Survey of benzodiazepine and antidepressant use in outpatients with mood disorders in Japan


Hiroyuki Uchida, MD, PhD, Centre for Addiction and Mental Health, PET Centre, 250 College Street, Toronto, Ontario, M5T 1R8, Canada. Email:


Data on benzodiazepine use in mood disorders are still limited, especially among seniors. A cross-sectional review of psychotropic prescriptions in 948 outpatients with mood disorders (405 male; mean ± SD age, 52 ± 17 years; age range, 16– 91 years) was conducted in Japan. The use of benzodiazepine-derivative anxiolytics was approximately 60% in all decades, including older patients, without a group difference. The frequent use of benzodiazepines is a cause for concern because they are not preferred treatment, given their well-known adverse effects especially in the elderly.

PSYCHOPHARMACOLOGICAL TREATMENT IS an integral part of the treatment for mood disorders. Of psychotropic drugs, benzodiazepines should be prescribed with special caution because of their well-known side-effects such as dependence1 and cognitive impairment.2 This is especially true for the elderly, given age-related physiological changes.3 This caution notwithstanding, several cross-sectional studies reported on the frequent use of benzodiazepines in older patients with depression or mixed diagnoses, ranging between 20% and 50% in the USA, Taiwan, and Italy.4–6 Prescription behavior is expected to be subject to physician background in training as well as direct and indirect influence of local standard of care, which indicates the necessity of further information from various clinical settings in order to provide a robust agreement on this issue.

In an effort to address this gap in the literature, we examined the effect of age on prescription patterns of benzodiazepines and antidepressants in outpatients with mood disorders across 30 sites in Japan.


A systematic chart review was conducted at 30 psychiatric hospitals and clinics in Tokyo, Japan. A total of 948 outpatients with mood disorders according to the ICD-107 who received care at least once at a participating site in September 2002 (mean ± SD age, 52 ± 17 years) were included. The information on age, gender, and prescribed psychotropic medications were collected by trained psychiatrists. Daily doses of benzodiazepine-derivative anxiolytics and antidepressants were converted to imipramine and diazepam equivalents, respectively.8 Thienodiazepines were not included. For prescription of pro re nata (as-needed) medication, patients were regarded as benzodiazepine users and an expected average daily dose of the medications was calculated as a total prescribed dose divided by days until their next scheduled appointment. For patients receiving more than one anxiolytic or antidepressant drugs, a sum of equivalents for all prescribed medications was calculated. The study was approved by the institutional review board or equivalent committee at all participating sites.

Differences between age groups were tested using one-way analysis of variance (anova) for parametric data, or χ2 test for categorical variables and the Tukey–Kramer honestly significant difference when applicable. P < 0.05 was considered statistically significant and all tests were two-tailed.


The use of benzodiazepine-derivative anxiolytics was approximately 60% in all age groups without an age group difference (Table 1). When the data from those antidepressant recipients (n = 799) were separately analyzed, benzodiazepine-derivative anxiolytics were prescribed for 64.6% of patients; selective serotonin re-uptake inhibitor/serotonin norepinephrine re-uptake inhibitor (SSRI/SNRI) was found to be associated with lower use of benzodiazepine-derivative anxiolytics (61.7%, n = 258/418) as compared to tricyclic antidepressants (TCA; 72.6%, n = 159/219) and sulpiride (73.4%, n = 185/252; χ2 = 9.91, d.f. = 2, P = 0.007). Similar to the results from all patients, no significant differences in the concomitant use rate of benzodiazepine-derivative anxiolytics were found among age groups in any of these drug categories. Benzodiazepine use was lower in concomitant antipsychotics users (48.7%, n = 73/150) than in those who were not (62.5%, n = 499/798; χ2 = 10.14, d.f. = 1, P = 0.002).

Table 1.  Prescription patterns of benzodiazepine-derivative anxiolytics and antidepressants
 TotalAge groups (years)30–3940–4950–5960–6970–79≥80Statistics
  • *a

    P < 0.05 for 4th vs 7th decades (Tukey–Kramer HSD);

  • *b

    P < 0.05 for 3rd vs 7th, 3rd vs 8th, 4th vs 7th, 4th vs 8th, 5th vs 8th, and 6th vs 8th decades (Tukey–Kramer HSD).

  • F = 4.6, d.f. = 9, 792 (one-way analysis of variance);

  • F = 2.8, d.f. = 6, 565 (one-way analysis of variance).

  • BZD, benzodiazepine-derivative anxiolytics; DZPE, diazepam equivalent; HSD, honestly significant difference; IMIE, imipramine equivalent.

n (male)948 (405)104 (43)171 (88)155 (85)173 (80)170 (67)133 (34)42 (8) 
Antidepressants (n = 799)         
 Prescription rate (%)84.389.487.183.280.984.782.083.3NS
 Dose (IMIE mg/day), mean ± SD*a114.0 ± 86.5112.2 ± 67.3131.1 ± 106.4126.6 ± 88.4114.5 ± 77.4113.8 ± 90.092.6 ± 76.064.5 ± 40.4P < 0.0001
BZD (n = 572)         
 Prescription rate (%)60.351.059.666.561.060.758.657.1NS
 Dose (DZPE mg/day), mean ± SD*b9.1 ± 8.08.1 ± 5.79.6 ± 8.310.7 ± 8.69.7 ± 8.09.4 ± 10.36.8 ± 4.35.9 ± 4.6P < 0.0001


Benzodiazepines were used in the majority of patients irrespective of age in the treatment of mood disorders in the present sample. Furthermore, this mode of treatment was prevalent also in older patients, similar to previous surveys conducted in other countries.4–6 Adverse effects of benzodiazepines are potentially very serious,1,2 especially in the elderly.3.In view of these detrimental adverse effects, benzodiazepines are listed in the 2002 criteria for potentially inappropriate medications used in older adults.9 The physician prescribing practice of favoring benzodiazepines for mood disorders may reflect psychological and empirical dependence on these drugs due to the rapid onset of clinical effect and/or a groundless sense of safety associated with the use of these potentially harmful drugs.10

There are several limitations to be noted. First, we analyzed the data from patients with all F3 diagnoses (ICD-10) combined together. A plausible heterogeneity among these disorders may be associated with a variation in terms of psychopharmacology, although we are not aware of any report that advocates the use of benzodiazepines for any mood disorder. Second, the cross-sectional nature of this survey did not allow us to evaluate the process of drug treatment or compliance to the treatment. If we had collected such information, it might have allowed us to answer various relevant questions, including reasons for lower benzodiazepine use in SSRI/SNRI users as compared to TCA users and age-related changes in prescribed doses.

In conclusion, the high prescription rate of benzodiazepines is a cause of concern given the well-known adverse effects of these medications, especially in the elderly. Future investigations are warranted to pursue safe methods of discontinuing (or at least tapering) benzodiazepines for already medicated patients to minimize the potentially serious but reversible side-effects.


The authors thank Drs K. Ishii, S. Katayama, and Y. Imasaka for their valuable comments.