Daniel L. Sontheimer, MD

BACKGROUND: Discontinuation of benzodiazepines in patients on long-term treatment may be associated with restlessness, agitation, increased anxiety, insomnia, irritability, palpitations, and many other troublesome symptoms. Thus, patients with anxiety disorders taking long-term benzodiazepine therapy have difficulty successfully discontinuing treatment. Approaches to discontinuation include a gradual taper, cognitive-behavioral therapy, and adjunctive pharmacologic treatment.

POPULATION STUDIED: A total of 107 adult patients with generalized anxiety disorder were recruited from physician offices and by notices in the media. The patients had taken benzodiazepines for an average of 8.5 years (range=1-31 years); 91% had previously attempted discontinuation (average=3.4 attempts). Interestingly, only 24% of the patients were satisfied with their benzodiazepine therapy. The mean age was 48 years (range=22-77 years); 45% were women. All were treated at a psychopharmacology research unit and returned to the care of their family physicians at the end of the study.

STUDY DESIGN AND VALIDITY: After 2 to 4 weeks taking a steady dose of their benzodiazepine, the patients were assigned to receive double-blind treatment with imipramine 25 mg, buspirone 5 mg, or placebo. Each study drug was titrated over 2 weeks to a goal of 6 capsules daily in divided doses. After 4 weeks of concomitant treatment with a benzodiazepine and assigned study drug, the benzodiazepine dose was tapered by approximately 25% each week for 4 to 6 weeks. This taper was followed by a 5-week benzodiazepine-free phase; treatment was continued for 3 weeks followed by 2 weeks of placebo treatment. Patients were monitored weekly for withdrawal symptoms, anxiety, and depression. There are several limitations in the validity of this study. There is no mention of the method of randomization or of concealed allocation of study patients. Thirty-two patients (30%) did not complete the taper phase, and these patients were not all accounted for, that is, there was no intention-to-treat analysis. Additionally, the total number of patients remaining in each treatment arm was small, and the study may not have been large enough to find differences if they actually exist. Also, there was no comparison of demographics or clinical characteristics of the 3 treatment groups.

OUTCOMES MEASURED: The main outcomes measured were: (1) whether the taper was successful as defined by a benzodiazepine-free state at 12 weeks post-taper and (2) severity of symptoms of benzodiazepine discontinuation as rated by both physicians and patients.

RESULTS: At 3 months after benzodiazepine discontinuation, 82.5% of the imipramine-treated patients were benzodiazepine-free compared with 37.5% of the placebo-treated patients (P >.01; number needed to treat=2). Buspirone was less effective, and the success rate was not statistically different from placebo. However, the severity of withdrawal symptoms was worse in the imipramine-treated patients than with the buspirone- and placebo-treated patients (16.6 vs 8.9 and 10.4, respectively) on the 38-item physician-rated checklist (P >.03). Similar results were obtained using the patient-rated checklist, although these results were not shown. Dry mouth occurred significantly more frequently in imipramine-treated patients than in buspirone- and placebo-treated patients (number needed to harm=2).

BACKGROUND: Discontinuation of benzodiazepines in patients on long-term treatment may be associated with restlessness, agitation, increased anxiety, insomnia, irritability, palpitations, and many other troublesome symptoms. Thus, patients with anxiety disorders taking long-term benzodiazepine therapy have difficulty successfully discontinuing treatment. Approaches to discontinuation include a gradual taper, cognitive-behavioral therapy, and adjunctive pharmacologic treatment.

POPULATION STUDIED: A total of 107 adult patients with generalized anxiety disorder were recruited from physician offices and by notices in the media. The patients had taken benzodiazepines for an average of 8.5 years (range=1-31 years); 91% had previously attempted discontinuation (average=3.4 attempts). Interestingly, only 24% of the patients were satisfied with their benzodiazepine therapy. The mean age was 48 years (range=22-77 years); 45% were women. All were treated at a psychopharmacology research unit and returned to the care of their family physicians at the end of the study.

