Generalised anxiety disorder and panic disorder in adults: management

Reasons for the decision

For further details and a summary of all evidence identified in surveillance, see appendix A.

New evidence was assessed as either consistent with existing recommendations, from small trials or was inconclusive.

Among the new evidence identified some studies suggested that internet- delivered cognitive behavioural therapy (CBT) was effective compared to waiting list controls for generalised anxiety disorder (GAD) and panic disorder. Evidence also suggested that digital therapies incorporating some type of therapist contact may be more effective than those that do not. The guideline recommendations 1.2.11 and 1.4.10 support both unguided and guided approaches to low intensity therapies and overall we did not find enough consistent evidence to recommend one approach over the other.

Topic experts informed us that there was new evidence for the effectiveness of transdiagnostic therapies and we identified evidence that suggested some advantage for transdiagnostic therapy in the short term but that outcomes were the same as those for disorder specific therapies after 6 months. The evidence was judged as insufficient to make specific recommendations for one approach over the other. We also found a small amount of evidence that appeared to suggest that CBT was less effective in older people than people of working age. It is possible that this conclusion was as a result of different CBT techniques being compared and we concluded further research is required before an assessment of impact can be made. During consultation a stakeholder agreed that further research is needed in this area.

A stakeholder commented that they thought recommendation 1.2.16 which advises that there is no evidence for superiority of either pharmacological or psychological therapies was misleading. They highlighted evidence that suggests larger effects sizes for pharmacological therapies compared with psychological therapies. They also highlighted evidence that suggests comparing the effects of these therapies is difficult because the effects size for each is frequently measured relative to different comparators. Because of the difficulties highlighted, as well as issues of side effects associated with drug therapies, it is difficult to assess head-to-head superiority. Overall the evidence reports effects for both pharmacological and psychological therapies, and concludes treatment choice should consider patient preference. Recommendation 1.2.16 aims to communicate this and is judged to still be valid.

Topic experts highlighted that the cost-effectiveness of sertraline, which is recommended as a first choice SSRI for GAD, may have been impacted by the availability of escitalopram as a generic. A stakeholder disagreed with the decision to not update saying that the original economic study including sertraline and escitalopram was dated and that there are no placebo-controlled trials investigating relapse prevention in sertraline. A large network meta-analysis highlighted by topic experts in addition to other new evidence for pharmacological interventions indicates that escitalopram and other SSRIs and SNRIs are reasonably well-tolerated and effective. Although escitalopram is available as a generic sertraline remains the cheaper drug by unit cost and evidence suggests it is as effective as escitalopram. For these reasons, recommendation 1.2.22 is judged to still be valid.

A stakeholder highlighted that there was new evidence for the effectiveness of benzodiazepines for panic disorder and that recommendation 1.4.21 needed revising. The evidence was seen during surveillance and it was judged to support the recommendation. Whilst it is acknowledged that benzodiazepines can be effective for panic disorder in the short term, the long-term use of these medicines is associated with dependence. The recommendation was assessed as still being valid.

For further details and a summary of all evidence identified in surveillance, see appendix A.