The Pharmacotherapy of Anxiety Disorders

Antidepressants

The selective reuptake inhibitors (SSRIs) and serotonin-norepinephrine serotonin reuptake inhibitors (SNRIs) have become first-line agents for the treatment of panic disorder as well as other anxiety disorders because of their broad spectrum of efficacy (including benefit for disorders commonly co-morbid with panic disorder, such as major depression), favorable side-effect profile, and lack of cardiotoxicity. Currently, paroxetine, both the immediate ([Paxil] and controlled-release formulations [Paxil-CR]), sertraline (Zoloft), fluoxetine (Prozac) and extended-release venlafaxine Effexor-XR) are Food and Drug Administration (FDA)-approved for the treatment of panic disorder, though other SSRIs including citalopram (Celexa), and escitalopram (Lexapro), and fluvoxamine (Luvox) have also demonstrated anti-panic efficacy in both open and double-blind trials. A recently introduced SNRI, duloxetine (Cymbalta), has also been reported effective for panic disorder in case reports, and in an open-label trial though no randomized controlled trials (RCTs) are currently reported.

A recent meta-analysis on 50 clinical trials (yielding over 5,000 participants) confirmed that citalopram, paroxetine, fluoxetine, and venlafaxine were superior to placebo in the treatment of panic disorder. Finally, while the majority of data supporting the efficacy of pharmacological agents for panic disorder derive from short-term trials, several long-term studies have also demonstrated sustained efficacy over time.

Because the SSRI/SNRIs have the potential to cause initial restlessness, insomnia, and increased anxiety, and because panic patients are commonly sensitive to somatic sensations, the starting doses should be low, typically half (or less) of the usual starting dose (e.g., fluoxetine 5 to 10 mg/d, sertraline 25 mg/d, paroxetine 10 mg/d [or 12.5 mg/d of the controlled-release formulation], controlled-release venlafaxine 37.5 mg/d), to minimize the early anxiogenic effect. Doses can usually begin to be raised, after about a week of acclimation, to achieve typical therapeutic levels, with further gradual titration based on clinical response and side effects, although even more gradual upward titration is sometimes necessary in particularly sensitive or somatically-focused individuals. Although the nature of the dose–response relationship for the SSRIs in panic is still being assessed, available data support doses for this indication in the typical antidepressant range, and sometimes higher, i.e., fluoxetine 20 to 40 mg/d, paroxetine 20 to 60 mg/d (25 to 72.5 mg/d of the controlled-release formulation), sertraline 100 to 200 mg/d, citalopram 20 to 60 mg/d, escitalopram 10 to 20 mg/d, fluvoxamine 150 to 250 mg/d, and controlled-release venlafaxine 75 to 225 mg/d (although some patients may respond at lower doses). In some cases of refractory panic, even higher doses may be clinically useful, although additional data examining such dosing is needed.

SSRI and SNRI administration may be associated with adverse effects that include sexual dysfunction, sleep disturbance, weight gain, headache, dose-dependent increases in blood pressure (with venlafaxine), gastrointestinal disturbance, potential risk of bleeding (with anticoagulants, aspirin or NSAIDs), and provocation of increased anxiety (particularly at initiation of therapy) that may make their administration problematic for some individuals. The SSRIs/SNRIs are usually administered in the morning (though for some individuals, agents such as paroxetine and others may be sedating and better tolerated with bedtime dosing); emergent sleep disruption can usually be managed by the addition of hypnotic agents. The typical 2–3 week lag in onset of therapeutic efficacy for the SSRI/SNRIs can be problematic for acutely distressed individuals. There is also an FDA class warning for risk of emergent suicidal thoughts and behaviors based on short-term studies that suggests close monitoring is advised for individuals age 24 or younger, with use balancing risk and clinical need. In addition, some data suggesting possibly dose-dependent QTc prolongation with citalopram in those aged 60 and up has led to recommendations to limit dose to 20 mg/day and EKG monitoring in some populations.

Although results from a recent meta-analysis of 50 trials suggest the following increasing order of effectiveness among SSRIs/SNRIs, citalopram, sertraline, paroxetine, fluoxetine, and venlafaxine, for panic symptoms, they found a different order for associated overall anxiety symptoms, and there is no clear evidence of a differential efficacy between agents in the SSRI or SNRI classes to guide selection. On the other hand, potentially relevant differences in their side-effect profiles (e.g., potential for weight gain and discontinuation-related symptomatology), differences in their potential for drug interactions, and the availability of generic formulations may be clinically relevant.

Benzodiazepines

Despite guidelines for the use of antidepressants as first-line anti-panic agents, benzodiazepines are still commonly prescribed for the treatment of panic disorder. Two high-potency benzodiazepines, alprazolam (immediate and extended-release forms) and clonazepam, are FDA-approved for panic disorder; however, other benzodiazepines of varying potency, such as diazepam, adinazolam, and lorazepam, at roughly equipotent doses have also demonstrated anti-panic efficacy in RCTs. Benzodiazepines remain widely used for panic and other anxiety disorders, likely due to their effectiveness, tolerability, rapid onset-of-action, and ability to be used on an “as needed” basis for situational anxiety. It should be noted; however, that “as needed” dosing for monotherapy of panic disorder is rarely appropriate, as this strategy generally exposes the patient to the risks associated with benzodiazepine use without the benefit of adequate and sustained dosing to achieve and maintain comprehensive efficacy. Further, from a cognitive-behavioral perspective, “as needed” dosing engenders dependency on the medication as a safety cue and interferes with exposure to and mastery of avoided situations.

