Regular Article Psychother Psychosom 2015;84:273–283 DOI: 10.1159/000434755

Received: September 16, 2014 Accepted after revision: June 2, 2015 Published online: August 6, 2015

Hypnotics and Triazolobenzodiazepines – Best Predictors of High-Dose Benzodiazepine Use: Results from the Luxembourg National Health Insurance Registry Jean-Marc Cloos a, b Valéry Bocquet c Isabelle Rolland-Portal d Paul Koch d Guy Chouinard e–h a

Department of Addictology, Hôpitaux Robert Schuman – ZithaKlinik, b Jugend- an Drogenhëllef, c Competence Centre in Methodology and Statistics, Luxembourg Institute of Health, d Contrôle Médical de la Sécurité Sociale, and e Pain Clinic, ZithaKlinik, Luxembourg, Grand Duchy of Luxembourg; f Clinical Pharmacology and Toxicology Program, McGill University, and g University Mental Health Institute of Montreal, Fernand Seguin Research Center, Montreal, Que., Canada; h Hôpital Saint-Antoine, Centre de recherches psychanalyse, médecine et société, Paris Diderot University, Paris, France

Abstract Background: Benzodiazepines are not all the same concerning their risk of high-dose use. Methods: We studied benzodiazepine use from the Luxembourg national records of all insured. We calculated the 12-year prevalence from 1995 to 2007. Benzodiazepine users were divided into 3 groups, short-term with no longer than 3-month intake, intermediate with multiple administration with at least a 1-year interruption, and continuous who never stopped. A high-dose user (HDU) was defined as a patient who received a higher dose than the yearly maximum usual therapeutic dose. Results: An average of 16.0% of the adult insured population received at least 1 benzodiazepine annually, 42.9% were older than 50, 55.9% were women, and 5.4% were HDUs. We found that 32.6% were short-term users, 49.0% intermediate and 18.4% continuous. Compared to diazepam, hypnotics had higher risks for high-dose use in at least 1 age group at first-benzodiazepine intake, the risks being greater in elder-

© 2015 S. Karger AG, Basel 0033–3190/15/0845–0273$39.50/0 E-Mail [email protected]

ly subjects and women, the highest risks being with triazolam (adjusted odds ratio = 215.85; 95% confidence interval = 133.75–348.35) in the 69- to 105-year-old group at first-benzodiazepine intake. Anxiolytics had a low risk except for alprazolam and prazepam in the 69- to 105-year-old group at first-benzodiazepine intake, clonazepam and clobazam had the lowest risk in 18- to 43-year-olds at first-benzodiazepine intake. Alprazolam had dispensed volumes increased by threefold over the 12-year period. Conclusion: All hypnotics had higher risks for high-dose use compared to diazepam in continuous users. Two anxiolytics, clonazepam and clobazam, had the lowest risks. Hypnotics and the triazolobenzodiazepines alprazolam and triazolam were most problematic. Elderly subjects and women are at greater risks. © 2015 S. Karger AG, Basel


For decades, benzodiazepines have been recommended as the standard treatment of anxiety and insomnia [1– 4]. Benzodiazepine withdrawal, rebound, overuse and abuse often limit their use in clinical practice [5]. ChouiDr. Jean-Marc Cloos Department of Addictology, Hôpitaux Robert Schuman ZithaKlinik, 36, rue Ste Zithe LU–2763 Luxembourg (Grand Duchy of Luxembourg) E-Mail jean-marc.cloos @

Downloaded by: G. Fava - 28048 - 8/7/2015 9:24:29 PM

Key Words Benzodiazepine long-term use · Benzodiazepine high-dose use · Benzodiazepine prevalence · Hypnotics · Anxiolytics

Methods Study Design We performed an analysis of benzodiazepine prescription claim data from 1995 to 2006. Using Anatomical Therapeutic Chemical classification (ATC) codes, we identified claims for all 24 benzodiazepines available for prescription during the study period. We classified persons into 3 groups according to their individual benzodiazepine use. We calculated the prevalence of long-term and high-dose use, and identified gender and age differences. We analysed the risks of high-dose use for each benzodiazepine.


