Benzodiazepine high‐doses: The need for an accurate definition

Abstract Objectives A clear definition of what we understand of high‐dose misuse or of a ‘markedly increased dose’ (as stated by the DSM‐5) is important and past definitions may be inadequate. The aim of this review is to describe the different definitions used and to test these definitions for their accuracy. Methods A narrative PubMed literature review was conducted based on articles published between 1 January 1990 and 31 December 2020 describing benzodiazepines (in MeSH Terms or MeSH Major Topic) and high‐dose (or high‐dosage). Specific definitions were applied to a population sample to show how definitions affect high‐dose benzodiazepine prevalence. Results Multiples of an equivalent‐diazepam dose or of the World Health Organization ‘defined daily dosage’ were used more frequently than the overstep of the recommended maximum therapeutic dosage as a cut‐off point. Conclusion High‐dose use is rare but the prevalence in the general population varies among studies, mainly due to different definitions, making both clinical and epidemiological comparisons between studies difficult. Defining a high‐dose user as a person who takes at least a higher dose than the maximum usual therapeutic dose over a defined period of time therefore appears to be clinically more consistent.

they are often used in general medicine, psychiatry, neurology, internal medicine, and in the field of anaesthesia.
For decades, benzodiazepines have been recommended as the standard treatment of anxiety and insomnia (Egan, 1999;Kurko et al., 2015;Laux & Puryear, 1984;Zandstra et al., 2002). Benzodiazepine withdrawal, rebound, overuse and abuse often limit their use in clinical practice (Moher et al., 2009). Several articles showed that the consequences of benzodiazepines use differs according to their type: anxiolytic or hypnotic (Cloos et al., 2015;Silberman et al., 2020), and to the dose a patient uses.
Benzodiazepines are generally well-tolerated, and their toxic threshold is very far from their therapeutic threshold. However, their use can cause several problems, depending on the dose and duration of the prescription, individual sensitivity, and their different characteristics. There are differences between abuse, misuse, addiction and dependence (physical, psychological and/or social), but they all have clinical significance (Silberman et al., 2020). There is a significant correlation between the dosage and duration of consumption and the risk of developing a physical dependence (Kan et al., 1997). Scales have been T A B L E 1 DSM-5 diagnostic criteria for sedative-, hypnotic-or anxiolytic-related disorders A problematic pattern of sedative, hypnotic, or anxiolytic use leading to clinically significant impairment or distress, as manifested by at least two of the following, occurring within a 12-month period: 1. Sedatives, hypnotics, or anxiolytics are often taken in larger amounts or over a longer period than was intended 2. There is a persistent desire or unsuccessful efforts to cut down or control sedative, hypnotic or anxiolytic use 3. A great deal of time is spent in activities necessary to obtain the sedative, hypnotic or anxiolytic; use the sedative, hypnotic or anxiolytic; or recover from its effects 4. Craving, or a strong desire or urge to use the sedative, hypnotic or anxiolytic 5. Recurrent sedative, hypnotic or anxiolytic use resulting in a failure to fulfil major role obligations at work, school or home (e.g., repeated absences from work or poor work performance related to sedative, hypnotic or anxiolytic use; sedative-, hypnotic-or anxiolytic-related absences, suspensions, or expulsions from school; neglect of children or household) 6. Continued sedative, hypnotic or anxiolytic use despite having persistent or recurrent social or interpersonal problems caused or exacerbated by the effects of sedatives, hypnotics or anxiolytics (e.g., arguments with a spouse about consequences of intoxication; physical fights) 7. Important social, occupational or recreational activities are given up or reduced because of sedative, hypnotic or anxiolytic use 8. Recurrent sedative, hypnotic or anxiolytic use in situations in which it is physically hazardous (e.g., driving an automobile or operating a machine when impaired by sedative, hypnotic or anxiolytic use) 9. Sedative, hypnotic or anxiolytic use is continued despite knowledge of having a persistent or recurrent physical or psychological problem that is likely to have been caused or exacerbated by the sedative, hypnotic or anxiolytic 10. Tolerance, as defined by either of the following: a) A need for markedly increased amounts of the sedative, hypnotic or anxiolytic to achieve intoxication or desired effect b) A markedly diminished effect with continued use of the same amount of the sedative, hypnotic or anxiolytic Note: This criterion is not considered to be met for individuals taking sedatives, hypnotics or anxiolytics under medical supervision.
11. Withdrawal, as manifested by either of the following: a) The characteristic withdrawal syndrome for sedatives, hypnotics or anxiolytics b) Sedatives, hypnotics or anxiolytics (or a closely related substance, such as alcohol) are taken to relieve or avoid withdrawal symptoms Note: This criterion is not considered to be met for individuals taking sedatives, hypnotics or anxiolytics under medical supervision.
Specify if: In early remission: After full criteria for sedative, hypnotic or anxiolytic use disorder were previously met, none of the criteria for sedative, hypnotic or anxiolytic use disorder have been met for at least 3 months but for less than 12 months (with the exception that Criterion A4, "Craving, or a strong desire or urge to use the sedative, hypnotic or anxiolytic," may be met) In sustained remission: After full criteria for sedative, hypnotic or anxiolytic use disorder were previously met, none of the criteria for sedative, hypnotic or anxiolytic use disorder have been met at any time during a period of 12 months or longer (with the exception that Criterion A4, "Craving, or a strong desire or urge to use the sedative, hypnotic or anxiolytic," may be met) developed to measure the severity of dependence (Cuevas et al., 2000).

