A scoping review of motor vehicle operator performance assessments for benzodiazepine receptor agonists

Abstract

Background

Benzodiazepines and Z-drugs (e.g., zopiclone, zolpidem) (benzodiazepine receptor agonists or BZRAs), are prescribed for anxiety and insomnia disorders. However, they are not indicated as first line therapies for long-term management due to harms and efficacy limitations. BZRAs have also been associated with traffic accident risks. Patients taking BZRAs are told to consult with health care providers regarding motor vehicle operation safety. However, advice on driving is variable. The objective of this scoping review is to identify, map, and characterize the evidence for assessments that measure driving performance in people taking BZRAs.

Methods

Embase (Elsevier), MEDLINE (Ovid), and PsycINFO (EBSCO) were searched. Covidence was used for screening. Each stage of screening included two independent reviewers. A REDCap database was used for data extraction by two independent reviewers. Results were tabulated and summarised as a narrative.

Results

Driving performance was assessed with 20 unique BZRAs across 183 studies (n = 92 experimental; n = 91 observational) in 178 publications. Zopiclone was the most studied. In experimental studies, the Standard Deviation of Lateral Position (SDLP) was used most often (n = 54, 62 %) and many studies (n = 35, 38 %) were conducted in the Netherlands. For observational studies, biological detection (e.g., urine, blood) (n = 73, 80 %) followed by prescription drug/dispensing records (n = 17, 19 %) were the most common impairment measures and Norway (n = 20) is where most studies took place. In experimental studies, most (n = 89, 97 %) were conducted using only one driving setting. Simulated driving in a car (n = 36) and road driving in traffic (n = 36) were common as compared to nontraffic driving course (n = 8) and simulated driving (n = 9). In experimental studies, seventy-eight of the 92 studies (85 %) had at least one measure that identified impairment.

Conclusions

BZRA effects on motor vehicle driving performance have been studied using heterogenous protocols with multiple measures and settings, ranging from simulation to authentic traffic situations in experimental studies to biological detection and dispensing records in observational studies. Many BZRAs have been studied but study representation does not match prescribing pattern prevalence. The interpretation and contextualization of results for clinical practice is challenging due to the complexity (i.e., protocols, measures, settings). Future work in this area should work to improve knowledge translation of results so information is more readily accessible and applicable to health care providers and patients.

1. Introduction

Benzodiazepine receptor agonists (BZRAs) (e.g., diazepam, lorazepam, zopiclone, zolpidem) are sedating medications with prevalent use of 3 % to 5 % in the general population.1, 2., 3, 4, 5, 6 Use increases with age, with estimates of chronic use ranging between 10 % to 20 %⁴ in older adults with these individuals more likely to use BZRAs more than prescribed.^7,8 BZRAs have a narrow role in the management of anxiety disorders and are of limited benefit in the treatment of insomnia.^9,10 Their adverse effects (e.g., falls,¹¹ dependence and withdrawal,¹² memory problems,¹³ pneumonia,¹⁴ sleepwalking and sleep driving,¹⁵ daytime sedation and impairment¹⁶) are well established. Reducing the use of BZRAs for anxiety-related disorders and insomnia across age groups has proven to be challenging despite numerous guidelines, policies, and campaigns to reduce their use, and it continues to be an active area of investigation.^9,10,17., 18, 19, 20., 21. BZRA-related impaired cognition and physical functioning is associated with driving impairment.22., 23, 24, 25 These risks can become greater when BZRAs are combined with other sedating medications (e.g., opioids²⁶) and non-prescription substances (e.g., alcohol).²⁷ Drivers can experience challenges with controlling speed, lane positioning, and interpretation of traffic signage.¹⁶ Driver impairment regulations and penalties vary globally²⁸ and in some places, depend on whether medications are prescribed by a physician.²⁹

Patients are encouraged to consult health care providers regarding safe operation of a motor vehicle while consuming BZRAs. However, the information available to health care providers can be vague and lack consensus regarding individual BZRAs and driving safety. Health care providers will seek a variety of sources to inform patient education and advice regarding BZRA adverse effects (e.g., impairment), including product monographs. For example, a zopiclone monograph includes a black box for “serious warnings and precautions” that includes, “Caution patients that impairment when driving or operating machinery can persist for up to 12 hours after an evening dose” and, “Risk for road traffic crashes appears to be highest during the first 2 weeks of therapy.”³⁰ Patient information for a brand product of zopiclone also indicates that people not drive under several conditions including, but not limited to, “if it has not been 12 hours or more…” since taking the medication, if the person is “elderly”, or taking, “the 7.5 mg dose”.³¹ The manufacturer warns patients about “…. driving a motor vehicle after ingesting the drug” as part of their section on “Driving and Operating Machinery”.³² Other resources make more general statements by telling patients to avoid driving until it is safe to do so³³ or indicating that patients should not operate a motor vehicle until they know how a medication affects them, especially with dosage changes and additions of new medications.³⁴ The Driving under the Influence of Drugs, Alcohol and Medicines (DRUID) project in Europe aimed to improve knowledge of medication-related level of impairment, building on their work of categorizing impairment ranging from none to severe.²⁵ Many centrally acting agents were categorized as a moderate or severe influence for impaired driving.²⁵ This would be expected, but how this work and its evidence base are interpreted by health care providers, and translated to explicit information in shared decision making with patients remains largely unknown.

The content and extent of education given to patients in clinical settings is relatively unexplored, and to our knowledge, there is no direct and objective evaluation in the current literature on practice-based assessments of patient education provided regarding BZRAs and the impact on driving performance. For example, studies using secret or mystery shoppers with insomnia have been used to evaluate pharmacists' patient education, but not specifically with BZRAs. Two studies evaluated pharmacists' advice to a mystery shopper regarding insomnia management with non-prescription products (e.g., sedating antihistamines) and identified areas of improvement including around information on medication-related adverse events.^35,36 Another study regarding sedating antihistamines, albeit for allergies, found only one pharmacist from the 88 included pharmacies cautioned about sedation and avoiding driving.³⁷ In general, pharmacists feel less confident about counselling around risks associated with use of alertness impairing medicines.³⁸ Patient understanding and perceptions of alertness impairment risks may also be inaccurate with respect to the timing of these effects (e.g., immediately after dosing versus several hours after dosing).³⁹

Health care providers seeking to critically appraise literature regarding medications impacting driving safety can also face challenges regarding the complexity of driving performance methods of assessment. Driving assessments vary widely from laboratory cognitive tests, driving simulators, to on-the-road driving tests.^40,41 On-the-road tests, viewed as one with highest fidelity, demonstrate dose-dependent impairment for alcohol, illicit drugs, as well as some sedative-hypnotic medications, antidepressants, and antihistamines.⁴² Driving simulators can assess performance but are criticized for limited realism.⁴⁰ Lastly, laboratory cognitive tasks are limited in their validity to compare to real-life driving and whether they can be accurately used to predict performance-related risk.⁴⁰

Given the prevalence of BZRA use, often with chronicity, and variability in assessment methods for impacts on driving, we conducted a scoping review to identify, map, and characterize the methods used to assess the impact of individual BZRAs on driving performance.

2. Methods

The scoping review protocol was developed a priori and published⁴³ and was adhered to with minor modifications as outlined. As part of the initial scoping review process, we conducted a preliminary search of PROSPERO, MEDLINE, the Cochrane Database of Systematic Reviews, and the JBI Database of Systematic Reviews and Implementation Reports and no scoping reviews related to the topic were identified. The review was conducted following the JBI guidance for the conduct of scoping reviews⁴⁴ and reported using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR).⁴⁵

We limited the scope of the review to the BZRAs for pragmatic and clinically-focused reasons. The BZRAs have a high rate of use across ages groups and are the most prevalent class of sedative-hypnotics prescribed for prevalent conditions (e.g., insomnia, anxiety). There are also numerous individual medications that fall within the BZRA group that have a range of pharmacodynamic and pharmacokinetic properties that can cause challenges to determine recommendations regarding their use and driving.

2.1. Search strategy

The search strategy aimed to locate both published and unpublished studies. An initial limited search of MEDLINE (Ovid) and Embase (Elsevier) was undertaken to identify articles on the topic. To identify benzodiazepine keywords for the search, benzodiazepines listed in various drug monograph databases (e.g., Micromedex, Lexicomp, and Martindale: The Complete Drug Reference) were searched. The text words contained in the titles and abstracts of relevant articles, and the index terms used to describe the articles were used to develop a full search strategy in MEDLINE (Ovid; see Appendix II). The search strategy, including all identified keywords and index terms, was adapted for each included database. The databases searched for published research, with a search dated June 9, 2020, included Embase (Elsevier), MEDLINE (Ovid), and PsycINFO (EBSCO). This search was rerun to update the results on May 3, 2022. Initial keywords used were “benzodiazepine,” “sedative-hypnotic,” “Z-drug,” “driving impairment,” and “driving performance.”

Grey literature (e.g., policy documents and legislation on driving and medication use, vital statistics) were used during the scoping review process to provide background information, contextualize the issues, and identify application of findings for clinical and policy settings. Other grey literature was not searched given the scoping review question focused on the body of evidence evaluating the methods of driving performance and BZRAs.

2.2. Participants

This review considered studies of participants who were new, intermittent, or chronic users of BZRAs (Appendix I, Inclusion and Exclusion Criteria). No exclusions were applied regarding the health status of participants or whether their BZRA use was for an approved indication, if known, as indicated by government agencies (e.g., Health Canada) or practice guidelines. Participants were required to be of licensed age to operate a motor vehicle. Fifteen years of age was chosen as the cut-off based on preliminary review of studies from various countries. Those hospitalized for reasons unrelated to driving or residing in long-term care or assisted living facilities were not considered in this review. All BZRAs that participants were consuming were eligible for inclusion provided there was evidence of one BZRA used and not multiple, and that it was not used intravenously or intramuscularly.

2.3. Concept

The current scoping review considered studies in which there is a demonstrated or inferred presence of a BZRA in association with the operation of a motor vehicle (e.g., cars, trucks, motorcycles). Operation of a motor vehicle could have been real or simulated. Outcomes could include direct or indirect objectives or standard subjective measures or indicators of impairment while operating a motor vehicle. However, indirect methods such as electroencephalogram and tests of proprioception were excluded. Experimental and observational studies that had links to outcomes, such as traffic accidents, fatalities, and hospital visits related to operation of a motor vehicle to BZRAs were included.

2.4. Context

The setting for studies was not limited by geography and included community-based settings; urban, rural, and remote settings; and international settings.

2.5. Types of sources

The scoping review considered experimental studies and observational studies. Example of study designs include randomized controlled trials, cohort and case control studies, cross-sectional studies, prevalence studies, and case-crossovers. Systematic reviews were used for citation chasing and examined for studies related to the scoping review objectives. Only studies published in English were included for data extraction as it was the prevalent language spoken and understood by all members of the team. No date limits were applied.

2.6. Study selection

Following the searches, all identified citations were collated and uploaded into Covidence 2019 (Veritas Health Innovation, Melbourne, Australia) and duplicates removed. Covidence was used for title and abstract screening conducted by two independent reviewers for assessment against the inclusion and exclusion criteria. The full texts of selected citations were then assessed in detail against the inclusion and exclusion criteria (Appendix I) by two independent reviewers using Covidence. Any disagreements that arose between the reviewers at each stage of the study selection process were resolved through discussion or via a third reviewer from the team. Reasons for exclusion of full text studies that did not meet the inclusion criteria are presented in the PRISMA flow diagram (Fig. 1).³⁰ All studies included for full text were managed using REDCap electronic data capture tools hosted at Dalhousie University. REDCap (Research Electronic Data Capture) is a secure, web-based software platform designed to support data capture for research studies, providing: 1) an intuitive interface for validated data capture; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for data integration and interoperability with external sources.^46,47 The data extraction in REDCap differed based on the design of the study being extracted (e.g., experimental, observational) (data dictionary code book and fields available above request). Microsoft Excel, available via our Dalhousie University Microsoft® 365 Subscription, was used in addition to REDCap for further analyses.

Fig. 1.

Search results and study selection and inclusion process.⁴⁵

2.7. Data extraction

Pairs of two independent reviewers (KS, DMG, ALM, SS, TS, IA, SP) completed data extraction of included studies in REDCap (Appendix III) following training and discussion as a team. Any disagreements that arose between the reviewers during data extraction were resolved through discussion or via a third reviewer. At least 10 % of publications were audited for accuracy by two (ALM, DMG) other team members. The data extracted included specific details about the population (e.g., age, sex), study design (experimental vs observational), year of publication, sample size, country of publication, BZRAs studied (inclusive of dosing regimen, calculated diazepam equivalents), driving assessment setting (on-the-road driving vs driving simulation), measures of driving impairment (e.g., standard deviation of lateral position, accident, road-side assessment), and key findings relevant to the review objective.

2.8. Modifications from the study protocol

The data extraction was completed in REDCap versus the JBI System for the Unified Management, Assessment and Review of Information (JBI SUMARI; JBI, Adelaide, Australia) as originally outlined in the protocol. REDCap was chosen due to the research team's experience and familiarity with it as a project and data management tool, and the flexibility with data extraction options (e.g., pull-down menus, branching), and other options available. As per the protocol,⁴³ the data extraction template submitted with the protocol was revised extensively.

In the protocol,⁴³ we described including key findings relevant to the review objective. We decided to report impairment findings for both experimental and observational studies as part of these key findings. In the observational studies, the impairment outcomes (e.g., motor vehicle accident, suspected impairment based on driving behaviours) were often used to identify issues with people's motor vehicle driving performance. Essentially, the impairment findings were often used to identify potential cases in case-control studies. For experimental studies, we included the authors' interpretations regarding impairment findings, without critical appraisal of their validity given this was not the purpose of the scoping review.

2.9. Data analysis

The data were analysed to determine how BZRA driving impairment has been assessed and are presented visually in figures and in tabular format. The data in text are presented with experimental studies reported first, followed by observational studies. For readability related to lengthy in-text citations, when citations were over 35 in number, readers are referred to specific appendices for specific citations versus listing lengthy citations.

3. Results

3.1. Study inclusion

We identified a total of 5886 studies through our searches of EMBASE (Elsevier), MEDLINE (Ovid), and PsycINFO (EBSCO) (Fig. 1), which were imported to Covidence. Duplicates (n = 1430) were removed before starting title/abstract screening, which resulted in 535 studies moving to full text review. Full texts were then screened and a total of 178 publications were included, with data extraction for 183 studies given some individual publications reported more than one distinct study. Eighteen systematic reviews48, 49, 50, 51., 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62., 63., 64, 65 were identified, and their citations examined to find other relevant studies. No additional relevant studies were identified in the systematic reviews' citation lists compared to our own searches.