Psychother Psychosom 2015;84:273–283 DOI: 10.1159/000434755

Data Sources A 12-year study from January 1995 to January 2007 included all residents aged 18 and older insured in the national health system (n = 387,862 in 1995; n = 449,972 in 2006). The cohort, from the national compulsory health insurance, represented approximately 95% of the population in a specific year (n = 411,600 in 1995; n = 476,200 in 2006). Outpatient prescription reimbursements from all practicing physicians (n = 771 in 1995 and n = 1,074 in 2006) were examined. Data confidentiality was maintained throughout the study. 214,170 persons were identified as having received at least 1 benzodiazepine dispensed within the study period. Ethical permissions from the Ministry of Social Security were obtained. Measures The cohort was divided into 3 groups according to benzodiazepine exposure: (1) a short-term-use group (G1) including patients who received, throughout the 12-year study period, either a one-time benzodiazepine for less than 3 months or had a total length of dispensing of less than 3 months [19]; (2) an intermittentuse group (G2) including patients who received several administrations but with at least a 1-year interruption; (3) a continuoususe group (G3) including patients who never stopped taking benzodiazepines once prescribed, either until the end of the study or until death. In addition, benzodiazepine users were divided into high-dose and non-high-dose users. We defined a high-dose user (HDU) as a person who received at least a higher dose than the yearly maximum usual therapeutic dose (table 1). In cases of multiple benzodiazepines, we divided yearly doses by the yearly equivalent maximum usual therapeutic dose of each benzodiazepine (table 1) and took the sum of the fractions (fig. 1). If the sum was higher than 1, we considered the person as an HDU for that year. Non-HDU is defined as not meeting HDU definitions. Available Benzodiazepines Table 1 lists the 24 benzodiazepines approved for medical use. Of those, 11 are registered in the USA. Benzodiazepines were identified by Anatomical Therapeutic Chemical codes (N03AE, N05BA and N05CD) and converted into defined daily dose (DDD). Statistical Analysis Patient visits were discretized in years to measure trends of other variables. Data were censored on January 1, 2007. Univariate statistics were used for population characteristics, 2-tailed 2-sample t tests for comparisons between HDUs and nonHDUs and Fisher exact tests for proportion comparisons between men and women. Total benzodiazepine dispensing prevalence was calculated throughout the study. Before analyses, continuous variable means were centred to interpret them as meaningful values. In accordance with a repeated measure design, a generalized estimating equation (GEE) regression [20] was used to determine attributes associated with yearly HDU status. The method person-years was used, since it gives the sum of individual time units that persons have been exposed to a benzodiazepine, and a person can enter or leave the cohort at any time during a given year. Groups (G1, G2 and G3) were significantly different for HDUs (p < 0.01). Only the continuous-use G3 group was studied since it was the most homogeneous group of sufficient size and

Cloos/Bocquet/Rolland-Portal/Koch/ Chouinard

Downloaded by: G. Fava - 28048 - 8/7/2015 9:24:29 PM

nard [5] proposed major differences between benzodiazepines in terms of rebound syndrome, potency, risk of abuse and pharmacokinetics. Despite these limitations, benzodiazepines remain widely prescribed as anxiolytics and hypnotics [6]. Recently, Fava [7] suggested revising the practice guidelines for the pharmacological treatment of patients with anxiety disorders, more specifically the use of antidepressants as first-line treatment of anxiety without depression as comorbidity. The critical reviews by Rickels [8] and Balon [9], and the meta-analysis by Offidani et al. [4] are in agreement with Fava’s recommendations [7]. Furthermore, the new selective serotonin reuptake inhibitor (SSRI) and serotonin norepinephrine reuptake inhibitor (SNRI) classification of withdrawal syndromes, which includes persistent postwithdrawal disorders [10], and a recent systematic review of SSRI withdrawal symptoms [11] provide additional evidence for a revision of SSRI and SNRI guidelines as first-line pharmacological treatment of anxiety disorders without depression as comorbidity [12]. The annual prevalence of prescribed benzodiazepines in the general population varies greatly from 2.2 to 17.6% [13–18]. The 2:1 female/male ratio of benzodiazepine intake appears constant [13, 18]. The Danish prescription registry study including 4,614,807 subjects found a female/male ratio of benzodiazepine intake (expressed in rates per 100 person-years) of 17.0/8.6 in 1997 and 14.7/7.9 in 2008 [16]. The Danish study also confirmed the highest prevalence of benzodiazepine intake in elderly subjects, 25.6% aged 75–85 years and 30.3% over 85 years. The US 2008 IMS LifeLink annual prescription prevalence also increased with age [18]. The aim of this study was to estimate the prevalence and patterns of benzodiazepine use from the Luxembourg National Health Insurance Registry, assess the risks of high-dose use and identify which benzodiazepines show the highest risk for continuous and high-dose use.

Table 1. Characteristics of the 24 benzodiazepines available for prescription Drug


Anxiolytics Alprazolam1 Bromazepam Clobazam1 Clonazepam1 Clorazepate1 Clotiazepam Cloxazolam Diazepam1 Ketazolam Loflazepate Lorazepam1 Nordiazepam Oxazepam1 Prazepam Tetrazepam

0.5 6 20 0.5 – 2 15 – 20 5 – 10 1–2 10 15 – 30 1–2 1–2 10 – 20 20 10 – 20 20 – 50

1 10 20 8 20 15 9 10 30 2 2.5 15 50 30 100

4 18 30 8 60 15 4 40 60 3 6 15 120 60 150

Hypnotics Brotizolam 0.25 – 0.5 Flunitrazepam 0.5 – 1 15 – 30 Flurazepam1 Loprazolam 1–2 Lormetazepam 1–2 Nitrazepam 5 – 10 20 Temazepam1 0.25 – 0.5 Triazolam1

0.25 1 30 1 1 5 20 0.25

0.25 1 30 1 2 10 20 0.25



Other Midazolam1 (IV)

5 – 7.5


UTDmax Potency

Onset of action (after oral dose)

Length of action Active metabolite with a (including metabolites) half-life >48 h

high high medium high medium medium medium medium medium medium high medium low medium low

rapid (15 – 30 min) rapid rapid intermediate intermediate rapid slow (>60 min) rapid slow rapid rapid rapid slow slow rapid

short (6 – 24 h) short long medium (24 – 48 h) long short long long long long short long short long medium

