Journal of Psychiatric Research xxx (2017) xxx-xxx

Contents lists available at ScienceDirect

Journal of Psychiatric Research

journal homepage: www.elsevier.com

F
OO
What can predict and prevent the long-term use of benzodiazepines?
Takeo Hataa⁠ ,⁠ b⁠ ,⁠ ∗⁠ , Tetsufumi Kanazawac⁠ ,⁠ d⁠ ,⁠ e⁠ , Takeshi Hamadaa⁠ , Masami Nishiharaa⁠ , Ashley Ian Bushd⁠ ,⁠ e⁠ ,
Hiroshi Yonedac⁠ , Miki Nakajimab⁠ , Takahiro Katsumataa⁠

PR
a
Department of Pharmacy, Osaka Medical College Hospital, 2-7 Daigaku-machi, Takatsuki, Osaka 569-8686, Japan
b
Drug Metabolism and Toxicology, Faculty of Pharmaceutical Sciences, Kanazawa University, Kakuma-machi, Kanazawa, Ishikawa 920-1192, Japan
c
Department of Neuropsychiatry, Osaka Medical College, 2-7 Daigaku-machi, Takatsuki, Osaka 569-8686, Japan
d
Department of Psychiatry, The University of Melbourne, VIC 3010, Australia
e
Melbourne Dementia Research Centre, The Florey Institute of Neuroscience and Mental Health, 30 Royal Parade, Parkville, VIC 3052, Australia

ARTICLE INFO ABSTRACT

D
Keywords: Although benzodiazepines (BZDs) are commonly prescribed for insomnia or anxiety, long-term use of BZDs
Benzodiazepines (BZDs) causes serious adverse effects such as daytime drowsiness and cognitive decline. In the current study, we eval-
Long-term use
TE uated the predictors and preventers of long-term usage of BZDs from a retrospective survey by utilizing the
Sleep medicine 12-year prescription record of a university hospital. From the prescription data of 92,005 people, users of BZDs
Pharmacy
(n = 3,470, male = 39.2%, mean age = 60 ± 17.5) were analyzed. During this period, both the number of pre-
Psychiatry
Retrospective survey scriptions (2722 in 2004 to 1019 in 2016) and the number of BZDs (1.73 in 2004 to 1.36 in 2016) gradually
decreased, although more than half of the patients continued to take BZDs for over three years. High risk factors
for long-term use of BZDs include elderly patients (>65 years old), high dosage (>5 mg diazepam per day), psy-
chiatrist-prescribers, and users with polytherapy. Discontinuation is significantly found in users of hypnotic BZDs
EC

and alternative psychotropic medical drugs (including antipsychotics, serotonergic drugs, or newer types of sleep
medicine). Future studies should focus on elucidating interventions that are more effective against long-term us-
age of BZDs.

2003), and cognitive decline (Shorr and Robin, 1994), as well as socioe-
RR

1. Introduction conomic costs (Panneman et al., 2003) has been described. A nested,
case-control study estimating the impact of injurious falls associated
Benzodiazepines (BZDs) have been widely prescribed in clinical with BZDs in a population of elderly persons reported a statistically sig-
practice around the world for over four decades to treat various dis- nificant risk of serious falls (Pariente et al., 2008).
orders (Charlson et al., 2009). BZDs have been used for conditions in- Although short-term use of BZDs is considered safe and effective
cluding insomnia, anxiety, panic disorder, muscle spasticity, and vari- (Martin et al., 2007; Pfeiffer et al., 2011), various studies suggest that
ous mental disorders (Cloos and Ferreira, 2009; Fisher et al., 2012). Al- long-term use may be harmful. Long-term use carries the risk of de-
CO

though all BZDs have very similar mechanisms of action, specific BZDs pendence, withdrawal syndrome, cognitive impairment, diminishing ef-
are used for specific purposes, as they differ based on the duration of ac- fect, tolerance, and difficulty in discontinuing treatment (Charlson et al.,
tion, metabolic half-life, and the presence/absence of psychotropically 2009; Lader, 2011; Lader and Petursson, 1981). Some patients may ex-
active metabolites (Lader, 2011). perience withdrawal symptoms such as rebound insomnia and anxiety
It has been well-documented that the use of BZDs presents poten- after only 2–4 weeks of treatment (Urru et al., 2015). Receiving hyp-
tially serious adverse effects, including daytime drowsiness, light-head- notic prescriptions is also associated with excess mortality (Belleville,
edness, ataxia, psychomotor disturbance, and anterograde amnesia 2010) and cancer incidence (Kripke et al., 2012).
UN

(Barker et al., 2004; Holbrook et al., 2000). Especially in the elderly, a In response to these risks, although several international clinical
relationship between BZD use and an increased risk of road traffic ac- guidelines and expert consensus statements have been published that
cidents (Smink et al., 2010), falls, fractures (Cumming and Le Couteur, recommend limiting the long-term use of BZDs, especially in older pa

