England
Journal of Medicine, 350(3), 259e265.
Kozak, M. J., & Foa, E. B. (1997). Mastery of obsessiveecompulsive disorder: A cogni-
tiveebehavioral approach. San Antonio: Psychological Corporation.
Kushner, M. G., Kim, S. W., Donahue, C., Thuras, P., Adson, D., Kotlyar, M., et al.
(2007). D-cycloserine augmented exposure therapy for obsessiveecompulsive
disorder. Biological Psychiatry, 62(8), 835e838.
Norberg, M. M., Krystal, J. H., & Tolin, D. F. (2008). A meta-analysis of D-cycloserine
and the facilitation of fear extinction and exposure therapy. Biological Psychi-
atry, 63(12), 1118e1126.
Pollard, C. A. (2008). Behavior Therapy Institute: handbook and program descrip-
Ressler, K. J., Rothbaum, B. O., Tannenbaum, L., Anderson, P., Graap, K., Zimand, E.,
et al. (2004). Cognitive enhancers as adjuncts to psychotherapy: use of
D-cycloserine in phobic individuals to facilitate extinction of fear. Archives in
General Psychiatry, 61, 1136e1144.
Rothbaum, B. O. (2008). Critical parameters for D-cycloserine enhancement of
cognitive-behaviorial therapy for obsessiveecompulsive disorder. American
Journal of Psychiatry, 165(3), 293e296.
Schruers, K., Koning, K., Luermans, J., Haack, M. J., & Griez, E. (2005). Obsessi-
veecompulsive disorder: a critical review of therapeutic perspectives. Acta
Psychiatrica Scandinavica, 111, 261e271.
Steketee, G., Frost, R., & Bogart, K. (1996). The Yale-Brown obsessive compulsive
scale: interview versus self-report. Behaviour Research and Therapy, 34(8),
Storch, E. A., Merlo, L. J., Bengtson, M., Murphy, T. K., Lewis, M. H., Yang, M. C., et al.
(2007). D-cycloserine does not enhance exposure-response prevention therapy
in obsessiveecompulsive disorder. Internaional Clinical Psychopharmacol, 22(4),
Van Oppen, P., De Haan, E., Van Balkom, A. J. L. M., Spinhoven, P., Hoogduin, K., &
Van Dyck, R. (1995). Cognitive therapy and exposure in vivo in the treatment of
obsessive compulsive disorder. Behaviour Research and Therapy, 33(4), 379e390.
Wilhelm, S., Buhlmann, U., Tolin, D. F., Meunier, S. A., Pearlson, G. D., Reese, H. E.,
et al. (2008). Augmentation of behavior therapy with D-cycloserine for obses-
siveecompulsive disorder. American Journal of Psychiatry, 165(3), 335e341, quiz
G.S. Chasson et al. / Behaviour Research and Therapy 48 (2010) 675e679