Boston University

OpenBU http://open.bu.edu
BU Open Access Articles BU Open Access Articles

2017-02

A meta-analysis of
pharmacotherapy for social anxiety
disorder: an examination of
efficacy, moderators, and
mediators

Curtiss, Joshua

Joshua Curtiss, Leigh Andrews, Michelle Davis, Jasper Smits, Stefan G Hofmann.
2017. "A meta-analysis of pharmacotherapy for social anxiety disorder: an
examination of efficacy, moderators, and mediators.." Expert Opin Pharmacother,
Volume 18, Issue 3, pp. 243 - 251.
https://hdl.handle.net/2144/21703
Boston University
 Pharmacotherapy for SAD 1

Running Head: Pharmacotherapy for SAD

A meta-analysis of pharmacotherapy for social anxiety disorder: An examination of efficacy,

moderators, and mediators

Joshua Curtiss1, Leigh Andrews1, Michelle Davis2, Jasper Smits2, and Stefan G. Hofmann1

1
Department of Psychological and Brain Sciences, Boston University
2
The University of Texas, Department of Psychology and Institute for Mental Health

Research

In press: Expert Opinion on Pharmacotherapy

Keywords: pharmacotherapy, social anxiety disorder, quality of life, meta-analysis

 Pharmacotherapy for SAD 2

Abstract

Introduction

Social anxiety disorder (SAD) is among the most prevalent mental disorders, associated

with impaired functioning and poor quality of life. Pharmacotherapy is the most widely utilized

treatment option. The current study provides an updated meta-analytic review of the efficacy of

pharmacotherapy and examines moderators and mediators of treatment efficacy.

Areas Covered

A comprehensive search of the current literature yielded 52 randomized, pill placebo-

controlled trials of pharmacotherapy for adults diagnosed with SAD. Data on potential mediators

of treatment outcome were collected, as well as data necessary to calculate pooled correlation

matrices to compute indirect effects.

Expert Opinion

The overall effect size of pharmacotherapy for SAD is small to medium (Hedges’ g =

0.41). Effect sizes were not moderated by age, sex, length of treatment, initial severity, risk of

study bias, or publication year. Furthermore, reductions in symptoms mediated

pharmacotherapy’s effect on quality of life. Support was found for reverse mediation. Future

directions may include sustained efforts to examine treatment mechanisms of pharmacotherapy

using rigorous longitudinal methodology to better establish temporal precedence.

1. Introduction
 Pharmacotherapy for SAD 3

Social anxiety disorder (SAD) is among the most ubiquitous emotional disorders with a

lifetime prevalence of 12.1% [1]. Individuals with SAD often experience an unremitting course

associated with numerous functional impairments and diminished quality of life [1-4]. Although

cognitive behavior therapy (CBT) is regarded as one of the most efficacious treatments for SAD

[5-6], pharmacological interventions are more prevalent and widely utilized. Of the various

pharmacological interventions that have been applied to SAD, most published randomized

controlled trials have examined the efficacy of antidepressants (e.g., selective serotonin reuptake

inhibitors (SSRIs) and serotonin and norepinephrine reuptake inhibitors (SNRIs)), monoamine

oxidase inhibitors (MAOIs), and benzodiazepines. Despite the fact that prior research has

substantiated the efficacy of these pharmacological treatments, relatively little is known about

the potential mechanisms underlying pharmacotherapy [7].

Indeed, a number of individual treatment mechanisms have been investigated in CBT,

such as changes in estimated social cost and extinction learning [8-9]. Extant literature suggests

that antidepressants and MAOIs are associated with improvement in more general mental health

domains such as quality of life (QOL) [10-11]. As regards pharmacotherapy, it may be the case

that it exerts its therapeutic effect on QOL by way of reducing symptoms of social anxiety.

Because individuals with SAD experience impaired QOL, treatments that enhance overall QOL

may do so through remission of social anxiety symptoms [12].

Thus, the principle objective of the current meta-analysis is to review and analyze extant

randomized, placebo controlled trials of pharmacotherapy for SAD to: 1) provide an updated

controlled effect-size of pharmacotherapy for symptoms of anxiety; 2) examine potential

moderators of treatment efficacy; and 3) investigate the mediation hypothesis that

pharmacotherapy confers its indirect effect on QOL by way of symptom reduction.

 Pharmacotherapy for SAD 4

2. Method

2.1 Search Strategy

A comprehensive search was conducted in the PsycINFO (1840 to May 2016),

MEDLINE (1966 to October 2016), and Scopus (1869 to October 2016) databases for

randomized placebo-controlled pharmacotherapy trials for SAD. The following search terms

were used to search key words, title, abstract, and Medical Subject Headings:

“pharmacotherapy”, “medication”, or “pharmacology”; “social anxiety”, “social anxiety

disorder”, “SAD”, or “social phobia”; and “trial”, or “clinical trial”. Furthermore, we also

examined citation maps and used “cited by” search strategies to locate studies which were then

cross-referenced with references from reviews.

2.2 Inclusion/Exclusion Criteria

For the current meta-analysis, study eligibility was determined by the following inclusion

criteria: (a) adult participants meeting diagnostic criteria for SAD; (b) pharmacotherapy for SAD

including more than two doses (i.e., administrations) of medication; (c) a pill placebo control

condition; (d) inclusion of specific measures of SAD treatment outcome, including: the

Liebowitz Social Anxiety Scale (LSAS) total score, the Clinical Global Impression of Severity

Scale (CGI-S), the Clinical Global Impression – Social Phobia Scale (CGI-SP), the Clinical

Impression of Severity – Social Phobia Scale (CIS-SP), or the Social Phobia Disorders Severity

and Change Form (SPDSC). Exclusion criteria included: (a) studies lacking sufficient data to

conduct analyses; (b) treatment studies using only medication responders after a medication trial

period; (c) single case studies; and (d) treatment conditions based on augmentation or studies

aimed at treating comorbid diagnoses.

2.3 Data Collection and Synthesis

 Pharmacotherapy for SAD 5

We collected data on pharmacotherapy class, number of participants, length of

pharmacotherapy (in weeks), mean age, percentage of male/female participants, and initial

severity. Study bias was assessed with the Cochrane risk of bias instrument. This tool involves

assessing each study as containing a high, low or unclear level of bias risk in a number of

domains using pre-specified criteria. The domains used in our assessment were: 1) Sequence

Generation, which assesses whether allocation of participants to treatment condition are

adequately generated in a random manner; 2) Allocation Concealment, which assesses whether

treatment assignment was adequately concealed from investigators and participants prior to

randomization; 3) Incomplete Outcome Data, which assesses whether outcome data are missing

at random and imputed using appropriate methods; and 4) Selective Outcome Reporting, which

determines whether pre-specified outcome variables of interest are adequately and completely

reported. Following recommendations from the Cochrane guidelines, a total bias assessment was

created for each study such that an ‘unclear’ rating in any category meant an ‘unclear risk’

overall rating unless there was a ‘high’ rating in any category, which lead to a ‘high risk’ overall

rating. ‘Low risk’ studies had to be rated as ‘low’ in all four categories.

We estimated controlled effect sizes using Hedges’ g, which corrects for parameter bias

due to small sample size [13]. To compute Hedges’ g, we extracted means and standard

deviations, as well as information from significance tests (e.g., t, F) [13]. The pooled effect sizes

were estimated using random effects models, which assume significant heterogeneity of the

included studies. All standard analyses were completed with Comprehensive Meta-Analysis [14].