∗ Corresponding author. Department of Pharmacy, Osaka Medical College Hospital, 2-7 Daigaku-machi, Takatsuki, Osaka 569-8686, Japan.
Email address: [email protected] (T. Hata)

https://doi.org/10.1016/j.jpsychires.2017.11.012
Received 28 September 2017; Received in revised form 14 November 2017; Accepted 22 November 2017
Available online xxx
0022-3956/ © 2017.
T. Hata et al. Journal of Psychiatric Research xxx (2017) xxx-xxx

tients (American Geriatrics Society 2012 Beers Criteria Update Expert ethics committee of Osaka Medical College (Approval ID: RIN-486,
Panel, 2012; McLeod et al., 1997), the prevalence of long-term use re- 2160).
mains widespread (Vicens et al., 2014). Thus, reducing the long-term
use of BZD is an important worldwide issue. 2.2. Study population
Many studies have been performed to identify the prevalence and
risk factors of long-term BZD use. Older age has been found to be a sig- To describe the tendency to prescribe BZDs using the medical ac-

F
nificant factor in predicting the continuation of BZD use (Luijendijk et counting data, we used the data of outpatients from the Department
al., 2008; Manthey et al., 2011). Patients who may have previously used of Psychiatry. To describe the predictors of BZD continuation using
BZDs tend to become long-term BZD users. The quantity of BZD used is the prescription data, we defined the following inclusion criteria: age

OO
also a strong predictor of long-term use (Kjosavik et al., 2012), whereas older than 18 years, outpatients of all clinical departments, oral ad-
the impact of factors such as lower level of education (Manthey et al., ministration, no BZD prescription in the last year or more in order to
2011) and sex (Taipale et al., 2015) is still controversial. Although psy- adopt to the new user design. We included drugs approved for medical
chiatric history, such as anxiety, tension, schizophrenia, bipolar disor- prescription in Japan coded as N03A, N05B, and N05C, according to
der, depression, and Alzheimer's disease, are positively associated with the Anatomical Therapeutic and Chemical (ATC) classification system
BZD continuation (Manthey et al., 2011; Taipale et al., 2015), concur- (Table 1).
rent use of antipsychotics and mood stabilizers is significantly associ-

PR
ated with BZD discontinuation (Veronese et al., 2007). 2.3. Continuation of BZD use
Several alternative medications to BZDs are prescribed for psychi-
atric conditions. Quetiapine is approved and marketed in many coun- Similar to previous reports, we defined BZD continuation as the re-
tries for schizophrenia, and is commonly used off-label for the treatment ceipt of at least one prescription for BZD within three months (Ishigooka
of insomnia and anxiety in older adults, owing to its sedative properties et al., 1998; Takeshima et al., 2016). The three-month time window was
when used at low doses (Iaboni et al., 2016). Trazodone, a serotonin chosen because the prescription of BZDs is restricted to once per 30 or
antagonist and reuptake inhibitor used for depressive disorders, is also 90 days in Japan, depending on the product (Takeshima et al., 2016).
prescribed off-label for the treatment of these conditions (Bossini et al.,

D
2015; Iaboni et al., 2016). 2.4. Potential predictors of BZD continuation
Furthermore, a new class of hypnotics has recently begun to be dis-
tributed in Japan. Ramelteon, a selective melatonin receptor type 1 We investigated the following potential predictors: sex, age (18–45,
and 2 (MT1/MT2) agonist, was approved in 2010 to treat sleep disor-
TE
45–65, 65–85, >85), medical specialty (psychiatrist-prescriber or
ders; it has a good safety profile. Ramelteon has no measurable affin-
non-psychiatrist-prescriber) (Takeshima et al., 2016), dosage of BZD [di-
ity for a large number of ligand binding sites (including BZD receptors,
azepam (DZP) equivalent daily dose (Inada and Inagaki, 2015); ≤5 mg,
dopamine receptors, opiate receptors) associated with cognitive impair-
5 mg <], concurrent use of BZD (monotherapy, polytherapy), indica-
ment, dependence, and withdrawal effects (Kato et al., 2005). Suvorex-
ant, an orexin receptor antagonist, has also been available since 2014, tion of initial prescription (only at bedtime, others), duration of initial
and was found to be well tolerated in clinical trials, with some mild ad- prescription (<21 days, ≥ 21 days), half-life of initial BZD [ultra-short
EC

verse effects (Schroeck et al., 2016). Rebound or withdrawal symptoms (<6 h), short (6–12 h), medium (12–24 h), or long (>24 h)] (Passaro
were not observed with the discontinuation of suvorexant (Schroeck et et al., 2000), type of BZD (hypnotic, anxiolytic, others), and concurrent
al., 2016). psychotropic drugs (antidepressant, antipsychotics, or alternative drug
These alternative hypnotics may effectively prevent the long-term to BZD).
use of BZDs while maintaining good sleep quality. Nevertheless, little In case of patients who had visited doctors from multiple clinical de-
information is available about the effects of alternative hypnotics on partments including psychiatry, we assigned them to a psychiatrist-pre-
RR

reducing or discontinuing BZD use. Recently, a study based on a na- scriber. Note that the psychiatry department in our hospital sees vari-
tionwide claims-database identified that concurrent use of antipsychotic ous psychiatric patients, including those with a diagnosis of schizophre-
drugs was a predictor of long-term BZD use. However, the limitation nia, depression, bipolar disorder, dementia, anxiety disorder, and so on.
of this study was that the cohort analyzed included very few people We converted the total BZD daily dose into a DZP equivalent accord-
that were 75 years old and above, because the analyzed claims-database ing to Inada et al. (Inada and Inagaki, 2015), and a maximal dosage of
primarily consisted of corporate employees (Takeshima et al., 2016). DZP-equivalent dose in the observation period was chosen as the dosage
Therefore, the current study aimed to describe the tendency to prescribe of BZD. We referred to package inserts and booklets, also known as in-
CO

