The short answer
- SSRIs and SNRIs were the only drug class to beat pill placebo in the NICE network meta-analysis, and SSRIs are the only class with moderate-quality evidence that continuing them prevents relapse.
- MAOIs (phenelzine) and benzodiazepines show large effects, but on few trials, and both carry real risks.
- Pregabalin helped only at the highest dose tested. Propranolol has almost no trial evidence for social anxiety disorder.
- Stopping an antidepressant after response roughly triples the odds of relapse within a year (anxiety disorders pooled).
- Against an appropriate placebo, individual CBT and SSRIs are the two treatments that clearly work. On pre-post change, drugs look better than therapy. On controlled comparisons, individual CBT comes out ahead.
Medication decisions belong with a prescriber. If a client wants to stop or change a medicine, they should do it with their doctor, and taper rather than stop suddenly.
Table 1. NICE network meta-analysis (Mayo-Wilson 2014)
A network meta-analysis of 101 trials with 13,164 adults, done for the NICE guideline. Effects are standardized mean differences on social anxiety symptoms at the end of treatment. Negative values favour treatment.
| Intervention class | vs waiting list (SMD) | vs an appropriate placebo |
|---|---|---|
| Individual CBT | −1.19 | −0.56 vs psychological placebo |
| MAOIs | −1.01 | Not given |
| Benzodiazepines | −0.96 | Not given |
| Group CBT | −0.92 | Not given |
| SSRIs / SNRIs | −0.91 | −0.44 vs pill placebo |
| Anticonvulsants (gabapentinoids) | −0.81 | Not given |
| Psychodynamic therapy | −0.62 | Not given |
In the summary of this analysis, only individual CBT and SSRIs/SNRIs were reported as beating an appropriate placebo. “Not given” means no placebo-controlled figure is given here.
Table 2. Cochrane review of medication (Williams 2017)
The Cochrane review pooled 66 randomized trials (11,597 participants). It used treatment response, a yes/no outcome, so the figures are risk ratios: how many times more likely a person is to respond on the drug than on placebo.
| Class | Response vs placebo (risk ratio) | Notes |
|---|---|---|
| SSRIs | 1.65 (24 trials) | Evidence for response rated very low quality. The only class with moderate-quality evidence for relapse prevention |
| MAOIs | 2.36 | Phenelzine needs dietary and drug-interaction precautions |
| Benzodiazepines | 4.03 (2 trials) | Very small evidence base; dependence risk |
| GABA-analogue anticonvulsants (pregabalin, gabapentin) | 1.60 | Pregabalin: only 600 mg/day beat placebo in a fixed-dose trial of 329 people (LSAS −29.8 vs −19.7); gabapentin beat placebo in one trial of 69 people |
The large risk ratio for benzodiazepines comes from 2 trials. Treat it as a hint, not a ranking.
Table 3. What happens when people stop
| Study | What was done | Result |
|---|---|---|
| Batelaan 2017 | 28 relapse-prevention trials (5,233 people who had responded to an antidepressant), all anxiety disorders pooled | Relapse 36.4% when switched to placebo vs 16.4% when continued, within up to 1 year; odds ratio 3.11 |
| Connor 1998 | People who had responded to 6 months of clonazepam | Relapse 0% if continued vs 21.1% with a slow taper; withdrawal symptoms 12.5% vs 27.7% |
| Clark 2003 | Cognitive therapy vs fluoxetine plus self-exposure vs placebo plus self-exposure | Cognitive therapy ahead at 16 weeks and still at 12 months |
Batelaan pooled all anxiety disorders and followed people for a year at most. There is no good social-anxiety-specific data on stopping beyond a year. NICE advises continuing a medicine that has worked in the first 3 months for at least a further 6 months, then reducing the dose gradually. If symptoms come back, options include restarting, raising the dose, or offering individual CBT.
Table 4. Combining medication and therapy
| Trial | Combination | Result |
|---|---|---|
| Blanco 2010 (n = 128, US) | Phenelzine + group CBT (CBGT) vs each alone vs placebo | Remission at 24 weeks: 53.1% combined, 25.7% phenelzine, 23.5% CBGT, 14.8% placebo |
| Heimberg 1998 (n = 133, US) | Phenelzine vs CBGT vs two placebos (no combination arm) | Both beat placebo; phenelzine faster and ahead on some measures at 12 weeks |
| Clark 2003 (n = 60, UK) | Fluoxetine + self-exposure vs individual cognitive therapy | Fluoxetine + self-exposure no better than placebo + self-exposure; cognitive therapy better than both |
The phenelzine + CBGT result cannot be transferred to an SSRI + individual cognitive therapy. No trial has tested that combination directly. NICE suggests adding individual CBT for partial response to an SSRI after 10 to 12 weeks, and adding medication for partial response to CBT.
Why the rankings change with the comparison
Drugs and therapies swap places depending on the yardstick. This confuses clients and clinicians alike.
- Against waiting list, everything looks good. Waiting lists include no expectation of help, so the gap is large. Individual CBT, MAOIs, benzodiazepines and SSRIs all land between about −0.9 and −1.2.
- Against placebo, the bar is higher. Drug trials use a pill placebo, which itself helps many people. Therapy trials need a psychological placebo, which is harder to design. In the NICE analysis only individual CBT (−0.56) and SSRIs/SNRIs (−0.44) cleared it. Across anxiety disorders, CBT against placebo was g = 0.56.