To ascertain whether symptom remission mediated improvement in QOL, meta-analytic

structural equation modelling was employed to estimate the indirect effect. In this model, we

determined whether the effect of treatment condition (i.e., a dummy-coded variable denoting
 Pharmacotherapy for SAD 6

treatment versus pill placebo) on QOL was mediated by symptom reduction. In accordance with

established precedent, a two-stage structural equation modelling (TS-SEM) analysis was

conducted by: (i) synthesizing correlation matrices between the X, M, and Y variables; and (ii)

specifying a structural model on the pooled correlation matrix [14]. If a study did not explicitly

report correlation matrices, then means, standard deviations, t-statistics, F-statistics, and effect

sizes were extracted to compute correlation coefficients [16-17]. We also pursued a reverse

mediation model in which QOL mediates the relationship between treatment condition and

symptom reduction. The TS-SEM model was estimated in R with the metaSEM package.

3. Results

3.1 Study Characteristics

Our initial search yielded 2,202 abstracts (see Figure 1), of which 113 were deemed

relevant to the current study after review and evaluated for inclusion and exclusion criteria. Of

these, 67 did not meet inclusion criteria. Six studies were also identified from recently published

meta-analyses [6]. The remaining 52 studies included pill placebo controlled comparisons to

selective serotonin reuptake inhibitors (SSRIs; 22 studies), monoamine oxidase inhibitors

(MAOIs; 7 studies), monoamine oxidase A inhibitors (MAO-As; 6 studies), serotonin-

norepinephrine reuptake inhibitors (SNRIs; 3 studies), anticonvulsants (4 studies), antipsychotics

(1 study), benzodiazepines (2 studies), herbal supplements (1 study), and noradrenergic and

specific serotonin antidepressants (NaSSAs; 1 study). Additionally, seven studies included

multiple pill placebo controlled medication comparisons, including two studies comparing an

SNRI and an SSRI, one study comparing two types of SSRIs, one study comparing two types of

MAOIs, one study comparing an SSRI and an NK1 receptor agonist, one study comparing an
 Pharmacotherapy for SAD 7

SSRI and an anticonvulsant, and one study comparing an MAOI and a beta blocker. These 52

studies with a total sample size of 12,153 individuals are depicted in Table 1.

3.2 Overall Efficacy of Pharmacotherapy

Controlled effect sizes, confidence intervals, and significance values for each study are

presented in Table 2. The random effects meta-analysis yielded an overall, controlled effect size

of Hedges’ g = 0.41 (95% CI: 0.36-0.46, z = 16.34, p < .001), which suggests that

pharmacotherapy contributed to greater symptom remission than pill placebo. The fail-safe N

analysis was conducted to address the file drawer problem [18, 19]. If the number of studies

needed to reduce the effect size to a non-significant level (i.e., the fail-safe N) exceeds 5K+10,

the effect is considered robust [13]. Results of the fail-safe N analysis were robust, suggesting

that it would require 7,761 future or unpublished studies with an effect size of zero to attenuate

the overall effect size to non-significance. Inspection of the funnel plot reveals effect sizes

distributed roughly symmetrically around the mean effect size (see Figure 2). Using the Trim and

Fill method [20], we determined that 9 studies would need to fall to the left of the mean (i.e.,

have an effect size smaller than the mean) and 0 studies would need to fall to the right of the

mean (i.e., have an effect size larger than the mean) to make the plot symmetrical. The random-

effects model for the new imputed mean effect size revealed a Hedges’ g = 0.35 (95% CI: 0.28-

0.42).

3.3 Moderator Analyses

A number of moderator analyses were conducted to determine whether treatment efficacy

varies as a consequence of patient and study characteristics. A multiple meta-regression was

computed with mean age, overall percentage of female participants, treatment duration,

publication date, and initial severity as predictors. There was no significant relation between
 Pharmacotherapy for SAD 8

effect size and age (β = -0.03, p = 0.17), overall percentage of female participants (β = 0.01, p =

0.13), treatment duration (β = -0.01, p = 0.15), initial severity (β = -0.22, p = 0.06), or

publication year (β = 0.01, p = 0.48). No significant differences were revealed between high

(Hedges’ g = 0.39; 95% CI: 0.33-0.46, p < .001), low (Hedges’ g = 0.66; 95% CI: 0.29-1.05, p <

.001), and unclear (Hedges’ g = 0.42; 95% CI: 0.28-0.55, p < .001) risks of study bias (QB =

1.92, df = 2, p = 0.38). The overall class of pharmacotherapy did not moderate treatment efficacy

(QB = 17.25, df = 10, p = 0.07). Furthermore, no significant differences were revealed between

acute (Hedges’ g = 0.41; 95% CI: 0.36-0.46, p < .001) and long-term (Hedges’ g = 0.35; 95%

CI: 0.04-0.68, p < 0.05) effects of pharmacotherapy (QB = 0.09, df = 1, p = 0.76).

3.4 Mediator Analyses

Of all the studies included in the current meta-analysis, three studies included a

psychometrically validated QOL measure (i.e., the Quality of Life Enjoyment and Satisfaction

Questionnaire (QLESQ) and the Short Form Health Survey (SF-36)) [21-23]. Results of the

traditional random effects meta-analysis demonstrated that the controlled effect size for QOL

was Hedges’ g = 0.28 (95% CI: 0.13-0.44, z = 3.52, p < .001), which suggests that

pharmacotherapy contributed to greater improvements in QOL than pill placebo. Furthermore,

correlation matrices were synthesized from these three studies, and a structural equation model

was specified such that treatment condition exerted an indirect effect on QOL through reductions

in symptoms. Results of the TS-SEM analysis established that the indirect effect was significant

(Standardized β = 0.14; 95% CI: 0.09-0.19), as well as the direct effect (Standardized β = 0.19;

95% CI: 0.13-0.27), which indicates partial mediation. However, results of the reverse mediation

model indicated that the indirect effect was significant (Standardized β = 0.10; 95% CI: 0.07-
 Pharmacotherapy for SAD 9

0.13), as well as the direct effect (Standardized β = 0.09; 95% CI: 0.06-0.13), suggesting partial

mediation.

4. Discussion

The current meta-analysis of 52 randomized, pill placebo controlled trials of

pharmacotherapy for individuals diagnosed with SAD yielded an overall controlled effect size of

Hedges’s g = 0.41, which reflects a small to medium effect of pharmacotherapy for SAD. The

Fail-Safe N suggested that this effect size is robust, as 7,761 studies demonstrating null-effects

would be required to attenuate the overall effect size to non-significance. Indeed, inspection of

the funnel plot does not reveal substantial publication bias. Thus, the results of the current meta-

analysis substantiate pharmacotherapy as an efficacious intervention for SAD.

To further understand the conditions under which pharmacotherapy is maximally

efficacious, a number of treatment moderators were investigated. Consistent with a prior meta-

analysis [7], the moderator analyses provided no evidence that the efficacy of pharmacotherapy

varies as a function of treatment duration, initial symptom severity, age, sex, and year of study

publication. No differences in effect sizes were observed between different classes of study bias

(i.e., low, high, and uncertain). Furthermore, it was also revealed that there were no significant

differences between acute and long-term effects of pharmacotherapy.

In an effort to elucidate potential treatment mechanisms underlying pharmacotherapy,

meta-analytic structural equation modelling was employed to assess our mediation hypothesis.