– – norclobazam (50 – 100) – nordiazepam (36 – 200) – delorazepam (70 – 100) nordiazepam (36 – 200) nordiazepam (36 – 200) descarboxyloflazepate (70 – 120) – – – nordiazepam (36 – 200) –

high high medium high medium medium low high

rapid rapid rapid intermediate rapid rapid intermediate rapid

short short/medium long medium short medium short very short

– – desalkylflurazepam (50 – 160) – – – – –


ultrarapid (69 years at first intake, all hypnotics have a significantly higher risk for high-dose use (excluding temazepam with too few persons, n = 7); the nitrazepam ORa to become HDU is 1.80 (95% CI = 1.04–3.12), that of flurazepam Psychother Psychosom 2015;84:273–283 DOI: 10.1159/000434755


Downloaded by: G. Fava - 28048 - 8/7/2015 9:24:29 PM

year study, which could be calculated since we could follow each person through the 12-year period. The 0.16% annual incidence of new cases of HDUs was stable over the 12 years studied, averaging 0.8% of all insured.

Table 4. National studies reporting DDD per 1,000 inhabitants per day and/or annual prevalence of benzodiaz-

epine prescription use Country

Number of persons



Prevalence year, %

Portugal [14]

national national 411,600 476,200 national national national national 2,803 national national national 520,000 national national 4,614,807 national 874 (national sample) 919 (national sample) national

84.1 81.6 76.8 78.9 51.2 73.8 67.8 57.3 49.4 50 47.5 49.5 43.3 27.6 22.9 n.a. n.a. n.a. n.a. n.a.

1999 2003 1995 2006 2000 2011 2000 2010 1992 – 1993 1995 – 2003 2003 2011 2002 2007 2011 1997 2008 1992 2002 2008

n.a. n.a. 14.4 16.4 n.a. n.a. n.a. n.a. 8.6 n.a. n.a. n.a. 9.1 n.a. n.a. 12.0 10.7 7.8 7.9 5.2

Luxembourg Spain [17] France [32] Italy [35] Italy [36] Italy [34] Switzerland [37] Norway [38] Denmark [16] Netherlands [15] USA [18]

n.a. = Not available.


Psychother Psychosom 2015;84:273–283 DOI: 10.1159/000434755


In this 12-year National Registry study, we found 16% of the adult insured population received at least 1 benzodiazepine annually, 14.4% of the population having at least 1 benzodiazepine in 1995, and 16.4% in 2006. Most persons (81.6%) who received a benzodiazepine were short-term or intermittent users. 32.6% were patients who received a benzodiazepine for no longer than 3 months (G1 short-term group) and 49% had intermittent use with at least a 1-year interruption (G2 intermittent group). 18.4% of persons who received benzodiazepines were continuous users (G3 group). In the first 2 groups (short-term and intermittent), the proportion of persons with high-dose dispensing is small, in the G1 group 0.03%, and in the G2 group 3.6%. In contrast, 20.3% (n = 7,967) of patients, who received benzodiazepine dispensing continuously (G3 group), were also HDUs. If we define problematic use as either continuous or high-dose use, percentages of continuous users in the general population would be 2.7–3.0% and high-dose use 0.8–0.9%. Continuous use is not necessarily associated with dose escalation. In our study, the annual incidence Cloos/Bocquet/Rolland-Portal/Koch/ Chouinard

Downloaded by: G. Fava - 28048 - 8/7/2015 9:24:29 PM

5.95 (95% CI = 2.16–16.36), of lormetazepam 9.95 (95% CI = 7.09–13.98), of loprazolam 28.75 (95% CI = 19.66– 42.03), of flunitrazepam 50.62 (95% CI = 34.02–75.32), of brotizolam 181.40 (95% CI = 121.73–270.32) and finally of triazolam 215.85 (95% CI = 133.75–348.35), triazolam having the highest risk. In all age groups, hypnotics had an increased risk to induce high-dose use (see data in italics of table 3). In contrast, other benzodiazepines have significantly low ORa values to become an HDU, especially in the young age group (see bold data in table 3). As can be seen in the first column of table 3, at the age of 18–43 at first intake, the prazepam ORa to become an HDU is 0.66 (95% CI = 0.46–0.94), that of tetrazepam 0.59 (95% CI = 0.42–0.82), of clotiazepam 0.54 (95% CI = 0.38–0.76), of nordiazepam 0.50 (95% CI = 0.36–0.71), of clonazepam 0.42 (95% CI = 0.28–0.65) and of clobazam 0.26 (95% CI = 0.13–0.52), clonazepam and clobazam having the lowest risk. Only 2 anxiolytics, alprazolam (ORa = 1.67; 95% CI = 1.17–2.40) and prazepam (ORa = 1.77; 95% CI = 1.11–2.81), were found to have significantly increased risks for high-dose use, and this was in the >69-year-old group at first-benzodiazepine intake.