The most common reason for exclusion in 133 studies was a lack of single entity BZRA use. In these studies, participants were taking more than one substance with the potential to impair driving (e.g. alcohol in combination with a BZRA) or it was not evident that participants were taking a single-entity BZRA. As such, driving impairment observed in these studies may not be an accurate representation of the true impairment caused by a BZRA alone. Other notable reasons for exclusion included wrong study design (n = 143). Studies excluded for wrong study design measured skills indirectly related to driving (e.g., electroencephalogram, hand foot proprioception), or investigated objectives where measuring impaired driving was not part of a focus of the study. Studies excluded for wrong publication type were abstracts, conference proceedings, or letters to the editor. A complete list of reasons for exclusion are provided in the PRISMA flow diagram (Fig. 1).

3.2. Characteristics of included studies

3.2.1. Experimental

There were 89 publications categorized as experimental studies, and these included 92 experiments in total (Appendix III).66, 67, 68, 69., 70, 71, 72, 73, 74, 75, 76., 77, 78., 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92., 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111., 112, 113, 114, 115, 116, 117, 118., 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131., 132, 133, 134., 135, 136, 137, 138, 139, 140, 141, 142, 143, 144., 145, 146, 147, 148, 149., 150., 151, 152, 153., 154 O'Hanlon et al.¹¹⁷ published three experiments in one paper and Vanakoski et al.¹³³ reported on two different experiments. We treated the experiments within the O'Hanlon and Vanakoski papers as distinct studies and completed data extraction from the two papers as five extractions. When these are reported in the text, the same citation is repeated. There were also some publications with common investigators that used data from previously published studies and reanalysed these data using a different metric. For example, Verster et al. in their 2014¹⁴⁷ publication used data from Verster et al. in 2002¹⁴⁵ and Mets et al.¹⁰⁸ from 2011. We treated the Verster et al. from 2014¹⁴⁷ as one extraction since the authors combined the sample from the original papers published in 2002 and 2011.

3.2.2. Observational

There were 89 publications^23,93,155, 156, 157., 158, 159, 160., 161, 162, 163, 164, 165, 166, 167, 168, 169., 170, 171, 172, 173, 174, 175, 176, 177., 178., 179, 180, 181, 182., 183, 184, 185, 186, 187, 188., 189, 190, 191, 192., 193., 194, 195, 196., 197, 198., 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219., 220., 221, 222., 223., 224, 225, 226, 227, 228., 229, 230., 231, 232, 233, 234, 235, 236, 237., 238, 239., 240, 241 classified as observational and two studies^220,223 within this group that were extracted as four separate studies, leaving 91 studies in total (Appendix IV).For observational studies, we divided them into two major categories: comparative studies, and non-comparative studies. Within comparative studies, we categorized them using the following taxonomy: cohort-study (prospective or retrospective); case-crossover; and case-control. Non-comparative studies were grouped into one of the following: cross-sectional or prevalence.

4. Review findings – Experimental

4.1. Country and date of publication

The country of publication varied between experimental and observational study designs (Table 1). Experimental studies were conducted across 12 different countries in North America, Europe, and Asia with the majority being in the Netherlands (Appendix III).

Table 1.

Country of publication and number of experimental studies.

Country	Number of Experiments within Experimental Studies
Netherlands	35
Finland	11
United Kingdoma	9
France	8
United States	8
Germany	6
Sweden	5
Australia	3
Japan	3
Canada	2
Belgium	1
Italy	1

^aUnited Kingdom includes studies from Scotland, England, Wales, and/or Northern Ireland.

For experimental studies (Appendix III), we found the highest number (26/92, 28 %) published between 2010 and 2019 inclusive, with fewer in the decades before (n = 17, 2000s; n = 18, 1990s; n = 13, 1980s; n = 11, 1970s). Since 2020 to the time of the search, there were 7 experimental studies published.

4.2. Population demographics

A total of 2944 participants were included across the experimental studies. Subject ages in the experimental studies ranged from 18 to 77 years old with a mean age of 38.2 years. Females comprised 31 %, 36 % were males, and for the remaining 33 % sex of participants was not provided or unknown.

4.3. BZRAs studied

Across the 92 experimental studies within 89 papers, 20 unique BZRAs were studied (Appendix III). Some BZRAs have been studied more frequently than others. For example, within the Z-drugs, zopiclone (n = 30)^{67,71,72,77,78,89,92,93,}96, 97, 98, 99, 100^{,104,105,107,108,113,121,132,134,138,}140, 141, 142, 143, 144.^,147, 148, 149. and zolpidem (n = 16)^{67,71,72,97,99,105,107,118,120,134,137,141,145,}147, 148, 149. were studied more often compared to zaleplon (n = 5).^{138,140,145,147,148} When any one of the Z-drugs was a medication being investigated, 1897, 98, 99, 100^{,108,113,121,134,137,138,}140, 141, 142^,144,145,147, 148, 149. were conducted in the Netherlands. Diazepam (n = 24)^{70,80,81,83,90,93,}101, 102, 103^{,105,111,114,117,119,125,129,}133, 134., 135^{,149,151,153,154} was the most studied non-Z-drug BZRA, followed by lorazepam (n = 13)^{79,85,99,106,110,112,116,117,128,134,136,149} and alprazolam (n = 10).^76,86,87,95,126, 127, 128^,134,146,149 Overall, zopiclone was the most frequently studied of all drugs.

Some studies compared different BZRAs or used the BZRA as a comparator, including studies that had another non-BZRA being researched. For example, Vermeeren et al.¹³⁷ compared zolpidem, flunitrazepam, and placebo in a double-blind crossover trial. Mercier-Guyonx et al.¹⁰⁶ examined lorazepam versus an antihistamine.

4.4. Driving impairment setting and measures overview

When examining the study design and methods of experimental studies, we found that two defining features are important: where the study takes place, denoted as the driving impairment assessment setting (Table 2); and driving impairment measures (Table 3). These two features appeared to guide how studies are conducted.

Table 2.

Driving setting, description, and approximation of actual driving conditions in experimental studies.

Driving Setting	Description	Approximation of driving conditions
Road traffic driving	Participants drive a real car on roads with other drivers.	Highest fidelity
Driving course (non-traffic)	Participants drive a real car on closed roads with no other drivers.
Simulated car driving	Participants are placed in a specially instrumented car indoors that mimics real road driving.
Simulated computer driving	Participants use a computer to simulate driving.	Lowest fidelity

Table 3.

Four most common measures to assess driving performance and impairment in experimental studies.

Measure of driving impairment	Description	Interpretation
Standard Deviation of Lateral Position (SDLP)90,142	The participants' task is to drive in the center of the designated traffic lane while maintaining a constant speed. Deviations from the center of the lane are recorded.	Higher SDLP values are associated with a greater degree of driving impairment.
Standard Deviation of Speed (SDS)142	The participants' task is to drive at a constant predetermined speed. Deviations from the instructed driving speed are recorded.	Higher SDS values are associated with a greater degree of driving impairment.
Number of Road Exits (#RE)148	Road exits are events where the participant driving leaves the lane unintentionally, either to another traffic lane or the shoulder of the road.	Higher number of road exits are associated with a greater degree of driving impairment.
Brake Reaction Time (BRT)73	The participants' task is to brake when instructed. The participants reaction time is recorded.	Longer brake reaction times are associated with a greater degree of driving impairment.

With respect to Standard Deviation of Lateral Position (SDLP; weaving of the car) and Standard Deviation of Speed (SDS), the mean lateral position and the mean speed are continuously recorded and SDLP and the SDS are computed, respectively. The SDLP test has been shown to demonstrate dose-dependent impairment for alcohol, illicit drugs, as well as some sedative-hypnotic medications, antidepressants, and antihistamines.⁴² The Number of Road Exits (#RE) are counts of the events where those driving leave the lane unintentionally, either to another traffic lane or the shoulder of the road.¹⁴⁸ This may represent loss of control of the vehicle. Other commonly used and related terminology includes excursions out-of-lane¹⁴⁸ and inappropriate line crossings.^79,89

Within driving setting, it was noted that there was use of two broad categories: real road driving, and simulated driving. Nuances exist within these categories. For example, road driving could occur in real traffic or on a driving course. Studies using simulated driving could include a simulated car or a computer-based mechanism to simulate the driving experience. Driving simulators range from fully interactive systems using instrumented vehicles with 360-degree view to computer desktop systems. They are capable of assessing performance in a safe and controlled manner, however, are criticized for limited realism.⁴⁰ Lastly, laboratory cognitive tasks permit an assessment of a single isolated aspect of driving performance in a controlled environment. However, these are limited in their validity to compare to real-life driving and are questionable in whether they can be accurately used to predict accident risk.⁴⁰ Other assessments used infrequently in the studies were varied (e.g., gap estimation/acceptance task, weaving task, parking task, speed exceedances, mean speed, etc.)

4.5. Driving impairment setting

Except for three studies by Daurat et al.⁷⁹ and Volkerts et al.,¹⁵⁰ both of which used road driving (traffic) and simulated driving (car), and Laurell and Tornros⁹⁴ who used driving course (non-traffic) and simulated driving (car), all experimental studies used only one driving setting. Nontraffic driving course was used in eight studies,^68,69,73,84, 85, 86^,106,151 and simulated driving with a computer was used in nine.^{67,77,78,88,93,105,124,128,152} Simulated driving in a car and road driving in traffic were most used with 36 studies each (Appendix III). Most driving course studies (nontraffic) have been conducted in the UK.^68,69,73,84, 85, 86^,151 Simulated car driving tests and simulated computer driving tests have been conducted in multiple countries without a clear trend.

4.6. Driving impairment measures

The ways in which driving impairment were measured or quantified differed across the experimental studies. Most employed multiple driving impairment measures (Appendix III). For example, SDLP was most commonly used in 54 of 92 (62 %) trials of which 12 used it as the sole measure.^74,87,90,115, 116, 117^{,121,134,137,141,144} All were conducted in the Netherlands (Appendix IV). Country trends were less clear regarding some measures used to assess performance. For example, in driving course studies (nontraffic) from the UK,^68,69,73,84, 85, 86^,151 three used an individual measure such as Brake Reaction Time (BRT)^73,86 and gap estimation/acceptance task,¹⁵¹ while the remainder used combinations of methods.

There were other patterns of measures co-occurring frequently together (Appendix IV). For example, SDS was almost exclusively used alongside SDLP. Only one study was conducted by Otmani et al.,¹¹⁸ employing SDS without SDLP. Other measures were used with SDS including mean speed and number of collisions.¹¹⁸

4.7. Driving impairment settings' and measures' trends

Certain driving impairment measures were used more commonly in specific driving impairment assessment settings. For example, SDLP and SDS are heavily used in road traffic driving and simulated car driving (Appendix III), while they are less frequently used in driving course or simulated computer driving settings. Conversely, other measures such as number of Road Exits (#REs) were used in 19 studies^{71,76,79,83,89,91,92,}101, 102, 103^{,109,111,112,126,127,132,136,146,148} and infrequently used in road traffic driving or driving course settings. Number of road exits was more commonly used in both car and computer simulated driving setting (Appendix III). The metric #REs was never used as the sole measure of impairment.

Brake reaction time was used as the sole metric in four studies^66,73,86,130 and used for assessment on a driving course (non-traffic)^73,86 and in two studies with simulated driving in a car.^66,130 Eighteen studies used BRT in combination with other measures, and most (n = 11) studies^66,81, 82, 83^{,89,102,109,122,}129, 130, 131. used the measure to assess performance in simulated driving (car). Three studies used a non-traffic driving course.^73,85 Laurell and Tornros⁹⁴ used both driving course (non-traffic) and simulated driving in a car with the BRT, whilst road driving was used by Leufkens et al.¹⁰⁰ Two studies by Carter et al.^77,78 used simulated driving with a computer.

Similarly, certain BZRAs have been studied more frequently using driving settings and driving impairment measures (Fig. 2, Fig. 3, Appendix IV). For example, zopiclone has been studied most using road traffic driving and the SDLP.

Fig. 2.

Driving impairment measures and benzodiazepine receptor agonists in experimental studies.

SDLP: Standard Deviation of Lateral Position; SDS: Standard Deviation of Speed; BRT: Brake Reaction Time; #RE: Number of Road Exits.

Fig. 3.

Driving settings and benzodiazepine receptor agonists in experimental studies.

4.8. Impairment

Authors' interpretation related to driving impairment, focusing on primary outcomes, was grouped into three categories: 1) no driving impairment identified; 2) driving impairment identified; and 3) mixed results – driving impairment identified with some measures. Seventy-eight of the 92 (85 %) studies had at least one measure that identified impairment. Forty-one studies found impairment in the study measures and 37 studies (Appendix III) were categorized as having mixed results, with impairment identified at least one or more, but not all measures. There were multiple measures used, except for four studies,^{73,130,137,151} where driving impairment was found.

No driving impairment was reported by the authors for the primary endpoint in 14 studies in a total of 2529 people.^73,77,78,81, 82, 83, 84^{,88,99,103,120,130,137,151} Six^{73,77,78,99,120,142} of these studies explored Z-drugs. Diazepam^83,103,151 and zolpidem^99,120,137 were the most common individual drugs studied in three studies each (Appendix IV).

4.9. BZRA dosing, dose timing, and duration of therapy

In experimental studies, the BZRA dosing compared to placebo (Appendix V) varied from 5 mg to 30 mg diazepam-equivalents, and the time between administration and impairment testing ranged from 0.5 h to 36 h among studies. The consecutive days of dosing was most often one (n = 34^66,68,73,76,78., 79, 80^{,87,89,91,92,98,100,}102, 103, 104^,108, 109, 110^{,112,113,121,122,126,127,129,}131., 132, 133^,142, 143, 144.^,146) and up to 30 days.⁷⁷

For studies comparing BZRAs to active comparators and placebo (Appendix VI), dosing of diazepam-equivalents ranged from 1.75 mg up to 30 mg. Törnros et al.¹³¹ included one participant with 60 mg diazepam-equivalents, but data were grouped and not presented for individual BZRAs in their analyses. The time between BZRA administration and impairment testing varied (e.g., 0.5 to 17 h), and similar to placebo-only comparative studies, the number of consecutive days of dosing was most often one (Appendix VI). Several studies (e.g., Törnros et al.,¹³¹ van der Sluiszen et al.,¹³⁴ Vinckenbosch et al.¹⁴⁹) included people who had been prescribed BZRAs for lengthy time periods (e.g. more than 6 months to “years”).

5. Review findings – Observational studies

5.1. Country and date of publication

Observational studies (Appendix VII) were conducted in 24 different countries from around the world with the majority being in Norway (Table 4).

Table 4.

Country of publication and number of observational studies^a.