Consistent with our prediction, reductions in anxiety mediated pharmacotherapy’s effect on

improvement in QOL. However, results of the reverse mediation were also significant, which

prohibits definitive conclusions regarding temporal precedence of the mediator. Although prior

research has examined neurobiological mechanisms targeted by pharmacotherapy (e.g., serotonin

 Pharmacotherapy for SAD 10

receptors in the case of SSRIs) [24], there has been a relative dearth of research investigating

psychological mechanisms underlying pharmacological interventions. A large literature on CBT

suggests that this intervention ameliorates symptoms through a variety of treatment mechanisms

(e.g., extinction learning, changes in cognitions, etc.) [8]. Despite the limited research on

psychological mechanisms underlying pharmacotherapy, results of the current study indicate that

it might confer its benefit on QOL by way of reductions in anxiety symptoms.

Certain limitations warrant mention. Relatively few studies examined head-to-head

comparisons between different classes of active pharmacological interventions, precluding

conclusions about whether a particular class of pharmacotherapy substantially outperforms

another one. Because only three studies examined the effect of pharmacotherapy on QOL in

SAD, it remains uncertain as to whether pharmacotherapy leads to substantial improvement in

QOL. Furthermore, the results of the mediation analysis were derived from pre- to post-treatment

changes in anxiety symptoms and in QOL, which do not yield the most robust test of mediation.

Future studies should better establish temporal precedence of the mediator by including more

frequent assessment points to ascertain whether decreases in symptoms actually precede

increases in QOL. That notwithstanding, the current meta-analysis represents a crucial first effort

in determining whether changes in social anxiety mediates the efficacy of pharmacotherapy on

QOL. Another important limitation concerns the significant indirect effect of the reverse

mediation analysis, which prohibits firm conclusions about the directionality of mediation. Thus,

it may be the case that reductions in anxiety symptoms account for improvements in QOL and

vice versa. To further enhance our understanding of pharmacotherapy for SAD, future research

should employ more rigorous experimental designs to better delineate the psychological

mechanisms underlying pharmacological interventions. For instance, a randomized controlled

 Pharmacotherapy for SAD 11

trial in which both the outcome and mediator measures are assessed frequently (e.g., weekly)

could enable richer conclusions as to whether changes in the mediator temporally precede

changes in the outcome variable.

5. Conclusion

In summary, results of this meta-analysis indicate that pharmacotherapy produces greater

reduction in social anxiety symptoms relative to pill placebo. Evidence of efficacy was revealed

for several classes of pharmacotherapy (e.g., SSRIs, SNRIs, MAOIs, etc.). Results also

suggested that pharmacotherapy may also enhance QOL. Although the mediation analyses

provided evidence that pharmacotherapy produces better QOL through symptom reduction,

results of the reverse mediation analysis preclude definitive interpretations regarding

directionality.

6. Expert Opinion

Although pharmacotherapy was associated with small to medium controlled effect sizes

relative to pill placebo, this suggests that there is still considerable room for improvement above

and beyond potential placebo effects. Results of the current meta-analysis are in accord with

prior meta-analytic research that examined the efficacy of pharmacotherapy for SAD [6]. The

largest number of studies supporting the efficacy of any given class of pharmacotherapy was

found for SSRIs (Hedge’s g = 0.44) and MAOIs (Hedge’s g = 0.36). Consistent with prior

research [6], available evidence indicates that SSRIs and MAOIs are the most reliably

efficacious forms of pharmacotherapy for SAD. The current meta-analysis, however, enables us

to better appreciate the underlying mechanisms that may be responsible for the efficacy of

pharmacotherapy. Even though symptom reduction and improvement in QOL were corroborated

as possible treatment mechanisms in the current meta-analysis, it is still possible that

 Pharmacotherapy for SAD 12

pharmacotherapy exerts its effect through other mechanisms. Because the studies included in the

meta-analysis reported relatively few outcome variables, as almost none of them tested

mediation hypotheses, it was only possible to examine a limited subset of mediators. Better

understanding of the psychological mechanisms underlying pharmacotherapy is of great

importance.

That notwithstanding, results of the current meta-analysis accord well with prior literature

suggesting that this treatment enhances QOL in individuals diagnosed with a variety of

emotional disorders [25-26], which tend to be very comorbid [27-28]. This might suggest that

pharmacological interventions such as SSRIs have broad spectrum effects that contribute to

symptom remission across several conditions. Indeed, research attests to a strong inverse

relationship between quality of life and anxiety symptoms [26], as the presence of a disease

symptom (e.g., fear about speaking up at a meeting) can be often associated with the absence of

some aspect of quality of life (e.g., employment satisfaction). Therefore, it would make

conceptual sense that treatment response to pharmacotherapy results from mutually reinforcing

improvements in both anxiety symptoms and quality of life.

Another important consideration is that pharmacotherapy constitutes a heterogeneous

class of interventions. The pharmacological interventions investigated in the current meta-

analysis are intrinsically anxiolytic, ameliorating symptoms of anxiety directly. Another class of

pharmacotherapy is cognitive enhancers [29]. Rather than conferring anxiolytic effects by

themselves, cognitive enhancers, such as d-cycloserine, facilitate learning processes. When used

in combination with CBT, cognitive enhancers may enhance extinction learning that occurs

throughout the context of exposure exercises, thereby augmenting CBT specific mechanisms

[30-33]. In particular, d-cycloserine is characterized by well-defined psychological (i.e.,

 Pharmacotherapy for SAD 13

extinction learning) and neurobiological mechanisms (i.e., partial agonist of NMDA receptors).

Consequently, these novel combination approaches may better enable researchers to develop

maximally efficacious treatments. As our cumulative knowledge of treatment mechanisms

increases, we will enrich our understanding of these interventions and optimize their

implementation.
 Pharmacotherapy for SAD 14

Author Note

Corresponding author: Stefan G, Hofmann, Ph.D., Department of Psychological and Brain

Sciences, Boston University, 648 Beacon Street, 6th Floor, Boston, MA 02215, USA, Tel: 617 353

9233, email: shofmann@bu.edu. Dr. Hofmann receives support from NIH/NCCIH

(R01AT007257), NIH/NIMH (R01MH099021, R34MH099311, R34MH086668, R21MH102646,

R21MH101567, K23MH100259), the James S. McDonnell Fundation 21st Century Science

Initiative in Understanding Human Cognition – Special Initiative, and the Department of the

Army for work unrelated to the studies reported in this article. He receives compensation for

his work as an advisor from the Palo Alto Health Sciences and Otsuka Digital Health, Inc., and

for his work as a Subject Matter Expert from John Wiley & Sons, Inc. and SilverCloud Health, Inc.

He also receives royalties and payments for his editorial work from various publishers.
 Pharmacotherapy for SAD 15

References

Papers of special note have been highlighted as either of interest (*) or of considerable interest

(**) to readers.

1. Kessler RC, Berglund P, Demler O, et al. Lifetime prevalence and age-of-onset distributions

of DSM-IV disorders in the national comorbidity survey replication. Arch Gen Psychait

2005; 62(6): 593-602.

2. Furmark T. Social phobia: overview of community surveys. Acta Psychiat Scand

2002;105(2):84-93.

3. Katzelnick DJ, Kobak KA, DeLeire T, et al. Impact of generalized social anxiety disorder in

managed care. Am J Psychiat 2001;158(12):1999-2007.

4. Kushnir J, Marom S, Mazar M, et al. The link between social anxiety disorder, treatment

outcome, and sleep difficulties among patients receiving cognitive behavioral group

therapy. Sleep Med 2014;15(5):515-521.

5. Hofmann SG, Smits JAJ. Cognitive-behavioral therapy for adult anxiety disorders: A meta-

analysis of randomized placebo-controlled trials. J Clin Psychiat 2008;69(4):621-632.

