Several kinds of anxiety medication have real evidence behind them, but the research is strongest for one condition, generalised anxiety disorder, and answers the question of which drug suits you only in part. A 2019 analysis of 89 trials found that duloxetine, pregabalin, venlafaxine and escitalopram were more efficacious than placebo with relatively good acceptability. [slee-2019] This guide covers what the main reviews found for each family of medication, what the evidence says about stopping, what it could not show and what to ask a prescriber. Our overview of anxiety covers the conditions themselves.
Which medications work for anxiety?
For generalised anxiety disorder (GAD), several do, across more than one class of drug. A 2019 network meta-analysis compared 22 drugs with placebo across 89 trials and 25,441 adult outpatients, and found that duloxetine, pregabalin, venlafaxine and escitalopram were more efficacious than placebo, with relatively good acceptability. [slee-2019] Escitalopram is an SSRI, venlafaxine and duloxetine are SNRIs, and pregabalin belongs to a different group again.
The authors described it as the largest contemporary review of drugs for GAD to use network analysis, which allows comparisons across drugs that were never tested head to head. Their conclusion was that there are several effective treatment choices across classes of medication, and that failure of initial drug therapy might not be a reason to abandon a drug-based strategy. [slee-2019] In practice that means a first drug that does not suit you is not evidence that medication cannot help.
The same paper’s introduction points to two practical problems. It cites an analysis of 29,131 people with GAD in a UK clinical research database in which 46.0% stopped SSRIs, tricyclic antidepressants or related drugs after a mean of 3.7 months, and it lists sexual dysfunction, weight gain and sleep disturbance as common problems with SSRIs and SNRIs. [slee-2019] Those figures come from earlier work the paper cites, not from its own trials.
Do beta blockers like propranolol work for anxiety?
The evidence is too weak to say so. Propranolol is a beta blocker that slows the heart and is often prescribed for anxiety, but a systematic review of randomised trials found only 8 that met its criteria: panic disorder with or without agoraphobia (four studies, 130 people), specific phobia (two studies, 37), social phobia (one study, 16) and post-traumatic stress disorder (one study, 19). [steenen-2016]
Pooled analyses of three panic disorder trials found no statistically significant difference between propranolol and benzodiazepines for short-term treatment, and no evidence that propranolol reduced PTSD symptom severity. The authors concluded that the quality of evidence is insufficient to support the routine use of propranolol in any of the anxiety disorders. [steenen-2016] That is a statement about evidence, not a finding that it never helps: the trials were tiny, and the review covered anxiety disorders, not one-off situations such as a performance.
- Panic disorder (4 studies) 64% people
- Specific phobia (2 studies) 18% people
- PTSD (1 study) 9% people
- Social phobia (1 study) 8% people
Participants by condition in the 8 randomised trials of oral propranolol for anxiety disorders in the systematic review and meta-analysis by Steenen and colleagues (2016): panic disorder with or without agoraphobia 130 (four studies), specific phobia 37 (two studies), post-traumatic stress disorder 19 (one study), social phobia 16 (one study). The total of 202 is calculated here.
What happens when you stop anxiety medication?
Some people have withdrawal effects, and how common and how long they last is disputed. A systematic review of 24 studies of antidepressant withdrawal (not limited to anxiety) found incidence rates from 27% to 86% in 14 studies, with a weighted average of 56%. In four large studies of severity, a weighted average of 46% of those with withdrawal effects chose the most extreme severity rating on offer. [davies-2019]
On duration, seven of the ten studies that reported it contradicted the guidance that withdrawal typically clears in one to two weeks: a significant share of people were affected for more than two weeks, and effects lasting several months were not uncommon. The four studies that calculated a mean duration found 5 days, 10 days, 43 days and 79 weeks, in quite different populations. [davies-2019]
The authors called for guidelines to be updated and for prescribers to tell patients about withdrawal. Published comments, including a critique by other researchers, followed, so read the figures as a reason to plan a gradual stop with a prescriber and not as a settled estimate. [davies-2019]
Incidence of antidepressant withdrawal effects across 14 studies in a systematic review by Davies and Read (2019): the lowest and highest study rates and the weighted average. Antidepressants in general, not only those prescribed for anxiety.
| Medication family | What the reviews found | What is uncertain |
|---|---|---|
| Escitalopram, venlafaxine, duloxetine, pregabalin | More efficacious than placebo for GAD (89 trials) | Order of the other 18 drugs is not covered here |
| Propranolol (beta blocker) | No significant difference from benzodiazepines in short-term panic (3 trials) | Evidence insufficient for routine use |
| Stopping antidepressants | Withdrawal effects in 27% to 86% of people, 56% on average | Disputed, and not specific to anxiety |
What can this research not show?
It cannot show which medication suits you, how it compares with therapy or what happens over years. Five limits matter most:
- Mostly one condition. The largest review studied generalised anxiety disorder in adult outpatients, so it says little about panic disorder, social anxiety, phobias, OCD or PTSD, or about children. [slee-2019]
- The drug ranking is partial. Of the 22 drugs compared, we could verify results for four, and the paper has since had published corrections, so check the current text before relying on a figure. [slee-2019]
- No head-to-head with therapy. None of these reviews compared medication with psychological treatment, so they cannot say which to try first.
- The propranolol evidence is tiny. Eight small trials, with 16 people in the only social phobia study, support no firm conclusion either way. [steenen-2016]
- The withdrawal figures are contested. The review pooled very different studies and was criticised by other researchers, and it covers antidepressants in general. [davies-2019]
What should you ask a prescriber?
Ask which drug fits your diagnosis, what side-effects to expect, how long to give it before judging and what the plan is if it does not work. This section is our suggestion and is not something the reviews tested, although the first-drug point follows from the 2019 authors’ conclusion that an initial failure is not a reason to give up on medication. [slee-2019]
Useful questions are whether a gradual reduction plan can be agreed from the start, how alcohol and other medicines interact with it, and whether therapy alongside medication is an option. Our guide to therapy for anxiety covers the psychological side, and natural remedies for anxiety explains why supplements are a different question from prescribed drugs. Do not stop a prescribed medication abruptly without speaking to the prescriber.
If you want a rough picture of your symptoms before an appointment, our GAD-7 anxiety screener is a short self-check. It is not a diagnosis, and a high score is a reason to talk to a doctor, not a reason to start medication.
When to seek help
Speak to a doctor if anxiety is taking over your days, sleep or work, if you have panic attacks, or if you are already taking medication and are thinking of stopping it or are having troublesome side-effects. If you feel unable to stay safe, contact your local emergency services or a crisis helpline.
How MyFreud can help
A record makes a medication review easier. Daily mood tracking in MyFreud shows how your days run over weeks, and the notepad is somewhere to note sleep, side-effects and when anxiety is at its worst. It does not replace a doctor or a prescriber.
Download MyFreud and start today: App Store or Google Play.