A Placebo-Controlled Trial of Phenelzine, Cognitive Behavioral Group Therapy, and Their Combination For Social Anxiety Disorder
A Placebo-Controlled Trial of Phenelzine, Cognitive Behavioral Group Therapy, and Their Combination For Social Anxiety Disorder
A Placebo-Controlled Trial of Phenelzine, Cognitive Behavioral Group Therapy, and Their Combination For Social Anxiety Disorder
Context: Medication and cognitive behavioral treat- Results: Linear mixed-effects models showed a specific
ment are the best-established treatments for social anxi- order of effects, with steepest reductions in Liebowitz So-
ety disorder, yet many individuals remain symptomatic cial Anxiety Scale scores for the combined group, fol-
after treatment. lowed by the monotherapies, and the least reduction in
the placebo group (Williams test=4.97, P.01). The CGI
Objective: To determine whether combined medica- response rates in the intention-to-treat sample at week
tion and cognitive behavioral treatment is superior to 12 were 9 of 27 (33.3%) (placebo), 16 of 34 (47.1%)
either monotherapy or pill placebo. (CBGT), 19 of 35 (54.3%) (phenelzine), and 23 of 32
Design: Randomized, double-blind, placebo-controlled (71.9%) (combined treatment) (12 =8.76, P .01). Cor-
trial. responding remission rates (CGI=1) were 2 of 27 (7.4%),
3 of 34 (8.8%), 8 of 35 (22.9%), and 15 of 32 (46.9%)
Setting: Research clinics at Columbia University and (12 =15.92, P .01). At week 24, response rates were 9
Temple University. of 27 (33.3%), 18 of 34 (52.9%), 17 of 35 (48.6%), and
25 of 32 (78.1%) (12 =12.02, P =.001). Remission rates
Participants: One hundred twenty-eight individuals with were 4 of 27 (14.8%), 8 of 34 (23.5%), 9 of 35 (25.7%),
a primary DSM-IV diagnosis of social anxiety disorder. and 17 of 32 (53.1%) (12 =10.72, P =.001).
Interventions: Cognitive behavioral group therapy
(CBGT), phenelzine sulfate, pill placebo, and combined Conclusion: Combined phenelzine and CBGT treat-
CBGT plus phenelzine. ment is superior to either treatment alone and to pla-
cebo on dimensional measures and on rates of response
Main Outcome Measures: Liebowitz Social Anxiety and remission.
Scale and Clinical Global Impression (CGI) scale scores
at weeks 12 and 24. Arch Gen Psychiatry. 2010;67(3):286-295
S
OCIAL ANXIETY DISORDER significant differences in efficacy be-
(SAD) is a highly preva- tween the groups. The second study18 com-
lent1-3 chronic and disabling pared buspirone hydrochloride, placebo,
anxiety disorder associated CBT plus buspirone, and CBT plus pla-
with substantial impair- cebo; CBT resulted in improvement in SAD
ment, decreased quality of life,4-7 and psy- symptoms, but buspirone alone was not
chiatric comorbidity.8,9 Although cogni- superior to placebo and did not augment
tive behavioral therapy (CBT) and the efficacy of CBT.
pharmacotherapy are the most effica- In the third study,19 patients were ran-
cious treatments for SAD,10-15 only two- domized to receive sertraline hydrochlo-
thirds of patients who receive these treat- ride or placebo and separately to receive ex-
ments are considered responders, of which posure therapy or general medical care.
only half are considered remitters.16 Most Sertraline was associated with greater effi-
patients remain symptomatic after initial cacy than was placebo, whereas exposure
treatment. alone was not. The fourth study15 exam-
Six controlled trials have examined the ined the efficacy of fluoxetine hydrochlo-
efficacy of combining medication and psy- ride, pill placebo, group CBT, CBT plus
chosocial treatments for SAD. The first fluoxetine, and CBT plus pill placebo. All
study17 compared social skills training plus active treatments had greater efficacy than
Author Affiliations are listed at propranolol hydrochloride with social did pill placebo, but there were no differ-
the end of this article. skills training plus placebo. There were no ences among the active treatments.