6. Mayo-Wilson E, Dias S, Mavranezouli I, et al., Psychological and pharmacological

interventions for social anxiety disorder in adults: a systematic review and network meta-

analysis. Lancet Psychiat 2014; 1(5): 368-376.

7. Davis ML, Smits JAJ, Hofmann SG. Update on the efficacy of pharmacotherapy for social

anxiety disorder: a meta-analysis. Exp Op Pharm 2014;15(16):2281-2291.

8. Hofmann SG. Cognitive mediation of treatment change in social phobia. J Consult Clin

Psychol 2004;72(3):392-399.
 Pharmacotherapy for SAD 16

9. Hofmann SG. Cognitive processes during fear acquisition and extinction in animals and

humans: Implications for exposure therapy of anxiety disorders. Clin Psychol

Rev 2008;28(2):199-210.

10. Llorca PM. Review: The antidepressant agomelatine improves the quality of life of depressed

patients: implications for remission. J Psychopharmacol 2010;24:21-26.

11. Stein MB, Liebowitz MR, Lydiard RB, et al. Paroxetine treatment of generalized social

phobia (social anxiety disorder): a randomized controlled trial. JAMA 1998;280(8):708-713.

12. Olatunji BO, Cisler JM, Tolin DF. Quality of life in the anxiety disorders: a meta-analytic

review. Clin Psychol Rev 2007;27(5);572-581.

13. Rosenthal R. Meta-analytic procedures for social research. Thousand Oaks: Sage, 1991.

14. Borenstein M, Rothstein H. Comprehensive meta-analysis: A computer program for research

synthesis. Englewood: Biostat Inc., 1999.

15. Cheung MWL. metaSEM: An R package for meta-analysis using structural equation

modeling. Front Psychol, 2015;5:1-7.

16. Gu J, Strauss C, Bond R, et al. How do mindfulness-based cognitive therapy and

mindfulness-based stress reduction improve mental health and wellbeing? A systematic

review and meta-analysis of mediation studies. Clin Psychol Rev, 2015;37:1-12.

17. Lipsey MW, Wilson DB. Practical meta-analysis. 2001 Thousand Oaks, CA: Sage

Publications.

18. Rosenthal R. The “file drawer problem” and tolerance for null results. Psychol Bull

1979;85:638-641.

19. Orwin RG. A fail-safe N for effect size in meta-analysis. J Educ Stat 1983;8(2):157-159.
 Pharmacotherapy for SAD 17

20. Duval S, Tweedie R. Trim and fill: a simple funnel plot–based method of testing and

adjusting for publication bias in meta-analysis. Biometrics 2000;56(2):455-463.

21. Kobak K, Greist J, Jefferson J, Katzelnick D. Fluoxetine in social phobia: a double-blind,

placebo-controlled pilot study. J Clin Psychopharmacol 2002;22(3):257-262.

* Denotes studies included in the meta-analysis.

22. Liebowitz M, Demartinis N, Weihs K, et al. Efficacy of sertraline in severe generalized

social anxiety disorder: results of a double-blind, placebo-controlled study. J Clin Psychiatry

2003;64(7):785-792.

* Denotes studies included in the meta-analysis.

23. Pande AC, Feltner DE, Jefferson JW, et al. Efficacy of the novel anxiolytic pregabalin in

social anxiety disorder: a placebo-controlled, multicenter study. J Clin Psychopharmacol

2004;24(2),141-149.

* Denotes studies included in the meta-analysis.

24. Stahl SM. Mechanism of action of serotonin selective reuptake inhibitors: serotonin

receptors and pathways mediate therapeutic effects and side effects. J Affect

Disord 1998;51(3):215-235.

25. Newhouse PA, Krishnan KRR, Doraiswamy PM, et al. A double-blind comparison of

sertraline and fluoxetine in depressed elderly outpatients. J Clin Psychiat 2000;61(8):559-

568.

26. Mendlowicz MV, Stein MB. Quality of life in individuals with anxiety disorders. A J

Psychiat 2000;157(5):669-682.
 Pharmacotherapy for SAD 18

27. Curtiss J, Klemanski DH. Taxonicity and network structure of generalized anxiety disorder

and major depressive disorder: An admixture analysis and complex network analysis. J

Affect Disord 2016;199:99-105.

28. Curtiss J, Klemanski DH. Identifying individuals with generalised anxiety disorder: A

receiver operator characteristic analysis of theoretically relevant measures. Behav

Change 2015;32(04):255-272.

29. Hofmann SG, Mundy EA, Curtiss J. Neuroenhancement of exposure therapy in anxiety

disorders. AIMS Neurosc 2014;2(3):123-138.

30. Hofmann SG, Meuret AE, Smits JA, et al. Augmentation of exposure therapy with D-

cycloserine for social anxiety disorder. Arch Gen Psychiat 2006;63(3):298-304.

31. Hofmann SG, Wu JQ, Boettcher H. D-Cycloserine as an augmentation strategy for cognitive

behavioral therapy of anxiety disorders. Biol Mood Anxiety Disord 2013;3(11):1-10.

32. Guastella AJ, Richardson R, Lovibond PF, et al. A randomized controlled trial of D-

cycloserine enhancement of exposure therapy for social anxiety disorder. Biol Psychiat

2008;63(6):544-549.

33. Smits JA, Rosenfield D, Otto MW, et al. D-Cycloserine enhancement of fear extinction is

specific to successful exposure sessions: evidence from the treatment of height phobia. Biol

Psychiat 2013;73(11):1054-1058.

34. Allgulander, C. Paroxetine in social anxiety disorder: a randomized placebo-controlled study.

Acta Psychiatr Scand 1999;100(3):193-198.

* Denotes studies included in the meta-analysis.

 Pharmacotherapy for SAD 19

35. Allgulander C, Mangano R, Zhang J, et al. Efficacy of Venlafaxine ER in patients with social

anxiety disorder: a double-blind, placebo-controlled, parallel-group comparison with

paroxetine. Hum Psychopharm Clin 2004;19(6):387-396.

* Denotes studies included in the meta-analysis.

36. Asakura S, Tajima O, Koyama T. Fluvoxamine treatment of generalized social anxiety

disorder in Japan: a randomized double-blind, placebo-controlled study. The Int J

Neuropsychop 2007;10(2):263-274.

* Denotes studies included in the meta-analysis.

37. Baldwin D, Bobes J, Stein D, et al. Paroxetine in social phobia/social anxiety disorder.

Randomised, double-blind, placebo-controlled study. Paroxetine Study Group. Br J

Psychiatry 1999;175(2):120-126.

* Denotes studies included in the meta-analysis.

38. Barnett SD, Kramer ML, Casat CD, et al. Efficacy of olanzapine in social anxiety disorder: A

pilot study. J Psychopharmacol 2002;16(4):365-368.

* Denotes studies included in the meta-analysis.

39. Blanco C, Heimberg R, Schneier F, et al. A placebo-controlled trial of phenelzine, cognitive

behavioral group therapy, and their combination for social anxiety disorder. Arch Gen

Psychiatry 2010;67(3):286-295.

* Denotes studies included in the meta-analysis.

40. Blomhoff S, Haug T, Hellström K, et al. Randomised controlled general practice trial of

sertraline, exposure therapy and combined treatment in generalised social phobia. Br J

Psychiatry 2001;179(1):23-30.

* Denotes studies included in the meta-analysis.

 Pharmacotherapy for SAD 20

41. Clark D, Ehlers A, Mcmanus F, et al. Cognitive therapy versus fluoxetine in generalized

social phobia: a randomized placebo-controlled trial. J Consult Clin Psychol

2003;71(6):1058-1067.

* Denotes studies included in the meta-analysis.