- Before and after treatment (pre-post), drugs win. A meta-analysis of 234 trials across panic, GAD and social phobia found pre-post d = 2.02 for medication and 1.22 for psychotherapy. Pre-post change includes placebo response, time and regression to the mean. In the same analysis, pill placebo arms alone showed a pre-post d of 1.29, against 0.20 for waitlists.
- Few trials, big numbers. The benzodiazepine and MAOI estimates rest on a handful of small, older trials. A small number of trials makes extreme estimates more likely.
- End of treatment only. Most comparisons stop when treatment ends. Therapy gains tend to last after it ends; medication gains depend on continuing (Table 3).
The NICE economic model found individual cognitive therapy the most cost-effective option overall, with phenelzine also among the more cost-effective choices. Side effects of drugs were not included in that model.
Classes to be careful with
- Benzodiazepines: large short-term effects in 2 trials, but dependence, withdrawal and interaction with alcohol, which many people with social anxiety use to cope. NICE says not to offer them routinely.
- Propranolol and other beta-blockers: only one small trial in social phobia (16 people), and not enough evidence for routine use. Some guidelines accept them for one-off performance situations. In online forums, people with social anxiety commonly report that propranolol calms shaking and a racing heart but not the fearful thoughts, which fits a drug that acts only on the body.
- SSRIs and emotional blunting: in a qualitative study, some SSRI users described feeling emotionally flattened and linked it strongly to the drug. Ask about it at follow-up.
For guideline differences on these classes, see guidelines compared. For clients who also drink heavily or are depressed, see adapting treatment.
Sources
- Mayo-Wilson E, Dias S, Mavranezouli I, Kew K, Clark DM, Ades AE, Pilling S (2014). Psychological and pharmacological interventions for social anxiety disorder in adults: a systematic review and network meta-analysis.Lancet Psychiatry.
- Williams T, Hattingh CJ, Kariuki CM, Tromp SA, van Balkom AJ, Ipser JC, Stein DJ (2017). Pharmacotherapy for social anxiety disorder (SAnD).Cochrane Database of Systematic Reviews.
- Batelaan NM, Bosman RC, Muntingh A, Scholten WD, Huijbregts KM, van Balkom AJLM (2017). Risk of relapse after antidepressant discontinuation in anxiety disorders, obsessive-compulsive disorder, and post-traumatic stress disorder: systematic review and meta-analysis of relapse prevention trials.BMJ.
- Blanco C, Heimberg RG, Schneier FR, Fresco DM, Chen H, et al. (2010). A placebo-controlled trial of phenelzine, cognitive behavioral group therapy, and their combination for social anxiety disorder.Archives of General Psychiatry.
- Bandelow B, Reitt M, Röver C, Michaelis S, Görlich Y, Wedekind D (2015). Efficacy of treatments for anxiety disorders: a meta-analysis.International Clinical Psychopharmacology.
- Heimberg RG, Liebowitz MR, Hope DA, Schneier FR, et al. (1998). Cognitive behavioral group therapy vs phenelzine therapy for social phobia: 12-week outcome.Archives of General Psychiatry.
- Clark DM, Ehlers A, McManus F, Hackmann A, Fennell M, Campbell H, Flower T, Davenport C, Louis B (2003). Cognitive therapy versus fluoxetine in generalized social phobia: a randomized placebo-controlled trial.Journal of Consulting and Clinical Psychology.
- Feltner DE, Liu-Dumaw M, Schweizer E, Bielski R (2011). Efficacy of pregabalin in generalized social anxiety disorder: results of a double-blind, placebo-controlled, fixed-dose study.International Clinical Psychopharmacology.
- Pande AC, Davidson JR, Jefferson JW, Janney CA, Katzelnick DJ, Weisler RH, Greist JH, Sutherland SM (1999). Treatment of social phobia with gabapentin: a placebo-controlled study.Journal of Clinical Psychopharmacology.
- Connor KM, Davidson JR, Potts NL, Tupler LA, et al. (1998). Discontinuation of clonazepam in the treatment of social phobia.Journal of Clinical Psychopharmacology.
- Steenen SA, van Wijk AJ, van der Heijden GJ, van Westrhenen R, de Lange J, de Jongh A (2016). Propranolol for the treatment of anxiety disorders: Systematic review and meta-analysis.Journal of Psychopharmacology.
- Mavranezouli I, Mayo-Wilson E, Dias S, Kew K, Clark DM, Ades AE, Pilling S (2015). The Cost Effectiveness of Psychological and Pharmacological Interventions for Social Anxiety Disorder: A Model-Based Economic Analysis.PLoS One.
- Carpenter JK, Andrews LA, Witcraft SM, Powers MB, Smits JAJ, Hofmann SG (2018). Cognitive behavioral therapy for anxiety and related disorders: A meta-analysis of randomized placebo-controlled trials.Depression and Anxiety.
- Price J, Cole V, Goodwin GM (2009). Emotional side-effects of selective serotonin reuptake inhibitors: qualitative study.British Journal of Psychiatry.
- National Institute for Health and Care Excellence (NICE) (2013). Social anxiety disorder: recognition, assessment and treatment (CG159).NICE clinical guideline.
- NICE CG159: recommendations on medication for adults (1.3.5 to 1.3.25, 1.6)