BZDs in the last 12 years, and to elucidate the predictors of BZD dis- terview forms, prepared by the manufacturers, to obtain the informa-
continuation in an extended age group of patients, who were prescribed tion regarding the half-life of the particular BZD. When patients used
BZDs in our university hospital located in a city area, using data derived more than one BZD with different half-lives, the BZD with the longest
from an electronic health-record database. half-life was chosen (Takeshima et al., 2016). When patients received
both a hypnotic and an anxiolytic, we assigned them to both the hyp-
2. Methods notic and anxiolytic groups to define the type of BZD. Alternative drugs
to BZD that were analyzed in this study are listed in Table 2.
UN

2.1. Data source

2.5. Statistical analysis
We used accounting data from the medical accounting system dur-
ing the period from March 2004 to September 2016 to analyze BZD The Kaplan-Meier event-free curve was employed to calculate the
prescription tendencies for psychiatric patients in the past 12 years, event-free probability. Multivariate Cox regression analysis was used to
and the prescription data from the electronic health record during explore independent predictors associated with the time-to-event. The
the period from January 2012 to September 2016 for describing the event was the discontinuation of BZD prescription. When the partici-
predictors of BZD continuation in our university hospital. The hospi- pants were not prescribed BZD during the course of three months, we
tal serves as a general hospital for the city of Takatsuki (population: took the month of the last BZD prescription as the BZD discontinuation
350,000) in Osaka. It has 882 beds and 29 clinical departments in- date (Takeshima et al., 2016). Variance inflation factors (VIF) of 10 or
cluding the Department of Psychiatry. This study was approved by the

2
T. Hata et al. Journal of Psychiatric Research xxx (2017) xxx-xxx

Table 1 Table 2
The 31 benzodiazepines analyzed in this study. Alternative drugs for BZD.

Diazepam Escitalopram Oxalate

equivalent dose Fluvoxamine Maleate
Benzodiazepines (mg) Half-life Type Mianserin Hydrochloride
Mirtazapine

F
Tofisopam 125 Ultra Others Paroxetine Hydrochloride Hydrate
short Quetiapine Fumarate
(<6 h) Ramelteon
Triazolam 0.25 Ultra Hypnotic Suvorexant

OO
short Sertraline Hydrochloride
(<6 h) Tandospirone Citrate
Zolpidem Tartrate 10 Ultra Hypnotic Trazodone Hydrochloride
short
BZD, Benzodiazepine.
(<6 h)
Zopiclone 7.5 Ultra Hypnotic
short more were considered evidence that multi-collinearly existed. The pro-
(<6 h) portional-hazards assumption was assessed using log-minus-log plots.
Alprazolam 0.8 Short Anxiolytic All reported P values are two-tailed and the level of significance was set

PR
(6–12 h)
at P < 0.05. Statistical analyses were performed using R version 3.4.0.
Brotizolam 0.25 Short Hypnotic
(6–12 h)
Clotiazepam 10 Short Anxiolytic 3. Results
(6–12 h)
Eszopiclone 2.5 Short Hypnotic 3.1. Characteristics and the disposition of patients
(6–12 h)
Etizolam 1.5 Short Others
We extracted the data of 92,005 people from the electronic health
(6–12 h)

D
Lormetazepam 1 Short Hypnotic records, which included outpatients who were prescribed drugs from
(6–12 h) any clinical department in the hospital during the observational period.
Rilmazafone 2 Short Hypnotic Of the total number of outpatients, 8006 (8.7%) were prescribed BZDs
Hydrochloride (6–12 h) at least once; of these, 3639 patients were identified as new BZD users.
TE
Hydrate
Upon excluding 169 patients who were less than 18 years old, 3470 pa-
Amobarbital 50 Medium Hypnotic
(12–24 h) tients remained in our cohort (Fig. 1). Demographic and clinical charac-
Bromazepam 2.5 Medium Anxiolytic teristics of our cohort are presented in Table 3. The mean age ± standard
(12–24 h) deviation (SD) was 60 ± 17.5 years old, and 50.5% of the cohort was 65
Estazolam 2 Medium Hypnotic years old or more. Additionally, 22.8% of the patients were prescribed
(12–24 h)
BZD by psychiatrists. The mean dosage of BZD ±SD was 7.2 ± 8.7 mg of
EC

Flunitrazepam 1 Medium Hypnotic

(12–24 h) a DZP-equivalent daily dose, and 32.2% of the cohort belonged to the
Lorazepam 1.2 Medium Anxiolytic group with more than 5 mg dosage. Furthermore, 20.3% of patients had
(12–24 h) concurrent use of BZD, 64.4% of patients were prescribed a hypnotic as
Nitrazepam 5 Medium Hypnotic
a type of BZD, and 24.4% of patients had concurrent use of an alterna-
(12–24 h)
Chlordiazepoxide 10 Long Anxiolytic
tive drug to BZDs.
(24 h ≤)
RR

