9

Injury and Death

This chapter discusses the association between cannabis use and all-cause mortality, occupational injury, motor vehicle accidents, and overdose injuries and death. These health endpoints are distinguished not only by their status as significant public health issues but also by the extent to which directed public health actions and policy changes hold the potential for lessening their detrimental impacts on population health. Motor vehicle accidents are a leading cause of death and injury in the United States, and occupational injuries, especially those that permanently limit an individual’s capacity to perform tasks at home and in the workplace, impose substantial economic burdens on workers, employers, and communities. If research indicates that cannabis use is positively associated with either occupational injury or motor vehicle accidents, evidence-based policies limiting the use of cannabis while driving or in the workplace could potentially reduce the incidence of cannabis-related

accidents and injury. Similarly, research suggesting that cannabis use is linked to mortality could prompt the development of programs to educate health professionals and the general public on the effects of cannabis use and positively influence cannabis-related mortality rates.

In this chapter, the committee reviews and draws conclusions from the findings of six good- to fair-quality systematic reviews and 18 primary literature articles that best address the committee’s research questions of interest. Study limitations and research gaps are noted, and the strength of the available evidence is weighed in five formal conclusions.

ALL-CAUSE MORTALITY

The Institute of Medicine (IOM) report Marijuana and Medicine: Assessing the Science Base states that “epidemiological data indicate that in the general population marijuana use is not associated with increased mortality” (IOM, 1999, p. 109). More recently, modeling studies have estimated that a substantial disease burden—and the associated decrements in the quality and length of life—can be attributed to cannabis use (Degenhardt et al., 2013; Imtiaz et al., 2016). By contrast, a recent systematic review informed by epidemiological data did not report a statistically significant association between cannabis use and mortality (Calabria et al., 2010). This section reviews the available literature to assess the evidence and develop conclusions about cannabis-related mortality.

Is There an Association Between Cannabis Use and All-Cause Mortality?

Systematic Reviews

Calabria et al. (2010) conducted a systematic review to determine the association between cannabis use and all-cause mortality in the general population, and they identified two prospective epidemiological cohort studies relevant to this health endpoint.¹ A meta-analysis of these studies was not performed; consequently, the results of the individual studies are presented below.

Sidney et al. (1997) assessed the risk of mortality associated with cannabis use in a cohort of 65,171 individuals ages 15 to 49 years who were enrolled in the Kaiser Permanente Medical Care Program and followed

___________________

¹ The review also addressed the association between cannabis use and health endpoints that are often or always fatal, such as motor vehicle accidents, cancer, and suicide. These health endpoints are not reviewed in this section, as they are discussed elsewhere in the report.

for a mean length of 10 years. Compared to men who never smoked or who smoked experimentally (i.e., cannabis use on one to six occasions), those who were current smokers were at a significantly increased risk of all-cause mortality after adjusting for several potential confounders, including cigarette smoking, alcohol use, and demographic and socioeconomic factors (relative risk [RR], 1.33, 95% confidence interval [CI] = 1.11–1.59). Notably, among men who currently smoked cannabis, the relative risk of mortality due to AIDS was significantly elevated (RR, 1.90, 95% CI = 1.33–2.73), while the risk of mortality due to known causes other than AIDS was not significantly elevated (RR, 1.12, 95% CI = 0.89–1.39). After accounting for potential confounders, women who currently smoked cannabis were not at a significantly increased risk of all-cause mortality compared to those who had never smoked or who had smoked experimentally (RR, 1.09, 95% CI = 0.80–1.48). Among men who currently smoked cannabis, the frequency of use had only a small effect on the risk of all-cause mortality: those who smoked at least once per week and those who smoked daily were at, respectively, 46 percent (RR, 1.46, 95% CI = 1.19–1.79) and 43 percent (RR, 1.43, 95% CI = 1.08–1.90) greater relative risk of all-cause mortality than nonusers and experimental users. In women, the frequency of use among current smokers had a larger impact on the risk of mortality: those who smoked at least once a week had a less elevated risk of mortality than those who smoked daily as compared to nonusers and experimental users (RR, 1.23, 95% CI = 0.84–1.80 versus RR, 1.44, 95% CI = 0.80–2.56).

Andreasson and Allebeck (1990) reported that among 45,540 Swedish male military conscripts followed for 15 years, the relative risk of mortality was elevated for those who reported having smoked cannabis more than 50 times by the time of conscription compared to nonsmokers (RR, 2.8, 95% CI = 1.9–4.1). After adjusting for multiple confounders, including smoking tobacco, alcohol use, and other drug use, the relative risk of mortality for heavy cannabis smokers was no longer significantly elevated compared with nonsmokers (RR, 1.2, 95% CI = 0.7–1.9). Similarly, participants who reported having smoked cannabis on fewer than 50 occasions by the time of conscription were not at significantly greater risk than nonsmokers after adjustments (RR, 0.7, 95% CI = 0.4–1.2).