42. Davidson JR, Potts N, Richichi E, et al. Treatment of social phobia with clonazepam and

placebo. J Clin Psychopharmacol 1993;13(6):423-428.

* Denotes studies included in the meta-analysis.

43. Davidson J, Foa E, Huppert J, et al. Fluoxetine, comprehensive cognitive behavioral therapy,

and placebo in generalized social phobia. Arch Gen Psychiat 2004a;61(10):1005-1013.

* Denotes studies included in the meta-analysis.

44. Davidson J, Yaryura-Tobias J, DuPont R, et al. Fluvoxamine-controlled release formulation

for the treatment of generalized social anxiety disorder. J Clin Psychopharmacol 2004b; 24:

118–25.

* Denotes studies included in the meta-analysis.

45. Fahlén T, Nilsson HL, Borg K, et al. Social phobia: the clinical efficacy and tolerability of

the monoamine oxidase-A and serotonin uptake inhibitor brofaromine. A double-blind

placebo-controlled study. Acta Psychiat Scand. 1995;92(5):351-8.

* Denotes studies included in the meta-analysis.

46. Feltner DE, Liu-Dumaw M, Schweizer E, Bielski R. Efficacy of pregabalin in generalized

social anxiety disorder: results of a double-blind, placebo-controlled, fixed-dose study. Int

Clin Psychopharmacol 2011; 26: 213–220.

* Denotes studies included in the meta-analysis.

 Pharmacotherapy for SAD 21

47. Furmark T, Appel L, Michelgåard A, et al. Cerebral blood flow changes after treatment of

social phobia with the neurokinin-1 antagonist GR205171, citalopram, or placebo. Biol

Psychiat. 2005;58(2):132-142.

* Denotes studies included in the meta-analysis.

48. Gelernter CS, Uhde TW, Cimbolic P, et al. Cognitive–behavioral and pharmacological

treatments of social phobia: a controlled study. Arch Gen Psychiatry 1991; 48: 938–945.

* Denotes studies included in the meta-analysis.

49. Gergel I, Pitts C, Oakes R, et al. 14-27-Significant improvement in symptoms of social

phobia after paroxetine treatment. Biol Psychiat. 1997;42(1):26S.

* Denotes studies included in the meta-analysis.

50. Giménez M, Ortiz H, Soriano-Mas C, et al. Functional effects of chronic paroxetine versus

placebo on the fear, stress and anxiety brain circuit in Social Anxiety Disorder: initial

validation of an imaging protocol for drug discovery. Euro Neuropsychopharmacol.

2014;24(1):105-16.

* Denotes studies included in the meta-analysis.

51. GlaxoSmithKline. A randomized, double-blind, parallel-group, placebo-controlled, forced-

dose titration study evaluating the efficacy and safety of a New Chemical Entity (NCE) and

paroxetine in subjects with social anxiety disorder. GSKClinical Study Register

[wwwgskclinicalstudyregistercom]. 2006.

* Denotes studies included in the meta-analysis.

52. Heimberg RG, Liebowitz MR, Hope DA, et al. Cognitive behavioral group therapy vs

phenelzine therapy for social phobia: 12 week outcome. Arch Gen Psychiat 1998;55:1133-

1141.
 Pharmacotherapy for SAD 22

* Denotes studies included in the meta-analysis.

53. Kasper S, Stein D, Loft H, Nil R. Escitalopram in the treatment of social anxiety disorder

Randomised, placebo-controlled, flexible-dosage study. Br J Psychiatry 2005;186(3):222-

226.

* Denotes studies included in the meta-analysis.

54. Katschnig H. Moclobemide in social phobia. Euro Arch Psychiat Clin Neurosci.

1997;247(2):71-80.

* Denotes studies included in the meta-analysis.

55. Katzelnick D, Kobak K, Greist J, et al. Sertraline for social phobia: a double-blind, placebo-

controlled crossover study. Am J Psychiatry 1995;152(9):1368-1371.

* Denotes studies included in the meta-analysis.

56. Kobak KA, Taylor LV, Warner G, et al. St John’s wort versus placebo in social phobia:

Results from a placebo-controlled pilot study. J Clin Psychopharmacol 2005;25(1):51-58.

* Denotes studies included in the meta-analysis.

57. Lader M, Stender K, Bürger V, Nil R. Efficacy and tolerability of escitalopram in 12-and 24-

week treatment of social anxiety disorder: Randomised, double-blind, placebo-controlled,

fixed-dose study. Depress Anxiety 2004;19(4):241-248.

* Denotes studies included in the meta-analysis.

58. Lepola U, Bergtholdt B, St Lambert J, et al. Controlled-release paroxetine in the treatment of

patietns with social anxiety disorder. J Clin Psychiat 2004;65(2):222-229.

* Denotes studies included in the meta-analysis.

59. Liebowitz M, Schneier F, Campeas R, et al. Phenelzine vs atenolol in social phobia: a

placebo-controlled comparison. Arch Gen Psychiatry 1992;49(4):290-300.

 Pharmacotherapy for SAD 23

* Denotes studies included in the meta-analysis.

60. Liebowitz M, Stein M, Tancer M, Carpenter D, Oakes R, Pitts C. A randomized, double-

blind, fixed-dose comparison of paroxetine and placebo in treatment of generalized social

anxiety disorder. J Clin Psychiatry 2002;63(1):66-74.

* Denotes studies included in the meta-analysis.

61. Liebowitz M, Gelenberg A, Munjack D. Venlafaxine extended release vs placebo and

paroxetine in social anxiety disorder. Arch Gen Psychiatry 2005a;62(2):190-198.

* Denotes studies included in the meta-analysis.

62. Liebowitz MR, Mangano RM, Bradwejn J, et al. A randomized controlled trial of

venlafaxine extended release in generalized social anxiety disorder. J Clin Psychiat

2005b;66:238-247.

* Denotes studies included in the meta-analysis.

63. Lott M, Greist JH, Jefferson JW, et al. Brofaromine for social phobia: a multicenter, placebo-

controlled, double-blind study. J Clin Psychopharmacol 1997;17(4):255-260.

* Denotes studies included in the meta-analysis.

64. Noyes R, Moroz G, Davidson J, et al. Moclobemide in social phobia: a controlled dose-

response trial. J Clin Psychopharmacol 1997;17(4):247-254.

* Denotes studies included in the meta-analysis.

65. Oosterbaan DB, van Balkom AJ, Spinhoven P, et al. Cognitive therapy versus moclobemide

in social phobia: a controlled study. Clin Psychol Psychothe 2001;8(4):263-73.

* Denotes studies included in the meta-analysis.

66. Pande AC, Davidson JR, Jefferson JW, et al. Treatment of social phobia with gabapentin: a

placebo-controlled study. J Clin Psychopharmacol 1999;19(4):341-348.

 Pharmacotherapy for SAD 24

* Denotes studies included in the meta-analysis.

67. Pfizer. A 10-week, randomized, double-blind, placebo-controlled study of paroxetine and

pregabalin in patients with social phobia (1008-081 and 1008-153). Web Synopsis Protocol

1008-081/153 – 21 June 2007 – Final; 2007.

* Denotes studies included in the meta-analysis.

68. Rickels K, Mangano R, Khan A. A double-blind, placebo-controlled study of a flexible dose

of venlafaxine ER in adult outpatients with generalized social anxiety disorder. J Clin

Psychopharmacol 2004;24(5):488-496.

* Denotes studies included in the meta-analysis.

69. Schneier F, Goetz D, Campeas R, et al. Placebo-controlled trial of moclobemide in social

phobia. Br J Psychiatry 1998;172(1):70-77.