Clobazam 10 Long Others 3.2. Tendency to prescribe BZDs

(24 h ≤)
Clonazepam 0.25 Long Hypnotic The tendency to prescribe BZDs for psychiatric patients from March
(24 h ≤)
2004 to September 2016 is shown in Fig. 2. The number of prescrip-
Clorazepate 7.5 Long Anxiolytic
Dipotassium (24 h ≤) tions containing BZDs decreased gradually from 2722 to 1019 during
Cloxazolam 1.5 Long Anxiolytic the whole period that we observed (Fig. 2A). The reduction in these val-
(24 h ≤) ues also occurred as the proportion of patients per 1000 gradually de-
CO

Diazepam 5 Long Others creased from 992 to 482 during our period of observation (Fig. 2B). Fur-
(24 h ≤) thermore, we calculated the number of BZDs per a prescription, which
Ethyl Loflazepate 1.67 Long Anxiolytic
decreased slowly from 1.73 to 1.36 (Fig. 3).
(24 h ≤)
Fludiazepam 0.5 Long Anxiolytic
(24 h ≤) 3.3. Proportion of BZD continuation
Flurazepam 15 Long Hypnotic
Hydrochloride (24 h ≤) The Kaplan-Meier event-free curve, censored at 45-months, describes
UN

Medazepam 10 Long Anxiolytic the time to discontinuation of BZD use; it is shown in Fig. 4. We found
(24 h ≤)
that 2901 (83.6%) people were censored. The estimated probabilities of
Oxazolam 20 Long Anxiolytic
(24 h ≤) the continuation of BZD use for new BZD users were 90.0% for 2-month
Phenobarbital 15 Long Others continuation, 86.2% for 3-month continuation, 80.7% for 6-month con-
(24 h ≤) tinuation, and 71.1% for 12-month continuation, respectively. Notably,
Quazepam 15 Long Hypnotic 57.8% represented continuation for 36-months or higher. The median
(24 h ≤)
discontinuation time was not calculated because the proportion of the
Tandospirone Citrate 25 Long Anxiolytic
(24 h ≤)
event was less than 50%.

3
T. Hata et al. Journal of Psychiatric Research xxx (2017) xxx-xxx

Table 3
Demographic and clinical characteristics of the patients.

Characteristics (n = 3470)

Sex, no. (%)

Male 1360 (39.2)

F
Female 2110 (60.8)
Age, year
Mean (SD) 60 (17.5)
Range 18–96

OO
Age group, no. (%)
18–45 789 (22.7)
45–65 928 (26.8)
65–85 1677 (48.3)
85 ≤ 76 (2.2)
Medical specialty, no. (%)
Non-psychiatry 2679 (77.2)
Psychiatry 791 (22.8)

PR
Dosage of BZD, mga⁠
Mean (SD) 7.2 (8.7)
Range 0.8–120
Dosage of BZD group, no. (%)a⁠
≤ 5 mg 2354 (67.8)
5 mg < 1116 (32.2)
Concurrent use of BZD, no. (%)
Monotherapy 2765 (79.7)
Polytherapy 705 (20.3)
Direction of initial prescription, no. (%)

D
Only at bedtime 2426 (69.9)
Others 1044 (30.1)
Duration of initial prescription, no. (%)
< 21 days 2002 (57.7)
TE
21 days ≤ 1468 (42.3)
Half-life of initial BZD, no. (%)b⁠
Ultra short (<6 h) 629 (18.1)
Short (6–12 h) 1768 (51.0)
Medium (12–24 h) 272 (7.8)
Long (24 h ≤) 801 (23.1)
Fig. 1. The study population.
Type of BZD, no. (%)c⁠
EC

Hypnotic 2234 (64.4)

3.4. Independent predictors of BZD continuation Anxiolytic 743 (21.4)
Others 713 (20.5)
Multivariate Cox regression analysis revealed seven independent Concurrent psychotropic drug, no. (%)c⁠
Antidepressants 770 (22.1)
predictors (Fig. 5). Continuation of BZDs was significantly associated Antipsychotics 549 (15.8)
with age 65 years or greater, psychiatrist-prescribers, dosage of 5 mg Alternative drug of BZD 846 (24.4)
or more of a DZP-equivalent daily dose, and concurrent use of BZDs.
RR

BZD, Benzodiazepine; DZP, diazepam.

In contrast, the prescription of hypnotics (rather than only anxiolytics), a Data expressed as the maximal dosage of a DZP equivalent daily dose.
and concurrent use of antipsychotics or alternative drugs to BZDs were b When patients used more than one BZD with different half-lives, the one with the

significantly associated with the discontinuation of BZD use. We cal- longest half-life was chosen.
culated VIF as a measure of multicollinearity; we found that each VIF c As some patients receiving no drugs or multiple drugs, the sum is not 100%.

did not exceed 10. The proportional-hazards assumption was confirmed

through a log-minus-log plot. arisen due to a difference in the proportion of elderly patients aged 65
or more. Our current study revealed older age as a significant predictor
CO