Despite their generally favorable tolerability, benzodiazepines may be associated with side effects that include sedation, ataxia, and memory impairment (particularly problematic in the elderly and those with prior cognitive impairment). Despite concerns that ongoing benzodiazepine administration will result in the development of therapeutic tolerance (i.e., loss or therapeutic efficacy or dose escalation), available studies of their long-term use suggest that benzodiazepines remain generally effective for panic disorder over time, and do not lead to reports of significant dose escalation. Of interest, a recent randomized, naturalistic, parallel-group study, found that after 3 years of treatment, individuals receiving clonazepam were slightly better, and reported fewer side effects, than those receiving paroxetine. However, even after a relatively brief period of regular dosing, rapid discontinuation of benzodiazepines may result in significant withdrawal symptoms (including increased anxiety and agitation) ; for instance, in one study, over two-thirds of patients with panic disorder, discontinuing alprazolam, experienced a discontinuation syndrome. Discontinuation of longer-acting agents (such as clonazepam) may result in fewer and less intense withdrawal symptoms with an abrupt taper. Patients with a high level of sensitivity to somatic sensations may find withdrawal-related symptoms particularly distressing, and a slow taper as well as the addition of CBT during discontinuation may be helpful to reduce distress associated with benzodiazepine discontinuation. A gradual taper is recommended for all patients treated with daily benzodiazepines for more than a few weeks, to reduce the likelihood of withdrawal symptoms (including in rare cases, seizures). Though individuals with a predilection for substance abuse are at risk for abuse of benzodiazepines, those without this diathesis do not appear to share this risk. However, benzodiazepines and alcohol may negatively interact in combination, and the concomitant use of benzodiazepines in patients with current co-morbid alcohol abuse or dependence can be problematic (thus further supporting the use of antidepressants as first-line anti-panic agents in this population with co-morbid illness). In addition, given the high rates of co-morbid depression associated with panic disorder, it is worth noting that benzodiazepines are not in general effective for treatment of depression and may in fact induce or intensify depressive symptoms in those with co-morbid depression. A meta-analysis reported that, although benzodiazepines may be as effective as antidepressants on PD symptoms, they might be less so on depressive symptoms.

Although benzodiazepines are commonly prescribed for the treatment of panic disorder, benzodiazepine monotherapy has decreased somewhat. Treatment with a combination of an antidepressant and a benzodiazepine compared to an antidepressant alone results in acceleration of therapeutic effects as early as the first week, although by weeks 4 or 5 of treatment, combined treatment (whether maintained or tapered and discontinued) shows no advantage over monotherapy. Thus, the data suggest that co-administration improves the rapidity of response when co-initiated with antidepressants, but that ongoing use may not be necessary after the initial weeks of antidepressant pharmacotherapy. Recent data suggest some benefit of the augmentation with a benzodiazepine for individuals who remain symptomatic on antidepressant monotherapy.

Other Agents

The data addressing the potential efficacy of bupropion (a relatively weak reuptake inhibitor of norepinephrine [noradrenaline] and dopamine) for the treatment of panic disorder is mixed, with a small study of the immediate-release formulation administered at high doses demonstrating no benefit, but a more recent open-label study employing standard doses of the extended-release formulation suggesting potential benefit. Similarly, there is mixed support for the potential efficacy for panic disorder of another noradrenergic agent, reboxetine, with a meta-analysis reporting that this agent may be ineffective in treating both panic and anxiety symptoms in panic-disordered patients.

There is suggestive evidence from case reports that buspirone (an azapirone 5-HT _1A partial agonist) may be useful as an adjunct to antidepressants and benzodiazepines and acutely, although not over the long term, to CBT for panic disorder, but appears ineffective as monotherapy.

Beta-blockers reduce the somatic symptoms of arousal associated with panic and anxiety, but may be more useful as augmentation for incomplete response rather than as initial monotherapy. Pindolol, a beta-blocker with partial antagonist effects at the 5-HT _1A receptor, was effective in a small double-blind RCT of patients with panic disorder remaining symptomatic despite initial treatment.

Atypical antipsychotics, including olanzapine, risperidone, and aripiprazole, have demonstrated potential efficacy as monotherapy or as augmentation for the treatment of patients with panic disorder refractory to standard interventions in a number of small, open-label trials or case series. More recently, a randomized, single-blind comparison of low-dose risperidone to paroxetine in the treatment of panic attacks failed to show any significant difference. However, evidence for treatment-emergent weight gain, hyperlipidemia, and diabetes with some of the atypical agents, as well as the lack of large RCTs examining their efficacy and safety in panic disorder to date, do not support the routine first-line use of these agents for panic disorder, but rather consideration for patients whose panic disorder has not sufficiently responded to standard interventions.

On the basis of limited data, some anticonvulsant s appear to have a potential role in the treatment of panic disorder, in individuals with co-morbid disorders (such as bipolar disorder and substance abuse), for which the use of antidepressants and benzodiazepines, respectively, are associated with additional risk. Small studies support the potential efficacy of valproic acid, but not carbamazepine for the treatment of panic disorder. Gabapentin did not demonstrate significant benefit compared to placebo for the overall sample of patients with panic disorder in a large RCT, but a post-hoc analysis found efficacy for those with at least moderate panic severity. Another related compound, the alpha ₂ delta calcium channel antagonist pregabalin, has demonstrated utility for GAD, but there are no published reports to date in panic disorder.

Antidepressants