A Single-Blind Trial of Risperidone vs. Paroxetine for Treatment of Panic Attacks

Rationale

Risperidone is an atypical antipsychotic drug that acts as an antagonist at 5-HT2A/2C and D2 receptors (Chouignard et al, 1993; Marder \& Meibach, 1994). In its serotonergic activity, risperidone resembles both the atypical antidepressant and 5HT2-antagonist nefazodone (Bakish et al, 1993) and a 5HT2-antagonist ritanserine (Lima \& Moncrieff, 2000; Preskorn 1995). Nefazodone treats anxiety symptoms and induces less agitation than other reuptake blockers (Fawcett et al, 1995) while ritanserin shows effectiveness in treatment of dysthymia in clinical trials. Moreover, risperidone increases release of dopamine and norepinephrine in the prefrontal cortex (Zhang et al, 2000), an activity that in other pharmacological agents is associated with their antidepressant, anxiolytic, and antipanic properties (Thase, 2002).

Therefore, risperidone has a potential to be effective for treatment of either Generalized Anxiety Disorder, or Panic Disorder, or Major Depressive Disorder with Panic Attacks (DSM-IV, 1994) as a single agent. Moreover, due to its D2 and 5-HT2A/2C antagonism and proven antipsychotic efficacy, the drug could be uniquely effective for severe panic attacks with psychotic features or para-psychotic features (Galynker et al, 1996). Consistent with this supposition, Marder et al (1997) reported that in patients with schizophrenia, risperidone produced significantly greater improvements than haloperidol in PANSS scores measuring anxiety and depression.

Our clinical experience with over 50 patients with panic attacks showed that low dose risperidone, similar to lorazepam, reduced or eliminated anxiety and panic, often after the first dose. Risperidone was most effective for patients experiencing obsessive, intrusive, anxiety-inducing thoughts. In contrast to lorazepam, risperidone treatment did not result in building of tolerance to the drug or disinhibition. We did not observe extrapyramidal side effects in patients treated with risperidone for panic attacks. Thus, our preliminary clinical data indicate that due to its efficacy, quick onset of action, absence of addictive potential, and lack of side effects, risperidone could be superior to both benzodiazepines and SSRIs for treatment of panic attacks, particularly for severe panic attacks. Moreover, risperidone has a potential to be the first line drug for treatment of these symptoms.

Consequently, at present, a clinical study of risperidone for treatment of panic attacks is indicated and is of great potential value. We propose an exploratory, open-label, single-blind trial of risperidone vs. paroxetine for treatment of panic attacks.

Hypothesis/Objectives

Objective One: To compare the efficacy of risperidone to that of paroxetine in treatment of panic attacks in patients with Panic Disorder and with Major Depressive Disorder with Panic attacks.

• Hypothesis One: Risperidone and will be more efficacious in treatment of panic symptoms in both patient groups.

Objective Two: To compare the side effect profile of risperidone vs. paroxetine in treatment of panic attacks.

• Hypothesis Two: Risperidone, compared to paroxetine, will cause less sexual side effects, anxiety, and restlessness. There will be no difference in extrapyramidal side effects between the paroxetine and the risperidone groups.

Objective Three: To compare response rates of risperidone vs. paroxetine in treatment of panic attacks.

• Hypothesis Three: Patients with panic attacks will respond significantly faster to risperidone than to paroxetine.

Objective Four: To examine demographic and clinical predictors of robust response to risperidone treatment using appropriate statistical analyses of individual scales items and treatment response rates.

• Hypothesis Four: The subgroup of participants who have high scores on items reflecting somatic anxiety, and ruminative-intrusive thinking, and those with nocturnal panic attacks will show better response to risperidone therapy than patients with low scores on those items.

If study hypotheses are confirmed, future research could include a randomized, double-blind, fixed-dose study of risperidone as the sole agent for treatment of panic disorder and Phase IV trials aimed at determining the effectiveness of 0.125 mg and 0.0675 mg doses of risperidone for the treatment of anxiety and panic attacks.

Study Population

The patient population will consist of psychiatric outpatients who meet criteria for the DSM-IV diagnosis of Major Depressive Disorder with Panic Attacks or Panic Disorder. The subjects will be primarily recruited from the Psychiatric Outpatient Services for Adults clinic (POSA) located at Beth Israel Medical center that provides care to approximately 900 patients per year. At present, the patient ethnic and racial mix at POSA is 50% Hispanic, 15% African American, 32.5% Caucasian and 2.5% Asian. The gender distribution is approximately 33% male and 67 % female. The outpatient payer mix is 20 % Medicare, 74 % Medicaid and 6% combined. Of those patients, approximately 40% (360) manifest panic attacks. In addition, patients will be recruited from the Emergency Department and from inpatient psychiatric units at Beth Israel Medical Center and from newspaper advertisements. Based on our previous experience with recruitment for research studies, we expect that 10-12% of patients (36-42) who meet diagnostic criteria will also meet inclusion and exclusion criteria, will sign informed consent and will participate in the study. We expect to recruit 40 patients per year and to complete the study in 4 years.