pine intake, we found that all 7 benzodiazepines registered as hypnotics (nitrazepam, flurazepam, lormetazepam, loprazolam, flunitrazepam, brotizolam and triazolam) were at significantly higher risk for high-dose use (temazepam excluded, n = 7). In contrast, 2 anxiolytic benzodiazepines, clobazam and clonazepam, had the lowest risk for high-dose use in the 18- to 43-year-old group at first-benzodiazepine intake. Significant risk factors for high-dose use escalations are older age at firstbenzodiazepine dispensing, the prescriber’s choice of first-time molecule and continuous-benzodiazepine use [29]. In our study, all benzodiazepines approved as hypnotics were identified to present an increased risk of highdose use, and 5 short- to medium-acting drugs (triazolam, brotizolam, flunitrazepam, loprazolam and lormetazepam) were problematic in all age groups. It is worth noting that flunitrazepam showed a rapid decline of annual dispensed volume after Luxembourg changed prescription regulations in 2000 and classified it as a narcotic [30]. The Luxembourg regulation was efficacious and explains this decline from 7.7 DDD in 1995 to 0.04 DDD in 2007. Thus, our study shows clearly that patients suffering from insomnia and treated with hypnotic benzodiazepines are at greater risks to become HDUs. Alternative cognitive behavioral therapies are preferred by some authors [31] to hypnotic medications in the treatment of insomnia, but the functional impairment caused by insomnia needs to be recognized and insomnia be corrected. Six benzodiazepines registered as anxiolytics had lower risks of high-dose use in the 18- to 43-year-old age group at first-benzodiazepine dispensing. All benzodiazepines registered as hypnotics (table 1) presented a significantly increased risk of high-dose use in at least 1 age group, compared to only 2 benzodiazepines registered as anxiolytics (alprazolam and prazepam) in the 68- to 105-year-old group. Physicians should be cautious when prescribing hypnotics for insomnia since they are associated with continuous and high-dose use. Precautions should also apply to populations at risk, elderly and women, and to drugs registered as hypnotics and prescribed for insomnia. Our study confirms the predominant use of benzodiazepines in women and the elderly. It has detected major differences in the risk of developing high-dose use with benzodiazepine hypnotics. We found an average of 80.1 DDD/1,000 inhabitants/ day for the 12-year period, representing 0.8 mg of an equivalent dose of diazepam per inhabitant per day. As

High-Dose Benzodiazepine Use

Psychother Psychosom 2015;84:273–283 DOI: 10.1159/000434755


Downloaded by: G. Fava - 28048 - 8/7/2015 9:24:29 PM

of high-dose use remained stable at 0.16%. Soumerai et al. [23] found an incidence of 1.6% of escalation to high-dose use, in 2,440 patients over a 39-month period, while Martinez-Cano et al. [24] conclude that it is the combination of high potency and short elimination half-life that is critical for the occurrence of benzodiazepine high-dose use. Our results would be in agreement with this conclusion, but also provide evidence that the chemical structure of the benzodiazepine and the hypnotic properties are important. While the female/male ratio is 48.9% in the G1 group and 57.6% in the G2 group, 63.6% continuous users are women (63.0% non-HDU; 65.8% HDU). Higher uses by women are seen in the Danish National study [16] and others [13], reaching 72.8% in a study by Neutel [25]. Women appear particularly vulnerable to become continuous benzodiazepine prescription users. In our study, we found significantly more women to be HDUs compared to men, and this was even more so in the continuous benzodiazepine group (table 2). These findings show that large studies are necessary to examine continuous users and those persons at risk, since they may represent 20% of persons dispensed benzodiazepines. The elderly also appear most vulnerable to become continuous benzodiazepine prescription users. Increased sensitivity of this population to the effects of psychotropic medications including benzodiazepines has been reported [26–28]. In the Danish National registry database study of 1997–2008 [16], which included 4,614,807 subjects, 32% (1,480,017) received at least 1 benzodiazepine. There was a significant decrease in benzodiazepine prescriptions from 12.0% (1997) to 10.7% (2008) in all age groups, but the 2008 prevalence of benzodiazepine use remained high in the elderly and increased significantly with age: 19.0% in the 65- to 75-year-olds, 25.6% in the 75- to 84-year-olds and 30.3% in those older than 85 [16]. While the Danish prevalence of benzodiazepine use decreased from 1997 to 2008, the use of zolpidem/zopiclone increased to finally start to decrease in 2008. While the Danish study [16] shows clearly the increase in benzodiazepine use with age, our study found in addition that the risk to become an HDU or continuous user increased with age at first-benzodiazepine intake. We found that 37.9% of long-term users are aged 69 and more, and 19.4% of these >69-year-old continuous users are HDUs. Shorr and Robin [26] reported comparable results with 47.8% of long-term users aged 65 and more, and 12.7% of these >65-year-old continuous users used high doses. In the elderly aged >69 at first-benzodiaze-