Country	Number of Observational Studies
Norway	20
United States, Sweden	10
Canada	9
Australia	8
Finland	7
Denmark	6
Italy, Netherlands	5
Poland, UKb	4
Belgium, Spain, Taiwan	2
Slovakia, Switzerland, Jordan, Brazil, France, New Zealand, Greece, Iceland, Lithuania, Portugal	1

^aThe numbers do not sum to 91 because some studies involved more than one country.

^bUnited Kingdom was used as the code for studies from Scotland, England, Wales, and/or Northern Ireland.

Trends in publications per decade were similar to experimental studies. There were 33/91 (36 %) published between 2010 and 2019 inclusive (n = 26, 2000s; n = 13, 1990s; n = 4, 1980s; n = 3, 1970s) (Appendix V). Since 2020 to the time of the search, there were 12 observational studies published.

5.2. Population demographics

There was a total of 14,968,248 people in the observational studies. The sex was known for 11,483,838 participants with females comprising 44 % of participants and males were 56 %. Ages of participants in the observational studies ranged from 16 to 93 years old with a mean age of 35.9 years. There were eight analyses that included drivers 50 years of age and older.^{165,185,193,195,199,220,227}

5.3. Study design

There were 21 case-control,^{159,164,169,175,}180, 181, 182., 183^{,185,188,190,193,}197, 198., 199^{,211,213,215,216,220,221} 12 cohort,^23,165, 166, 167^{,191,195,211,}217, 218, 219.^,223,227 4 case-crossover^{160,220,223,241} studies, and the remainder were non-comparative (Appendix VII).

5.4. BZRAs studied

Most studies had multiple BZRAs represented in the data, prescription drugs, or alcohol as part of the assessment. When examining Z-drugs specifically, there were 18 publications with 20 studies.^{23,159,164,165,169,175,178,190,191,196,198,201,}208, 209, 210^,220,223,241

5.5. Driving impairment measures

The most common method reported in the observational studies was biological detection (e.g., urine, blood) (n = 73) (Appendix VII) followed by prescription drug records (e.g., dispensing records for prescriptions) (n = 17).^{23,160,169,175,185,193,199,}217, 218, 219., 220., 221^,223,227,241 A pill bottle review occurred in one study.¹⁶⁵

5.6. Driving impairment measures' and outcomes' trends

Unlike experimental studies, impairment was described in observational studies (Appendix VII) in multiple ways, and with many studies including multiple different outcomes. For example, motor vehicle accidents (n = 59), driving while impaired (DWI) suspected (n = 42), fatal MVA (n = 23)155, 156, 157., 158^,173,179, 180, 181^,183,187, 188., 189, 190^{,209,213,214,}226, 227, 228.^{,230,231,238,239} and 10 or fewer studies having others (e.g., accessing emergency services, hospitalization, injuries, etc). Of the 17^{23,160,169,175,185,193,199,}217, 218, 219., 220., 221^,223,227,241 studies using prescription records, all had MVA involvement (e.g., MVA, MVA with injury, fatal MVA, etc.). Of all the studies that used suspected DWI as at least one of their measures (n = 42), all but four^160,169,220 used biological detection.

6. Discussion

Driving performance has been assessed with 20 unique BZRAs across 183 studies in 178 publications based on the findings of this scoping review. The methods to assess driving impairment differed whether the study design was experimental or observational, which was expected. In experimental studies, the SDLP was the most used measure alone or in combination with other measures to examine BZRA impact on performance, whilst in observational studies, biological detection through analysis of bodily fluids was most common. The Netherlands was the country where the SDLP was most frequently used as at least one of the assessment measures. Generally, most experimental studies occurred in the Netherlands and observational studies were from Norway. Road driving in traffic and simulated car driving were used more often as the setting in experimental studies than alternatives. Most studies reported driving impairment with BZRAs.

6.1. BZRAs studied and prevalence of use in practice

Zopiclone was the most studied medication in experimental studies and investigated using the SDLP method. The representation of zopiclone in the literature is beneficial given it is commonly prescribed.^4,5,242,243 However, the number of times a BZRA was investigated for driving impairment was not always consistent with its frequency of clinical use in various regions representing a gap in the literature. For example, in Canada's largest province (Ontario; population over 14.7 million in 2020²⁴⁴), 14.5 % of patients who were prescribed a new BZRA in 2019 were given clonazepam.²⁴⁵ Clonazepam impairment was reported in two people in the included experimental studies.^99,149 Further, 71 % of Ontarians prescribed a new BZRA in 2019 received lorazepam, and in our review, the number of studies (13 of 92) for lorazepam was limited compared to its prescribing prevalence. Similarly, temazepam remains one of the most commonly prescribed medications for insomnia in Australia²⁴⁶ and yet it was studied in only eight experimental papers.

6.2. Heterogeneity of BZRA dosing and timing in protocols

There was considerable heterogeneity in the experimental study designs with respect to dosing protocols including the dose amount and diazepam dosing equivalency, timing of BZRA dose with respect to the timing of the assessments, variability in the type of assessment, and duration of therapy in relation to driving assessments. Each of these factors are important considerations in study design when considering whether medications were compared equally and when considering the pharmacodynamics and pharmacokinetics of individual BZRAs.

For dosing, the diazepam equivalents were different within and across many experimental studies thus limiting direct comparisons among agents and across studies. The amount given is relevant as sedative effects can be dose dependent. With respect to timing of the dose and the driving assessment, there was also considerable variability ranging from under an hour to more than a day for experimental studies. It was unclear in many of these studies as to whether relevant pharmacokinetic parameters such as time to peak drug concentration, peak drug concentration, elimination half-life, and other factors (e.g., active metabolites and their respective pharmacokinetic parameters) were considered when choosing the interval between dosing and time to conduct driving assessments. With observational studies, timing of dosing in relation to measuring outcomes was uncontrolled due to the nature of their design.

Most of the experimental studies did not have consecutive days of dosing (i.e., single dose studies), and very few were of longer duration (e.g., 30 days with Carter et al.,⁷⁷ 6 months with van der Sluiszen et al.¹³⁴). Therefore, considerations related to pharmacodynamic tolerance from repeated administration and its impact on driving ability is relatively understudied.

Additionally, there are also considerations with respect to pharmacokinetic and pharmacodynamic inter- and intra-subject variability in responses to medications. Partinen et al.,¹²⁰ for example, found no difference with mean time to collision in the primary outcome, and the general conclusion was that no impairment was found. However, there were considerable interindividual differences in the driving performance in the study. For example, there were numerous off-road accidents with one patient having five accidents after ingesting zolpidem and another patient had five accidents after taking temazepam.

For observational studies, similar issues of heterogeneity among studies were found with respect to doses, lengths of therapy, reliability of dosing, and the timing of reported outcomes in relation to dosing. Using the length of the exposure period to demonstrate, Gustavsen et al.²³ examined injuries during an exposure period of one week from dispensation of specific BZRAs (i.e., zopiclone, zolpidem, flunitrazepam, and nitrazepam), whereas Neutel et al.²¹⁹ reported injuries due to traffic accidents within 28 days of the index BZRA prescription for one of triazolam, flurazepam, oxazepam, lorazepam, or diazepam.

7. Knowledge translation and practical application

7.1. Measures of impairment

Through this scoping review process, the review team discussed the challenges with whether health care providers can readily understand and interpret the various measures and settings in which BZRA-related impairment is explored given the number of tests and measures. It also raised questions about how information from the various measures can translate to shared decision making with patients regarding the effects of BZRAs on driving. Specifically, how can results from measures, such as SDLP, SDS, BRT, and #REs, be translated into language and advice that is relevant to both health care providers and patients? The practicalities of knowledge translation become important when health care providers must advise and counsel patients on medication use. Notably, organizations like the Canadian Council of Motor Transport Administrators acknowledge that physicians, pharmacists, and nurses can play key roles in the education regarding medications' effects on driving performance.²⁴⁷ Within the profession of pharmacy, the International Pharmaceutical Federation (FIP) has published guidelines on medications affecting driving performance²⁴⁸ and indicate, among several recommendations, that clarifications on manufacturers' warnings about any impacts on driving performance is conveyed and patients' experiences of driving while taking the medication are monitored. Additionally, FIP indicates patients should be informed by pharmacists of how to detect side effects such as psychomotor impairment. No information is provided on when and how this assessment should be completed in terms of factors that may be related to psychomotor impairment. Aside from limited information available from research published using simulated patients and pharmacy encounters, the advice given to patients by pharmacists is not adequately documented. Similarly, there is little to no objective evidence that reports other health care prescribers' advice giving. Health care providers cannot conduct tests that were encountered in this review with patients, and a referral pathway to such services would be required. It is unlikely that access to such services that conduct these tests would be readily available to match the prevalence of BZRA prescribing across the population.

7.2. Comparison to alcohol

One approach that has been used to describe BZRA impairment with extrapolating from various measures in studies, is to relate or compare impairment caused by BZRAs against alcohol. Alcohol impairment as a concept is likely more recognizable by health care providers and patients; although for patients, inquiring about their alcohol use is important as alcohol patterns of consumption can change with age and time.^249,250 The International Council on Alcohol, Drugs and Traffic Safety (ICADTS) System for Medicinal Drugs Affecting Driving Performance,²⁵¹ categories can be leveraged in this area. The ICADTS,²⁵¹ published in 2007, has categories for medications based on experiments using on-the-road driving tests in the Netherlands and makes comparisons with alcohol. These categories include category I as equivalent to blood alcohol concentration (BAC) of <0.5 g/L (0.05 %), category II with 0.5–0.8 g/L (0.05–0.08 %), and III as BAC >0.8 g/L (>0.08 %). Per Verster and Mets,⁴² Category I drugs are, “presumed to be safe or unlikely to produce an effect”, with advice to patients to, “be careful not to drive before having read the warnings in the package insert”; Category II are, “likely to produce minor or moderate effects”, with advice to the patient as, “do not drive without consulting a healthcare professional about the possible impairing effects”; and lastly, Category III are, “likely to produce severe effects or presumed to be potentially dangerous”, with advice to the patient including, “do not drive when this drug is taken and consult a healthcare professional when to start driving again after evaluation of the treatment outcomes.”⁴² Verster and Mets⁴² provided additional information incorporating SDLP measures alongside levels of impairment based on BAC using the work of Louwerens presented at the International Conference on Alcohol, Drugs and Traffic Safety.²⁵² The SDLP increments after alcohol consumption (BAC) were + 2.4 cm (0.05 %), +4.1 cm (0.08 %), and + 5.3 cm (0.10 %). Some publications in our scoping review applied these comparisons to enhance their interpretation of their results. For example, Jongen et al. 2018⁹⁰ used the same information from Louwerens et al.²⁵² and related their SDLP results to BACs in their discussion of their results. They indicated that after administration of oxazepam 10 mg, there was a significant SDLP increase of 1.83 cm, which would correspond to minor impairment (i.e. a BAC of <0.5 g/L). After diazepam 10 mg, the SDLP increased with 3.03 cm, corresponding to moderate impairment (i.e. BAC 0.5–0.8 g/L), and after oxazepam 30 mg, SDLP increased with 7.57 cm, corresponding to severe impairment (i.e. BAC of >1.0 g/L).⁹⁰ Further, Verster and Mets⁴² also report the various BZRAs as per the ICADTS categories, with the majority falling in category III, which would indicate that studies of road tests produce results of a BAC >0.08 % for most BZRAs. Only three medications (zolpidem, medazepam, clobazam) were listed in category II. Based on the observational studies retrieved from this scoping review, there is prevalent use of category II and III medications while driving. While the advice to patients in these categories provides clear guidance to avoid driving without consultation of a healthcare providers, the onus is on the health care provider to determine when driving may be safe again. Using knowledge related to BAC, ICADTS categories, and measures such as SDLP may be an appropriate reference for prescribers with patients to explain impairment. These comparisons to SDLP are useful, but as seen in this review, it is not the only measure of assessing impairment extracted from studies, and although it was common, it was not used in all experimental studies.

8. Conclusion

The effects of individual benzodiazepine receptor agonists (BZRAs) on people's motor vehicle driving performance have been assessed using heterogenous protocols with multiple approaches to dosing and timing of testing, measures, and settings that ranged from simulated to authentic traffic situations. The scoping review findings indicate that many BZRAs were studied but representation in the studies does not always match with prevalence in prescribing patterns. The interpretation and contextualization of results for use in clinical practice is challenging due to the complexity of protocols, measures, and variety of settings used to assess performance. Comparisons to other substances such as alcohol may be valuable, but challenges remain with limited driving assessments compared against BAC. Future work in this area should seek to improve knowledge translation of results so that the information is more readily accessible and applicable to health care providers and patients in shared decision-making processes. The findings will also serve to inform future syntheses regarding BZRA-related impairment assessments and driving.

Funding

The research is unfunded. Student research support was provided to K.S. by the Drug Evaluation Alliance of Nova Scotia (DEANS). The funder providing the student research support was not involved in the design, data collection and analysis, decision to publish, or preparation of the manuscript.