4. Discussion of the continuation of BZD use, similar to a previous report (Luijendijk

et al., 2008; Manthey et al., 2011). Thus, the high frequency of BZD
In the current study, we investigated the independent predictors of continuation is expected because of a high proportion of older patients
discontinuation of BZD use in our university hospital by extracting data in our cohort. Psychiatric prescribers were also a significantly strong
from the electronic health record database, with a new -user design. predictor of long-term BZD use. This is potentially because the popula-
We found that 8006 (8.7%) of all outpatients at all clinical departments tion in the current study contained many patients (22.8%) whose symp-
in the hospital were prescribed BZDs during our observational period, toms tended to be severe, and who were prescribed BZDs by psychi-
UN

and continuation of BZD use for more than 36 months was observed in atrists. Although a decrease was observed in the number of prescrip-
57.8% of 3470 new users. Concurrent use of alternative drugs was an tions containing BZDs provided by psychiatrists over the entire exper-
independent predictor of the discontinuation of long-term use of BZDs. imental period, psychiatric prescribers were found to be a predictor
Continuation of BZDs for more than 36 months was observed in of BZD continuation. This implies that only decreasing the number of
57.8% of 3470 new users. A previous study investigating the long-term prescriptions has no contribution to a reduction in long-term BZD use.
use of BZDs reported that 15.2% of new BZD-users continued to use Several studies have suggested that general practitioners, rather than
BZDs in the first year (Takeshima et al., 2016). The frequency of BZD psychiatrists, are responsible for writing BZD prescriptions (Cascade
continuation found in our results was high. The discrepancy between and Kalali, 2008; Lagnaoui et al., 2004; Takeshima et al., 2016). This
the results of our study and those of the previous study may have suggestion stems from the observation that half of the general practi-
tioners (47%) think that patients expect prescriptions, and as many as
18% think that not writing prescriptions would threaten their doctor-

4
T. Hata et al. Journal of Psychiatric Research xxx (2017) xxx-xxx

F
OO
PR
Fig. 4. The Kaplan–Meier survival curve describes the time to discontinuation of benzodi-
azepine.

associated with BZD discontinuation. It has been reported that concur-

rent use of antipsychotics potentially increases the risk of long-term BZD

D
use (Takeshima et al., 2016), although our current data showed the op-
posite result. This is potentially because of differences in the analyzed
samples. The previous study by Takeshima et al. (2016) analyzed pri-
TE
marily working-age adults and their dependent, therefore, the average
age of the sample was 40 years, while our sample had a mean age of
Fig. 2. Monthly outpatient prescriptions containing BZD in the Department of Psychiatry, 60. In addition, the prior data were derived predominately from clin-
from March 2004 to September 2016.
ics, in which many patients with psychiatric diagnoses were diagnosed
with anxiety disorder or affective disorder (93.3%). The current out-
patient samples were obtained from a university hospital with a psy-
EC

chiatric inpatient ward; therefore, a considerable proportion of patients

with schizophrenia was included in the younger age group. Short-term
BZD use is frequently permitted for schizophrenia treatment. This could
explain the discontinuation of BZDs in those prescribed antipsychotics.
Thus, there is a possibility of reducing the quantity of BZDs and initiat-
ing antipsychotics while maintaining hypnotic efficacy. In contrast, it is
unknown if the use of antidepressants or BZDs has beneficial net effects,
RR

especially in the treatment of schizophrenia (Tiihonen et al., 2012). As

previously reported (Isacson, 1997; Kjosavik et al., 2012), we found the
quantity of BZD used to be a predictor of the long-term use of BZDs in
the present study. BZD polytherapy was a strong predictor of BZD con-
tinuation; therefore, it is ethically and practically correct to utilize BZD
monotherapy for treatment. One group reported that the type of BZDs
CO

used, including either hypnotics or sedatives, or both combined, is not a

statistically significant predictor of discontinuation/continuation, as de-
Fig. 3. The number of BZDs per a prescription of outpatients in the Department of Psychi-
atry, from March 2004 to September 2016. termined through multivariate analysis (Isacson, 1997). Another group
reported that hypnotic BZDs are associated with a higher risk of chronic
patient relationship (Anthierens et al., 2010). Unfortunately, we could use compared to anxiolytic BZDs (Luijendijk et al., 2008). In the current
not investigate this in our single-center study. study, we found that the use of hypnotic BZDs was a statistically signifi-
Concurrent use of alternative drugs was found to be an independent cant predictor of discontinuation. Thus, the effect of the type of BZD on
long-term use is still controversial.
UN

predictor of discontinuation of long-term use of BZDs, using multivari-

ate Cox regression analysis. As shown in Table 2, in addition to inves- Contrary to a previous report (Takeshima et al., 2016), the half-life
tigating the influence of antidepressants, anxiolytics, and antipsychotics of BZDs was not found to be a predictor of continued use. We think
as alternative hypnotics to BZD, we also tested novel hypnotics such as this disagreement is due to the different age ranges of the study popula-
ramelteon and suvorexant. Our results suggested that replacing BZDs tion in the two studies. The duration of initial prescription was also not
with these alternative hypnotics might lead to discontinuation of BZD found to be a predictor of continued use. In Japan, BZDs are restricted
use. To initiate the prescription of these alternative drugs to BZD, it is to a single prescription with a duration of 30 days in order to regu-
important to consider contraindications and drug interactions. Prescrib- late chronic BZD use; however, it is possible to prescribe BZDs repeat-
ing alternative drugs could be a countermeasure against long-term BZD edly, resulting in long-term prescription. This likely renders the 30-day
use. Along with the results of a previous report (Veronese et al., 2007), duration of prescription restriction ineffective, and permits BZD con-
our results suggest that concurrent use of antipsychotics is significantly tinuation. The PMDA (Pharmaceuticals and Medical Devices Agency, a

5
T. Hata et al. Journal of Psychiatric Research xxx (2017) xxx-xxx

F
OO
PR
D
TE
Fig. 5. Independent predictors of BZD continuation.

