A multiple treatments meta-analysis has found that sertraline is the best tolerated therapy for generalised anxiety disorder (GAD).

There were 27 randomised controlled trials which were suitable for inclusion in the analysis. Response was defined as a reduction of 50% or more from the baseline Hamilton anxiety scale (HAM-A) score, remission as a final HAM-A score ≤7 and tolerability as the withdrawal rate due to adverse events. For each outcome measure, drugs were ranked according to the probability of their being the most effective/well tolerated treatment.

Nine drugs (duloxetine, escitalopram, fluoxetine, lorazepam, paroxetine, pregabalin, sertraline, tiagabine, and venlafaxine) were compared with one another. Fluoxetine was ranked first for response (probability 62.9%) and remission (60.6%) whereas sertraline was ranked first for tolerability (49.3%). Of those drugs which are licensed in the UK for the treatment of GAD, duloxetine was ranked first for response (probability 2.7%), escitalopram was ranked first for remission (26.7%) and pregabalin was ranked first for tolerability (7.7%).

All treatments were more effective than placebo and the difference was significant except in the case of tiagabine and escitalopram (response only). When compared with one another, however, there were no significant differences except for the comparisons between duloxetine and tiagabine (response) which favoured duloxetine, and venlafaxine and tiagabine (remission) favouring venlafaxine.

These findings need to be treated with caution. All studies were sponsored by drug companies, unpublished studies were not specifically looked for, and publication bias is therefore likely. The absence of tricyclics from the list of drugs included in the analysis does not mean that SSRIs are superior to tricyclics in the treatment of GAD. It simply reflects the fact that most clinical trials of tricylics were carried out before GAD became a recognised diagnosis.1

There was only one trial involving fluoxetine. This was a post hoc analysis from a double-blind placebo-controlled trial comparing venlafaxine and fluoxetine in patients with major depression. There were 92 participants, of whom only 33 received fluoxetine.2 The participants had comorbid GAD and depression, whereas most participants in other RCTs had pure GAD.3 The trial itself reported response rates of 66% for venlafaxine, 45% for fluoxetine and 24% for placebo.4 Clearly, it would be unwise to recommend fluoxetine based on this isolated study.

NICE did not include the fluoxetine study in its own multiple treatments meta-analysis. It ranked duloxetine first for response (probability 36.6%), followed by sertraline (27.7%) and venlafaxine (17.7%). Sertraline was ranked first for tolerability (60.6%), followed by pregabalin (22.3%) and escitalopram (14.5%).3

NICE recommends drug treatment with an SSRI as an alternative to high intensity treatment at step 3. I do not think there is any convincing evidence to favour one drug over another on the grounds of efficacy. Choice is therefore determined by tolerability. If patients are restless and not sleeping, a sedative tricyclic such as amitriptyline may be more appropriate as SSRIs may cause increased anxiety and restlessness at the onset of treatment.1 Otherwise, SSRIs are the treatment of choice, and both NICE and the current meta-analysis have concluded that sertraline is the best tolerated drug.

Dr Phillip Bland

• Baldwin D, Woods R, Lawson R et al. Efficacy of drug treatments for generalised anxiety disorder: systematic review and meta-analysis. BMJ 2011;342:d1199