Inclusion Criteria

1. Males and females, ages 18-55.
2. Ability to sign an informed consent
3. Diagnosis of Panic Disorder, or MDD with Panic attacks, single episode, recurrent, or chronic
4. HAM-A score >17

Exclusion Criteria

1. Alcohol or substance abuse within the last 6 months
2. Current diagnosis of Obsessive-Compulsive Disorder
3. Current diagnosis of Schizophrenia, Schizoaffective Disorder, or Bipolar Mood Disorder
4. Use of antipsychotic medications over the two months preceding enrollment in the study
5. Changes in antidepressant or mood stabilizer dosing over the two months preceding enrollment in the study
6. Previous adverse reaction to risperidone or paroxetine.

Study Design and Drug Regimens

The study will be an 8-week parallel comparison (with a blinded rater) of risperidone and paroxetine for treatment of three groups of patients with panic attacks. Each group will consist of 30 subjects.

After signing an informed consent, the subjects will be randomly assigned in an open manner to treatment with risperidone or paroxetine alone.

Arm 1: Treatment with risperidone will be flexible-dose and will be initiated at 0.25 mg po qhs. For the subjects who did not achieve a remission of panic symptoms, the dose will be increased to 0.5 mg po qhs on day 3. For subjects who experienced morning sedation, the dose will be decreased to 0.125-mg po qhs on day 3.

Arm 2: Treatment with paroxetine will be flexible-dose and will be initiated at 10-mg po qhs. For subjects who experienced morning sedation, the dose will be decreased to 5-mg po qhs on day 3. For the subjects who did not achieve a remission of panic symptoms, the dose will be increased to 20-mg po qhs. For the subjects who did not achieve a remission of panic symptoms, the dose will be increased to 40-mg po qhs as needed.

Arm 3: Treatment with paroxetine on a fixed dose schedule will be initiated at 10 mg po qhs on day 1. On day 3 it will be increased to 20 mg po qhs. On day 7 it will be increased to 30 mg po qhs. Most patients respond to an average therapeutic dose of 30 mg of paroxetine. Many patients in Arm 2, since it was flexible-dose, never achieved a therapeutic dosage, which could potentially create a bias in the comparison. The third arm was added to ensure that a subgroup of study participants would receive an optimum therapeutic dose of paroxetine, to allow for an unbiased comparison between the two study drugs.

Patients randomized on any of the above arms, and not responding or having side effects will be taken off the study and referred for regular psychiatric treatment.

Subjects receiving fixed dose antidepressants or mood stabilizers prior to their enrollment in the study will be continued on these medications throughout the study. No dose adjustment of antidepressants or mood stabilizers will be allowed. At the end of study, participants can choose to continue their study medications under the care of one of study physicians or one of the other psychiatrists at POSA. Those wishing to stop study medications at the end of the study will be able to do so under the supervision of one of the study physicians. Paroxetine will be tapered over a two-week period to avoid SSRI withdrawal. Risperidone will be stopped abruptly since no withdrawal symptoms have been reported in the literature.

Assessments to evaluate mood and severity of panic attacks will be performed by a rater blinded to medication status on a weekly basis.

All study participants' will receive psychiatric treatment, free of charge, for at least one month following the end of the study. At the end of that period, participants will be assisted in finding further psychiatric care. Subjects will receive $10 (paid in cash) as reimbursement for travel expenses for each assessment that they attend. Subjects will have to sign a receipt of reimbursement in order to receive the cash.

Data Analysis

Pearson Product Moment correlation coefficients (r) will be calculated for each of the treatment outcome variable groups. This will enable determination of the degree of association between treatment and outcome. More importantly, (r) can be interpreted with the Binomial Effect Size Display (BESD) (Rosenthal, 1991). The BESD will enable estimation of a clinically meaningful improvement rate and the number of people that would improve as a result of treatment.

Power Analysis and Precision Analysis

As stated above, the measure of effect size will be the Pearson Product Moment correlation coefficient (r). For this study, the effect size of r = .50 was chosen as the smallest effect size that would be clinically meaningful, and a sample size of 26 was calculated as necessary to obtain power of 81% and a 95% confidence interval (r = .13 to r = .75) that would not include zero (Rosenthal, 1991).