Clinical application of paroxetine for tapering benzodiazepine use in non-major-depressive outpatients visiting an internal medicine clinic
Mutsuhiro Nakao, MD, MPH, PhD, Department of Hygiene and Public Health, Teikyo University School of Medicine, 2-11-1 Kaga, Itabashi-ku, Tokyo, Japan. Email: firstname.lastname@example.org
Abstract Chronic benzodiazepine (BDZ) users often have difficulty with BDZ withdrawal. To examine clinical effects of selective serotonin reuptake inhibitor (SSRI) on tapering BDZ use in non-depressive patients, 97 outpatients with a history of BDZ use for at least 3 months were recruited at an internal medicine clinic of a university hospital. After the 4th edition of the Diagnostic and Statistical Manual (DSM-IV) clinical interviews for screening major depression, 66 outpatients (68%) without the DSM-IV major depression were randomly assigned to one of three groups: SSRI-assisted BDZ-reduction group (10–20 mg of paroxetine, n = 22), simple BDZ-reduction group (no paroxetine, n = 23), and reference group (no BDZ-reduction, n = 21). A standardized 8-week program involving gradual BDZ discontinuation was performed in the two BDZ-reduction groups. The Hamilton Rating Scales for Depression (HAM-D) and Anxiety (HAM-A) and the BDZ Withdrawal Symptom Questionnaire were assessed during the intervention period. Those with major depression were excluded from the BDZ-reduction intervention and treated with a different protocol of medication. In total, 10 (45.5%) in the SSRI-assisted BDZ-reduction group (n = 22) succeeded in becoming BDZ-free after completing the program, whereas only four (17.4%) in the simple BDZ-reduction group (n = 23) succeeded. The assistance of the SSRI significantly predicted the success of becoming BDZ-free (P = 0.023), controlling for the effects of age, gender, period of BDZ use, and baseline HAM-D and HAM-A scores. The score changes on the three questionnaires were comparable (all P > 0.05) among the three groups during the intervention period. The use of SSRI may have beneficial effects on BDZ withdrawal without the worsening of mood states in cases without major depression.
A substantial proportion of chronic benzodiazepine (BDZ) users suffer from depression not only at psychiatric departments but also at internal medicine and other departments in hospitals.1,2 Although single BDZ prescriptions may be unsuitable or inadequate for depressive patients, BDZ prescriptions could be substituted or reduced in the absence of severe depression, since the long-term use of BDZ has risks of dependency, rebound anxiety, memory impairment, and withdrawal.3,4 Many clinicians, especially primary care physicians, have experienced that reducing the number of BDZ prescriptions for chronic users is difficult and, therefore, a standardized BDZ discontinuation program with established guidelines is required for daily clinical practice.5
Although several studies6–8 have shown some effectiveness of tricyclic antidepressants and other psychotropic agents for the treatment of BDZ withdrawal, the use of selective serotonin reuptake inhibitors (SSRI) for BDZ withdrawal remains ambiguous.1,6,9 The Dutch Chronic BDZ Working Group’s study10 investigated chronic BDZ users with major depression attending a BDZ discontinuation program in the primary care setting, and reported that the use of paroxetine, an SSRI, was of limited value in discontinuing chronic BDZ use.10 The lack of SSRI effect in tapering the BDZ intake could result from the fact that the target subjects had suffered major depression and had been previously treated with SSRI. If non-depressive patients had received a BDZ discontinuation program with an SSRI from the start of the study intervention, the results might have been different.
Therefore, the authors hypothesized that a minimum dosage of paroxetine would increase the success rate of the BDZ discontinuation program in chronic BDZ users without major depression. To test the hypothesis, non-major-depressive subjects were randomly allocated to a group receiving a standardized BDZ discontinuation program with paroxetine, a comparable group receiving the program without paroxetine, or a reference group not receiving the program.