STUDY DESIGN AND VALIDITY: After 2 to 4 weeks taking a steady dose of their benzodiazepine, the patients were assigned to receive double-blind treatment with imipramine 25 mg, buspirone 5 mg, or placebo. Each study drug was titrated over 2 weeks to a goal of 6 capsules daily in divided doses. After 4 weeks of concomitant treatment with a benzodiazepine and assigned study drug, the benzodiazepine dose was tapered by approximately 25% each week for 4 to 6 weeks. This taper was followed by a 5-week benzodiazepine-free phase; treatment was continued for 3 weeks followed by 2 weeks of placebo treatment. Patients were monitored weekly for withdrawal symptoms, anxiety, and depression. There are several limitations in the validity of this study. There is no mention of the method of randomization or of concealed allocation of study patients. Thirty-two patients (30%) did not complete the taper phase, and these patients were not all accounted for, that is, there was no intention-to-treat analysis. Additionally, the total number of patients remaining in each treatment arm was small, and the study may not have been large enough to find differences if they actually exist. Also, there was no comparison of demographics or clinical characteristics of the 3 treatment groups.

OUTCOMES MEASURED: The main outcomes measured were: (1) whether the taper was successful as defined by a benzodiazepine-free state at 12 weeks post-taper and (2) severity of symptoms of benzodiazepine discontinuation as rated by both physicians and patients.

RESULTS: At 3 months after benzodiazepine discontinuation, 82.5% of the imipramine-treated patients were benzodiazepine-free compared with 37.5% of the placebo-treated patients (P >.01; number needed to treat=2). Buspirone was less effective, and the success rate was not statistically different from placebo. However, the severity of withdrawal symptoms was worse in the imipramine-treated patients than with the buspirone- and placebo-treated patients (16.6 vs 8.9 and 10.4, respectively) on the 38-item physician-rated checklist (P >.03). Similar results were obtained using the patient-rated checklist, although these results were not shown. Dry mouth occurred significantly more frequently in imipramine-treated patients than in buspirone- and placebo-treated patients (number needed to harm=2).

BACKGROUND: Discontinuation of benzodiazepines in patients on long-term treatment may be associated with restlessness, agitation, increased anxiety, insomnia, irritability, palpitations, and many other troublesome symptoms. Thus, patients with anxiety disorders taking long-term benzodiazepine therapy have difficulty successfully discontinuing treatment. Approaches to discontinuation include a gradual taper, cognitive-behavioral therapy, and adjunctive pharmacologic treatment.

POPULATION STUDIED: A total of 107 adult patients with generalized anxiety disorder were recruited from physician offices and by notices in the media. The patients had taken benzodiazepines for an average of 8.5 years (range=1-31 years); 91% had previously attempted discontinuation (average=3.4 attempts). Interestingly, only 24% of the patients were satisfied with their benzodiazepine therapy. The mean age was 48 years (range=22-77 years); 45% were women. All were treated at a psychopharmacology research unit and returned to the care of their family physicians at the end of the study.

STUDY DESIGN AND VALIDITY: After 2 to 4 weeks taking a steady dose of their benzodiazepine, the patients were assigned to receive double-blind treatment with imipramine 25 mg, buspirone 5 mg, or placebo. Each study drug was titrated over 2 weeks to a goal of 6 capsules daily in divided doses. After 4 weeks of concomitant treatment with a benzodiazepine and assigned study drug, the benzodiazepine dose was tapered by approximately 25% each week for 4 to 6 weeks. This taper was followed by a 5-week benzodiazepine-free phase; treatment was continued for 3 weeks followed by 2 weeks of placebo treatment. Patients were monitored weekly for withdrawal symptoms, anxiety, and depression. There are several limitations in the validity of this study. There is no mention of the method of randomization or of concealed allocation of study patients. Thirty-two patients (30%) did not complete the taper phase, and these patients were not all accounted for, that is, there was no intention-to-treat analysis. Additionally, the total number of patients remaining in each treatment arm was small, and the study may not have been large enough to find differences if they actually exist. Also, there was no comparison of demographics or clinical characteristics of the 3 treatment groups.

OUTCOMES MEASURED: The main outcomes measured were: (1) whether the taper was successful as defined by a benzodiazepine-free state at 12 weeks post-taper and (2) severity of symptoms of benzodiazepine discontinuation as rated by both physicians and patients.

RESULTS: At 3 months after benzodiazepine discontinuation, 82.5% of the imipramine-treated patients were benzodiazepine-free compared with 37.5% of the placebo-treated patients (P >.01; number needed to treat=2). Buspirone was less effective, and the success rate was not statistically different from placebo. However, the severity of withdrawal symptoms was worse in the imipramine-treated patients than with the buspirone- and placebo-treated patients (16.6 vs 8.9 and 10.4, respectively) on the 38-item physician-rated checklist (P >.03). Similar results were obtained using the patient-rated checklist, although these results were not shown. Dry mouth occurred significantly more frequently in imipramine-treated patients than in buspirone- and placebo-treated patients (number needed to harm=2).