CRediT authorship contribution statement

Andrea L. Murphy: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Korolos Sawires: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Methodology, Formal analysis, Data curation. Sophie M. Peltekian: Writing – review & editing, Writing – original draft, Visualization, Validation, Investigation, Formal analysis, Data curation. Melissa Helwig: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Methodology, Formal analysis, Data curation. Marilyn Macdonald: Writing – review & editing, Writing – original draft, Methodology, Formal analysis, Data curation. Ruth Martin-Misener: Writing – review & editing, Writing – original draft, Methodology, Data curation. Bandana Saini: Writing – review & editing, Writing – original draft, Methodology, Data curation. Heather Neyedli: Writing – review & editing, Writing – original draft, Methodology, Data curation. Chris Giacomantonio: Writing – review & editing, Writing – original draft, Methodology, Data curation. David M. Gardner: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix 1. Inclusion and exclusion criteria

Inclusion criteria	Exclusion criteria
BZRAs
Study participants taking a BZRA	No BZRAs Studied (Study did not include any BZRAs), No Single-Entity BZRA use (i.e., exclude those exclusively co-administration cases with no single entity BZRA use, or unclear if mono-use of BZRAs) Wrong route of administration (i.e., intravenous, intramuscular). BZRA must be regulated in one or more countries (i.e. exclude “designer” BZRAs for illicit use).
Participants
Of legal age to drive (e.g., 15 & up)	Wrong Patient Population a. Not of legal age to be driving b. Study subjects in hospital/hospitalized for reasons other than motor vehicle accident or residing in long-term care or assisted living facilities
Study design: (i) experimental design (randomized or non-randomized prospective study) with one or more study groups; (ii) observational/epidemiological design (cross-sectional study, ecological study, time-series analysis, case control study, cohort study, case series, surveillance, vital statistics) conducted prospectively or retrospectively; (iii)*Systematic Reviews, Scoping Review, Rapid Review, Meta-Analysis (*citation chasing)	Wrong Study Design a. Studies that measure skills indirectly related to driving (i.e. flicker fusion test, EEG, reaction time tests, hand and foot proprioception). b. Drug-detection sensitivity studies in which impaired performance was not the focus of the study c. Use of non-standard subjective measures and non-objective self-reported measures of driving impairment d. Random sampling of drivers (no suspected impairment) Wrong Publication Type (exclude anecdotes, testimonials, editorials, commentary, conference abstracts, unpublished reports, letters, narrative reviews)
Context for driving assessment
Operating/driving a vehicle in real or simulated conditions and assessment of driving performance, driving score (includes crashes, ER visits, sobriety test, police ride checks, etc).	Wrong Setting/Context a. Overdose/Toxicity b. Recreational vehicles (ATV, Bicycles, Boats, etc) c. Not in the context of driving
Language
English language publications	Non-English Language

Appendix 2. Search strategy

1. MEDLINE (Ovid)

(May 3, 2022)

(((Automobile Driving/ or motor vehicles/ or automobiles/ or motorcycles/ or (Automobile* or car or cars or truck* or motorcycle* or motor cycle* or motor vehicle* or vehicle* or traffic or driver* or driving or crash or crashes or crashing or collision*).ti,ab,kw,kf.) and ((Impair* or drug* or intoxicat* or under the influence or somnolen* or fatigue* or sleep* or drows* or hungover or hangover* or drunk*).ti,ab,kw,kf. or Fatigue/) and (“hypnotics and sedatives”/ or sleep aids, pharmaceutical/ or Tranquilizing Agents/ or Anti-Anxiety Agents/ or benzodiazepines/ or alprazolam/ or bromazepam/ or clonazepam/ or exp. diazepam/ or flunitrazepam/ or flurazepam/ or lorazepam/ or nitrazepam/ or oxazepam/ or prazepam/ or temazepam/ or chlordiazepoxide/ or clobazam/ or clorazepate dipotassium/ or estazolam/ or medazepam/ or midazolam/ or triazolam/ or Eszopiclone/ or Nordazepam/ or Zolazepam/ or Zolpidem/ or exp. GABA Agonists/ or (Sleep medication* or sleep aid* or hypnotic drug* or hypnotic agent* or sleep promoting agent* or sedative-hypnotic* or benzodiazepine* or benzodiazepinone* or anxiolytic*).ti,ab,kw,kf. or ((anxiolytic* or anti anxiety or anti-anxiety or antianxiety or ataractic*) adj2 (agent* or drug*)).ti,ab,kw,kf. or (minor adj2 tranquiliz*).ti,ab,kw,kf. or ((Benzodiazepine or gaba) adj3 (agonist* or receptor* or stimulant* or stimulator* or stimulating or agent*)).ti,ab,kw,kf. or (benzodiazepine adj2 derivative*).ti,ab,kw,kf. or (alprazolam or bromazepam or brotizolam or clonazepam or cloxazolam or chlordiazepoxide or clobazam or chlorazepate or cinolazepam or clotiazepam or delorazepam or diazepam or estazolam or eszopiclone or etizolam or flunitrazepam or fludiazepam or flutazolam or flutoprazepam or flurazepam or halazepam or haloxazolam or Ketazolam or loprazolam or lorazepam or lormetazepam or midazolam or medazepam or mexazolam or nitrazepam or nimetazepam or nordazepam or oxazepam or Oxazolam or prazepam or phenazepam or pinazepam or quazepam or temazepam or tetrazepam or triazolam or tofisopam or zolazepam or zapizolam or zopiclone or zaleplon or zolpidem or bzra).ti,ab,kw,kf.)) or (((Driver* or driving) adj3 (health or behavio?r* or abilit* or perform* or accident* or crash or crashes or crashing or injur* or collision* or fatalit* or safety or risk* or fatigue* or attention or inattention or sleep* or drows* or alert or alertness or fitness or assess* or evaluat*)).ti,ab,kw,kf. or ((Road* or highway or motorway or traffic) adj3 (accident* or crash or crashes or crashing or injur* or collision* or fatalit* or safety or risk* or fatigue* or attention or inattention or sleep* or drows* or alert or alertness or fitness or assess* or evaluat*)).ti,ab,kw,kf or (fitness to drive or fitness-to-drive).ti,ab,kw,kf. or driving under the influence/ or Accidents, Traffic/) and (“hypnotics and sedatives”/ or sleep aids, pharmaceutical/ or Tranquilizing Agents/ or Anti-Anxiety Agents/ or benzodiazepines/ or alprazolam/ or bromazepam/ or clonazepam/ or exp. diazepam/ or flunitrazepam/ or flurazepam/ or lorazepam/ or nitrazepam/ or oxazepam/ or prazepam/ or temazepam/ or chlordiazepoxide/ or clobazam/ or clorazepate dipotassium/ or estazolam/ or medazepam/ or midazolam/ or triazolam/ or Eszopiclone/ or Nordazepam/ or Zolazepam/ or Zolpidem/ or exp. GABA Agonists/ or (Sleep medication* or sleep aid* or hypnotic drug* or hypnotic agent* or sleep promoting agent* or sedative-hypnotic* or benzodiazepine* or benzodiazepinone* or anxiolytic*).ti,ab,kw,kf. or ((anxiolytic* or anti anxiety or anti-anxiety or antianxiety or ataractic*) adj2 (agent* or drug*)).ti,ab,kw,kf. or (minor adj2 tranquiliz*).ti,ab,kw,kf. or ((Benzodiazepine or gaba) adj3 (agonist* or receptor* or stimulant* or stimulator* or stimulating or agent*)).ti,ab,kw,kf. or (benzodiazepine adj2 derivative*).ti,ab,kw,kf. or (alprazolam or bromazepam or brotizolam or clonazepam or cloxazolam or chlordiazepoxide or clobazam or chlorazepate or cinolazepam or clotiazepam or delorazepam or diazepam or estazolam or eszopiclone or etizolam or flunitrazepam or fludiazepam or flutazolam or flutoprazepam or flurazepam or halazepam or haloxazolam or Ketazolam or loprazolam or lorazepam or lormetazepam or midazolam or medazepam or mexazolam or nitrazepam or nimetazepam or nordazepam or oxazepam or Oxazolam or prazepam or phenazepam or pinazepam or quazepam or temazepam or tetrazepam or triazolam or tofisopam or zolazepam or zapizolam or zopiclone or zaleplon or zolpidem or bzra).ti,ab,kw,kf.)) not (exp Animals/ not Humans/)

Appendix 3. Experimental studies

Year of publication	First author	Experimental study designa	number of subjects	number of females	Mean age (years)	Age range (years)	Consecu-tive days of BZRA use (Yes/No)	Outcome measures	Authors' conclusions regarding BZRA and driving impairment
Single driving impairment assessment setting
Road Driving (traffic)
2019	Vermeeren, A	R, DB, PC, AC, CO	48	22	58.5	21–65	No	Standard deviation of lateral position (SDLP)	Driving impairment identified
2018	Jongen, S	R, DB, PC, CO	23	11	36.8		No	Standard deviation of lateral position (SDLP)	Driving impairment identified
2018	Verster, J	R, DB, PC, AC, CO	60	30	24	21–65	No	# Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs) Lapses	Driving impairment identified
2016	Vermeeren, A	R, DB, PC, CO	24	14	68.8	65–76	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Driving impairment identified
2015	Vermeeren, A	R, DB, PC, AC, CO	28	15	45.6	21–64	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Driving impairment identified
2014	Leufkens, T	R, DB, PC, CO	48	22	62.6	52–71	No	Standard deviation of lateral position (SDLP) Braking reaction time (BRT) Standard deviation of speed (SDS) Gain	Mixed results - driving impairment identified with some measures
2014	Motak, L	R, DB, PC, CO	16	–	29.69	25–35	No	Standard deviation of lateral position (SDLP) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Driving impairment identified
2014	Leufkens, T	R, DB, PC, CO	101	51	39.6	21–73	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Driving impairment identified
2014	Verster, J	R, DB, PC, AC, CO	60	30	24	21–65	No	Standard deviation of lateral position (SDLP) Lapses	Driving impairment identified
2011	Vermeeren, A	R, DB, PC, CO	40	20	37.3	21–64	No	Standard deviation of lateral position (SDLP)	Mixed results - driving impairment identified with some measures
2011	Mets, M	R, DB, PC, CO	30	15	–	21–55	No	Standard deviation of lateral position (SDLP) Average lateral position (ALP) Standard deviation of speed (SDS) Mean speed (MS)	Mixed results - driving impairment identified with some measures
2011	Ramaekers, JG	R, DB, PC, AC, CO	32	16	33	–	No	Standard deviation of lateral position (SDLP)	Driving impairment identified
2009	Leufkens, T	R, DB, PC, CO	18	10	64.3	55–75	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Mixed results - driving impairment identified with some measures
2009	Leufkens, T	R, DB, PC, AC, CO	25	12	31.4	22–44	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Driving impairment identified
2007	Leufkens, T	R, DB, PC, CO	18	9	32.3	20–45	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Mixed results - driving impairment identified with some measures
2002	Verster, J	R, DB, PC, CO	30	15	24	22–26	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Mixed results - driving impairment identified with some measures
2002	Verster J	R, DB, PC, CO	20	12	25	–	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS) Mean speed (MS) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Driving impairment identified
2002	Vermeeren, A	R, DB, PC, AC, CO	30	15	31.6	21–45	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Mixed results - driving impairment identified with some measures
2001	van Laar, M	R, DB, PC, CO	18	–	28	25–36	Yes	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Driving impairment identified
1998	Vermeeren, A	R, DB, PC, CO	24	–	29	22–45	Yes	Standard deviation of lateral position (SDLP) Average reaction time (ART) Standard deviation of speed (SDS)	Driving impairment identified
1998	Vermeeren, A	R, DB, PC, AC, CO	28	14	31	23–40	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Mixed results - driving impairment identified with some measures
1995	O'Hanlon, JF	R, DB, PC, Parallel	56	36	43	24–64	Yes	Standard deviation of lateral position (SDLP)	Driving impairment identified
1993	O'Hanlon, JF	R, DB, PC, Parallel	56	–	–	21–65	Yes	Standard deviation of lateral position (SDLP)	Driving impairment identified
1986	O'Hanlon, JF	R, DB, PC, CO	11	–	–	26–38	Yes	Standard deviation of lateral position (SDLP)	Mixed results - driving impairment identified with some measures
1986	Schmidt, U	R, DB, AC, Parallel	32	12	–	–	Yes	Steering angle velocity Optimization Quotient	Mixed results - driving impairment identified with some measures
1982	O'Hanlon, JF	R, DB, PC, CO	9	–	–	24–34	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS) Mean speed (MS)	Mixed results - driving impairment identified with some measures
1979	Biehl, B	R, DB, PC, Parallel	24	–	–	18–24	Yes	Observer rated behavioural measures of driver impairment	Mixed results - driving impairment identified with some measures
2021	Vinckenbosch, FRJ	NR, Open study, PC, parallel	126	–	55	21–75	–	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Driving impairment identified
2019	van der Sluiszen, N	NR, Open study, Parallel	109	54	56.2	–	–	Standard deviation of lateral position (SDLP)	Mixed results - driving impairment identified with some measures
2014	Leufkens, T	NR, Open study, Parallel	63	29	61.6	50–75	No	Standard deviation of lateral position (SDLP) Average reaction time (ART) Standard deviation of speed (SDS) Mean speed (MS) Gain	No driving impairment identified
1995	O'Hanlon, JF	NR, DB, PC, CO	16	8	34	25–43	Yes	Standard deviation of lateral position (SDLP)	Driving impairment identified
1995	O'Hanlon, JF	NR, DB, PC, CO	18	9	25	22–34	Yes	Standard deviation of lateral position (SDLP Average reaction time (ART) Standard deviation of following distance (SDFD)	Driving impairment identified
1995	Vermeeren, A	NR, DB, PC, AC, CO	17	–	40.8	25–51	No	Standard deviation of lateral position (SDLP)	No driving impairment identified
1992	vanLaar, M	NR, DB, PC, Parallel	24	12	40	18–50	Yes	Standard deviation of lateral position (SDLP),Standard deviation of speed (SDS)	Driving impairment identified
1990	Brookhuis, KA	NR, DB, PC, CO	16	10	–	26–41	Yes	Standard deviation of lateral position (SDLP)	Mixed results - driving impairment identified with some measures
1981	deGier, JJ	NR, Open study, Parallel	22	–	43.1	–	No	110 item driving behaviour checklist	Driving impairment identified
Driving Course (non-traffic)
2008	Boyle, J	R, DB, PC, CO	64	25	27	19–47	No	Braking reaction time (BRT)	No driving impairment identified
2005	Hindmarch, I	R, DB, PC, AC, CO	23	11	29	–	Yes	Braking reaction time (BRT)	Mixed results - driving impairment identified with some measures
1999	Mercier-Guyon, C	R, DB, AC, CO	16	–	40	29–44	Yes	Gap estimation/acceptance task Weaving task	Mixed results - driving impairment identified with some measures
1980	Hindmarch, I	R, TB, PC, CO	12	–	34	26–40	Yes	Braking reaction time (BRT) Gap estimation/acceptance task Weaving task Parking task	Mixed results - driving impairment identified with some measures
1979	Wetherell, A	R, DB, PC, Parallel	20	–	25.5	19–41	No	Gap estimation/acceptance task	No driving impairment identified
1972	Betts, TA	R, DB, PC	100	50	–	18–30	No	Gap estimation/acceptance task, Weaving task Parking task	Driving impairment identified
1982	Betts, TA	NR, DB, PC, CO	12	–	–	–	No	Gap estimation/acceptance task Weaving task	Driving impairment identified
1977	Hindmarch, I	NR, DB, PC, CO	10	5	27	–	Yes	Gap estimation/acceptance task Weaving task Parking task	No driving impairment identified
Simulated driving (car)
2022	Su, S	R, DB, PC, AC, CO	100	41	38.1	22–59	No	Standard deviation of lateral position (SDLP) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs) # of collisions	Driving impairment identified
2022	Torres, R	R, DB, PC, AC, CO	48	20	–	21–55	No	Standard deviation of lateral position (SDLP) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Driving impairment identified
2022	Muehlan, C	R, DB, PC, AC, CO	60	30	64.6	–	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS) Mean speed (MS)	Driving impairment identified
2022	Iwamoto, K	R, DB, PC,CO	28	–	33.1	22–59	No	Standard deviation of lateral position (SDLP) Braking reaction time (BRT) Standard deviation of speed (SDS) Following distance coefficient of variation (CV) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs) # of collisions	Mixed results - driving impairment identified with some measures
2021	Moller, HJ	R, DB, PC, AC, CO	25	14	33.1	25–55	No	Standard deviation of lateral position (SDLP) Average reaction time (ART) # of collisions	Driving impairment identified
2019	Huizinga, RH	R, DB, PC, CO	24	12	26	20–43	No	Standard deviation of lateral position (SDLP)	Driving impairment identified
2018	Brown, TL	R, DB, PC, CO	8	4	30	–	No	Standard deviation of lateral position (SDLP) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Driving impairment identified
2017	Kay, GG	R, DB, PC, CO	72	–	32.4	18–50	No	Standard deviation of lateral position (SDLP) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Driving impairment identified
2016	Kay, GG	R, DB, PC, CO	59	13	41.1	25–55	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Driving impairment identified
2014	Miyata, A	R, DB, PC, CO	17	–	34.1	23–44	No	Standard deviation of lateral position (SDLP) Braking reaction time (BRT) Following distance coefficient of variation (CV) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Mixed results - driving impairment identified with some measures
2013	Brown, T	R, DB, PC, Parallel	7	–	–	19–50	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS)	Mixed results - driving impairment identified with some measures
2013	Sarris, J	R, DB, PC, CO	22	15	33	18–65	No	Standard deviation of lateral position (SDLP) Braking reaction time (BRT) Standard deviation of speed (SDS) Lapses # of collisions	Mixed results - driving impairment identified with some measures
2010	Takahashi, M	R, DB, PC, CO	18	–	37.1	32–44	No	Standard deviation of lateral position (SDLP) Braking reaction time (BRT) Following distance coefficient of variation (CV)	Mixed results - driving impairment identified with some measures
2008	Otmani, S	R, DB, PC, CO	16	4	59.4	55–65	Yes	Standard deviation of speed (SDS) Mean speed (MS) # of collisions	Mixed results - driving impairment identified with some measures
2003	Partinen, M	R, DB, PC, CO	23	–	49.5	35–60	No	Standard deviation of lateral position (SDLP) Average lateral position (ALP) Average reaction time (ART) Mean time to collision (MTTC)	No driving impairment identified
2000	Vanakoski, J	R, DB, PC, CO	9	4	–	22–24	No	Cumulative reaction time (CRT) Tracking error severity index	Driving impairment identified
2000	Vanakoski, J	R, DB, PC, CO	9	3	–	55–77	No	Cumulative reaction time (CRT) Tracking error severity index	Mixed results - driving impairment identified with some measures
1994	Mattila, MJ	R, DB, PC, CO	12	6	–	19–32	No	Cumulative reaction time (CRT) Tracking error severity index	Driving impairment identified
1993	Mattila, MJ	R, DB, PC, CO	12	6	–	22–26	No	# Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs) Tracking error severity index	No driving impairment identified
1990	Tornros, J	R, DB, PC, CO	18	–	–	20–35	Yes	Braking reaction time (BRT)	No driving impairment identified
1986	Smiley, A	R, DB, PC, Parallel	48	24	–	21–40	Yes	Lane position variability	Driving impairment identified
1984	Willumeit, H	R, DB, PC, CO	16	6	26.4	20–33	No	Average reaction time (ART)	Driving impairment identified
1984	Willumeit, H	R, DB, PC, CO	28	–	–	–	Yes	Average reaction time (ART)	Mixed results - driving impairment identified with some measures
1978	Dureman, I	R, DB, PC, CO	12	–	–	19–30	Yes	Braking reaction time (BRT) Number of errors in car simulator	No driving impairment identified
1975	Dureman, I	R, DB, PC, CO	42	–	–	19–29	No	Braking reaction time (BRT) Mean speed (MS) Steering precision	No driving impairment identified
1974	Berry, P A	R, DB, PC and AC, Parallel	6	–	–	25–44	No	Braking reaction time (BRT)	Driving impairment identified
2015	Stone, BT	NR, DB, PC, CO	19	6	25.3	18–38	No	Standard deviation of lateral position (SDLP) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Driving impairment identified
2011	Bocca, ML	NR, DB, PC, AC, CO	16	8	60.3	55–65	No	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Mixed results - driving impairment identified with some measures
2009	Meskali, M	NR, DB, PC, AC, CO	16	8	60.3	55–65	No	Average lateral position (ALP) Mean speed (MS) # of collisions	Mixed results - driving impairment identified with some measures
2001	Törnros, J	NR, Open study, PC, AC, Parallel	40	10	42	28–55	No	Standard deviation of lateral position (SDLP) Braking reaction time (BRT) Standard deviation of speed (SDS)	Mixed results - driving impairment identified with some measures
1999	Bocca, ML	NR, DB, PC, AC, CO	16	7	24.5	20–30	No	Variance of lateral position (VLP) Variance of velocity (VV)	Mixed results - driving impairment identified with some measures
1991	Friedel, B	NR, Open study, Parallel	60	–	–	22–26	No	Braking reaction time (BRT) Mean speed (MS) Following distance coefficient of variation (CV) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs) # of collisions	No driving impairment identified
1982	Moskowitz, H	NR, DB, PC, Parallel	48	24	–	21–40	Yes	Standard deviation of lateral position (SDLP) Standard deviation of speed (SDS) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Driving impairment identified
1978	Palva, ES	NR, DB, PC, CO	38	21	22.9	20–29	Yes	Cumulative reaction time (CRT) Co-ordination test	Mixed results - driving impairment identified with some measures
1974	Linnoila, M	NR, DB, PC, Parallel	70	–	–	19–22	No	Braking reaction time (BRT) Mean speed (MS) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs) # of collisions	Mixed results - driving impairment identified with some measures
1973	Linnoila, M	NR, DB, PC, Parallel	90	–	–	19–21	No	# Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs), # of collisions Instructions ignored by driver	Mixed results - driving impairment identified with some measures
Simulated driving (computer)
2020	Carter, SG	R, DB, PC, CO	28	7	45	18–65	No	Standard deviation of lateral position (SDLP) Braking reaction time (BRT) Standard deviation of speed (SDS) % of individuals that crashed	No driving impairment identified
2017	Carter	R, DB, PC, Parallel	30	5	48	18–65	Yes	Standard deviation of lateral position (SDLP) Braking reaction time (BRT) Standard deviation of speed (SDS)	No driving impairment identified
2002	Ludice, A	R, DB, PC, CO	12	7	31	27–38	Yes	Speed exceedances Length of run, infractions, time to collision	No driving impairment identified
1998	Mattila, MJ	R, DB, PC, CO	12	5	–	21–28	No	Cumulative reaction time (CRT) Tracking error severity index	Driving impairment identified
1994	Kuitunen, T	R, DB, PC, CO	24	11	–	20–28	No	Cumulative reaction time (CRT) Tracking error severity index	Driving impairment identified
1986	Subhan, Z	R, DB, CO	12	–	–	28–45	Yes	Average reaction time (ART)	Mixed results - driving impairment identified with some measures
1976	Seppala, T	R, DB, PC, CO	37	–	–	–	Yes	Choice reaction test: Mistake Percentage	Driving impairment identified
2003	Berthelon, C	NR, DB, PC, AC, CO	10	6	28	23–42	No	Error rates (ERs) response times (RTs) points of subjective equalization (PESs)	Mixed results - driving impairment identified with some measures
1983	Willumeit, HP	NR, DB, PC, AC, CO	12	1	–	21–30	Yes	Cumulative reaction time (CRT) Choice reaction test	Mixed results - driving impairment identified with some measures
Multiple driving impairment assessment settings
2013	Daurat, A	R, DB, PC, CO	14	–	29.79	25–35	No	Standard deviation of lateral position (SDLP) # Road exits (REs) or # Inappropriate lane crossings (ILCs) or # lane departures (LDs)	Driving impairment identified
1992	Volkerts E.R	R, DB, PC, CO	18	–	26.3	25–31	Yes	Standard deviation of lateral position (SDLP) Average reaction time (ART) Standard deviation of speed (SDS)	Mixed results - driving impairment identified with some measures
1986	Laurell, H	R, DB, PC, CO	18	–	–	20–34	Yes	Braking reaction time (BRT), Weaving task	Mixed results - driving impairment identified with some measures