Psychother Psychosom 2015;84:273–283 DOI: 10.1159/000434755

tries [38], hydroxyzine in France [33] and SSRI/SNRI in Italy [39]. Benzodiazepine actions are due to their potentiation of γ-aminobutyric acid (GABA) inhibition, through their indirect effects on the ionotropic GABAA receptors, since they require GABA to have an action through the benzodiazepine receptor. The pharmacological effects depend on the ligands, which bind to the benzodiazepine receptor. They have 4 main CNS effects, sedation, hypnosis, anxiolysis and anterograde amnesia, in addition to their muscle-relaxant and anticonvulsant properties [40]. Three benzodiazepine agonists, midazolam, diazepam and lorazepam, are given for induction of anaesthesia due to their sedation and anterograde amnesia effects [40]. The present study is concerned with the benzodiazepine agonists registered as hypnotics and anxiolytics. Subtypes of the GABAA receptor have been identified [41]. The α1 GABAA receptors would be responsible for sedation and anterograde amnesia, while anxiolysis and muscle relaxation would be due to their benzodiazepine agonist effects on α2 GABAA receptors [40]. The marked differences between anxiolytics and hypnotics regarding the risks for continuous and highdose use found in our study could be explained by their differences in ligand binding with high affinity to the α1subtype GABAA receptors for hypnotics versus high affinities to α2- and α3-subtypes and low-affinity α1-subtype for anxiolytics [42]. The benzodiazepine associated with the greatest risks of high-dose use is the triazolobenzodiazepine hypnotic triazolam (ORa = 215.85; 95% CI = 133.75–348.35), in the >69-year-olds at first-benzodiazepine intake, and with the highest risks of high-dose use in most age groups. Triazolam became a controversial drug, soon after its introduction. Hong Kong restricted triazolam in 1991 and extended these restrictions to all benzodiazepines in 1992 [43]. Triazolam was first suspended in the Netherlands (1979) and later definitively in the UK (1993) [44]. In the USA, dose restrictions of triazolam were imposed [45]. Triazolam is still marketed in the USA, with a dose labelling revised in 1991 [45], as it is in several countries. The other triazolobenzodiazepine with a molecular structure similar to triazolam included in our study is the anxiolytic alprazolam, which is also problematic. All anxiolytics had a low risk of high-dose use except alprazolam and prazepam in the 69- to 105-year-old age group at first-benzodiazepine intake. Alprazolam had dispensed volumes increased by threefold over the 12-year period. In our study, alprazolam dispensing tripled DDD from 1995 (3.8 DDD/1,000 inhabitants/day) to 2006 (10.4 Cloos/Bocquet/Rolland-Portal/Koch/ Chouinard

Downloaded by: G. Fava - 28048 - 8/7/2015 9:24:29 PM

can be seen in table 4 for Luxembourg, a small increase occurs from 76.8 DDD/1,000 inhabitants/day in 1995 to 78.9 DDD in 2006, but the DDD data remain overall high. Compared to other national studies (table 4), Portugal [14] showed the highest annual DDD rates, with an average of 85.8 DDD/1,000 inhabitants/day for a 5-year period (1999–2003). The use of benzodiazepines increased by 44.1% in Spain [17] between 2000 (51.2 DDD/1,000 inhabitants/day) and 2011 (73.8 DDD/1,000 inhabitants/ day). It decreased by 15.5% in France [32] from 2000 to 2010, with an average of 61.2 DDD/1,000 inhabitants/day in the 11-year period, and one explanation may be the switch to hydroxyzine [33]. Recently, Beuscart et al. [33] analysed all prescriptions dispensed in community pharmacies to patients aged 75 and older between January 1 and March 31, 2012, in the Nord-Pas-de-Calais Region. They reported that 32.6% (n = 67,863) of patients received at least 1 anticholinergic drug and found that the H1 antihistamine anticholinergic non-selective inverse agonist hydroxyzine (n = 10,792) ranked first among the anticholinergic drugs as the most potent and the most often prescribed of potentially inappropriate medications. In Italy [34], DDD increases from 47.5 DDD/1,000 inhabitants/ day between 2003 and 2011 to 49.5 DDD/1,000 inhabitants/day, but still remains lower than in Luxembourg. However for Italy, it should be noted that benzodiazepines are no more reimbursed by the National Care, which may lead to unchanged DDD over time since fewer new patients are prescribed benzodiazepines to avoid patients to pay. In an earlier national Italian study (1992– 1993), the prevalence of benzodiazepine use was found to be 8.6% in 1993 and doses to be 49.4 DDD/1,000 inhabitants/day in 1991; 28 benzodiazepines were available (60 brands and 133 products) [35]. Another Italian national study found benzodiazepine use to be stable with 50 DDD/1,000 inhabitants/day from 1995 to 2003 as compared to the marked increase in DDDs for SSRI antidepressants [36]. In Switzerland (table 4) with 17 different benzodiazepines available in 2002, 43.3 DDD/1,000 inhabitants/day are reported with a 6-month prevalence of 9.1% [37]. From examining these studies (table  4), we explain the higher DDD/1,000 inhabitants/day in Luxembourg compared to some countries such as Scandinavian countries, France and Italy, by the non-substitution of benzodiazepines by other drugs, since benzodiazepines did not get a bad press in Luxembourg compared to those three countries. Z drugs (zopiclone) appear to have replaced in part benzodiazepines in Scandinavian coun-