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 67 (NO. 3), MAR 2010 WWW.ARCHGENPSYCHIATRY.COM
286
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 67 (NO. 3), MAR 2010 WWW.ARCHGENPSYCHIATRY.COM
287
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 67 (NO. 3), MAR 2010 WWW.ARCHGENPSYCHIATRY.COM
288
560 Excluded
141 Did not meet the inclusion criteria
242 Refused to participate
177 Other reasons
166 Randomized
45 Patients assigned to phenelzine sulfate 39 Patients assigned to placebo 40 Patients assigned to CBGT 42 Patients assigned to combination therapy
35 Received phenelzine sulfate 27 Received placebo 34 Received CBGT 32 Received combination therapy
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 67 (NO. 3), MAR 2010 WWW.ARCHGENPSYCHIATRY.COM
289
Treatment Group a
Abbreviations: CBGT, cognitive behavioral group therapy; SAD, social anxiety disorder; SSRI, selective serotonin reuptake inhibitor.
a Because of rounding, percentages may not total 100.
b Differences across groups were compared using analysis of variance for continuous variables and 2 tests for categorical variables.
were seen between the CBGT and placebo groups on any percentage of CGI-I remitters was 22.1% in the phenel-
of the outcome measures. Effect sizes were generally small zine group (OR, 3.70; 95% CI, 0.82-19.14) and 8.8% in
for CBGT, medium for phenelzine, and large for com- the CBGT group (1.21, 0.19-7.81), neither significantly
bined treatment. different from the rate in the pill placebo group. When
Categorical measures yielded similar results. Pa- remission was defined by an LSAS score of 30 or less,
tients randomized to receive combined treatment were 59.4% of patients in the combined treatment group and
significantly more likely than those randomized to re- 11.1% in the placebo group were classified as remitters
ceive placebo to be classified as responders (OR, 5.11; (OR, 11.69; 95% CI, 2.91-47.05). The percentage of re-
95% CI, 1.68-15.52). There were no significant differ- mitters was 20.0% in the phenelzine group (OR, 2.00;
ences in the probability of response in patients random- 95% CI, 0.47-8.60) and 20.6% in the CBGT group (2.07,
ized to receive phenelzine (OR, 2.38; 95% CI, 0.84- 0.48-8.93). Here again, the monotherapies did not dis-
6.72), CBGT (1.78, 0.63-5.06), or placebo. Rates of tinguish themselves from placebo.
remission were also significantly higher for patients ran-
domized to combined treatment than for those random- Hypothesis 2
ized to placebo (Figure 2). Using the definition of CGI-
I=1, 46.9% of patients who received combined treatment Table 4 provides mean slopes of change for all continu-
were classified as remitters compared with 7.4% taking ous measures and rates of response and remission for all
placebo (OR, 11.03; 95% CI, 2.23-54.57). In contrast, the the treatment groups. Across all measures, the results of
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 67 (NO. 3), MAR 2010 WWW.ARCHGENPSYCHIATRY.COM
290
Treatment Group a
Abbreviations: ADIS, Clinicians Severity Rating of the Anxiety Disorders Interview Schedule for DSM-IV; CBGT, cognitive behavioral group therapy;
CGI-I, Clinical Global Impression Improvement Scale; CGI-S, Clinical Global Impression Severity Scale; FQ, Fear Questionnaire Social Phobia Subscale;
HAM-D, 29-item Hamilton Rating Scale for Depression; LSAS, Liebowitz Social Anxiety Scale (total score); SDS, Sheehan Disability Scale; SIAS, Social Interaction
Anxiety Scale; SPS, Social Phobia Scale.
a All values are given as mean (SD) score.
b Differences in the mean at baseline were compared using analysis of variance. Differences between baseline and week 12 were compared using linear
mixed-effects models.
Table 3. 12-Week Pairwise Differences Between the Placebo Group and Patients
Receiving CBGT, Phenelzine, and Combined Treatment
Treatment Group
Abbreviations: ES, effect size (Cohen d). For other definitions, see Table 2.
a P .01.
b P .05.
the Williams test were highly significant. Examination cebo. The results also support this ordering of treat-
of categorical measures produced similar results. ment effects (Table 4). Additional analyses restricted the
Because in a previous study10 phenelzine was supe- sample to responders to examine whether responders to
rior to CBGT on several continuous measures after acute each treatment differed in magnitude of improvement.
treatment, in the exploratory analyses of the present study, The mean slope of the LSAS score change was signifi-
we reexamined the models hypothesizing the following cantly larger for combined treatment than for the mono-
order: combined treatment, phenelzine, CBGT, and pla- therapies considered separately or pooled (Table 5).
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 67 (NO. 3), MAR 2010 WWW.ARCHGENPSYCHIATRY.COM
291
At week 24, mean (SD) LSAS scores were 59.3 (23.5) for
the placebo group, 51.0 (22.9) for the CBGT group, 52.6 COMMENT
(24.0) for the phenelzine group, and 32.0 (19.6) for the
combined group, resulting in effect sizes (Cohen d) of 0.36, To our knowledge, this is the first study to show the su-
periority of a combined treatment over medication, psy-
chotherapy, and placebo in the acute treatment of SAD.
Placebo In addition, we found that phenelzine, but not CBGT
80 CGBT
Phenelzine sulfate
alone, was superior to placebo. These results were con-
70 Combined treatment sistent across several outcome measures and analytic strat-
egies and were maintained throughout the 12-week con-
60
tinuation phase.