^aAC: active control; DB: double blind; CO: cross-over; NR: non-randomized; PC: placebo controlled; R: randomized.

Appendix 4. Experimental studies' impairment measures and medications

		Impairment measures					Medications
Author, Year	Country	SDLP	SDS	RE	BRT	Other	ZC	DZ	ZD	LO	AL	LT	OX	TM	FT	FL	NT	TR	ZA	BT	CX	CB	CR	BR	CL	EZ	MX
Road Driving (traffic)
Randomized controlled trials
Vermeeren, A 2019	Netherlands	✓					✓
Jongen, S 2018	Netherlands	✓						✓					✓
Verster, J 2018	Netherlands			✓		✓	✓		✓										✓
Vermeeren, A 2016	Belgium	✓	✓				✓
Vermeeren, A 2015	Netherlands	✓	✓				✓
Leufkens, T 2014	Netherlands	✓	✓		✓	✓	✓
Motak, L 2014	France	✓		✓						✓
Leufkens, T 2014	Netherlands	✓	✓				✓
Verster, J 2014	Netherlands	✓				✓	✓		✓										✓
Vermeeren, A 2011	Netherlands	✓					✓		✓
Mets, M 2011	Netherlands	✓	✓			✓	✓
Ramaekers, JG 2011	Netherlands	✓					✓
Leufkens, T 2009	Netherlands	✓	✓				✓							✓
Leufkens, T 2009	Netherlands	✓	✓				✓		✓
Leufkens, T 2007	Netherlands	✓	✓								✓
Verster, J 2002	Netherlands	✓	✓						✓										✓
Verster J 2002	Netherlands	✓	✓	✓							✓
Vermeeren, A 2002	Netherlands	✓	✓				✓												✓
van Laar, M 2001	Netherlands	✓	✓	✓						✓
Vermeeren, A 1998	Netherlands	✓	✓													✓
Vermeeren, A 1998	Netherlands	✓	✓				✓												✓
O'Hanlon, JF 1995	Netherlands	✓								✓
O'Hanlon, JF 1993	Netherlands	✓								✓
O'Hanlon, JF 1986	Netherlands	✓												✓			✓
Schmidt, U 1986	Germany					✓								✓	✓
O'Hanlon, JF 1982	Netherlands	✓	✓			✓		✓
Biehl, B 1979	Germany					✓		✓														✓
Non-randomized controlled trials
Vinckenbosch, F.R.J. 2021	Netherlands	✓	✓				✓	✓	✓	✓	✓	✓	✓	✓			✓			✓			✓		✓		1
vanderSluiszen, N 2019	Netherlands	✓					✓	✓	✓	✓	✓	✓	✓	✓			✓			✓				✓			2
Leufkens, T 2014	Netherlands	✓	✓			✓	✓		✓	✓		✓	✓	✓		✓	✓								✓		3
O'Hanlon, JF 1995	Netherlands	✓						✓
O'Hanlon, JF 1995	Netherlands	✓				✓				✓
Vermeeren, A 1995	Netherlands	✓							✓						✓
vanLaar, M 1992	Netherlands	✓	✓					✓
Brookhuis, KA 1990	Netherlands	✓										✓				✓
deGier, JJ 1981	Netherlands					✓		✓
Driving course (non-traffic)
Randomized controlled trials
Boyle, J 2008	United Kingdom				✓																					✓
Hindmarch, I 2005	United Kingdom				✓						✓
Mercier-Guyon, C 1999	France					✓				✓
Hindmarch, I 1980	United Kingdom				✓	✓				✓												✓
Wetherell, A 1979	United Kingdom					✓		✓
Betts, TA 1972	United Kingdom					✓															✓
Non-randomized controlled trials
Betts, TA 1982	United Kingdom					✓								✓		✓						✓
Hindmarch, I 1977	United Kingdom					✓
Simulated driving (car)
Randomized controlled trials
Su, S 2022	United States & Canada	✓		✓		✓					✓
Torres, R 2022	United States	✓		✓			✓
Muehlan, C 2022	Netherlands	✓	✓			✓	✓
Iwamoto, K 2022	Japan	✓	✓	✓	✓	✓	✓
Moller, HJ 2021	Germany	✓				✓				✓
Huizinga, RH 2019	Netherlands	✓									✓
Brown, TL 2018	United States	✓		✓							✓
Kay, GG 2017	Canada	✓		✓			✓
Kay, GG 2016	United States	✓	✓	✓														✓
Miyata, A 2014	Japan	✓		✓	✓	✓												✓
Brown, T 2013	United States	✓	✓															✓
Sarris, J 2013	Australia	✓	✓		✓	✓							✓
Takahashi, M 2010	Japan	✓			✓	✓		✓
Otmani, S 2008	France		✓			✓			✓
Partinen, M 2003	Finland	✓				✓			✓					✓
Vanakoski, J 2000	Finland					✓		✓
Vanakoski, J 2000	Finland					✓		✓
Mattila, MJ 1994	Finland					✓	✓
Mattila, MJ 1993	Finland			✓		✓		✓
Tornros, J 1990	Sweden				✓												✓			✓
Smiley, A 1986	United States					✓		✓
Willumeit, H 1984	Germany					✓		✓				✓
Willumeit, H 1984	Germany					✓		✓				✓				✓
Dureman, I 1978	Sweden				✓	✓																	✓
Dureman, I 1975	Sweden				✓	✓		✓															✓
Berry, PA 1974	United Kingdom				✓																✓
Non-randomized controlled trials
Stone, BT 2015	United States	✓		✓							✓
Bocca, ML 2011	France	✓	✓	✓			✓		✓						✓
Meskali, M 2009	France					✓	✓		✓						✓
Törnros, J 2001	Sweden	✓	✓		✓
Bocca, ML 1999	France					✓	✓		✓						✓
Friedel, B 1991	Germany			✓	✓	✓		✓
Moskowitz, H 1982	United States	✓	✓	✓				✓
Palva, ES 1978	Finland					✓		✓					✓								✓
Linnoila, M 1974	Finland			✓	✓	✓		✓
Linnoila, M 1973	Finland			✓		✓		✓
Simulated driving (computer)
Randomized controlled trials
Carter, SG 2020	Australia	✓	✓		✓	✓	✓
Carter, SG 2017	Australia	✓	✓		✓		✓
Ludice, A 2002	Italy					✓						✓
Mattila, MJ 1998	Finland					✓	✓	✓	✓				✓
Kuitunen, T 1994	Finland					✓	✓	✓										✓
Subhan, Z 1986	United Kingdom					✓				✓	✓
Seppala, T 1976	Finland					✓																		✓
Non-randomized controlled trials
Berthelon, C 2003	France					✓	✓		✓						✓
Willumeit, H P 1983	Germany					✓						✓				✓
Multiple driving impairment assessment settings
Randomized controlled trials
Daurat, A 2013a	France	✓		✓						✓
Volkerts E.R 1992b	Netherlands	✓	✓			✓						✓	✓
Laurell, H 1986c	Sweden				✓	✓											✓	✓
Total count		54	30	19	18	49	30	24	16	13	10	9	8	8	6	6	6	5	5	3	3	3	3	2	2	1	1

Table abbreviations: Impairment measures: SDLP: standard deviation of lateral position; SDS: standard deviation of speed; RE: Number of road exits; BRT: brake reaction time; ZC: zopiclone; DZ: diazepam; ZD: zolpidem; LO: lorazepam; AL: alprazolam; LT: lormetazepam; FT: flunitrazepam; FL: flurazepam; OX: oxazepam; TM: temazepam; TR: triazolam; ZA: zaleplon; NT: nitrazepam; CX: chlordiazepoxide; CB: clobazam; CR: clorazepate; BR: bromazepam; BT: brotizolam; CL: clonazepam; EZ: eszopiclone; MX: Mixed BZRAs; MX1: groups based on BZRA concentration; MX2: groups based on BZRA concentration; MX3: groups based frequency of use.

