A meta-analysis of 101 trials with more than 13,000 participants with the disorder showed that individual CBT had the largest effect sizes and “should be regarded as the best intervention for initial treatment,” write the investigators. This form of talk therapy showed significantly better outcomes than psychological placebo, psychodynamic psychotherapy, interpersonal psychotherapy, and mindfulness training.

Selective serotonin-reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs) made up the only class of pharmacologic agents that had greater outcome effects than pill placebos.

“The good news is we found a lot of interventions were better than doing nothing. Specifically, there was very good evidence for one-on-one cognitive-behavioral therapy,” lead author Evan Mayo-Wilson, DPhil, research scientist in the Department of Epidemiology at Johns Hopkins University Bloomberg School of Public Health, in Baltimore, Maryland, told Medscape Medical News..

“There’s a push in a lot of healthcare services to move towards a stepped-care model, where you’re trying to save money by doing therapy in groups. But we found that individual therapy was by far the most beneficial and worth the added cost. And compared with medication taken over a long period of time, talk therapy is actually more cost-effective.”

Dr Mayo-Wilson reported that the findings have led to new treatment guidance in the United Kingdom, where he previously worked.

The study was published online September 26 in Lancet Psychiatry.

First Study of Its Kind

Dr Mayo-Wilson noted that although pharmacologic, psychological, and self-help interventions can be used to treat social anxiety disorder, no previous study has compared all of these interventions to determine which is most effective.

He also noted that previous studies suggest that individual CBT has a lower risk for treatment-related adverse events than pharmacotherapy.

The investigators reviewed data for 101 studies conducted between 1988 and 2013. Of the 13,164 adult participants (52% women; mean age, 36 years), all had severe, long-standing social anxiety.

The primary outcomes for the meta-analysis were effects on symptoms of social anxiety, as shown in validated measures, vs “a common reference condition” (wait-list reference). Of the trials that included a control group, 28 used a wait-list reference, 42 used placebo pills, and 6 used psychological placebo.

Results showed that psychological interventions that were efficacious compared with wait-list references were individual CBT (standardized mean difference [SMD], -1.19), group CBT (SMD, -0.92), exposure and social skills (SMD, -0.86), self-help with support from a book or the Internet (SMD, -0.86), self-help without support (SMD, -0.75), and psychodynamic psychotherapy (SMD, -0.62).

Promotion of exercise and “other psychological therapies” (which included supportive therapy, mindfulness training, and interpersonal psychotherapy) did not show greater outcome effects than the wait-list reference.

Individual CBT was the only class of psychological intervention to perform better than psychological placebo (SMD, -0.56) and pill placebo (SMD, -0.72). It also had greater effects in head-to- head comparisons with psychodynamic psychotherapy (SMD, -0.56) and the class made up of interpersonal psychotherapy/mindfulness/supportive therapy (SMD, -0.82).

Classes of pharmacologic agemts that showed significantly greater outcome effects than wait-list references were monoamine oxidase inhibitors (MOIs; SM, -1.01), benzodiazepines (SMD, -0.96), SSRIs and SNRIs (SMD, -0.91), and anticonvulsants (SMD, -0.81). Noradrenergic and specific serotonergic antidepressants, which only included mirtazapine (Remeron, Organon Pharmaceuticals USA Inc), did not show greater outcome effects.

Although the MOIs and benzodiazepines had the largest class effects, the investigators note that the evidence for these agents was limited compared with that for SSRIs and SNRIs. In addition, more of the participants received SSRIs and SNRIs (n = 4043) than any other intervention class.

SSRIs and SNRIs also showed greater effects than pill placebo (SMD, -0.44).

Preferred Treatment

When looking at individual medications, the SSRIs citalopram (multiple brands), escitalopram (Lexapro, Forest Laboratories, Inc) fluoxetine (multiple brands), fluvoxamine (Luvox, ANI Pharmaceuticals, Inc, and Jazz Pharmaceuticals, Inc) paroxetine (multiple brands), and sertraline (Zoloft, Pfizer Inc), as well as the SNRI venlafaxine (mutiple brands), all showed greater outcome effects than the wait-list reference. With the exception of citalopram, all of these drugs, as well as clonazepam (Klonapin, Roche), moclobemide (multiple brands), and phenelzine (Nardil, Parke-Davis), were more effective than pill placebo.

There was no evidence that combining medication and talk therapy was better than talk therapy alone.

Because there was little evidence of differential efficacy within or between classes of drugs, differences in tolerability and side effects are especially important when choosing treatment, note the investigators.

“SSRIs and SNRIs with a short half-life…are associated with the greatest risk of discontinuation effects,” they write. “Some side effects such as increased agitation and sexual dysfunction can be especially distressing for people with social anxiety disorder.”

The authors add that these issues should be discussed with patients before starting treatment. In addition, they note that most patients who respond to an SSRI will relapse within a few months if the drug is discontinued after acute treatment ― and 25% of those who continue treatment will still relapse within 6 months.

On the other hand, effects from psychological interventions are generally well maintained. “For this reason, and because of the lower risk of side-effects, psychological interventions should be preferred over pharmacological interventions for initial treatment,” they write.

However, if a patient declines talk therapy, the investigators recommend treatment with an SSRI. “Other drugs are recommended only for people who do not respond to initial treatment.”

Dr Mayo-Wilson noted that his investigative team examined cost-effectiveness of the therapies, the results of which will be reported in a separate article, and found that group therapy “was not worth the money” vs individual therapy.

“The conclusion we reached was that individual CBT was the most cost-effective thing you could do for patients with social anxiety disorder in the long term, and it doesn’t have the side effects you have from taking medication,” he said.

Good News for Patients

Pim Cuijpers, PhD, from the Department of Clinical Psychology at the VU University Amsterdam, the Netherlands, and Prof Dr Annemieke van Straten, from the EMGO Institute for Health and Care Research in Amsterdam, write in anaccompanying editorial that the meta-analysis is “an important next step in improving our knowledge” about social anxiety disorder treatments.

“The findings confirm that both psychological and pharmacological treatments are effective, and that the effects of these treatments are large,” they write. “The message that social anxiety disorder can often be treated successfully is important from a public health perspective, and would certainly be a good subject for an educational public campaign.”

The editorialists note that although the investigators write that individual CBT is better than SSRIs/SNRIs because of the latter group’s high relapse rates, “the long-term effects of treatment could not be included in this meta-analysis because the number of studies examining this was too small.”

“To really help patients and reduce their distress, long-term trials on drugs and psychotherapy are urgently needed,” the editorialists write.

Still, they conclude that the promising findings from the current analysis are “good news for our patients!”

Dr Mayo-Wilson, five of the other study authors, and the editorialists have reported no relevant financial relationships. The remaining study author reports being the developer of one of the versions of individual CBT examined.

Lancet Psychiatry. Published online September 26, 2014. Full article, Editorial