Japanese regulatory agency), in collaboration with the Japanese Min- develop complementary therapies that may be effective without the use
EC

istry of Health, Labour and Welfare warned to avoid long-term use of BZDs for disorders such as refractory insomnia.
of BZDs in March 2017 (http://www.pmda.go.jp/files/000217228.pdf),
although there exists no punitive legislation in Japan. Other poten- 5. Conclusion
tial predictors have also been surveyed previously. Earlier studies have
shown that the use of BZDs is higher in women, and it increases steadily This study revealed the factors associated with the long-term use of
with age (Petitjean et al., 2007; Simon et al., 1996), although, the effect BZDs, including elderly patients, psychiatrist prescribers, high dosage of
of sex on BZD continuation is still controversial (Taipale et al., 2015). BZD, and concurrent use of BZD. In contrast, factors associated with the
RR

These suggestions are consistent with our results. discontinuation of BZDs included using hypnotic types of BZDs and con-
Our study is not without limitations. A relatively small sample size current use of psychotropic drugs such as antipsychotics or alternative
was used in our study, compared to other studies based on nationwide drugs to BZDs. These results suggest that replacing BZDs with other psy-
claims-databases. There was a lack of prescriptions written by primary chotropic drugs, such as antipsychotics, selective serotonin reuptake in-
care physicians. It is uncertain whether the patients actually took the hibitors (SSRIs), tetracyclic antidepressants, or novel hypnotics such as
BZDs prescribed to them, as our data source only provided prescription ramelteon and suvorexant, may help to avoid long-term use of BZD.
CO

data from the electronic medical records. However, our dataset is su-
perior to a claims-database with respect to its impartial composition of Funding
sex, age, and occupation. We consider that our dataset, containing pre-
scriptions from several clinical departments and many kinds of drugs, This research did not receive any specific grant from funding agen-
is advantageous. The use of data from an unselected, population-based cies in the public, commercial, or not-for-profit sectors.
sample of patients may be an important step in avoiding selection bias
(Wu et al., 2011). Conflicts of interest
UN

We conclude that the prescription of alternative drugs to BZDs, as

surveyed in the present study, is one of the contributing factors that All authors do not have any relevant conflicts of interest to disclose
may help solve the problem of long-term BZD use. Novel hypnotics such with respect to this manuscript.
as ramelteon and suvorexant are more expensive than BZDs, and an-
tipsychotics have a risk of severe adverse effects, impaired glucose tol- Appendix A. Supplementary data
erance, neuroleptic malignant syndrome, and hematopathy. Therefore,
we currently cannot fully rely on these alternative drugs for treatment. Supplementary data related to this article can be found at https://
Further studies in form of clinical trials are necessary to clarify the doi.org/10.1016/j.jpsychires.2017.11.012.
efficacy and safety of alternative drugs to BZDs. This will help us to

6
T. Hata et al. Journal of Psychiatric Research xxx (2017) xxx-xxx

References lands study of depression and anxiety. J. Clin. Psychopharmacol. 31, 774–779. https://
doi.org/10.1097/JCP.0b013e3182362484.
Martin, J.L.R., Sainz-Pardo, M., Furukawa, T.A., Martín-Sánchez, E., Seoane, T., Galán,
American Geriatrics Society 2012 Beers Criteria Update Expert Panel, 2012. American
C., 2007. Benzodiazepines in generalized anxiety disorder: heterogeneity of outcomes
Geriatrics Society updated Beers Criteria for potentially inappropriate medication
based on a systematic review and meta-analysis of clinical trials. J. Psychopharmacol.
use in older adults. J. Am. Geriatr. Soc. 60, 616–631. https://doi.org/10.1111/j.
21, 774–782. https://doi.org/10.1177/0269881107077355.
1532-5415.2012.03923.x.
McLeod, P.J., Huang, A.R., Tamblyn, R.M., Gayton, D.C., 1997. Defining inappropriate
Anthierens, S., Pasteels, I., Habraken, H., Steinberg, P., Declercq, T., Christiaens, T.,
practices in prescribing for elderly people: a national consensus panel. CMAJ 156,

F
2010. Barriers to nonpharmacologic treatments for stress, anxiety, and insomnia: fam-
385–391.
ily physicians' attitudes toward benzodiazepine prescribing. Can. Fam. Physician 56,
Panneman, M.J.M., Goettsch, W.G., Kramarz, P., Herings, R.M.C., 2003. The costs of ben-
e398–406, doi:56/11/e398 (pii).
zodiazepine-associated hospital-treated fall Injuries in the EU: a Pharmo study. Drugs
Barker, M.J., Greenwood, K.M., Jackson, M., Crowe, S.F., 2004. Cognitive effects of
Aging 20, 833–839.