The study population included 97 outpatients (63 women and 34 men) who consecutively attended the Internal Medicine Clinic at the Teikyo University Hospital in Tokyo, Japan. None of the patients had an experience taking SSRI and were under the treatment of the Department of Psychiatry of the hospital or other psychiatric clinics. The subject selection criteria were as follows: (i) those aged 20–70 years; (ii) those whose medical condition was stable and drug regimens unchanged for longer than 3 months; (iii) those who had been prescribed either alprazolam, bromazepam, etizolam, or lorazepam for at least 3 months prior to visiting the clinic; and (iv) those who were able to visit the clinic for an 8-week intervention (or control) period. In regard to criterion (iii), the four BDZ drugs were specifically selected because they are commonly prescribed in the hospital and because high-potency BDZ with a short duration of action are generally more likely to cause BDZ abuse and dependency.11,12 Concerning criterion (iv), the whole content of the study was explained to all of the candidate participants with written information on the study protocol, and only those who provided written consent were entered into the study. The study protocol was approved by the university ethics committee. A total of 101 patients met the criteria but four (4.0%) did not visit the clinic on their appointed day, therefore, 97 patients were assessed according to the study protocol.
Then, structured clinical interviews for the 4th edition of the Diagnostic and Statistical Manual (DSM-IV) axis I disorders (SCID-I) were conducted to detect major depression.13,14 Because it is important to obtain the information on psychiatric comorbidity of anxiety disorders,15 the prevalence of anxiety disorders was also assessed through the SCID-I interviews.13,14 To confirm the severities of depression and anxiety, the Hamilton Rating Scale for Depression (HAM-D) as well as the Hamilton Rating Scale for Anxiety (HAM-A) was given to all patients after the interview.16,17
As a result, 31 (32%) outpatients diagnosed with DSM-IV major depression were excluded and the remaining 66 (68%) without such disorder were selected as the study subjects. Mean age (standard deviation [SD]) was 58.5 (8.4) years old, and 44 (67%) of the subjects were women. Concerning the comorbidity of anxiety disorders, there were two subjects with generalized anxiety disorder (3.0%) and one with agoraphobia without history of panic disorder (1.5%) in those without major depression. Mean period (SD) of visit to the clinic and BDZ use were 132.0 (79.9) months; the ICD-10 diagnoses at the internal medicine clinic were hypertension (n = 15), diabetes mellitus (n = 10), premature ventricular contraction (n = 5), angina pectoris (n = 5), hyperlipidemia (n = 4), and other (n = 27). Mean period (SD) of BDZ use was 72.4 (49.4) months; they were prescribed BDZ including etizolam (n = 40), alprazolam (n = 11), lorazepam (n = 8), and bromazepam (n = 7).
Benzodiazepine discontinuation program
Those without major depression were randomly allocated to one of three groups: the SSRI-assisted BDZ-reduction group (additional 10–20 mg of paroxetine, SSRI (+) group), simple BDZ-reduction group (no paroxetine, SSRI (–) group), or no BDZ-reduction group (reference group). No placebo was administered for those in the SSRI (–) group. The allocation of reference group was to show quantitatively that those in the two BDZ-reduction groups stepped up the BDZ discontinuation program without the worsening of depression and anxiety, compared with those who did not participate in the program. Both the HAM-D and HAM-A, as well as the BDZ Withdrawal Symptom Questionnaire18 consisting of 20 items with a 5-point scale, were assessed at weeks 0 (baseline), 1, 2, 4, and 8 in the two BDZ-reduction groups.
An 8-week program involving gradual BDZ discontinuation was performed in the two BDZ-reduction groups. The program was a modified version of the previous Dutch study10 (Table 1). The period of BDZ tapering was prolonged to 8 weeks to minimize the number of dropouts owing to rapid withdrawal during the program in the present study. When any of the scores on the BDZ Withdrawal Scale, HAM-D, and HAM-A increased by more than 10%, the research physician inquired about the patient’s symptoms in detail and decided not to step up the stage. A fulfilment of BDZ reduction with a frequency of more than 5 days/week was regarded as successful tapering-off. The patients defined as being successful were those who satisfactorily stepped up their stages (i.e. those classified as stage 1 at 1-week assessment, stepped up to stages 2 [2 week], 3 [4 week], and 4 [8 week]), allowing to back their former dose of BDZ within 2 days a week.