50 Supporting the main hypothesis, combined treat-
Patients, %
40
ment was superior to both monotherapies and to pla-
cebo. Two mechanisms could explain the higher effi-
30 cacy of combined treatment: (1) distinct groups of patients
20 with SAD could respond to only phenelzine or CBGT (by
receiving both, patients in the combined treatment group
10
would have increased their chances of receiving at least
0 1 treatment that was efficacious for them) and (2) com-
Response Remission Remission
(CGI-I = 1 or 2) (CGI-I = 1) (LSAS 30) bined treatment may exert a truly additive or synergis-
tic effect in the treatment of SAD beyond the effects of
either monotherapy alone.
Figure 2. Response and remission rates by treatment group at week 12.
CBGT indicates cognitive behavioral group therapy; CGI-I, Clinical Global If only the first mechanism was at work, responders
Impression Improvement Scale; and LSAS, Liebowitz Social Anxiety Scale. in the combined group would not have had larger aver-
Table 4. Statistical Inference Under Order Restrictions for Patients With SAD a
Phenelzine
Sulfate P P
Placebo or CBGT Combined Statistic b Value Placebo CBGT Phenelzine Combined Statistic b Value
CGI-I score, mean (SD) 3.00 (1.10) 2.41 (0.88) 1.79 (0.86) 4.45 .01 3.00 (1.10) NA NA 1.79 (0.86) NA NA
Slope, mean (SE), change
per mo
LSAS 1.82 (0.45) 2.07 (0.59) 2.59 (0.48) 4.97 .01 1.82 (0.45) 1.93 (0.51) 2.24 (0.60) 2.59 (0.48) 5.07 .01
ADIS 0.12 (0.06) 0.14 (0.06) 0.18 (0.07) 3.15 .01 0.12 (0.06) 0.13 (0.04) 0.14 (0.07) 0.18 (0.07) 3.21 .01
CGI-S 0.07 (0.06) 0.08 (0.06) 0.12 (0.06) 2.50 .01 0.07 (0.06) 0.07 (0.05) 0.09 (0.06) 0.12 (0.06) 2.53 .01
HAM-D 0.10 (0.05) 0.09 (0.05) 0.12 (0.05) NA NA 0.10 (0.05) 0.09 (0.06) 0.09 (0.04) 0.12 (0.05) NA NA
FQ 0.62 (0.19) 0.72 (0.18) 0.81 (0.11) 3.78 .01 0.62 (0.19) 0.68 (0.17) 0.75 (0.18) 0.81 (0.11) 3.80 .01
SIAS 1.03 (0.36) 1.26 (0.42) 1.40 (0.35) 3.14 .01 1.03 (0.36) 1.15 (0.29) 1.33 (0.48) 1.40 (0.35) 3.22 .01
SPS 1.09 (0.16) 1.17 (0.17) 1.23 (0.14) 2.91 .01 1.09 (0.16) 1.16 (0.15) 1.18 (0.18) 1.23 (0.14) 2.99 .01
SDS 0.58 (0.05) 0.60 (0.06) 0.62 (0.06) 2.15 .02 0.58 (0.05) 0.59 (0.05) 0.61 (0.07) 0.62 (0.06) 2.18 .02
Rate, No. (%)
Response 9/27 (33.3) 35/69 (50.7) 23/32 (71.9) 8.92 .01 9/27 (33.3) 16/34 (47.1) 19/35 (54.3) 23/32 (71.9) 8.76 .01
Remission, CGI-I=1 2/27 (7.4) 11/69 (15.9) 15/32 (46.9) 14.00 .01 2/27 (7.4) 3/34 (8.8) 8/35 (22.9) 15/32 (46.9) 15.92 .01
Remission, LSAS 30 3/27 (11.1) 14/69 (20.3) 19/32 (59.4) 17.78 .01 3/27 (11.1) 7/34 (20.6) 7/35 (20.0) 19/32 (59.4) 15.53 .01
Abbreviations: CBGT, cognitive behavioral group therapy; NA, not applicable; SAD, social anxiety disorder. For other definitions, see Table 2.
a The left side of the table describes the main analyses, that is, collapsing the phenelzine and CBGT groups into one. The right side of the table describes the
exploratory analyses, in which the phenelzine and CBGT groups are examined separately.
b Williams test for slope and linear-by-linear 2 tests for categorical measures.
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 67 (NO. 3), MAR 2010 WWW.ARCHGENPSYCHIATRY.COM
292
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 67 (NO. 3), MAR 2010 WWW.ARCHGENPSYCHIATRY.COM
293
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 67 (NO. 3), MAR 2010 WWW.ARCHGENPSYCHIATRY.COM
294
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 67 (NO. 3), MAR 2010 WWW.ARCHGENPSYCHIATRY.COM
295