| Eligibility criteria
The eligibility of studies was estimated using standardized inclusion and exclusion criteria. We included papers in the dates of publication and languages described earlier. In a first step, we read the titles and the abstracts to select the papers to be included in our review. Only studies on chronic high-dose usage associated with BDZs in humans were selected. Articles concerning intravenous (IV) midazolam used only in anaesthesia were also filtered out. In case of doubt on the inclusion of a specific reference in the final selection, in a second step, the decision to select or reject the article for our review was made by reading its full text.
CLOOS ET AL.

| Data extraction and synthesis
We classified the articles according to the definitions of high-dose BDZ used and discussed the different cut-off points found in the literature. Selected definitions were applied to population data from our published study (Cloos et al., 2015) to show their effects on highdose use prevalence rates.

| RESULTS
A total of 696 articles were found using our PubMed Boolean search string. 520 articles were excluded by reading the abstract and the title. Of the total, 122 further articles were excluded after full-text reading, leaving a total of 54 articles ( Figure 1). Four more articles listed in the references of the retained selection did not meet our Boolean search criteria, but have been included since they were of particular additional interest.
The BDZ high-dose definitions proposed in the selected articles were then classified according to their methodology. Characteristics of the 58 selected articles are described in Table 2 Japan, South Korea), 4 from Switzerland, 3 from Canada, 2 from the United Kingdom, 1 from South America, 1 from Australia, and 2 are not applicable. These papers came mainly from epidemiological studies (n = 13) and selected patients (n = 20). Length of follow-up is mainly at least one year (n = 30) or from a 3-month to 1-year duration (n = 16). High-dose definitions usually come from studies comprising at least 50 people (n = 43), largely adults (n = 43) in outpatient treatment (n = 28) using several BDZ (n = 53). Table 2 also highlights the studies examining the subgroups of benzodiazepines (anxiolytics and hypnotics) separately, and not as a whole. Not all benzodiazepines are the same and undifferentiated studies may provide insufficient data of potential risk of a specific BDZ (Cosci et al., 2015;Zito, 2015).

| Benzodiazepine high-dose definitions used in research
The following major definitions were identified by examining the research articles: � A specific equivalent dose of diazepam.
� A specific daily, weekly or yearly World Health Organization (WHO) defined daily dose (DDD) fraction.
� A percentile of an average daily dose.
� A combination of length and dose used.
� A maximum therapeutic dose recommended by the manufacturer.
F I G U R E 1 PRISMA flowchart of search strategy and abstract screening T A B L E 2 Description of the characteristics of the included studies
-5 of 19 T A B L E 2 (Continued)

| A specific equivalent dose of diazepam
Converting BDZs into an equivalent dose of diazepam (EDD) is common in clinical practice and useful in BDZ withdrawal. There is often a range rather than a unique value for a specific BDZ, for example, 10 mg of diazepam is equivalent to 20-50 mg of tetraze- pam, yet there is no standard conversion table, and diazepam dose equivalence tables can only be found in some of the articles using the EDD method (e.g., Alexander & Perry, 1991;Brinkers et al., 2016;Kaendler et al., 1996).

