Intended for healthcare professionals

Clinical Review Extracts from “Best Treatments”

Treating generalised anxiety disorderCommentary: Putting research into practiceCommentary: A patient's story of living with anxietyCommentary: Information for patients receiving cognitive therapy

BMJ 2003; 326 doi: (Published 29 March 2003) Cite this as: BMJ 2003;326:700

Treating generalised anxiety disorder

  1. Alison Tonks (atonks{at}, freelance medical editor
  1. BMJ Unified, London WC1H 9JR
  2. Faculty of Medicine and Health Sciences, University of Auckland, Auckland, New Zealand
  3. BMJ Unified, London WC1H 9JR

    Description Generalised anxiety disorder is a chronic, disabling mental illness affecting 1-2% of adults.1 It is characterised by worry and anxiety that are hard to control and that interfere with daily functioning. Although common and treatable, it remains underdiagnosed and undertreated.2 Only half of people with the disorder consult a doctor for it. Those who do are rarely offered the most effective treatment—cognitive therapy. Most patients are seen and treated by primary care doctors, who may be unfamiliar with the condition. About 8% of patients consulting primary care professionals have generalised anxiety disorder.3 Treatment involves cognitive therapy or medication. There are large gaps in the evidence on treatments, such as on their long term effects or on the value of combining cognitive therapy with medication.

    Research Several factors complicate research into treatment of generalised anxiety disorder:

    • Drug trials have high dropout rates, partly because of side effects4

    • The disorder has a relapsing and remitting course; improvements are not always due to treatment

    • The placebo effect is powerful—trials have to be large to show a treatment effect5

    • Most treatments take several weeks or even months to work2

    • The definition of generalised anxiety disorder has changed four times since 1980.6

    Prognosis Complete recovery is rare. Explain to your patients that you are aiming for remission rather than cure. A reasonable goal is to reduce symptoms by about 50% and improve quality of life.7 Patients shouldn't expect any benefit until eight to 12 weeks after starting treatment.2


    These articles, and the patients' version, will help you and your patients discuss treatment options together and choose treatments that are supported by evidence from randomised controlled trials. Clear, concise information is important for people with generalised anxiety disorder, who may have difficulty concentrating or making decisions.

    Treatments that work

    Cognitive therapy

    Cognitive therapy is a brief, practical form of psychotherapy focused on patients' current problems, not their past. Treatment typically consists of 16 weekly sessions. It is the best psychological treatment for generalised anxiety disorder, at least in the short term. It works for over half of patients, reducing symptoms and improving daily functioning. There are no trials reporting outcomes more than one year after the start of treatment. It is uncertain how it compares with drug treatments.

    Treatments that are likely to work


    Buspirone is an anxiolytic drug in a class of its own that can improve symptoms in the short term. It is unclear how it compares with cognitive therapy, or with other medications. The starting dose is 7.5 mg twice daily or 5 mg three times a day, increasing by 5 mg per day at two to three day intervals as needed. In trials, patients were treated with 15–45 mg daily in divided doses.


    Imipramine, paroxetine, venlafaxine, opipromol, and trazodone are the five antidepressants that have been studied in randomised controlled trials. They can reduce symptoms in the short term. There is no significant difference between them. Unlike cognitive therapy, they have side effects, including sedation, confusion, and an increased risk of falls. In trials, antidepressants were effective in about half to two thirds of people who completed the course. Compliance was generally poor—almost a third of participants dropped out in some trials, often because of side effects. See commentary below for dosing.

    Treatments that may do more harm than good


    Benzodiazepines can reduce symptoms in the short term. There is little difference in effectiveness among different agents. About 70% of patients experience drowsiness and up to a third get rebound anxiety when they stop taking them. Benzodiazepines have been implicated in about 5-10% of road traffic crashes. Longer term use can cause dependency. A dosing schedule in one trial was alprazolam 0.5 mg three times daily for one week, followed by reducing and then tapering off the dose over three weeks.


    Kava is a herbal treatment derived from a pepper plant, Piper methysticum. There is good evidence that it can reduce anxiety symptoms in people with generalised anxiety disorder, but its use has been associated with liver damage, and in Britain it has been voluntarily withdrawn from the market. The dose used in the trials was 60–240 mg daily of standardised extract.

    Treatments that need further study

    Applied relaxation

    In applied relaxation, patients learn to relax different muscle groups systematically. The only research on this treatment for generalised anxiety disorder compared it with cognitive therapy, but the results were inconclusive.


    Abecarnil is a new anxiolytic, developed as an alternative to benzodiazepines. There have been only two trials of this drug in generalised anxiety disorder, which had conflicting results.

    Antipsychotic drugs

    There has only been one trial of an antipsychotic drug, trifluoperazine, for generalised anxiety disorder. The drug improved symptoms, but 84 out of 244 (43%) patients taking it noticed drowsiness and about 1 in 6 developed extrapyramidal reactions and other movement disorders.


    Even though they are sometimes given to people with anxiety, there have been no trials of β blockers for generalised anxiety disorder.


    Hydroxyzine is a sedating antihistamine. An overview of two placebo controlled trials found an effect in one but not the other. Of 81 people taking it, 32 (40%) reported side effects, including drowsiness, headache, and gastrointestinal upset.


    Based on the Clinical Evidence chapter on generalised anxiety disorder written by Christopher Gale and Mark Oakley-Browne from the Department of Psychiatry, University of Auckland, New Zealand.


    • Editorial by Nash et al


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.

    Commentary: Putting research into practice

    1. Christopher Gale (kiwidoc{at}, consultant psychiatrist
    1. BMJ Unified, London WC1H 9JR
    2. Faculty of Medicine and Health Sciences, University of Auckland, Auckland, New Zealand
    3. BMJ Unified, London WC1H 9JR

      Making the diagnosis

      Generalised anxiety disorder is common and chronic, and it causes as much disability as depression.1 It is a serious mental disorder.