^aDaurat = road traffic and simulated car.

^bVolkerts = road traffic and simulated car.

^cLaurell = driving course and simulated car

Appendix 5. Benzodiazepine receptor agonist dose, dose timing, and consecutive days of dosing when assessed in driving studies with placebo comparators

Year of study publication	First author (surname, initials)	BZRA	BZRA dose (mg/24 h)	BZRA consecutive days of use	Interval between BZRA administration and assessment of impairment (hrs)	Diazepam Equivalent dose
Single driving impairment assessment setting
Road Driving (traffic)
1981	deGier, JJ	Diazepam	15	1	–	–
2014	Leufkens, T	Zopiclone	7.5	1	10	5
2014	Leufkens, T	Zopiclone	7.5	1	10	5
2011	Mets, M	Zopiclone	7.5	1	8.5	5
2014	Motak, L	Lorazepam	2	1	2	10
1993	O'Hanlon, JF	Lorazepam	2	7	–	10
1995	O'Hanlon, JF	Lorazepam	4	8	4	20
1995	O'Hanlon, JF	Diazepam	15	8	2	15
1995	O'Hanlon, JF	Lorazepam	1.5	9	3	7.5
2011	Ramaekers, JG	Zopiclone	7.5	1	11	5
1992	vanLaar, M	Diazepam	15	28	1.5	15
2001	vanLaar, M	Lorazepam	3	7	3	15
1998	Vermeeren, A	Flurazepam	30	3	10	15
2015	Vermeeren, A	Zopiclone	7.5	1	9	5
2016	Vermeeren, A	Zopiclone	7.5	1	9	5
2019	Vermeeren, A	Zopiclone	7.5	1	9	5
2002	Verster, J	Alprazolam	1	1	1	10
Driving Course (non-traffic)
1974	Berry, PA	Chlordiazepoxide	10	1	1, 2, 3, 4	2
1972	Betts, TA	Chlordiazepoxide	50	1	36	10
2008	Boyle, J	Eszopiclone	3	1	9.5–10.25	5
2018	Brown, TL	Alprazolam	1	1	3	10
1977	Hindmarch, I	Clobazam	20	6	–	5
2005	Hindmarch, I	Alprazolam	3	3	14–16	30
2019	Huizinga, RH	Alprazolam	1	1	–	10
2022	Iwamoto, K	Zopiclone	7.5	1	9	5
2016	Kay, GG	Triazolam	0.5	1	7.25	10
2017	Kay, GG	Zopiclone	7.5	1	9	5
1974	Linnoila, M	Diazepam	10	1	0.5	10
1993	Mattila, MJ	Diazepam	15	1	1.5, 3, 4.5, 6	15
1994	Mattila, MJ	Zopiclone	7.5	1	1.5, 3.5, 6	5
2014	Miyata, A	Triazolam	0.125	1	1, 4	2.5
2021	Moller, HJ	Lorazepam	1	1	1.5	5
1982	Moskowitz, H	Diazepam	15	8	1	15
2022	Muehlan, C	Zopiclone	7.5	1	9	5
2008	Otmani, S	Zolpidem	10	2	2, 13	5
2013	Sarris, J	Oxazepam	30	1	1.5	10
1986	Smiley, A	Diazepam	15	9	1	15
2015	Stone, BT	Alprazolam	1	1	2	10
2022	Su, S	Alprazolam	0.75	1	1.5	7.5
2010	Takahashi, M	Diazepam	5	1	1, 4	5
2001	Törnros, J	BZRA general	–	1	1	–
2022	Torres, R	Zopiclone	7.5	1	9.5	–
2000	Vanakoski, J	Diazepam	15	1	1.5, 4	15
2000	Vanakoski, J	Diazepam	10	1	1.5, 4	10
Simulated driving (computer)
2017	Carter, SG	Zopiclone	7.5	30	–	5
2020	Carter, SG	Zopiclone	15	1	8.5	10
2002	Ludice, A	Lormetazepam	1	3	–	5
1976	Seppala, T	Bromazepam	18	14	0.5, 1, 1.5	30
Multiple driving impairment assessment settings
2013	Daurat, A	Lorazepam	2	1	2	10

Appendix 6. Benzodiazepine receptor agonist dose, dose timing, and consecutive days of dosing when assessed in driving studies with multiple active comparators and placebo

Year of study publication	First author (surname, initials)a	Active comparators	BZRA dose (mg/24 h)	BZRA consecutive days of use	Interval between BZRA administration and assessment of impairment (hrs)	Diazepam Equivalent dose
Single driving impairment assessment setting
Road Driving (traffic)
1979	Biehl, B	Clobazam	20	3	–	5
		Diazepam	10	3	–	10
1990	Brookhuis, KA	Flurazepam	30	8	10–11, 16–17	10
		Lormetazepam	1	8	10–11, 16–17	5
		Lormetazepam	2	8	10–11, 16–17	10
2018	Jongen, S	Oxazepam	10	1	4	3.33
		Oxazepam	30	1	4	10
		Diazepam	10	1	4	10
2007	Leufkens, T	Alprazolam XR	1	1	4	10
		Alprazolam XR	1	1	4	10
2009	Leufkens, T	Temazepam	20	1	10	10
		Zopiclone	7.5	1	10	5
2009	Leufkens, T	Zopiclone	7.5	1	10	5
		Zolpidem	10	1	5	5
2014	Leufkens, T	Zolpidem	10	1	10	5
		Temazepam	10 20	1 1	10 10	5 10
		Zopiclone	3.75 7.5	1 1	10 10	2.5 5
		Nitrazepam	5	1	10	10
		Oxazepam	10 20 50	1 1 1	10 10 10	3.33 6.67 16.67
		Flurazepam	15	1	10	5
		Clonazepam	0.5	1	10	10
		Lormetazepam	0.5 2	1 1	10 10	2.5 10
1982	O'Hanlon, JF	Diazepam	5	1	1	5
		Diazepam	10	1	1	10
1986	O'Hanlon, JF	Temazepam	20	8	8–11, 14–17	10
		Nitrazepam	10	8	8–11, 14–17	20
1986	Schmidt, U	Temazepam	20	7	10	10
		Flunitrazepam	2	7	10	10
1995	Vermeeren, A	Zolpidem	10	1	10–11	5
		Flunitrazepam	2	1	10–11	10
1998	Vermeeren, A	Zaleplon	10	1	5–6, 10–11	5
		Zaleplon	20	1	5–6, 10–11	10
		Zopiclone	7.5	1	5–6, 10–11	5
2002	Vermeeren, A	Zaleplon	10	1	10	5
		Zopiclone	7.5	1	10	5
2011	Vermeeren, A	Zolpidem	3.5	1	3	1.75
		Zolpidem	3.5	1	4	1.75
		Zopiclone	7.5	1	9	5
2002	Verster, J	Zaleplon	10	1	4	5
		Zaleplon	20	1	4	10
		Zolpidem	10	1	4	5
		Zolpidem	20	1	4	10
2014	Verster, J	Zaleplon	10	1	4	5
		Zaleplon	20	1	4	10
		Zolpidem	10	1	4	5
		Zolpidem	20	1	4	10
		Zopiclone	7.5	1	8.5	5
2018	Verster, J	Zaleplon	10	1	4	5
		Zaleplon	20	1	4	10
		Zolpidem	10	1	4	5
		Zolpidem	20	1	4	10
		Zopiclone	7.5	1	8.5	5
Driving Course (non-traffic)
1982	Betts, TA	Flurazepam	15	1	12	5
		Temazepam	20	1	12	10
1998	Hindmarch, I	Clobazam	30	3	0.5	7.5
		Lorazepam	3	3	0.5	15
1999	Mercier-Guyon, Cb	Lorazepam	2	7	–	10
Simulated driving (car)
1999	Bocca, ML	Zopiclone	7.5	1	10–12	5
		Zolpidem	10	1	10–12	5
		Flunitrazepam	1	1	10–12	5
2011	Bocca, ML	Zopiclone	7.5	1	8	5
		Zolpidem	10	1	8	5
		Flunitrazepam	1	1	8	5
2013	Brown, T	Triazolam	0.125	1	2.5	2.5
		Triazolam	0.25	1	2.5	5
1975	Dureman, I	Clorazepate	10	1	1	5
		Clorazepate	20	1	1	10
		Clorazepate	40	1	1	20
		Diazepam	5	1	1	5
		Diazepam	10	1	1	10
		Diazepam	20	1	1	20
1978	Dureman, I	Clorazepate	20	14	8	10
		Clorazepate	20	14	8	10
1991	Friedel, B	Diazepam	7	1	–	7
		Diazepam	14	1	–	14
1973	Linnoila, Mb	Diazepam	10	1	0.5	10
2009	Meskali, M	Flunitrazepam	1	1	10	5
		Zolpidem	10	1	10	5
		Zopiclone	7.5	1	10	5
1978	Palva, ES	Oxazepam	45	14	0.5, 1, 2.5	15
		Diazepam	15	14	0.5, 1, 2.5	15
		Chlordiazepoxide	30	14	0.5, 1, 2.5	6
2003	Partinen, M	Zolpidem	10	1	5.5	5
		Temazepam	20	1	5.5	10
1990	Törnros, J	Nitrazepam	5	3	9.5	10
		Brotizolam	0.25	3	9.5	5
1984	Willumeit, H	Lormetazepam	2	1	1, 2, 3	10
		Diazepam	10	1	1, 2, 3	10
1984	Willumeit, H	Diazepam	10	7	1, 2, 3	10
		Flurazepam	30	7	1, 2, 3	10
		Lormetazepam	2	7	1, 2, 3	10
Simulated driving (computer)
2003	Berthelon, C	Flunitrazepam	1	1	10	5
		Zolpidem	10	1	10	5
		Zopiclone	7.5	1	10	5
1994	Kuitunen, T	Diazepam	15	1	1.5–4.5	15
		Zopiclone	7.5	1	1.5–4.5	5
		Triazolam	0.25	1	1.5–4.5	5
1998	Mattila, MJ	Zolpidem	15	1	1, 3.5, 5	7.5
		Zopiclone	7.5	1	1, 3.5, 5	5
		Diazepam	15	1	1, 3.5, 5	15
		Oxazepam	30	1	1, 3.5, 5	10
1976	Seppala, T	Bromazepam	18	14	0.5, 1, 1.5	30
1986	Subhan, Z	Alprazolam	0.5	7	1	5
		Lorazepam	2	7	1	10
1983	Willumeit, HP	Lormetazepam	2	7	–	10
		Flurazepam	30	7	–	10
Multiple driving impairment assessment settings
1986	Laurell, H	Triazolam	0.25	3	9	5
		Nitrazepam	5	3	9	10
1992	Volkerts, ER	Lormetazepam	1	2	10–11.5	5
		Oxazepam	50	2	10–11.5	16.67

^aThree studies (Törnros, 2001; van der Sluiszen, 2019; and Vinckenbosch, 2021) grouped BZRAs together and individual data cannot be presented).

^bNon-BZRAs (e.g., barbiturates, antipsychotics, alcohol, codeine) were used for comparators.