T A B L E 3 (Continued)
Definition for highdose use (number of studies)

Specific population
Patients' consequences Liebrenz et al. ( -2016 definition: • took BDZs for an extended period of time (without defining "longterm") • for a higher than 40 mg EDD, and/ or • those who had an otherwise problematic use of BDZs (such as substance mixing, repeated dose escalation, recreational use, euphoric effect enhancement, illegal acquisition strategies or experiences of negative social consequences).
Users with dosage taken by these patients had been at least a 40 mg EDD prior to sudden cessation (Seivewright & Dougal, 1993).
In Italy, the maximum approved dose of diazepam for extrahospital use is 10 mg/day. A research team from Verona considers an intake exceeding at least five times the maximum daily recommended dose as "high-dose" (Lugoboni et al., 2014), that is, 50-mg diazepam. "High-dose abuse": • >50 mg diazepam/day intake over an extended period of time (>6 months), (Faccini et al., 2019;Federico et al., 2020;Lugoboni et al., 2020Lugoboni et al., , 2018
• HDU as the average daily dose for the longest treatment episode in the top 10th percentile for each agent (Hermos et al., 2007).
/ / / A combination of length and dose used (n = 1) (Holzbach et al., 2010) Dependent: • Orange code attributed if the equivalent daily dose of diazepam was between 5 and 10 mg of diazepam.
• Red code if between 10 and 15 mg.  (Semla et al., 2016). Low dose treatment was then defined as under one SDD and high-dose as over three SDD.
A maximum therapeutic dose recommended by the manufacturer (n = 5) BDZ over-prescription: users receiving a higher dose than the manufacturers' recommended maximum daily dosage over a certain period of time (Sketris et al., 1985). Use of the British National Formulary to identify high-dose users taking the maximum or doses higher than recommended for more than 6 months (Perera & Jenner, 1987). For those taking several BDZs, the fractions of the dose of each BDZ taken were divided by the maximum recommended dose for each one of the BDZ or using the maximum dose recommendations of the Greek National Pharmaceutical Organization (Brayfield, 2017;Lekka et al., 2002) (Ellinwood et al., 1990), 15 mg of oxazepam (Bajwah et al., 2018), 20 mg of oxazepam (Imbert et al., 2016), 2 mg of alprazolam (Rintoul et al., 2013), and 5 mg of triazolam (Sullivan & Sellers, 1992 equivalent, authors use different values to define high-dosage, from a greater than one EDD of 10 mg (Voshaar et al., 2003) up to 50 mg (Faccini et al., 2016(Faccini et al., , 2019Federico et al., 2017Federico et al., , 2020Kaendler et al., 1996;Lugoboni et al., 2020Lugoboni et al., , 2014Lugoboni et al., , 2018Tamburin et al., 2017).

| A specific daily, weekly or yearly DDD fraction
According to the WHO, the DDD is a unit developed to measure 'the average maintenance dose per day for a drug used for its main indication in adults' (WHO, 2003). It does not necessarily correspond to the recommended or prescribed daily dose and is often a compromise between different doses in various countries. There may be insufficient clinical support and the equivalent dose of 10 mg of diazepam may represent 6.25%-200% of the DDD, depending on the specific BDZ. The DDD definitions may be arbitrary (e.g., due to the prescription habits in a specific country) and they are sometimes reconsidered, thus varying in time. The DDD for clonazepam, for example, may be inappropriate (Islam et al., 2014), since DDD of 8 mg is listed for that drug, where the usual clinical dose is 0.5 mg. To define high-dose use, the authors rely on a specific daily, weekly or yearly DDD fraction in their studies, as described in Table 3.
To reflect the mean BDZ dose of all the different BDZs by one parameter, Kan et al. (1997)  Specific attention should be given to the ageing population (≥65 years) in which a 50% reduction of the recommended adult dose is generally recommended (Etchepare et al., 2016;Nakra & Grossberg, 1986;Teboul & Chouinard, 1991) or a dose selection at the low end of the adult dosage range (Hanlon et al., 2009 Sometimes specific psychiatric or neurological disorders require, for a short period of time (days or 1-2 weeks), a dosage substantially higher than the maximum recommended therapeutic dosage (see Table 5), thus creating a bias when the study period is set too short (e.g., 1 or 3 months) and these short-term very-high-dose users risk being included in an abuse group. A 1-year period to study long-term or high-dose will reduce such biases (Kurko et al., 2015;Zandstra et al., 2002).
Several chronic psychiatric disorders may also require a pro- A 'higher than the usual therapeutic dosage' therefore appears to be a better way to define a 'markedly increased dosage' and a usage longer than a year should become the standard for studying 'longterm use'. Problematic high-dose use may then be defined as an 'over 1-year usage of a markedly increased dose (i.e., higher than the usual therapeutic dosage)'.