* Denotes studies included in the meta-analysis.

70. Schutters SI, Van Megen HJ, Van Veen JF, et al. Mirtazapine in generalized social anxiety

disorder: a randomized, double-blind, placebo-controlled study. Int Clin Psychopharmacol

2010;25(5):302-304.

* Denotes studies included in the meta-analysis.

71. Stein MB, Liebowitz MR, Lydiard RB, et al. Paroxetine treatment of generalized social

phobia (social anxiety disorder). JAMA 1998;280(8):708-713.

72. Stein MB, Fyer AJ, Davidson JR, et al. Fluvoxamine treatment of social phobia (social

anxiety disorder): a double-blind, placebo-controlled study. Am J Psychiat 1999;156(5):756-

760.

* Denotes studies included in the meta-analysis.

 Pharmacotherapy for SAD 25

73. Stein D, Cameron A, Amrein R, et al. Moclobemide is effective and well tolerated in the

long-term pharmacotherapy of social anxiety disorder with or without comorbid anxiety

disorder. Int Clin Psychopharmacol 2002;17(4):161-170.

* Denotes studies included in the meta-analysis.

74. Stein MB, Pollack MH, Bystritsky A, et al. Efficacy of low and higher dose extended-release

venlafaxine in generalized social anxiety disorder: a 6-month randomized controlled trial.

Psychopharmacology 2005;177(3):280-288.

* Denotes studies included in the meta-analysis.

75. Sutherland SM, Tupler LA, Colket JT, et al. A 2-year follow-up of social phobia: Status after

a brief medication trial. J Nerv Ment Diseas. 1996;184(12):731-8.

76. Tauscher J, Kielbasa W, Iyengar S, et al. Development of the 2nd generation neurokinin-1

receptor antagonist LY686017 for social anxiety disorder. Eur Neuropsychopharmacol 2010;

20: 80–87.

* Denotes studies included in the meta-analysis.

77. The International Multicenter Clinical Trial Group on Moclobemide in Social Phobia.

Moclobemide in social phobia. A double-blind, placebo-controlled clinical study. Eur Arch

Psychiatry Clin Neurosci 1997;247(2),71-80.

* Denotes studies included in the meta-analysis.

78. Van Ameringen M, Lane R, Walker J, et al. Sertraline treatment of generalized social phobia:

a 20-week, double-blind, placebo-controlled study. Am J Psychiatry 2001;158(2):275-281.

* Denotes studies included in the meta-analysis.

 Pharmacotherapy for SAD 26

79. van Vliet IM, den Boer JA, Westenberg HG. Psychopharmacological treatment of social

phobia: clinical and biochemical effects of brofaromine, a selective MAO-A inhibitor. Euro

Neuropsychopharmacol. 1992;2(1):21-9.

* Denotes studies included in the meta-analysis.

80. van Vliet IM, den Boer JA, Westenberg HG. Psychopharmacological treatment of social

phobia: A double blind placebo controlled study with fluvoxamine. Psychopharmacol.

1994;115(1-2):128-34.

* Denotes studies included in the meta-analysis.

81. Versiani M, Nardi AE, Mundim FD, et al. Pharmacotherapy of social phobia: A controlled

study with moclobemide and phenelzine. Brit J Psychiat 1992;161(3):353-360.

* Denotes studies included in the meta-analysis.

82. Westenberg H, Stein D, Yang H, et al. A double-blind placebo-controlled study of controlled

release fluvoxamine for the treatment of generalized social anxiety disorder. J Clin

Psychopharmacol 2004;24(1):49-55.

* Denotes studies included in the meta-analysis.

83. Zhang W, Connor KM, Davidson JR. Levetiracetam in social phobia: a placebo controlled

pilot study. J Psychopharmacol 2005;19(5):551-553.

* Denotes studies included in the meta-analysis.

 Pharmacotherapy for SAD 27

Table 1. Studies included in the meta-analysis Risk of Bias

Study Pharmacotherapy N Mean % Tx Outcome Diagnostic S A I S total

Type Age Female Weeks Measure Subtype G C D R
Allgulander et al., 1999 [34] paroxetine 92 41.0 NR 12 LSAS Mixed ✓ ? X ✓ X

Allgulander et al., 2004 [35] paroxetine; 389 38.8 53.4 12 LSAS GSAD ✓ ✓ X ✓ X
venlafaxine ER
Asakura et al., 2007 [36] fluvoxamine 265 32.4 38.8 10 LSAS GSAD ? ? ✓ ✓ ?
Baldwin et al., 1999 [37] paroxetine 290 36.1 54.2 12 LSAS; CGI-S Mixed ? ✓ X ✓ X

Barnett et al., 2002 [38] olanzapine 12 NR NR 8 LSAS Mixed ? ? ✓ ✓ ?

Blanco et al., 2010 [39] phenelzine 96 31.4 40.6 12 LSAS; CGI-S Mixed ✓ ✓ ✓ ✓ ✓

Blomhoff et al., 2001 [40] sertraline 277 40.4 60.5 24 CGI-SP GSAD ✓ ✓ ? ✓ ?

Clark et al., 2003 [41] fluoxetine 60 33.2 52.0 16 LSAS GSAD ? ✓ ✓ ✓ ?

Davidson et al., 1993 [42]* clonazepam 75 37.2 42.7 10 LSAS; CGI-S Mixed ? ? X ✓ X

Davidson et al., 2004a [43] fluoxetine 117 36.6 47.4 14 CGI-S GSAD ✓ ✓ ? ✓ ?

Davidson et al., 2004b [44] fluvoxamine 279 37.3 35.8 12 LSAS, CGI-S GSAD ? ? X ✓ X

Fahlen et al., 1995 [45] brofaromine 76 37.82 41.56 12 LSAS Mixed ? ? X ✓ X

 Pharmacotherapy for SAD 28

Feltner et al., 2011 [46] pregabalin**** 329 35.1 39.8 10 LSAS GSAD ? ? X ✓ X

Furmark et al., 2005 [47] citalopram; 36 31.6 52.8 6 CGI-S Mixed ? ✓ ? ✓ ?

GR205171
Gelernter et al., 1991 [48] alprazolam; 65 36.5 63 12 FQ Mixed ? ✓ ✓ ? ?
phenelzine
Gergel et al., 1997 [49] paroxetine 187 NR NR 12 LSAS Mixed ? ? X ✓ X

Gimenez et al., 2014 [50] paroxetine 33 23.0 84.8 8 LSAS Mixed ? ✓ ✓ ✓ ?

GlaxoSmithKline, 2013 [51] paroxetine 239 35.2 42 12 LSAS; CGI-S Mixed ✓ ✓ ✓ ✓ ✓

Heimberg et al., 1998 [52] phenelzine 81 34.9 49.6 12 SPDSC Mixed ? ? ✓ ✓ ?

Kasper et al., 2005 [53] escitalopram 358 37.5 45.5 12 LSAS GSAD ? ✓ X ✓ X

Katschnig et al., 1997 [54] moclobemide*** 578 36.4 43 12 LSAS Mixed ? ✓ X ✓ X

Katzelnick et al., 1995 [55] sertraline 12 42.6 33.3 10 LSAS Mixed ? ? X ✓ X

Kobak et al., 2002 [21]** fluoxetine 60 39.5 58.0 14 LSAS GSAD ? ? X ✓ X

Kobak et al., 2005 [56] St. John's Wort 40 37.5 52.5 12 LSAS GSAD ? ? X ✓ X

Lader et al., 2004 [57] escitalopram****; 820 36.9 47.2 24 LSAS; CGI-S GSAD ? ? X ✓ X
paroxetine
Lepola et al., 2004 [58] paroxetine 370 38.9 50.0 12 CGI-S Mixed ? ? ? ✓ ?