DDD/1,000 inhabitants/day). In Portugal, alprazolam was the second most prescribed benzodiazepine in 1999 with 15.8 DDD/1,000 inhabitants/day, and became the most prescribed in 2003 with 21.7 DDD/1,000 inhabitants/day) [14]. In Spain, alprazolam remains the second most prescribed [17], with a continuous increase from 11.4 DDD/1,000 inhabitants/day in 2000 to 15.7 DDD/1,000 inhabitants/day in 2011. In the USA, alprazolam-related emergency department visits involving non-medical use doubled between 2005 (57,419 visits) and 2010 (124,902 visits) according to the Dawn Report from the Substance Abuse and Mental Health Services Administration (SAMHSA) [46]. One explanation of the problematic use of triazolobenzodiazepines is their specific pharmacology, which is different from diazepam and other benzodiazepines [47– 49]. In 1986, Turmel and De Montigny [47] showed a ‘heightened’ effect of adinazolam on the serotonin neurons of the rat forebrain, compared to diazepam. Later, alprazolam was shown to increase serotonin release, in the hippocampus as opposed to diazepam [48], and in the CA1 region of the hippocampus, while diazepam decreased it [49]. National health insurance registry studies have strengths and limitations. The strengths are high external validity including all benzodiazepine dispensing, large sample size with high statistical power and availability of contiguous data on persons across 12 years so that initial intake is most likely to be true; additional strengths of our Luxembourg national study are the small number of prescribing physicians, the homogeneity of the sample (one small country), the homogeneity of the database, and the national registry follow-up. The use of WHO-defined DDD has strengths and limitations. WHO-defined DDDs of benzodiazepines [22] have been used in all European national studies including the most recent ones since 1992 (table 4), thus permit comparison with other national studies. The validity of DDD benzodiazepine comparative doses has been shown by their use in national studies of benzodiazepine and has been consistent over time, but they are a compromise between different countries regarding benzodiazepine prescribing practice. Ashton’s diazepam equivalent doses [50] and the Teboul/Chouinard usual daily dose (North American doses prescribed) [51] have also been proposed, but none of these has been shown as satisfactory as WHO-defined DDD. In this 12-year study using WHO-defined DDD and person-years, the benzodiazepine quantity dispensed, the dispensing date and total time exposure were used to calculate the number of DDD

per 1,000 inhabitants per day for the total dispensing period. The high-dose use status was calculated in a similar manner to define an HDU as a person who received at least a higher dose than the yearly maximum usual therapeutic dose. This approach is justified since we studied only the continuous user group. Total benzodiazepine dispensing prevalence was calculated throughout the study. The method of person-years was used, since it gives the sum of individual time units that persons have been exposed to a benzodiazepine, a person can enter or leave the cohort at any time during a given year. Groups (short-term, intermittent and continuous) were significantly different for HDUs. Only the continuous-use G3 group was studied since it was the most homogeneous group of sufficient size and included the majority of HDUs. Our results are thus limited to continuous benzodiazepine users. Our study has other limitations: the prevalence may be different than the actual prevalence due to drug noncompliance, especially in the short-term users. In addition, we did not include dispensing of benzodiazepine receptor agonists. Instead, we focus on one chemical class and not identical physiological effects. We are aware that patients treated with benzodiazepine receptor agonists could be treated with benzodiazepines, and believe our findings with benzodiazepine hypnotics also apply to benzodiazepine receptor agonists registered as hypnotics such as zolpidem. The SAMHSA reported that the visits to emergency departments due to overmedication with zolpidem doubled from 2005–2006 to 2009–2010 in the USA [46], leading to a warning from the US Food and Drug Administration to decrease the recommended dose by 50% in women [52].

High-Dose Benzodiazepine Use

Psychother Psychosom 2015;84:273–283 DOI: 10.1159/000434755



Downloaded by: G. Fava - 28048 - 8/7/2015 9:24:29 PM

In this 12-year national registry study, we find benzodiazepine hypnotics to be associated with increased risks of high-dose use among continuous users, compared to diazepam. Women and the elderly are at increased risks for high-dose use, especially the group of 69- to 105-yearolds at first-benzodiazepine intake. In contrast, anxiolytic benzodiazepines are associated with significantly lower risks of high-dose use, the lowest risks being with clobazam and clonazepam in the 18- to 43-year-old age group at first-benzodiazepine intake. We also find the chemical structure of the triazolobenzodiazepines to be an additional risk factor for high-dose use among continuous users.

Our nationwide cohort study provides additional evidence that hypnotics and the triazolobenzodiazepines alprazolam and triazolam are the most problematic, since they are associated with continuous and high-dose use. In contrast, judicious use of anxiolytic benzodiazepines, in particular clobazam and clonazepam, would be associated with a lower risk of high-dose use.

Disclosure Statement Within the last 5 years, J.-M. Cloos has received unrelated speaker honoraria from GlaxoSmithKline and G. Chouinard from Otsuka Pharmaceutical. No competing interests exist for V. Bocquet, I. Rolland-Portal and P. Koch.