Table 1. Benzodiazepine discontinuation program
|Stage 0:||No BDZ tapering|
|Stage 1:||BDZ tapering by 25%|
|Stage 2:||BDZ tapering by 50%|
|Stage 3:||BDZ tapering by 75%|
|Stage 4:||Free of BDZ|
|Baseline||Start tapering off from Stage 1|
|Week 1||Continue Stage 1 when judged successful or back to Stage 0 when unsuccessful (Go to Stage 0 or 1)|
|Week 2||Step up by one stage when judged successful or stay in the same stage (Go to Stage 0, 1, or 2)|
|Week 4||Step up by one stage when judged successful or stay in the same stage (Go to Stage 0, 1, 2, or 3)|
|Week 8||Step up by one stage when judged successful or stay in the same stage (Go to Stage 0, 1, 2, 3, or 4)|
In the SSRI (+) group, all the patients had started taking 10 mg of paroxetine from the beginning of the program. When the subjects failed to step up a stage after successively completing two assessments, they were asked to increase the dose of paroxetine by 10 mg, and only those who agreed were administered 20 mg a day.
Each participant was consistently assessed by one of three research physicians specializing in both psychiatry and internal medicine during the entire 8-week study protocol, under the continuous care of a physician-in-charge. After the completion of the program each participant was returned to the care of the physician-in-charge, except for those who wished to continue treatment with the research physicians.
All analyses were performed using the SAS statistical package.19 A P-value of <0.05 was regarded as significant.
Wilcoxon rank sum test was used to compare the progress of the steps (0–4) of the BDZ tapering between the two BDZ-reduction groups in relation to the effects of paroxetine after the completion of the 8-week program. Then, a multiple logistic regression analysis was used to examine the independent effects of SSRI assistance (SSRI (+) group or SSRI (–) group) on BDZ-free patients after the program completion, controlling for the effects of age, gender, period of visit to the internal medicine clinic, period of BDZ use, and HAM-D and HAM-A scores at baseline. The degree of change in the scores of the three questionnaires (HAM-D, HAM-A, and BDZ Withdrawal Symptom Questionnaire) between baseline and the 8-week assessment were estimated; one-way anova was used for group comparisons.
Of the 66 non-major-depressive patients, 22 were allocated to the SSRI (+) group, 23 to the SSRI (–) group, and 21 to the reference group. The results of the BDZ discontinuation program in the two BDZ-reduction groups are shown in Table 2. At the 1-week assessment, seven patients in the SSRI (+) group had not followed the drug regimen for tapering BDZ use; among them, three failed to follow the regimen because of the side-effects of paroxetine (two for nausea and one for headache). As a result, one patient did not reduce BDZ intake or take the 10-mg dose of paroxetine during the program and, therefore, is regarded as a stage 0 in all of the assessments up to 8 weeks (Table 2). Of the remaining 21 patients in the SSRI (+) group, 15 continued to be given 10 mg of paroxetine during the program, four were given 10 mg for the first 2 weeks and 20 mg for the remaining 6 weeks, and two were given 10 mg for the first 4 weeks and 20 mg for the latter 4 weeks, respectively.