OO
long-term benzodiazepine use: a meta-analysis. CNS Drugs 18, 37–48.
Pariente, A., Dartigues, J.-F., Benichou, J., Letenneur, L., Moore, N., Fourrier-Réglat, A.,
Belleville, G., 2010. Mortality hazard associated with anxiolytic and hypnotic drug use in
2008. Benzodiazepines and injurious falls in community dwelling elders. Drugs Aging
the National Population Health Survey. Can. J. Psychiatry 55, 558–567. https://doi.
25, 61–70.
org/10.1177/070674371005500904.
Passaro, A., Volpato, S., Romagnoni, F., Manzoli, N., Zuliani, G., Fellin, R., 2000. Ben-
Bossini, L., Coluccia, A., Casolaro, I., Benbow, J., Amodeo, G., De Giorgi, R., Fagiolini, A.,
zodiazepines with different half-life and falling in a hospitalized population: the
2015. Off-label trazodone prescription: evidence, benefits and risks. Curr. Pharm. Des.
GIFA study. Gruppo Italiano di Farmacovigilanza nell'Anziano. J. Clin. Epidemiol. 53,
21, 3343–3351.
1222–1229.
Cascade, E., Kalali, A.H., 2008. Use of benzodiazepines in the treatment of anxiety. Psy-
Petitjean, S., Ladewig, D., Meier, C.R., Amrein, R., Wiesbeck, G.A., 2007. Benzodiazepine
chiatry (Edgmont) 5, 21–22.
prescribing to the Swiss adult population: results from a national survey of commu-
Charlson, F., Degenhardt, L., McLaren, J., Hall, W., Lynskey, M., 2009. A systematic review

PR
nity pharmacies. Int. Clin. Psychopharmacol. 22, 292–298. https://doi.org/10.1097/
of research examining benzodiazepine-related mortality. Pharmacoepidemiol. Drug
YIC.0b013e328105e0f2.
Saf. 18, 93–103. https://doi.org/10.1002/pds.1694.
Pfeiffer, P.N., Ganoczy, D., Zivin, K., Valenstein, M., 2011. Benzodiazepines and ade-
Cloos, J.-M., Ferreira, V., 2009. Current use of benzodiazepines in anxiety disorders. Curr.
quacy of initial antidepressant treatment for depression. J. Clin. Psychopharmacol. 31,
Opin. Psychiatry 22, 90–95. https://doi.org/10.1097/YCO.0b013e32831a473d.
360–364. https://doi.org/10.1097/JCP.0b013e318217b4c4.
Cumming, R.G., Le Couteur, D.G., 2003. Benzodiazepines and risk of hip fractures in older
Schroeck, J.L., Ford, J., Conway, E.L., Kurtzhalts, K.E., Gee, M.E., Vollmer, K.A., Mergen-
people: a review of the evidence. CNS Drugs 17, 825–837.
hagen, K.A., 2016. Review of safety and efficacy of sleep medicines in older adults.
Fisher, J., Sanyal, C., Frail, D., Sketris, I., 2012. The intended and unintended conse-
Clin. Ther. 38, 2340–2372. https://doi.org/10.1016/j.clinthera.2016.09.010.
quences of benzodiazepine monitoring programmes: a review of the literature. J. Clin.
Shorr, R.I., Robin, D.W., 1994. Rational use of benzodiazepines in the elderly. Drugs Aging
Pharm. Ther. 37, 7–21. https://doi.org/10.1111/j.1365-2710.2011.01245.x.
4, 9–20.
Holbrook, A.M., Crowther, R., Lotter, A., Cheng, C., King, D., 2000. Meta-analysis of ben-
Simon, G.E., VonKorff, M., Barlow, W., Pabiniak, C., Wagner, E., 1996. Predictors of

D
zodiazepine use in the treatment of insomnia. CMAJ 162, 225–233.
chronic benzodiazepine use in a health maintenance organization sample. J. Clin. Epi-
Iaboni, A., Bronskill, S.E., Reynolds, K.B., Wang, X., Rochon, P.A., Herrmann, N., Flint,
demiol. 49, 1067–1073. https://doi.org/10.1016/0895-4356(96)00139-4.
A., 2016. Changing pattern of sedative use in older adults: a population-based cohort
Smink, B.E., Egberts, A.C.G., Lusthof, K.J., Uges, D.R.A., de Gier, J.J., 2010. The relation-
study. Drugs Aging 33, 523–533. https://doi.org/10.1007/s40266-016-0380-3.
ship between benzodiazepine use and traffic accidents: a systematic literature review.
Inada, T., Inagaki, A., 2015. Psychotropic dose equivalence in Japan. Psychiatry Clin. Neu-
TE
CNS Drugs 24, 639–653. https://doi.org/10.2165/11533170-000000000-00000.
rosci 69, 440–447. https://doi.org/10.1111/pcn.12275.
Taipale, H., Koponen, M., Tanskanen, A., Tolppanen, A.-M., Tiihonen, J., Hartikainen,
Isacson, D., 1997. Long-term benzodiazepine use: factors of importance and the devel-
S., 2015. Long-term use of benzodiazepines and related drugs among commu-
opment of individual use patterns over time - a 13-year follow-up in a Swedish
nity-dwelling individuals with and without Alzheimer's disease. Int. Clin. Psychophar-
community. Soc. Sci. Med. 44, 1871–1880. https://doi.org/10.1016/
macol. 30, 202–208. https://doi.org/10.1097/YIC.0000000000000080.
S0277-9536(96)00296-1.
Takeshima, N., Ogawa, Y., Hayasaka, Y., Furukawa, T.A., 2016. Continuation and dis-
Ishigooka, J., Sugiyama, T., Suzuki, M., Kobayashi, K., Takeuchi, H., Murasaki, M., 1998.
continuation of benzodiazepine prescriptions: a cohort study based on a large claims
Survival analytic approach to long-term prescription of benzodiazepine hypnotics.
database in Japan. Psychiatry Res. 237, 201–207. https://doi.org/10.1016/j.psychres.
EC

