Various drugs are effective and well-tolerated for generalised anxiety disorder

There are various drugs options for the treatment of generalised anxiety disorder, from different medication classes. This study found that most evidence was available for the serotonin and noradrenaline re-uptake inhibitors (SNRI) duloxetine and venlafaxine, the selective serotonin reuptake inhibitor (SSRI) escitalopram, and the anti-epileptic pregabalin. All appear well-tolerated and reduce symptoms compared with placebo. The SSRI sertraline, commonly used in practice, was equally effective and tolerated but it was evaluated in few studies.

This review is the largest to compare the drugs used for generalised anxiety disorder against each other. It included 89 trials in over 25,000 people randomised to 22 different drugs or placebo. Studies lasted two months on average, so only looked at the short-term effects of these drugs and did not compare them to psychological therapies.

Overall the review supports current guidance, that if a drug is to be used for anxiety, to consider an SSRI, followed by an SNRI or pregabalin if that is unsuitable.


Why was this study needed?

Feelings of anxiety or worry are normal when people face stressful events. Generalised anxiety disorder is persistent anxiety about many situations on most days, which can have considerable effect on quality of life and daily functioning. In 2014, generalised anxiety disorder was the most common mental health disorder in England, affecting 6% of people. It can be a difficult condition to diagnose and manage due to the range of co-existing physical and psychological symptoms, most often depression.

The recommended first-line treatment is psychological therapy, but not everyone will find it effective. People are less likely to respond if they are younger, live in deprived areas, or have disabilities. Therefore, drug treatment is often used alone or alongside. There are multiple drug options available. This review, therefore, aimed to compare the effectiveness and tolerability of the different drugs to inform practitioners in their choice.


What did this study do?

This systematic review included 89 randomised controlled trials assessing 22 drugs in 25,441 people with generalised anxiety disorder. Diagnosis was by recognised criteria in 84% of trials. Study duration was on average eight weeks (range 4 to 26 weeks).

Seventy per cent of trials (63) were placebo-controlled while half (45) included more than one active drug. Researchers performed network meta-analysis comparing treatment effects directly within trials, and indirectly across trials.

Trials were published from 1998 to 2016. Most were double-blind. The most likely risk of bias was through the possible drop-out of participants and incomplete or selective reporting of outcomes. Sixteen per cent of studies came from China. These tended to be lower quality, but their exclusion made little difference to results. Effects were measured using the Hamilton Anxiety Scale (HAM-A, score range 0-56, where a reduced score represents less anxiety).


What did it find?

  • Most evidence was available to support the effectiveness of the SNRIs duloxetine (HAM-A mean difference [MD] -3.13, 95% credible interval [CrI] -4.13 to -2.13) and venlafaxine (MD -2.69, 95% CrI -3.50 to -1.89), the SSRI escitalopram (MD -2.45, 95% CrI -3.27 to -1.63) and the anti-epileptic pregabalin (MD -2.79, 95% CrI -3.69 to -1.91). All reduced symptoms compared with placebo with good tolerability. People were no more likely to stop taking the drugs than placebo; in fact, people were slightly less likely to stop taking pregabalin.
  • The following drugs were also effective and tolerated but they had been assessed by fewer trials: SSRIs sertraline (MD -2.88, 95% CrI -4.17 to -1.59) and fluoxetine (MD -2.43, 95% CrI -3.74 to -1.16), tetracyclic antidepressant mirtazapine (MD -3.12, 95% CrI -4.43 to -1.80), anti-anxiety drug buspirone (MD -2.37, 95% CrI -3.83 to -0.91) and melatonin-receptor drug agomelatine (MD -3.55, 95% CrI -5.83 to -1.26).
  • The antipsychotic quetiapine improved anxiety to the greatest extent of all drugs (HAM-A MD -3.60, 95% CrI -4.83 to -2.39) but the drug was poorly tolerated, and people were more likely to drop out of the trials than people on placebo (odds ratio [OR] 1.44; 95% CrI 1.16 to 1.80).
  • The SSRI paroxetine (MD -2.29, 95% CrI -3.11 to -1.47) and anti-anxiety drugs benzodiazepines (MD -2.29, 95% CrI -3.19 to -1.39) were also effective but people were again more likely to drop-out (respective ORs 1.24, 95% CrI 1.03 to 1.50 and 1.43, 1.12 to 1.86).
  • The alternative drug vortioxetine had also been assessed by several studies but was ineffective (HAM-A MD -1.12, 95% CrI -2.47 to 0.24).


What does current guidance say on this issue?

The 2011 NICE guideline on generalised anxiety disorder recommends a stepped care approach, starting with education and monitoring, then low-intensity psychological interventions. Drug treatment is introduced in step 3 as an alternative to high-intensity psychological treatment.

Selective serotonin reuptake inhibitors are the recommended drugs of first-choice, starting with consideration of sertraline. If this is ineffective, an alternative SSRI or an SNRI is recommended. NICE highlight the risk of withdrawal syndrome with paroxetine and venlafaxine, and potential suicide and toxicity risks with the latter. If this is ineffective, pregabalin is considered third-line.

Antipsychotics, such as quetiapine, are not recommended. Benzodiazepines should only be used for short term crisis management.


What are the implications?

These findings are essentially in-line with current guidance around the management of generalised anxiety disorder. Selective serotonin reuptake inhibitors, SNRIs and pregabalin all seem to work. If one drug is ineffective or poorly tolerated, there are many other options from the same or a different class.

It’s worth noting that the evidence is based on small improvements of a few points on the anxiety scale, demonstrated in the short-term only. Psychological therapy also retains an essential place in the care of generalised anxiety disorder, and this was not evaluated by this review.  For people who have not responded to psychological therapies, drug treatment may help to reduce anxiety to a level at which they may respond to non-drug therapy.


Citation and Funding

Slee A, Nazareth I, Bondaronek P et al. Pharmacological treatments for generalised anxiety disorder: a systematic review and network meta-analysis. Lancet. 2019;393:768-77.

No funding was received by the authors for this work.



House of Commons Library. Mental health statistics for England: prevalence, services and funding. House of Commons Library Briefing Paper. London; 2018.

NHS website. Generalised anxiety disorder in adults. London: Department of Health and Social Care; updated 2018.

NICE. Generalised anxiety disorder and panic disorder in adults: management. CG113. London: National Institute for Health and Care Excellence; 2011.

Produced by the University of Southampton and Bazian on behalf of NIHR through the NIHR Dissemination Centre



Expert commentary

Diagnosis of generalised anxiety disorder is difficult because many patients present with mixed anxiety and depressive symptoms. Fortunately, antidepressants are often effective in generalised anxiety disorder, and the meta-analysis confirms the utility of selective serotonin reuptake inhibitors (SSRI) and serotonin and noradrenaline reuptake inhibitors (SNRI).

Of the SSRIs, escitalopram has the best combination of positive trial evidence and acceptability but is still black-listed by many primary care trusts, apparently unaware that it is now generic. The SNRI, venlafaxine, is also effective and relatively cheap but its unpleasant withdrawal syndrome is a drawback.

Pregabalin was the only compound whose overall acceptability was superior to placebo. Concerns about misuse, however, may limit its appeal, particularly now it has been classified as a Schedule 3 controlled drug. 

Philip Cowen, Professor of Psychopharmacology, University of Oxford

The commentator declares no conflicting interests