Table 2. Comparison of benzodiazepine reduction during the study period
|SSRI (+) group (n = 22):|
| Stage 0 (full dose of BDZ)||22 (100%)|| 7 (31.8%)|| 2 (9.1%)|| 1 (4.6%)|| 1 (4.6%)|
| Stage 1 (25% taper)||–||15 (68.2%)|| 7 (31.8%)|| 2 (9.1%)|| 0 (0%)|
| Stage 2 (50% taper)||–||–||13 (59.0%)|| 7 (31.8%)|| 3 (13.6%)|
| Stage 3 (75% taper)||–||–||–||12 (54.5%)|| 8 (36.4%)|
| Stage 4 (free of BDZ)||–||–||–||–||10 (45.5%)|
|SSRI (–) group (n = 23):|
| Stage 0 (full dose of BDZ)||23 (100%)|| 4 (17.4%)|| 4 (17.4%)|| 2 (8.7%)|| 1 (4.4%)|
| Stage 1 (25% taper)||–||19 (82.6%)|| 7 (30.4%)|| 7 (30.4%)|| 4 (17.4%)|
| Stage 2 (50% taper)||–||–||12 (52.2%)|| 9 (39.1%)||10 (43.5%)|
| Stage 3 (75% taper)||–||–||–|| 5 (21.7%)|| 4 (17.4%)|
| Stage 4 (free of BDZ)||–||–||–||–|| 4 (17.4%)|
At the baseline, 4-week, and 8-week assessments, there were no significant differences in the scores on the BDZ Withdrawal Symptom Questionnaire, the HAM-D scores, and HAM-A scores among the three groups. More specifically, the average (SD) baseline scores on the BDZ Withdrawal Symptom Questionnaire were 26.4 (3.7) for the SSRI (+) group, 24.1 (3.5) for the SSRI (–) group, and 25.7 (7.4) for the reference group (P = 0.375, among the three groups); those on the HAM-D were 8.8 (6.8), 5.6 (3.4), and 7.9 (6.6), respectively (P = 0.918); those on the HAM-A were 10.1 (9.2), 7.7 (6.8), and 9.0 (9.1), respectively (P = 0.344). At the 8-week assessment, the degrees of score changes in the BDZ Withdrawal Symptom Questionnaire were −2.7 (3.4) for the SSRI (+) group, 0.4 (3.9) for the SSRI (–) group, and −2.0 (2.4) for the reference group (P = 0.475, among the three groups); those on the HAM-D were −2.3 (5.2), 0.9 (3.2), and 0.8 (3.7), respectively (P = 0.087); those on the HAM-A were −0.5 (9.7), 4.7 (6.2), and −0.7 (6.4), respectively (P = 0.580).
At the 8-week assessment, 45% of the patients in the SSRI (+) group were successful, compared to 17% in the SSRI (–) group (Table 2). The results of multiple logistic regression analysis indicated that the use of SSRI significantly contributed to the success of becoming BDZ-free after program completion (odds ratio [95% confidence intervals]: 6.09 [1.14–32.49], P = 0.023), controlling for the effects of age, gender, period of BDZ use, period of visit to the internal medicine clinic, and HAM-D and HAM-A scores at baseline.
In the SSRI group, there were no significant differences in the scores on the BDZ Withdrawal Symptom Questionnaire, the HAM-D scores, and HAM-A scores between the successful cases (n = 10) and unsuccessful cases (n = 12) at the baseline, 1-week, 2-week, 4-week, and 8-week assessments. More specifically, the average (SD) baseline scores on the BDZ Withdrawal Symptom Questionnaire were 27.3 (4.4) for the successful cases and 25.8 (2.8) for the unsuccessful cases (P = 0.343); those on the HAM-D were 8.7 (5.2) and 8.8 (8.6), respectively (P = 0.975); those on the HAM-A were 9.3 (11.1) and 10.8 (7.8), respectively (P = 0.714). At the 8-week assessment, the degrees of score changes in the BDZ Withdrawal Symptom Questionnaire were −3.0 (4.8) for the successful cases and −2.4 (3.0) for the unsuccessful cases (P = 0.724); those on the HAM-D were −2.8 (5.1) and −1.8 (4.1), respectively (P = 0.615); those on the HAM-A were −0.9 (2.0) and −0.1 (1.1), respectively (P = 0.248).