Psychiatry Clin. Neurosci. 52, 541–545. https://doi.org/10.1046/j.1440-1819.1998.

2016.01.040.
00422.x.
Tiihonen, J., Suokas, J.T., Suvisaari, J.M., Haukka, J., Korhonen, P., 2012. Polypharmacy
Kato, K., Hirai, K., Nishiyama, K., Uchikawa, O., Fukatsu, K., Ohkawa, S., Kawamata, Y.,
with antipsychotics, antidepressants, or benzodiazepines and mortality in schizophre-
Hinuma, S., Miyamoto, M., 2005. Neurochemical properties of ramelteon (TAK-375),
nia. Arch. Gen. Psychiatry 69, 476–483. https://doi.org/10.1001/archgenpsychiatry.
a selective MT1/MT2 receptor agonist. Neuropharmacology 48, 301–310. https://doi.
2011.1532.
org/10.1016/j.neuropharm.2004.09.007.
Urru, S.A.M., Pasina, L., Minghetti, P., Giua, C., 2015. Role of community pharmacists in
Kjosavik, S.R., Ruths, S., Hunskaar, S., 2012. Use of addictive anxiolytics and hypnotics in
the detection of potentially inappropriate benzodiazepines prescriptions for insomnia.
a national cohort of incident users in Norway. Eur. J. Clin. Pharmacol. 68, 311–319.
Int. J. Clin. Pharm. 37, 1004–1008. https://doi.org/10.1007/s11096-015-0166-4.
https://doi.org/10.1007/s00228-011-1124-2.
Veronese, A., Garatti, M., Cipriani, A., Barbui, C., 2007. Benzodiazepine use in the real
RR

Kripke, D.F., Langer, R.D., Kline, L.E., 2012. Hypnotics' association with mortality or
world of psychiatric practice: low-dose, long-term drug taking and low rates of treat-
cancer: a matched cohort study. BMJ Open 2, e000850. https://doi.org/10.1136/
ment discontinuation. Eur. J. Clin. Pharmacol. 63, 867–873. https://doi.org/10.1007/
bmjopen-2012-000850.
s00228-007-0341-1.
Lader, M., 2011. Benzodiazepines revisited–will we ever learn?. Addiction 106,
Vicens, C., Bejarano, F., Sempere, E., Mateu, C., Fiol, F., Socias, I., Aragonès, E., Palop,
2086–2109. https://doi.org/10.1111/j.1360-0443.2011.03563.x.
V., Beltran, J.L., Piñol, J.L., Lera, G., Folch, S., Mengual, M., Basora, J., Esteva, M.,
Lader, M.H., Petursson, H., 1981. Benzodiazepine derivatives–side effects and dangers.
Llobera, J., Roca, M., Gili, M., Leiva, A., 2014. Comparative efficacy of two interven-
Biol. Psychiatry 16, 1195–1201.
tions to discontinue long-term benzodiazepine use: cluster randomised controlled trial
Lagnaoui, R., Depont, F., Fourrier, A., Abouelfath, A., Bégaud, B., Verdoux, H., Moore, N.,
in primary care. Br. J. Psychiatry 204, 471–479. https://doi.org/10.1192/bjp.bp.113.
2004. Patterns and correlates of benzodiazepine use in the French general population.
CO

134650.
Eur. J. Clin. Pharmacol. 60, 523–529. https://doi.org/10.1007/s00228-004-0808-2.
Wu, C.-S., Lin, Y.-J., Liu, S.-K., 2011. Benzodiazepine use among patients with schizophre-
Luijendijk, H.J., Tiemeier, H., Hofman, A., Heeringa, J., Stricker, B.H.C., 2008. Determi-
nia in Taiwan: a nationwide population-based survey. Psychiatr. Serv. 62, 908–914.
nants of chronic benzodiazepine use in the elderly: a longitudinal study. Br. J. Clin.
https://doi.org/10.1176/ps.62.8.pss6208_0908.
Pharmacol. 65, 593–599. https://doi.org/10.1111/j.1365-2125.2007.03060.x.
Manthey, L., Giltay, E.J., van Veen, T., Neven, A.K., Zitman, F.G., Penninx, B.W.J.H.,
2011. Determinants of initiated and continued benzodiazepine use in The Nether
UN