Appendix 7. Observational studies

Date, author	Country	Observation Period	Sample n / mean age (y) / age range (y)	Description	Groups	Drug Identification	Indicator of impairment
Comparative studies
Cohort studies
2022 Osler, M	Denmark	01/2002 to 12/2018	3,832,588 n/a ≥18	Cohort study using the Danish Civil Registration System of new users of BZRAs, pregabalin, or melatonin with no previous registered traffic accident followed for incident MVAs.	New BZRA user: 738,019 No BZRA use: 3,094,569	Prescription record	MVA
2016 Booth, JN	United States	10/2008 to 08/2011	2000 77 ≥70	In-person clinical assessment, interviewer questionnaire, and pill-bottle review of older adult drivers from north-central Alabama who responded to a general mailed invitation for study participants. Focus was on driving risk with z-drug zolpidem.	Zolpidem users: 75 Controls: 1925	Pill bottle review	MVA
2014 Hoiseth, G	Norway	03/2012 to 09/2013	13,225 n/a n/a	Compared probability of being assessed as impaired by CTI based on drug use status: BZRA alone, amphetamine alone, combination of BZRA and amphetamine. Blood samples were collected among apprehended drivers suspected of DWI (drug and alcohol).	BZRA alone: 196 Amphetamine alone: 322 Combination (BZRA and amphetamine only): 899	Biological detection	DWI suspect
2009 Gustavsen, I	Norway	01/2000 to 12/2007	3602 39 19–88	Association study of z-drug blood concentration (zopiclone, zolpidem) and level of impairment of Norwegian drivers with only one single drug detected who were apprehended due to suspicion of DUID. Alcohol group of drivers was used as a comparison.	Z-drug group: 122 (zolpidem 43, zopiclone 79) Blood test categories: i) possible therapeutic (9/122); ii) moderately elevated (57); iii) highly elevated (56) Alcohol group: 3480 alcohol-only Blood alcohol content: i) 0.004–0.05 %; ii) 0.051–0.1 %; iii) 0.101–0.15 %; iv) >0.15 %	Biological detection	DWI suspect
2008 Gustavsen, I	Norway	01/2004 to 12/2006	3,100,000 n/a 18–69	Comparison of MVA risk involving a personal injury within the 7 days of dispensing z-drugs (zopiclone, zolpidem), nitrazepam, and flunitrazepam of all Norwegian drivers aged 18–69 years.	Zopiclone: 28,435 person-years Zolpidem: 4537 person-years Nitrazepam: 4549 person-years Flunitrazepam: 1958 person-years	Prescription record	MVA with injury
2003 Bramness, JG	Norway	01/1987 to 12/1998	818 n/a n/a	Assessed relationship between BZRA drug concentration and clinical test for impairment sub-scores using data from Norwegian drivers suspected of being under the influence of non-alcoholic drugs with blood samples containing only one BZRA.	Therapeutic BZRA concentration: 351 (reference) Mildly elevated BZRA concentration: 198 Moderately elevated BZRA concentration: 116 Highly elevated BZRA concentration: 153	Biological detection	DWI suspect
2002 Bramness, JG	Norway	01/1987 to 12/1998	11,577 32.6 n/a	Drivers apprehended in Norway for suspected driving under the influence of drugs or alcohol.	Single benzodiazepine: 818 Alcohol only: 10,759	Biological detection	DWI suspect
2000 Longo, MC	Australia	04/1995 to 08/1996	2500 n/a n/a	Blood samples collected following MVAs in South Australia.	BZRA detected: 392 Prescription or illicit drug not detected: 1887	Biological detection	DWI suspect
1998 Neutel, I	Canada	01/1979 to 12/1986	323,658 n/a ≥20	Association of MVAs in the province of Saskatchewan with dispensing records for five BZRAs (triazolam, flurazepam, oxazepam, lorazepam or diazepam), new starts and repeats, sex, and age.	BZRA dispensed: 225,796 No BZRA dispensed: 97,862	Prescription record	MVA with hospitalization
1995 Neutel, CI	Canada	01/1979 to 12/1986	323,658 n/a ≥20	Rate of MVA-related hospitalizations in the province of Saskatchewan with new starts of BZRAs (triazolam, flurazepam, oxazepam, lorazepam or diazepam) within 2 weeks and 4 weeks of new start	BZRA dispensed: 225,796 No BZRA dispensed: 97,862	Prescription record	MVA with hospitalization
1995 Neutel, CI	Canada	01/1979 to 12/1986	323,658 n/a ≥20	Compared the risk of hospitalizations due to MVAs in the province of Saskatchewan in association with new starts of BZRA hypnotics, BZRA anxiolytics, and no BZRAs within 1 week, 2 weeks, and 4 weeks of receiving a new prescription in the province of Saskatchewan.	BZRA hypnotic dispensed: 78,000 BZRA anxiolytic dispensed: 148,000 Controls: 98,000	Prescription record	MVA
1977 Garriott, JC	United States	06/1974 to 12/1975	127 n/a n/a	Descriptive analysis of blood, urine, and other biological sample findings of drivers fatally injured in traffic accidents in the Dallas area.	–	Biological detection	Fatal MVA
Case-crossover studies
2022 Osler, M	Denmark	01/2002–12/2018	738,019 n/a ≥18	Case cross-over study using the Danish Civil Registration System of new users of BZRAs, pregabalin, or melatonin with no previous registered traffic accident followed for incident MVAs	Cases: 738,019 incident MVAs Controls: 738,019 with 3 selected time periods before MVA period Exposure period: any BZRA use 100 days prior to incident MVA or during control periods	Prescription record	MVA
2017 Nevriana, A	Sweden	02/2006–12/2009	26,586 n/a 50–80	Case-crossover analysis of Swedish drivers in a first MVA causing injury and suspected of DWI. Individuals were assessed for use of z-drugs (zolpidem, zopiclone). Case period was 4 weeks prior to MVA and an equal period at least 12 weeks earlier.	Cases: 26,586 MVAs causing injury Controls: 26,586 no MVA record Exposure period: 28 days prior to incident MVA or control period	Prescription record	MVA with injury DWI suspect
2011 Yang, Y	Taiwan	01/1998–12/2004	12,929 n/a ≥18	Association study among Taiwanese drivers experiencing hospitalization following a first-time MVA and dispensing of z-drugs (zolpidem, zopiclone), short half-life BZRAs, and long half-life BZRAs.	Cases: 12,929 MVAs causing hospitalization Controls: 12,929 3 earlier control periods per subject with no MVA record Case period: day before MVA and for controls the 91st, 182nd, and 273rd days before the case period Washout period: 13 weeks between the case period and each of the 3 control periods to reduce the likelihood of overlapping prescriptions between these periods	Prescription record	MVA with hospitalization
1998 Barbone, F	United Kingdom	08/1992–06/1995	1731 n/a ≥18	MVA involving police of all Tayside, United Kingdom residents (410,306) attached to a general practitioner. Exposures included 235 BZRAs among 1731 subjects taking psychotropics.	Cases: 235 MVAs involving police Controls: 235 no MVA record Exposure period: day of MVA and up to 18 previous matching weekdays	Prescription record	MVA DWI suspect
Case-control studies
2022 Olesen, AV	Denmark	01/1996–12/2018	779,844 n/a ≥17	Analysis of dispensing records of Danish drivers experiencing an MVA resulting in a police-recorded personal injury	Cases: 129,974 drivers with personal injury from MVA Controls: 649,870 (5:1 sex-, age-matched)	Prescription record	MVA
2021 Brubacher, JR	Canada	01/1997 to 12/2016	714,944 n/a ≥17	Drivers in the province of British Columbia with a first MVA that resulted in a police report. Driver accident culpability was categorized as indeterminate, responsible, or not responsible for the MVA. Numerous drugs categories examined, including: long-acting BZRAs, short-acting BZRAs, and z-drugs.	Cases: 382,685 drivers responsible for MVA Controls: 332,259 drivers not responsible for MVA	Prescription record	MVA DWI suspect Culpability assessment
2021 Asbridge, M	Canada	01/2010 to 07/2016	2318 44 16–93	Moderately injured drivers involved in MVAs leading to a police report and emergency visit in the province of British Columbia. Driver culpability was categorized as indeterminate, responsible, or not responsible for the MVA. Blood samples were collected within 6 h of the MVA. BZRA group combined BZRAs and z-drugs.	Cases: 1178 drivers responsible for MVA Controls: 647 drivers not responsible for MVA	Biological detection	MVA with A&E visit DWI suspect Culpability assessment
2020 Drummer, OH	Australia	07/2013 to 06/2018	4988 n/a n/a	Drivers of MVAs leading to hospital admission or emergency department visit with blood sample collected for toxicology analysis in state of Victoria. Multiple categories of drugs investigated for culpability that were scored as culpable, partial (contributory), or non-culpable.	Cases: 612 MVA with BZRA detected Controls: 1837 MVA with no drug substance	Biological detection	MVA DWI suspect Culpability assessment
2019 Jamt, R	Norway	01/2000 to 12/2015	3639 38 ≥18	Drivers in Finnmark, a northern county of Norway, involved in a MVA and arrested due to suspicion of impaired driving. Blood samples collected at time of arrest. Oral fluid samples collected from controls identified in routine driver survey. Z-drug and BZRAs reported: diazepam, clonazepam, alprazolam, zopiclone, oxazepam, flunitrazepam, zolpidem, nitrazepam, phenazepam.	Cases: 612 MVA arrested suspected DUI (alcohol or drugs) Controls: 3027 random drivers selected from police roadside survey	Biological detection	MVA DWI suspect
2017 Nevriana, A	Sweden	01/2006 to 12/2009	135,480 n/a 50–80	Older adult drivers of a first MVA with injury suspected of DWI (not due to alcohol). Focus was on zolpidem and zopiclone use. Z-drug exposure groups included: newly initiated use (1–30 days only); occasional use (1–2 dispensations 31–180 days before MVA); frequent use (>2 dispensings within 180 days); and non-Z-drug use (dispensing of other medications); and no medication use.	Cases: 27,096 with first MVA involving injury Controls: 108,384 4:1 matched for sex, age, and area of residence	Prescription record	MVA DWI suspect
1982 Cimbura, G	Canada	04/1978 to 03/1979	401 n/a n/a	Descriptive analysis of blood, urine, and other biological sample findings of all drivers fatally injured in traffic accidents in the province of Ontario.		Biological detection	Fatal MVA
2015 Fournier, JP	Canada	06/1990 to 06/2000	819,166 n/a 67–84	Association between BZRA use and first MVA causing injury or property damage involving older adult drivers of the province of Quebec.	Cases: 74,503 MVAs with injury or property damage Controls: 744,663 no MVA	Prescription record	MVA
2014 Bogstrand, ST	Norway	04/2008 to 03/2009	12,113 n/a n/a	Assessed drugs associated with being arrested for DWI. Included drivers arrested on suspicion of DWI due to drugs with blood samples collected (with BAC < 0.2 g/L) and unimpaired drivers randomly selected who provided an oral fluid sample. Reports individual BZRA risks of arrest (e.g., alprazolam, clonazepam, oxazepam, z-drugs (zolpidem, zopiclone)	Cases: 2738 arrested on suspicion of drugged driving (MVAs, dangerous driving, traffic control stop, citizen reports). Controls: 9375 randomly selected drivers from normal traffic	Biological detection	MVA DWI suspect
2014 Johnell, K	Sweden	07/2005 to 12/2009	154,225 n/a 50–80	Nationwide Swedish study of older adult drivers with a first MVA and associated injury with police suspicion of DWI (excluding alcohol). Crash and injury data extracted from the Swedish Traffic Accident Data Acquisition (STRADA) register.	Cases: 30,845 MVAs with injury Controls: 123,380 matched drivers with no MVA	Prescription record	MVA with injury
1992 Ray, WA	United States	01/1984 to 12/1988	16,262 n/a 65–84	Assessed the association between the use of BZRAs and 3 other medication categories and injurious MVAs involving older adult drivers (65–84 years) in the Tennessee state Medicaid program.	BZRA only: 2978 person-years Non-use of psychoactive prescriptions drugs: 21,578 person-years	Prescription record	Fatal MVA MVA with hospitalization MVA with A&E visit
2013 Chang, CM	Taiwan	01/2000 to 12/2009	36,276 38.4 ≥18	Association study evaluating the relationship between exposure to BDZs (all, short-, long-acting), z-drugs, antidepressants, and antipsychotics and risk of MVAs. Examined risk when medication dispensed within 1 month, 1 week, and 1 day of the MVA.	Cases: 5183 MVA drivers Controls: 31,093 matched outpatients with no MVA history	Prescription record	MVA
1993 Gjerde, H	Norway	01/1989 to 12/1990	159 n/a n/a	Descriptive report of blood sample analysis for alcohol and psychoactive drugs in fatally injured drivers involved in MVAs.	–	Biological detection	Fatal MVA
1995 Mercer, GW	Canada	10/1990 to 09/1991	227 36.8 n/a	Descriptive report using police traffic accident reports and toxicological analysis of blood samples from drivers involved in fatal MVAs in the province of British Columbia.	–	Biological detection	Fatal MVA
2000 Gerostamoulos, J	Australia	01/1995 to 12/1996	921 n/a n/a	Responsibility analysis of drivers involved in fatal accidents, categorized as culpable or non-culpable using a validated scoring system. Toxicological information on drug exposure was based on anti-mortem specimen analysis.	Cases: culpable drivers who were fatally injured in a MVA Controls: non-culpable drivers who were fatally injured in a MVA	Biological detection	Fatal MVA Culpability assessment
2004 Movig, K	Netherlands	05/2000 to 08/2001	926 38.6 n/a	Drivers involved in a MVA requiring hospitalization and drivers recruited at random while driving on public roads. Urine of blood samples were obtained from all participants.	Cases: 110 injured drivers admitted to A&E Controls: 816 drivers randomly selected for a roadside survey	Biological detection	MVA with A&E visit
2003 Mura, P	France	06/2000 to 09/2001	1800 n/a ≥18	Comparison of substances in blood samples of drivers injured in road accidents and a matched group of people attending emergency services for non-trauma reasons. Six emergency service sites from across France participated. BZRAs, among other substances, were included in the toxicological screen.	Cases: 900 drivers involved in a nonfatal MVA attending an A&E centre Controls: 900 patients attending the same A&E centres for any non-traumatic reasons	Biological detection	MVA with A&E visit
2000 Seymour, A	United Kingdom	01/1999 to 12/1999	214 n/a n/a	Descriptive analysis of psychoactive drugs identified from blood and urine samples from drivers in police custody suspected DUID and from fatally injured drivers involved in MVAs.	–	Biological detection	Fatal MVA DWI suspect
2000 Longo, MC	Australia	04/1995 to 08/1996	2500 31 n/a	Culpability analysis using an objective scoring system and measurement of drug concentration (cannabinoids, BZRAs and stimulants) and alcohol content of blood samples of injured drivers attending A&E centres in South Australia following MVAs. Driver responsibility was rates as culpable, contributory, and not culpable.	Cases: 55 % of 2029 drivers culpable of the MVA Controls: 39 % of 2029 drivers not culpable of the MVA	Biological detection	MVA with A&E visit Culpability assessment
1997 Hemmelgarn, B	Canada	01/1990 to 12/1993	61,369 n/a 67–84	Assessed BZRA prescription records of older adult drivers in the province of Quebec involved in a first MVA resulting in injury per police report compared to individuals with no MVA. BZRAs were categorized by half-life as short- (≤24 h) or long-acting (>24 h).	Cases: 5579 drivers with a MVA Controls: 55,790 matched controls with no MVA	Prescription record	MVA with injury
1980 Honkanen, R	Finland	04/1977 to 10/1977	526 35 ≥15	Blood sample analysis collected in A&E of injured drivers involved in MVAs in one urban centre were compared to a random selection of people screened for drugs at gas service stations.	Cases: 201 injured drivers presenting to a Helsinki emergency departments Controls: 325 drivers randomly selected at service stations	Biological detection	MVA with A&E visit