| The need of an international consensus
There is no international consensus on the maximum usual therapeutic doses for all substances. For example, the maximum dose recommended by the National Pharmaceutical Organization in Greece (used as cut-off points by Lekka et al., 2002) does not, at least for some substances, correspond to maximum dosages proposed by other sources (see Table 2). Standardization of the maximum therapeutic dosages is therefore also necessary when evaluating high-dose use by the UTD definition. In such a 'consensus dosage table', the cutoff-point is not necessarily the maximum annual dosage proposed by the manufacturers, but may be set lower by a panel of experts if they consider a longer high-dose use of the specific BDZ as being harmful.
Indeed, the dose level considered by a manufacturer as being safe for 4 weeks (e.g., in the treatment of an anxiety disorder) may eventually become harmful if the drug is taken at that dose for a longer period of time. Thus, one might also argue that a 365-UTDs-per-year cut-off could be as problematic as maximum levels of diazepam (e.g., 7300 mg per year) or DDDs. In the absence of an international 1- year-maximum-dosage-consensus, the UTD-cut-off-line could therefore be drawn at a lower level, for example, an annual intake of 100 times the manufacturers' maximum therapeutic dosages.

| The need of research on specific benzodiazepines
Instead of grouping all benzodiazepines together, as is done in most of the studies, there is a need to focus on each medication on its own in future research. Hypnotic benzodiazepines show a notably higher risk for high dose usage compared to anxiolytics, as shown by our own research (Cloos et al., 2015) and pointed out by other studies (Martínez-Cano et al., 1996;Takeshima et al., 2016;Wen et al., 2014).

| Strengths and limitations
Our study included only publications from PubMed. Other databases such as EMBASE, SCOPUS and PsycINFO and grey literature (theses and dissertations, research and committee reports, government reports, conference papers, and current research) were not included.
We believe that an article research of the PubMed database is sufficiently extensive to illustrate the various definitions of BDZ highdose that can be found in the scientific literature. Moreover, this research is the first narrative review of its kind to identify all the used definitions associated with benzodiazepine high-dose.

| CONCLUSION
There is a lack of an internationally recognized standard in the definition of a high daily BDZ dose, and the DSM-5 concept of a 'markedly increased dosage' remains vague. The DSM-5 diagnosis of a sedative-, hypnotic-or anxiolytic-related disorder needs several symptoms of compulsive, drug-seeking behaviour. Also, prescription medications can be used inappropriately, even in therapeutic dosages. However, studies relying solely on prescription or delivery data lack these additional clinical data allowing a more precise diagnosis.
In the past, researchers considered either a certain equivalent dosage of diazepam or an overrun of the WHO DDD as cut-off-points between 'normal' and 'high-dose' use. None of these methods is entirely satisfying, and prevalence rates of high-dose use may be overrated when using these definitions. Especially the 20-mg-diazepam-and the 2-DDDs-limit overestimate high-dose-use prevalence by including patients who actually respect dosages within therapeutic limits. These patients should not be considered inappropriate users, and they should not receive a diagnosis of a DSM-5 sedative-, hypnoticor anxiolytic-related disorder. The nonrespect of the maximum therapeutic dosage threshold over a period of 1 year appears to be clinically more adequate when defining BDZ high-dose abuse. We therefore propose to avoid the terminology of 'high-dose use' in future research.
Even the DSM-5 'markedly increased dosage' is unsatisfying. Referring to an 'above-therapeutic-limit' use would allow better comparisons between studies. Defining a problematic high-dose user as a person who received at least a higher dose than the usual yearly maximum therapeutic dose is appropriate when observing a population over a longer period of time and studying continuous users.
Furthermore, BDZs vary a lot as to potential dependence. Hypnotic BDZs show a higher risk for high dose dependence than anxiolytic BDZs, and the odds also vary largely inside the two subgroups.
For example, triazolobenzodiazepines, such as alprazolam and triazolam, may be particularly susceptible to long-term and high-dose use problems. The clinical differences among BDZs may be based on their chemical structure, their potency, the onset and length of action and/or their active metabolite. Based on these pharmacological aspects, future pharmacoepidemiological studies should therefore assess these substances separately, or at least differentiate between the two subgroups.