      The key symptom is pervasive, severe worry or anxiety about several different things. However, many patients will present with comorbid conditions such as depression, other anxiety disorders, and substance abuse. Careful, repeated review of the patient's symptoms should lead to the diagnosis.

      The disorder can and should be diagnosed in addition to other psychiatric conditions; it is an important independent cause of disability. You should consider the diagnosis particularly in anxious patients who remain handicapped despite a good response to the presenting symptom you are treating.

      Selecting treatments

      Cognitive therapy is the treatment of choice.24 It is effective but is expensive, time consuming, and not always accessible. There is no evidence that drugs are more effective than cognitive therapy.

      If cognitive therapy is not available, drugs may be indicated. All medications for the disorder, except benzodiazepines, take six to eight weeks to work, and they do not relieve all symptoms. The choice of drug depends on availability, cost, and side effects.

      A specific serotonin reuptake inhibitor is a reasonable choice. Because patients are acutely aware of bodily symptoms, and therefore sensitive to side effects, a useful strategy is to start with a low dose (the equivalent of 5 mg of paroxetine). The dose is then increased by 5–10 mg every week until it is within the effective dose range (equivalent of 40–60 mg of paroxetine).

      Many patients with generalised anxiety disorder are treated with benzodiazepines. These reduce disability, but the benefits have to be weighed against their long term side effects.

      There is no evidence that combination drug therapy is better than monotherapy.

      Pitfalls in using the evidence

      Providing evidence based mental health care can be difficult. In addition to the problem of publication bias, there is considerable debate in psychiatry as to how to define an acceptable outcome. Other pitfalls are:

      Not hearing the evidence—some services have traditionally not seen anxiety disorders as “serious mental disorders” despite the evidence about the disability they cause. Similarly, in many places effective psychotherapies are not available.

      Not considering the patient—for example by using “one size fits all” group therapy, or rigid protocols for treatment. Patients with generalised anxiety disorder are complex; comorbid conditions may also require treatment from the start or may recur during treatment. Patients may be slow in making changes to their life, and you may need to work at their pace. It is probably better to make small, slow changes than sudden, abrupt ones, particularly when prescribing medication.

      Using unproved treatments—all anxiety medications have a considerable placebo effect, and this can mislead you into considering that a treatment is effective.


      1. 1.
      2. 2.
      3. 3.
      4. 4.

      Commentary: A patient's story of living with anxiety

      By the time I was given the diagnosis, I was not surprised or upset because it had become quite clear to me that I had a problem. I was coping moderately well with life when a death in the family created extra stress. While I didn't feel any more depressed than seemed appropriate, I developed new symptoms, like chest pains. I became concerned that I may have developed a heart condition.

      We asked a 28 year old engineer about her experiences living with an anxiety disorder. Generalised anxiety disorder was diagnosed when she was 24 years old.

      What is it like to live with this condition?

      It is a big pain to live with anxiety disorder. I recently stopped taking medication, hoping that my calm feelings would continue, but that hasn't been the case. No matter how much I try to avoid it I find that my life is inhibited by my desire to avoid things that scare me. I don't like trains, airplanes, subways, tunnels, high places, and low places. When I go to the movies I worry about being trapped in a fire so I try to sit near the aisle. I live in England and I'm concerned about earthquakes!

      I've always been a nervous person—even as a child—so I've learned to make light of my fears, but there are times when I can't control it. It can be very embarrassing.

      Commentary: Information for patients receiving cognitive therapy

      1. Alison Tonks (atonks{at}, freelance medical editor
      1. BMJ Unified, London WC1H 9JR
      2. Faculty of Medicine and Health Sciences, University of Auckland, Auckland, New Zealand
      3. BMJ Unified, London WC1H 9JR

        Does it work?

        Yes. Studies show that cognitive therapy is the best form of psychotherapy (“talking treatment”) for generalised anxiety disorder.1

        What is it?

        Cognitive therapy is based on the idea that your anxiety happens because you have unreasonably negative thoughts about yourself and the world. During a course of treatment, you discuss these thoughts with a specially trained therapist and learn how to replace them with positive ones.2

        Cognitive therapy works faster than other kinds of psychotherapy, such as non-directive counselling (where a counsellor listens to your problems and reflects them back to you, without actively making suggestions).1 You typically see a therapist every week for several months. Most people have about 16 sessions, but treatment can go on for longer if you need it. Though we don't know how long the benefits last, there is some evidence that they last longer if you have one-on-one sessions rather than group therapy.1

        How can it help?

        Cognitive therapy works in different ways for different people. For you, getting better could mean feeling calmer, worrying about fewer things, or simply being able to answer the telephone again. It could also mean sleeping better, getting rid of headaches, or having more energy.

        When researchers asked people how they felt about cognitive therapy, here's what they said:3

        “I learned better ways of tackling difficult situations.”

        “It was reassuring to talk about my problems.”

        “I learned that I was able to cope.”

        Why should it work?

        Researchers think that anxiety disorder is linked to the way we think about ourselves and the world. If we can change the way we think, then we can control our anxiety. Doing this can also help us change the way we behave, so that we can do the things we need and want to do.2 Cognitive therapy is supposed to be a short, practical treatment. One goal is to teach you methods for handling anxiety. Then, if your problems return in the future, you can use those techniques to treat yourself and keep your symptoms under control.

        Can it be harmful?

        Cognitive therapy doesn't have any known side effects.


        For more information about generalised anxiety disorder, and for a list of other studies on this condition, see



        1. 1.
        2. 2.
        3. 3.
        View Abstract