15 Sonnenberg CM, Bierman EJ, Deeg DJ, Comijs HC, van Tilburg W, Beekman AT: Tenyear trends in benzodiazepine use in the Dutch population. Soc Psychiatry Psychiatr Epidemiol 2012;47:293–301. 16 Holm E, Fosbol E, Pedersen H, Jensen TB, Nielsen M, Weeke P, Brinth L, Torp-Pedersen C: Benzodiazepine use in Denmark 1997– 2008. Eur Geriatr Med 2012;3:299–303. 17 Vicente Sánchez MP, Macías Saint-Gerons D, de la Fuente Honrubia C, González Bermejo D, Montero Corominas D, Catalá-López F: Evolución del uso de medicamentos ansiolíticos e hipnóticos en España durante el período 2000–2011. Rev Esp Salud Publica 2013; 87: 247–255. 18 Olfson M, King M, Schoenbaum M: Benzodiazepine use in the United States. JAMA Psychiatry 2015;72:136–142. 19 Zandstra SM, Furer JW, van de Lisdonk EH, Bor JH, Zitman FG, van Weel C: Differences in health status between long-term and shortterm benzodiazepine users. Br J Gen Pract 2002;52:805–808. 20 Liang KY, Zeger SL: Longitudinal data analysis using generalized linear models. Biometrika 1986;73:13–22. 21 Cui J, Qian G: Selection of working correlation structure and best model in GEE analyses of longitudinal data. Commun Stat-Simul C 2007;36:987–996. 22 World Health Organization: Guidelines for ATC Classification and DDD Assignment. Oslo, WHO Collaborating Centre for Drug Statistics Methodology, 2012. 23 Soumerai SB, Simoni-Wastila L, Singer C, Mah C, Gao X, Salzman C, Ross-Degnan D: Lack of relationship between long-term use of benzodiazepines and escalation to high dosages. Psychiatr Serv 2003;54:1006–1011. 24 Martínez-Cano H, Vela-Bueno A, De Iceta M, Pomalima R, Martínez-Gras I, Sobrino MP: Benzodiazepine types in high versus therapeutic dose dependence. Addiction 1996; 91: 1179–1186. 25 Neutel CI: The epidemiology of long-term benzodiazepine use. Int Rev Psychiatry 2005; 17:189–197. 26 Shorr RI, Robin DW: Rational use of benzodiazepines in the elderly. Drugs Aging 1994;4: 9–20.

Psychother Psychosom 2015;84:273–283 DOI: 10.1159/000434755

27 Madhusoodanan S, Bogunovic OJ: Safety of benzodiazepines in the geriatric population. Expert Opin Drug Saf 2004;3:485–493. 28 Schweizer E, Case WG, Rickels K: Benzodiazepine dependence and withdrawal in elderly patients. Am J Psychiatry 1989;146:529–531. 29 Tvete IF, Bjørner T, Aursnes IA, Skomedal T: A 3-year survey quantifying the risk of dose escalation of benzodiazepines and congeners to identify risk factors to aid doctors to more rational prescribing. BMJ Open 2013; 3:e003296. 30 Mémorial A: règlement grand-ducal du 14 janvier 2000 modifiant l’annexe du règlement grand-ducal modifié du 4 mars 1974 concernant certaines substances toxiques. Journal Officiel du Grand-Duché de Luxembourg 2000;5:165. 31 Buysse DJ: Insomnia. JAMA 2013; 309: 706– 716. 32 Cavalié P, Richard N: Rapport d’expertise: état des lieux de la consommation des benzodiazépines en France. Saint-Denis, Agence française de sécurité sanitaire des produits de santé, 2012. 33 Beuscart JB, Dupont C, Defebvre MM, Puisieux F: Potentially inappropriate medications (PIMs) and anticholinergic levels in the elderly: a population based study in a French region. Arch Gerontol Geriatr 2014; 59: 630– 635. 34 Gruppo di lavoro OsMed: L’uso dei farmaci in Italia: rapporto nazionale anno 2011. Rome, Il Pensiero Scientifico, 2012. 35 Magrini N, Vaccheri A, Parma E, D’Alessandro R, Bottoni A, Occhionero M, Montanaro N: Use of benzodiazepines in the Italian general population: prevalence, pattern of use and risk factors for use. Eur J Clin Pharmacol 1996;50:19–25. 36 Ciuna A, Andretta M, Corbari L, Levi D, Mirandola M, Sorio A, Barbui C: Are we going to increase the use of antidepressants up to that of benzodiazepines? Eur J Clin Pharmacol 2004;60:629–634. 37 Petitjean S, Ladewig D, Meier CR, Amrein R, Wiesbeck GA: Benzodiazepine prescribing to the Swiss adult population: results from a national survey of community pharmacies. Int Clin Psychopharmacol 2007;22:292–298.