Non-comparative studies
2002 Carmen del Rio, M	Spain	01/1991 to 12/2000	5745 n/a n/a	Descriptive analysis of alcohol, illicit drugs, and medicines of fatally injured drivers involved in MVAs.	–	Biological detection	Fatal MVA
2022 Favretto, D	Italy	01/2014 to 01/2018	4066 n/a n/a	A retrospective blood sample analysis of sedative-hypnotic prevalence and concentrations in drivers involved in MVAs admitted to hospital in the province of Padova.	–	Biological detection	MVA with hospitalization
2004 Drummer, OH	Australia	01/1990 to 12/1999	3398 n/a n/a	MVAs with driver fatalities in 3 Australian states (i.e., Victoria, New South Wales, Western Australia). Responsibility assessment based on validated scale categorized drivers as culpable or non-culpable.	Cases: 1694 drug- and alcohol-positive fatally injured drivers Controls: 1704 drug- and alcohol-free fatally injured drivers	Biological detection	Fatal MVA Culpability assessment
2008 Dubois, S	United States	01/1993 to 12/2006	72,026 43.9 n/a	Fatal crashes (occupant or non-motorist) identified using the United States Fatality Analysis Reporting System database. Excluded drivers with non-zero blood alcohol concentration. Unsafe driver actions used as proxy of responsibility. BZRA categories based on half-life: short: <6 h; intermediate: 6 to <24 h as intermediate; long >24 h.	Cases: 44,066 drivers with ≥1 potentially unsafe driving actions in relation to the crash (e.g., speeding) Controls: 27,960 no unsafe driver actions recorded	Biological detection	Fatal MVA
2020 Christophersen, AS	Norway	01/2013 to 12/2015	17,201 n/a n/a	Comparison of BZRA blood concentrations in apprehended drug-impaired drivers and the maximal obtainable steady state concentrations if the drug had been used in accordance with Norwegian Health recommendations. BZRAs were present in 10,248 (60 %) of drivers suspected of DRUID. Reports rates of single-entity BZRAs (clonazepam, diazepam, alprazolam, oxazepam, nitrazepam) and z-drugs (zopiclone, zolpidem).	–	Biological detection	DWI suspect
2020 Herrera-Gómez, F	Spain	10/2018 to 11/2018	2881 39 16–84	Prevalence study for alcohol (breath) and psychoactive drugs (oral fluids) of a representative sample of Spanish drivers. Drivers were randomly selected during normal traffic conditions.	–	Biological detection	Drivers randomly selected
2019 Pelletti, G	Italy	01/2017 to 03/2018	1026 n/a n/a	Descriptive study of toxicological analysis of blood samples of drivers involved in MVAs in the Bologna area including 53 psychoactive substances (illicit and medical).	–	Biological detection	DWI suspect
2019 Bunn, T	United States	01/2010 to 12/2014	2235 n/a n/a	Prevalence of alcohol and drugs present in fatally injured drivers in the state of Kentucky.	–	Biological detection	MVA
2009 Alhm, K	Sweden	12/2004 to 02/2021	200 39 14–82	Prospective study from northern Sweden of drugs and alcohol identified from blood samples of drivers in a MVA who were hospitalized or fatally injured.	–	Biological detection	Fatal MVA MVA with hospitalization
2018 Carfora, A	Italy	01/2009 to 12/2016	1797 n/a n/a	Toxicological (blood, urine) reports of hospitalized drivers involved in a MVA suspected of DUID.	–	Biological detection	MVA with hospitalization
2017 Hoiseth, G	Norway	02/2012 to 05/2013	410 n/a 65–86	Prevalence and concentration of single entity z-drugs (zopiclone, zolpidem) and BZRAs (clonazepam, oxazepam, diazepam) measured from blood samples taken from older adult drivers apprehended under suspicion of DUID.	–	Biological detection	DWI suspect
2017 Valen, A	Norway	01/1990 to 12/2015	112,348 32 n/a	Toxicological data from apprehended Norwegian drivers investigated for illicit and psychoactive medicinal drugs.	–	Biological detection	DWI suspect
2009 Sidlo, J	Slovakia	01/2000 to 12/2007	75 n/a n/a	Descriptive analysis of 75 cases of road traffic fatalities (drivers and others) under the influence of psychoactive substances (not including alcohol). Includes psychoactive substances identified in fatally injured drivers.	–	Biological detection	Fatal MVA
2010 Maxwell, HG	United States	01/1993 to 12/2006	116,510 42 ≥20	Used the United States Fatality Analysis Reporting System to identify fatal crashes (occupant or non-motorist) with ≥1 unsafe driver action. BZRAs, identified by blood sample, were grouped as short-, intermediate- and long-acting.	Cases: 1179 drivers ≥1 unsafe driver action (e.g., weaving or speeding) recorded in relation to the fatal MVA Controls: 70,440 drivers who did not have an unsafe driver action (unsafe driver action; e.g., weaving or speeding) recorded in relation to the fatal MVA	Biological detection	Fatal MVA
2012 Poulsen, H	New Zealand	07/2004 to 06/2009	1046 39 14–92	Prevalence of alcohol and illicit and psychoactive drugs analysis of blood samples collected immediately after crash in fatally injured drivers involved in MVAs.	–	Biological detection	Fatal MVA
2013 Legrand, SA	Belgium, Denmark, Finland, Italy, Lithuania, Netherlands	10/2007 to 04/2010	2492 n/a ≥18	Analysis and characterization of alcohol and drug presence in blood samples of injured drivers admitted to emergency departments in the six countries involved (Belgium, Denmark, Finland, Italy, Lithuania, Netherlands). Reports rates of single-entity and combination psychoactive drugs for BZRAs and subgroups z-drugs and BZRAs.	–	Biological detection	MVA with A&E visit
2013 Jones, AW	Sweden	Not reported	1000 n/a n/a	Toxicological analysis of blood samples containing both diazepam and nordiazepam to infer rates of therapeutic and non-therapeutic use of diazepam among drivers apprehended under the suspicion of DUID in Sweden.	–	Biological detection	DWI suspect
2013 Jones, AW	Sweden	01/2001 to 12/2005	2205 32 n/a	Comparison of alprazolam concentration in blood samples from impaired drivers and from forensic autopsies classified as intoxication or other causes of death.	–	Biological detection	DWI suspect
2013 Burch, HJ	United Kingdom	02/2010 to 03/2011	376 n/a n/a	Measured concentrations of drugs in blood samples of drivers suspected DUIDs.	–	Biological detection	DWI suspect
2013 Stephenson, JB	United States	03/2010 to 08/2011	11 27 18–40	Case series of 11 phenazepam-positive drivers arrested due to suspicion of impaired driving.	–	Biological detection	DWI suspect
2012 Legrand, SA	Belgium, Netherlands	01/2008 to 05/2010	535 n/a n/a	Analysis and characterization of alcohol and drug presence in blood samples of injured drivers admitted to emergency departments in Belgium and the Netherlands.	–	Biological detection	MVA with A&E visit
2012 Kriikku, P	Finland, Norway, Portugal, Sweden	07/2010 to 06/2011	141 n/a n/a	Blood concentrations of phenazepam in drivers suspected of DUID.	–	Biological detection	DWI suspect
2013 Gjerde, H	Norway	01/2003 to 12/2010	9769 n/a n/a	Association study of alcohol and substance use from blood samples among car and van drivers with fatal injuries from MVAs in Norway compared with randomly selected drivers selected from normal traffic conditions who provided oral fluid samples. Risk estimates for BZRAs, diazepam-only, and zopiclone-only.	Cases: 508 driver fatalities with blood sample collection Controls: 9261 randomly identified drivers	Biological detection	Fatal MVA
2012 Jones, AW	Sweden	01/2001 to 12/2010	7302 41.5 n/a	Comparison of z-drug (zolpidem and zopiclone) concentrations in blood samples from impaired drivers and from forensic autopsies classified as intoxication or other causes of death.	–	Biological detection	DWI suspect
2011 Blencowe, T	Finland	05/2004 to 12/2008	224 n/a n/a	Testing of new oral sample technology to identify BZRAs and other drugs in cases of suspected DUID.	–	Biological detection	DWI suspect
2010 Senna, MC	Switzerland	01/2005 to 12/2005	4794 31 14–92	Nationwide study of blood and urine sample analysis for psychoactive drugs from drivers suspected of DUID (excluding drivers considered to be under the influence of alcohol).	–	Biological detection	DWI suspect
2009 Kulikowska, J	Poland	01/1997 to 12/2007	785 n/a n/a	Descriptive analysis of psychoactive drugs identified from blood samples of drivers causing MVAs.	–	Biological detection	MVA
2014 Legrand, S	Finland, Norway, Portugal, Sweden	01/2006 to 12/2009	1118 n/a n/a	Prevalence of alcohol, illicit drugs and psychoactive medicines in fatally injured drivers in four European countries (Finland, Norway, Portugal, Sweden). Reports poly- and single-entity BZRAs and z-drug fatalities per country.	–	Biological detection	Fatal MVA
2014 Szeremeta, M	Poland	01/2010 to 12/2011	274 28.6 17–69	Report of psychoactive medicines and substances identified in blood samples of drivers who died as the result of an MVA, survived an MVA, or were engaged in a roadside survey.	–	Biological detection	Fatal MVA MVA
2008 Smink, BE	Netherlands	01/1999 to 12/2004	171 n/a n/a	Descriptive retrospective descriptive study of driver impairment characteristics based on BZRA blood concentration taken at the time of apprehension for suspected DUID. BZRA concentration categories were: subtherapeutic, therapeutic, and elevated.	–	Biological detection	DWI suspect
2008 Christophersen, AS	Norway	01/2004 to 12/2004	796 n/a n/a	Drug concentration analysis of BZRAs in Norwegian drivers suspected of DUID.	–	Biological detection	DWI suspect
2007 Ch'ng, CW	Australia	12/2000 to 04/2002	436 n/a n/a	Descriptive analysis of psychoactive drugs identified from blood samples taken from drivers presenting to an emergency trauma centre following a MVA in the state of Victoria.	–	Biological detection	MVA with A&E visit
2006 Bramness, JG	Norway	01/1987 to 02/2003	415 30 n/a	Descriptive report of DUID cases in which flunitrazepam was detected from blood samples as the only psychoactive drug present.	–	Biological detection	DWI suspect
2005 Bernhoft, IM	Denmark	11/2002 to 03/2004	379 n/a n/a	Analysis of drugs identified from saliva and blood samples obtained from injured drivers involved in MVAs presenting to hospital emergency departments.	–	Biological detection	MVA with A&E visit
2005 Smink, BE	Netherlands	10/1998 to 09/1999	993 35 15–90	An analysis to assess the relationship between exposure to different classes of psychoactive drugs (alcohol, illicit drugs, medicines) and accident severity (severe: hospitalization or death) of drivers involved in MVAs from whom blood samples were obtained.	–	Biological detection	MVA
2005 Olszowy, Z	Poland	01/1997 to 12/2003	157 n/a n/a	Drug analysis of blood samples taken from drivers suspected of DUID.	–	Biological detection	DWI suspect
2004 Jones, AW	Sweden	01/2001 to 12/2002	94 n/a n/a	Retrospective descriptive analysis of drivers arrested for DUID with diazepam and nordiazepam identified from blood samples.	–	Biological detection	DWI suspect
2002 Skurtveit, S	Norway	01/1995 to 12/1995	3343 n/a n/a	Descriptive analysis of driver and drug characteristics of people apprehended for suspected DUID in which a BZRA was identified by blood sample.	–	Biological detection	DWI suspect
2016 Al-Abdallat, IM	Jordan	01/2008 to 12/2014	68 n/a n/a	Prevalence data of alcohol and psychotropic drugs in fatally injured drivers involved in a MVA in Jordan.	–	Biological detection	Fatal MVA
2001 Logan, BK	United States	01/1997 to 12/1999	29 n/a n/a	Descriptive report of drivers in Washington state arrested for DUID in which zolpidem was identified alone or in combination with other psychoactive substances.	–	Biological detection	DWI suspect
2016 Drummer, OH	Australia	01/2000 to 12/2013	2638 n/a n/a	Fatally injured drivers involved in MVAs in state of Victoria. Data collected from coroner's office records. Multiple categories of drugs investigated for culpability that were scored as culpable, contributory, or non-culpable.	Cases: 1263 MVA driver fatalities who were drug- and alcohol-positive (19 cases of BZRA-monotherapy) Controls: 1105 MVA driver fatalities who were drug- and alcohol-free	Biological detection	Fatal MVA Culpability assessment
1999 Christophersen, AS	Denmark, Finland, Iceland, Norway, Sweden	03/1996	800 n/a n/a	A 1-week survey for psychoactive drugs including alcohol identified in blood samples of drivers apprehended by police in 5 Nordic (Denmark, Finland, Iceland, Norway, Sweden) countries due to the suspicion of driving while impaired.	–	Biological detection	DWI suspect
1999 Sybirska, H	Poland	1990s	112 n/a n/a	Descriptive report of blood sample results of drug use profiles for drivers involved in MVAs in police custody under suspicion of DUID with negative alcohol screens.	–	Biological detection	MVA DWI suspect
1999 Ledingham, D	United Kingdom	01/1995 to 05/1997	72 23.3 16–36	Descriptive report of psychoactive drugs identified in blood and urine samples of drivers in police custody in south-east Scotland suspected of DUIDs.	–	Biological detection	DWI suspect
2018 Andreuccetti, G	Brazil	06/2014 to 12/2015	365 n/a n/a	Toxicological analyses of blood samples from sudden, unexpected fatalities (including from MVAs).	–	Biological detection	Fatal MVA
1995 Skurtveit, S	Norway	01/1992 to 12/1993	5642 n/a n/a	Descriptive analysis of alcohol and drug use from blood samples taken from drivers apprehended for suspected DWI.	–	Biological detection	DWI suspect
1994 Kuitunen, T	Finland	01/1977 to 12/1992	10,976 n/a n/a	Analysis of the association between diazepam and nordiazepam concentrations from blood samples and indications of impairment measured using the 13-item CTD in acute and chronic users of diazepam who were in police custody under suspicion of DUID. The analysis included people who screened negative for all other substances other than diazepam and nordiazepam.	–	Biological detection	DWI suspect
2020 Tsocha, A	Greece	01/2015 to 12/2018	169 n/a n/a	Report of the prevalence of psychoactive drugs detected in blood and urine samples of drivers involved in MVAs, both injurious and fatal, from northern Greece.	–	Biological detection	Fatal MVA DWI suspect
1987 Bjorneboe, A	Norway	01/1978 to 12/1983	393 n/a n/a	Descriptive analysis of psychoactive drugs identified from blood samples of drivers in police custody under suspicion of DUID.	–	Biological detection	DWI suspect
1987 Poklis, A	United States	06/1983 to 05/1986	137 n/a n/a	Descriptive analysis of psychoactive drugs detected in blood and urine samples taken from drivers apprehended by police who were suspected of driving while impaired who had a negative alcohol screen.	–	Biological detection	DWI suspect
2021 Schumann, J	Australia	07/2006 to 06/2016	2287 n/a n/a	Prevalence study of alcohol and other drugs (illicit and medications) from biological samples taken from people fatally injured in MVAs in the state of Victoria.	–	Biological detection	Fatal MVA
2022 Anzillotti, L	Italy	01/2009 to 12/2019	327 49 n/a	Descriptive report of blood sample analysis for alcohol and psychoactive drugs in fatally injured drivers involved in MVAs in one region of Italy.	–	Biological detection	Fatal MVA
1976 Garriott, JC	United States	06/1973 to 12/1974	135 n/a n/a	Descriptive analysis of blood sample findings of drivers apprehended under the suspicion of DUID (when alcohol ruled out).	–	Biological detection	DWI suspect
1974 Haffner, J	Norway	01/1973 to 12/1973	74 n/a n/a	Descriptive analysis of blood tests for alcohol and diazepam of drivers involved in MVAs and admitted to an Oslo hospital due to their injuries.	–	Biological detection	MVA with hospitalization

BZRA: benzodiazepine receptor agonist; DUID: driving under the influence of drugs; DWI: driving while impaired; MVA: motor vehicle accident.