A standardized BDZ discontinuation program was conducted in the patient group without major depression in the present study. The number of patients (45%) that succeeded in being free of BDZ use in the SSRI (+) group was significantly higher than in the SSRI (–) group. The results of the multiple regression analysis also indicated that being in the SSRI (+) group significantly contributed to the success of becoming BDZ-free when the effects of the important demographic and clinical factors of age, gender, period of BDZ use, period of visit to the internal medicine clinic, and degrees of depression and anxiety were controlled for in the analysis. These results are noteworthy because the previous Dutch study10 concluded that paroxetine is of limited value for the gradual BDZ withdrawal during a BDZ discontinuation program and because the higher success rate of being BDZ-free was observed in the SSRI (+) group without the worsening of scores on the BDZ Withdrawal Symptom Questionnaire, HAM-D, and HAM-A, compared with the SSRI (–) group in the present study. The excessive prescription of BDZ is one of important clinical issues in Japan; for example, the number of prescriptions of anxiolytic BDZ for neurotic disorders was estimated to be higher in Japan (15–18 million prescriptions per year) than in the USA (8–10 million) between 2000 and 2002, even though the number of prescriptions for antidepressants for such disorders was much higher in the USA (9–12 million) than in Japan (1.8–2.1 million) in the same period.20 Therefore, the prescription of BDZ should be monitored carefully and prescribed appropriately, particularly in general medicine hospitals and clinics in Japan.
In the present study, major depression was observed in 32.0% of the chronic BDZ users treated by the internists. The occurrence of major depression appears to be high in comparison with the occurrence in two of the authors’ previous studies (both n > 1000) that reported the prevalence of major depression, as assessed through the same DSM-IV interviews, to be 4.0–4.5% in a sample of Japanese patients visiting a psychosomatic medicine clinic21,22 and 2.5–3.0% in a sample of Japanese workers.23,24 Affective and cognitive symptoms of depression are not readily observable in some cases during internal medicine clinic visits,25,26 and may be difficult for internists to find so-called ‘masked depression’.27
The present study has several limitations. First, this was a non-blinded randomized controlled trial, and, therefore, the data need to be interpreted carefully. The authors avoided adopting a sham treatment of using placebo tablets of paroxetine because of ethical reasons.28 Although the authors obtained the promising findings based on daily clinical practice, the nature of the non-blinded study design requires them to consider the variability in both physician factors, such as the fluency of instructions for treatment, and patient factors, such as the willingness to accept instructions29 in future studies.
The second limitation of this study was that the type and dosage of BDZ were not controlled owing to the relatively small sample size. Instead, patients using high-potency BDZ with a short duration of action were specifically selected for the study. Third, because the patient sample was drawn from the Internal Medicine Clinic, a variety of disorders were included. This was intended to represent the problems of chronic BDZ use in general in the area of internal medicine practice and not to a specific disorder. Fourth, although three research physicians diagnosed the DSM-IV major depression, the interrate reliability was not discussed. However, the reliability would be better than that in the previous study, which consisted of 45 general practitioners for screening major depression.10 The authors also used the HAM-D to assess the severity of depressive symptoms and discussed patients’ diagnoses twice weekly in order to improve the reliability of the diagnoses.21 In future studies, it would be useful to consider other depressive states, such as minor depression and dysthymic disorders, in the withdrawal of BDZ.
In spite of these limitations, the authors showed that a minimum dosage of paroxetine might facilitate the discontinuation of chronic BDZ use among outpatients visiting an internal medicine clinic. The next step is to confirm the results at other institutes, taking into account the factors not controlled in the present study. A follow-up study is also required to confirm the long-term anxiolytic effect of SSRI after the successful BDZ discontinuation program.
The authors wish to thank the clinical staff at the Psychiatry Clinic of Teikyo University Hospital for their help of daily clinical activities at the Psychosomatic Medicine Clinic and for their valuable suggestions to organize the BDZ discontinuation program. Also, all the clinical staff at the Internal Medicine Clinic of Teikyo University Hospital for their help with patients’ data collection.