Liebowitz et al., 1992 [59] atenolol; 74 34.3 31.0 8 CGI-S Mixed ? ? ✓ ✓ ?

phenelzine
 Pharmacotherapy for SAD 29

Liebowitz et al., 2002 [60] paroxetine**** 360 37.0 41.4 12 LSAS; CGI-S GSAD ? ? X ✓ X

Liebowitz et al., 2003 [22]** sertraline 401 35.1 40.5 12 LSAS; CGI-S GSAD ? ? X ✓ X

Liebowitz et al., 2005a [61] paroxetine; 413 36.3 46.5 12 LSAS; CGI-S GSAD ? ? X ✓ X
venlafaxine ER
Liebowitz et al., 2005b [62] venlafaxine ER 271 35.4 45.0 12 LSAS GSAD ? ? X ✓ X

Lott et al., 1997 [63] brofaromine 102 36.5 39.2 10 LSAS Mixed ? ? X ✓ X

Noyes et al., 1997 [64] moclobemide***** 506 38.1 42.7 12 LSAS; CIS-SP Mixed ? ✓ X ✓ X

Oosterbaan et al., 2001 [65]* moclobemide 67 37 41.5 15 LSAS Mixed ? ✓ ? ✓ ?

Pande et al., 1999 [66] gabapentin 69 35.6 42.0 14 LSAS Mixed ? ? X ✓ X

Pande et al., 2004 [23]** pregabalin*** 135 38.4 41.5 10 CGI-S GSAD ? ? X ✓ X

Pfizer, 2007 [67] paroxetine; 371 37 38 10 LSAS; CGI-S GSAD ? ? X ✓ X

pregabalin***
Rickels et al., 2004 [68] venlafaxine ER 261 41.5 55.6 12 LSAS; CGI-S GSAD ? ? X ✓ X

Schneier et al., 1998 [69] moclobemide 77 35.5 44.9 8 LSAS Mixed ? ? ✓ ✓ ?

Schutters et al., 2010 [70] mirtazapine 60 38.6 56.7 12 LSAS GSAD ? ? ✓ ✓ ?

Stein et al., 1998 [71] paroxetine 187 NR NR 12 LSAS GSAD ✓ ✓ X ✓ X

Stein et al., 1999 [72] fluvoxamine 92 39.4 35.9 12 LSAS Mixed ? ? X ✓ X

 Pharmacotherapy for SAD 30

Stein et al., 2002 [73] moclobemide 377 34.4 47.0 12 LSAS Mixed ? ? X ✓ X

Stein et al., 2005 [74] venlafaxine ER*** 364 36.9 41.7 24 LSAS GSAD ✓ ✓ X ✓ X

Tauscher et al., 2009 [76] paroxetine 129 NR NR 12 LSAS GSAD ? ? X ✓ X

TIMCTG, 1997 [77] moclobemide*** 578 36.4 43.0 12 LSAS; CIS-SP Mixed ? ✓ ? ✓ ?

van Ameringen et al., 2001 [78] sertraline 203 35.7 44.4 20 CGI-S GSAD ? ? X ✓ X

van Vliet et al., 1992 [79] brofaromine 29 NR NR 12 LSAS Mixed ? ? ✓ ✓ ?

van Vliet et al., 1994 [80] fluvoxamine 28 35.2 56.67 12 LSAS Mixed ? ? ✓ ✓ ?

Versiani et al., 1992 [81] moclobemide; 78 NR NR 8 CGI-S Mixed ? ? X ✓ X

phenelzine
Westenberg et al., 2004 [82] fluvoxamine 294 38.0 52.0 12 LSAS; CGI-S GSAD ? ? X ✓ X

Zhang et al., 2005 [83] levetiracetam 16 37.5 53.0 7 LSAS Mixed ? ? X ✓ X

*Long term follow up data were reported in this study or a follow up study

**Study had quality of life outcomes reported

***Study included 2 doses of the same medication

****Study included 3 doses of the same medication

*****Study included 5 doses of the same medication

 Pharmacotherapy for SAD 31

Table 2. Effect sizes for pre-post, pre-FU, and QoL

95% 95%
Medication CI CI
Category Group Medication Study Hedges's g SE Low High Z p
Pre-post Anticonvulsant gabapentin Pande et al., 1999 0.53 0.24 0.05 1.00 2.18 0.03
pregabalin (150mg) Pande et al., 2004 0.04 0.21 -0.37 0.46 0.21 0.83
pregabalin (600 0.34 0.21 -0.06 0.75 1.66 0.10
mg) Pande et al., 2004
pregabalin (200mg) Pfizer, 2007 0.25 0.16 -0.07 0.57 1.56 0.13
pregabalin (400mg) Pfizer, 2007 0.03 0.17 -0.29 0.35 0.16 0.87
levetiracetam Zhang et al., 2005 0.39 0.48 -0.55 1.34 0.82 0.41
Pooled 0.21 0.08 0.05 0.38 2.52 0.01
anticonvulsant
Antipsychotics olanzapine Barnett et al., 2002 0.72 0.56 -0.38 1.82 1.29 0.20
Pooled 0.72 0.56 -0.38 1.82 1.29 0.20
antipsychotics
Benzodiazepines clonazepam Davidson et al., 1993 0.97 0.24 0.49 1.44 4.00 0.00
alprazolam Gelernter et al., 1991 0.49 0.38 -0.25 1.24 1.30 0.19
Pooled 0.82 0.22 0.40 1.25 3.78 0.00
benzodiazepines
Beta-blockers atenolol Liebowitz et al., 1992 0.08 0.28 -0.47 0.64 0.30 0.77
Pooled beta- 0.08 0.28 -0.47 0.64 0.30 0.77
blockers
Herbal St. John's Wort Kobak et al., 2005 -0.07 0.31 -0.68 0.54 -0.22 0.82
Pooled herbal -0.07 0.31 -0.68 0.54 -0.22 0.82
MAO-A brofaromine Fahlen et al., 1995 0.70 0.23 0.24 1.16 2.98 0.00
brofaromine Lott et al., 1997 0.36 0.20 -0.03 0.75 1.81 0.07
brofaromine van Vliet et al., 1992 1.02 0.39 0.26 1.78 2.62 0.01
Pooled MAO-A 0.60 0.17 0.26 0.94 3.44 0.00
 Pharmacotherapy for SAD 32