Cloos/Bocquet/Rolland-Portal/Koch/ Chouinard

Downloaded by: G. Fava - 28048 - 8/7/2015 9:24:29 PM

1 Rosenbaum JF: The drug treatment of anxiety. N Engl J Med 1982;306:401–404. 2 Shader RI, Greenblatt DJ: Use of benzodiazepines in anxiety disorders. N Engl J Med 1993;328:1398–1405. 3 Smith MT, Perlis ML, Park A, Smith MS, Pennington J, Giles DE, Buysse DJ: Comparative meta-analysis of pharmacotherapy and behavior therapy for persistent insomnia. Am J Psychiatry 2002;159:5–11. 4 Offidani E, Guidi J, Tomba E, Fava GA: Efficacy and tolerability of benzodiazepines versus antidepressants in anxiety disorders: a systematic review and meta-analysis. Psychother Psychosom 2013;82:355–362. 5 Chouinard G: Issues in the clinical use of benzodiazepines: potency, withdrawal, and rebound. J Clin Psychiatry 2004;65(suppl 5):7– 12. 6 Silber MH: Clinical practice. Chronic insomnia. N Engl J Med 2005;353:803–810. 7 Fava GA: Rational use of antidepressant drugs. Psychother Psychosom 2014; 83: 197– 204. 8 Rickels K: Should benzodiazepines be replaced by antidepressants in the treatment of anxiety disorders? Fact or fiction? Psychother Psychosom 2013;82:351–352. 9 Balon R: Benzodiazepines revisited. Psychother Psychosom 2013;82:353–354. 10 Chouinard G, Chouinard VA: New classification of selective serotonin reuptake inhibitor withdrawal. Psychother Psychosom 2015; 84: 63–71. 11 Fava GA, Gatti A, Belaise C, Guidi J, Offidani E: Withdrawal symptoms after selective serotonin reuptake inhibitor discontinuation: a systematic review. Psychother Psychosom 2015;84:72–81. 12 Fava GA, Balon R, Rickels K: Benzodiazepines in anxiety disorders. JAMA Psychiatry 2015;72:733–734. 13 Zandstra SM, Furer JW, van de Lisdonk EH, van’t Hof M, Bor JH, van Weel C, Zitman FG: Different study criteria affect the prevalence of benzodiazepine use. Soc Psychiatry Psychiatr Epidemiol 2002;37:139–144. 14 Furtado C, Teixeira I: Utilização de benzodiazepinas em Portugal continental (1999– 2003). Acta Med Port 2006;19:239–246.

High-Dose Benzodiazepine Use

43 Lee KK, Chan TY, Chan AW, Lau GS, Critchley JA: Use and abuse of benzodiazepines in Hong Kong 1990–1993 – the impact of regulatory changes. J Toxicol Clin Toxicol 1995; 33:597–602. 44 Dyer C: European court upholds UK ban on Halcion. BMJ 1999;318:418. 45 Leber P, Laughren T: The question of Halcion: overview and background for the 18 May 1992 PDAC Meeting. FDA Memo, 28 April 1992, pp 3–4. 46 CBHSQ: Drug Abuse Warning Network Methodology Report, 2011 Update. Rockville, Substance Abuse and Mental Health Services Administration, Center for Behavioral Health Statistics and Quality, 2013. 47 Turmel A, De Montigny C: Sensitization of rat forebrain neurons to serotonin by adinazolam, an antidepressant triazolobenzodiazepine. Eur J Pharmacol 1984;99:241–244. 48 Broderick PA: Alprazolam, diazepam, yohimbine, clonidine: in vivo CA1 hippocampal norepinephrine and serotonin release profiles under chloral hydrate anesthesia. Prog Neuropsychopharmacol Biol Psychiatry 1997; 21: 1117–1140.

49 Broderick PA, Hope O, Jeannot P: Mechanism of triazolo-benzodiazepine and benzodiazepine action in anxiety and depression: behavioral studies with concomitant in vivo CA1 hippocampal norepinephrine and serotonin release detection in the behaving animal. Prog Neuropsychopharmacol Biol Psychiatry 1998;22:353–386. 50 Ashton CH: The benzodiazepines: what they do in the body (the Ashton manual). Benzodiazepines: how they work and how to withdraw. 2002. (accessed May 28, 2013). 51 Teboul E, Chouinard G: A guide to benzodiazepine selection. I. Pharmacological aspects. Can J Psychiatry 1990;35:700–710. 52 Harward JL, Clinard VB, Jiroutek MR, Lingerfeldt BH, Muzyk AJ: Impact of a US Food and Drug Administration drug safety communication on zolpidem dosing: an observational retrospective cohort. Prim Care Companion CNS Disord 2015;17:94–98. 53 Brayfield A (ed): Martindale: The Complete Drug Reference, ed 38. London, Pharmaceutical Press, 2014.

Psychother Psychosom 2015;84:273–283 DOI: 10.1159/000434755


Downloaded by: G. Fava - 28048 - 8/7/2015 9:24:29 PM

38 Sakshaug S: Drug Consumption in Norway 2007–2011 (Legemiddelforbruket i Norge 2007–2011). Oslo, Norwegian Institute of Public Health, 2012. 39 Gualano MR, Bert F, Mannocci A, La Torre G, Zeppegno P, Siliquini R: Consumption of antidepressants in Italy: recent trends and their significance for public health. Psychiatr Serv 2014;65:1226–1231. 40 Olkkola KT, Ahonen J: Midazolam and other benzodiazepines; in Schüttler J, Schwilden H (eds): Handbook of Experimental Pharmacology. Heidelberg, Springer, 2008, vol 182: Modern Anesthetics, pp 335–360. 41 Möhler H, Fritschy JM, Rudolph U: A new benzodiazepine pharmacology. J Pharmacol Exp Ther 2002;300:2–8. 42 Te Beek ET, Chen X, Jacobs GE, Nahon KJ, de Kam ML, Lappalainen J, Cross AJ, van Gerven JMA, Hay JL: The effects of the nonselective benzodiazepine lorazepam and the α2/α3 subunit-selective GABAA receptor modulators AZD7325 and AZD6280 on plasma prolactin levels. Clin Pharm Drug Dev 2015; 4: 149–154.