MAOI phenelzine Blanco et al., 2010 0.49 0.26 -0.02 0.99 1.90 0.06
phenelzine Gelernter et al., 1991 0.33 0.37 -0.40 1.06 0.89 0.37
phenelzine Heimberg et al., 1998 1.15 0.29 0.57 1.72 3.91 0.00
moclobemide 0.22 0.10 0.02 0.42 2.15 0.03
(300mg) Katschnig et al., 1997
moclobemide 0.28 0.10 0.08 0.48 2.73 0.01
(600mg) Katschnig et al., 1997
phenelzine Liebowitz et al., 1992 0.86 0.29 0.30 1.43 2.99 0.00
moclobemide 0.17 0.15 -0.13 0.47 1.10 0.27
(150mg) Noyes et al., 1997
moclobemide 0.12 0.15 -0.18 0.42 0.78 0.44
(300mg) Noyes et al., 1997
moclobemide 0.13 0.15 -0.17 0.43 0.84 0.40
(600mg) Noyes et al., 1997
moclobemide 0.01 0.15 -0.29 0.31 0.08 0.93
(75mg) Noyes et al., 1997
moclobemide 0.19 0.15 -0.12 0.49 1.21 0.23
(900mg) Noyes et al., 1997
moclobemide Oosterbaan et al., 2001 -0.25 0.30 -0.85 0.34 -0.83 0.40
moclobemide Schneier et al., 1998 0.20 0.23 -0.25 0.65 0.89 0.37
moclobemide Stein et al., 2002 0.26 0.10 0.06 0.47 2.56 0.01
moclobemide 0.37 0.10 0.17 0.57 3.64 0.00
(300mg) TIMCTG, 1997
moclobemide 0.19 0.10 -0.01 0.39 1.84 0.07
(600mg) TIMCTG, 1997
moclobemide Versiani et al., 1992 1.22 0.30 0.64 1.81 4.10 0.00
phenelzine Versiani et al., 1992 2.16 0.35 1.48 2.84 6.25 0.00
Pooled MAOI 0.36 0.08 0.21 0.51 4.74 0.00
NaSSA mirtazapine Schutters et al., 2010 0.13 0.26 -0.37 0.63 0.51 0.61
Pooled NaSSA 0.13 0.26 -0.37 0.63 0.51 0.61
NK1 GR205171 Furmark et al., 2005 0.46 0.40 -0.33 1.24 1.14 0.25
 Pharmacotherapy for SAD 33

Pooled NK1 0.46 0.40 -0.33 1.24 1.14 0.25

SNRI venlafaxine ER Allgulander et al., 2004 0.62 0.13 0.37 0.87 4.92 0.00
venlafaxine ER Liebowitz et al., 2005a 0.40 0.12 0.16 0.63 3.29 0.00
venlafaxine ER Liebowitz et al., 2005b 0.48 0.12 0.23 0.72 3.87 0.00
venlafaxine ER Rickels et al., 2004 0.37 0.12 0.12 0.61 2.94 0.00
venlafaxine ER 0.42 0.13 0.17 0.68 3.29 0.00
(150mg) Stein et al., 2005
venlafaxine ER 0.42 0.13 0.17 0.68 3.29 0.00
(75mg) Stein et al., 2005
Pooled SNRI 0.45 0.05 0.35 0.55 8.82 0.00
SSRI paroxetine Allgulander et al., 1999 0.84 0.22 0.42 1.27 3.90 0.00
paroxetine Allgulander et al., 2004 0.59 0.13 0.34 0.83 4.64 0.00
fluvoxamine Asakura et al., 2007 0.30 0.13 0.05 0.56 2.33 0.02
paroxetine Baldwin et al., 1999 0.50 0.12 0.26 0.73 4.18 0.00
sertraline Blomhoff et al., 2001 0.27 0.15 -0.02 0.56 1.83 0.07
fluoxetine Clark et al., 2003 0.11 0.31 -0.50 0.72 0.35 0.72
fluoxetine Davidson et al., 2004a 0.56 0.19 0.19 0.92 2.96 0.00
fluvoxamine Davidson et al., 2004b 0.51 0.13 0.26 0.77 3.98 0.00
citalopram Furmark et al., 2005 0.32 0.40 -0.46 1.10 0.81 0.42
paroxetine Gergel et al., 1997 0.62 0.15 0.33 0.91 4.17 0.00
paroxetine Gimenez et al., 2014 1.15 0.37 0.43 1.87 3.12 0.00
paroxetine GSK, 2013 0.88 0.28 0.33 1.43 3.12 0.00
escitalopram Kasper et al., 2005 0.25 0.11 0.04 0.46 2.34 0.02
sertraline Katzelnick et al., 1995 0.80 0.56 -0.30 1.89 1.43 0.15
fluoxetine Kobak et al., 2002 -0.03 0.25 -0.53 0.47 -0.11 0.91
escitalopram 0.25 0.11 0.03 0.47 2.27 0.02
(10mg) Lader et al., 2004
escitalopram 0.37 0.11 0.15 0.59 3.29 0.00
(20mg) Lader et al., 2004
escitalopram (5mg) Lader et al., 2004 0.28 0.11 0.07 0.50 2.58 0.01
 Pharmacotherapy for SAD 34

paroxetine Lader et al., 2004 0.32 0.11 0.11 0.54 2.94 0.00
paroxetine Lepola et al., 2004 0.64 0.11 0.43 0.85 6.02 0.00
paroxetine (20 mg) Liebowitz et al., 2002 0.47 0.15 0.18 0.77 3.14 0.00
paroxetine (40 mg) Liebowitz et al., 2002 0.36 0.15 0.06 0.65 2.38 0.02
paroxetine (60 mg) Liebowitz et al., 2002 0.33 0.15 0.04 0.62 2.20 0.03
sertraline Liebowitz et al., 2003 0.33 0.10 0.14 0.53 3.31 0.00
paroxetine Liebowitz et al., 2005a 0.40 0.12 0.16 0.64 3.29 0.00
paroxetine Pfizer, 2007 0.44 0.16 0.12 0.76 2.73 0.01
paroxetine Stein et al., 1998 0.63 0.15 0.33 0.93 4.18 0.00
fluvoxamine Stein et al., 1999 0.67 0.22 0.24 1.10 3.04 0.00
paroxetine Tauscher et al., 2009 1.37 0.26 0.86 1.88 5.29 0.00
van Ameringen et al., 0.38 0.15 0.09 0.68 2.58 0.01
sertraline 2001
fluvoxamine van Vliet et al., 1994 0.75 0.38 0.00 1.50 1.95 0.05
fluvoxamine Westenberg et al., 2004 0.27 0.12 0.04 0.50 2.31 0.02
Pooled SSRI 0.44 0.04 0.37 0.51 12.17 0.00

Pooled Overall 0.41 0.03 0.36 0.46 16.34 0.00

Pre-follow up clonazepam Davidson et al., 1993 0.50 0.27 -0.04 1.04 1.81 0.07
moclobemide Oosterbaan et al., 2001 0.02 0.30 -0.57 0.61 0.06 0.95
Pooled Overall 0.27 0.24 -0.20 0.74 1.14 0.25
Quality of Life fluoxetine Kobak et al., 2002 0.20 0.26 -0.30 0.71 0.80 0.42
sertraline Liebowitz et al., 2003 0.39 0.01 0.20 0.58 3.94 0.00
pregabalin (150mg) Pande et al., 2004 0.07 0.21 -0.34 0.49 0.35 0.72
pregabalin (600mg) Pande et al., 2004 0.09 0.21 -0.31 0.50 0.46 0.65
Pooled Overall 0.28 0.08 0.13 0.44 3.52 0.00
 Pharmacotherapy for SAD 35

 Pharmacotherapy for SAD 36

Figure 1

Potentially
relevant abstracts identified
and screened for retrieval (n = 2208)

Abstracts excluded due to lack of

relevance to the study (n=2099)

Articles selected for further

screening (n = 119)

Articles excluded (n = 67) for the following
reasons:
• No active control group (n = 22)
• No DSM diagnosis of SAD (n = 11)
• Necessary measures not included (n =
12)
• Treatment responders only (n = 6)
Studies included in the meta-
• Augmentation studies (n = 5)
analysis (n = 52)
• Insufficient data (n = 4)
• < 2 treatment doses (n = 2)
• No pharmacotherapy (n = 5)
 Pharmacotherapy for SAD 37

Figure 2
Funnel Plot of Standard Error by Hedges's g

0.0

0.1

0.2
Standard Error

0.3

0.4

0.5

0.6

-3 -2 -1 0 1 2 3

Hedges's g