Primary Literature

Muhuri and Gfroerer (2011) assessed the risk of all-cause mortality associated with the use of cannabis and other illegal drugs among 20,983 adults over a 15-year follow-up period. After adjusting for confounders, including alcohol use, cigarette smoking, and demographic factors, individuals who reported using cannabis, but not other substances (i.e.,

cocaine, heroin, hallucinogens, inhalants), at baseline were not at increased risk of all-cause mortality compared with individuals who reported not using cannabis or other substances at baseline (hazard ratio [HR], 1.07, 95% CI = 0.85–1.33). Manrique-Garcia et al. (2016) conducted a follow-up study of a cohort of 50,373 Swedish male military conscripts to characterize the potential association between mortality and heavy cannabis use (i.e., using cannabis more than 50 times by 18 years of age). Among the cohort as a whole, heavy cannabis use was associated with a significantly increased risk of mortality compared with nonuse (HR, 1.4, 95% CI = 1.1–1.8). Notably, heavy cannabis use as compared with nonuse did not appreciably affect the risk of mortality among individuals with psychotic disorders—for whom the risk of mortality was particularly elevated (HR, 3.8, 95% CI = 2.6–6.2 versus HR, 3.7, 95% CI = 3.1–4.4).

Discussion of Findings

Sidney et al. (1997) found a statistically significant association between cannabis use and increased risk of all-cause mortality among men diagnosed with AIDS, but not among men without this diagnosis or among women. The authors suggest that the relationship between cannabis use and all-cause mortality among male AIDS patients was not causal; instead, it “most likely represented uncontrolled confounding by male homosexual behavior” (Sidney at al., 1997, p. 589). Limitations in Sidney et al. (1997) include the use of self-report without biological validation to assess patterns of cannabis use; the lack of post-baseline assessments of cannabis use, by which changes over time in the frequency of use could be documented; a lack of data on other substance use, creating the possibility for residual confounding; and, the inability to follow participants into later age, where potential long-term health effects of cannabis use may have emerged.

After accounting for potential confounders, Andreasson and Allebeck (1990) found no statistically significant association between cannabis use and mortality. Furthermore, although a high proportion of deaths among participants who reported smoking cannabis on 50 or more occasions by the time of conscription were due to suicide or uncertain suicide, use of narcotics was also common in these incidents, leading the authors to suggest that a “significant share of the mortality associated with cannabis abuse in this study is attributable to intravenous drug abuse” (Andreasson and Allebeck, 1990, p. 14). Limitations of the study include the use of non-anonymous self-report to collect data on patterns of cannabis use, and the lack of any post-baseline assessments of cannabis use.

Findings from Muhuri and Gfroerer (2011) are based on data from the 1991 National Health Interview Survey’s Drug and Alcohol Use supple-

mental questionnaire, and they indicate a lower prevalence of cannabis use than that seen in the 1991 National Household Survey on Drug Abuse (NHSDA) (45.2 percent versus 52.7 percent). If this discrepancy in the prevalence of cannabis use reported by two national surveys conducted in the same year is the result of underreporting by participants who died during the follow-up period, the mortality risk associated with cannabis use could have been underestimated. Other limitations include the use of self-report to collect data on patterns of cannabis use and the lack of post-baseline assessments to detect changes in cannabis use. Strengths of the study include a base population from a national household sample and an analysis that excluded users of other important illicit drug categories—heroin, cocaine, hallucinogens, and inhalants.

Findings from Manrique-Garcia et al. (2016) have several limitations. Risk estimates are based on cannabis use as of the time of conscription rather than lifetime cannabis exposure and therefore do not account for cannabis use during the ~40 year follow-up period. Similarly, data on potential confounders after the time of conscription is unavailable, so the extent to which they affected study participants and potentially impacted all-cause mortality risk is unknown. Finally, since data on cannabis use was collected by non-anonymous self-report without biological validation, cannabis use may have been underreported.

There is an overall dearth of cohort studies empirically assessing general population cannabis use and all-cause mortality. Although the available evidence suggests that cannabis use is not associated with an increased risk of all-cause mortality, the limited nature of that evidence makes it impossible to have confidence in these findings. These conclusions are not informed by the results of existing large-scale modeling studies that synthesized data from a variety of sources to estimate the burden of disease attributable to cannabis use (Degenhardt et al., 2013; Imtiaz et al., 2016). Although these studies were methodologically rigorous, their direct applicability to actual cannabis-related mortality rates in the United States is uncertain. Consequently, the committee chose not to include them in this review. Also excluded from review were studies of mortality among persons with known cannabis addiction or dependence, those who have been under medical treatment for these disorders, or those who were identified through a country’s criminal justice system, due to presence in these populations of important and often inadequately controlled confounders such as concurrent mental illness and poly-substance abuse.

OCCUPATIONAL INJURY

The Bureau of Labor Statistics reported that 4,821 fatal occupational injuries occurred in the United States in 2014, or about 3.4 fatal injuries for every 100,000 full-time equivalent workers (BLS, 2016). Private industry and state and local government employers reported another 3,486,400 nonfatal occupational injuries in the same year (BLS, 2015). The economic impact of these injuries is considerable. Leigh (2011) estimated that the average medical costs per nonfatal and fatal injury in 2007 were $5,369 and $55,595, respectively. Nationally, the medical and indirect costs of occupational injuries (fatal and nonfatal) totaled $191.83 billion in 2007 (Leigh, 2011). Marucci-Wellman et al. (2015) estimated that in the United States the direct workers’ compensation cost of the most severe, nonfatal occupational injuries was over $51 billion in 2010.²

Concurrent with this economic and public health burden is the increasing prevalence of cannabis use among employed U.S. adults ages 18 and older (Azofeifa et al., 2016). In 2015, 14.4 percent of U.S. adults ages 18 and older with full-time employment reported using cannabis during the previous year (CBHSQ, 2016, pp. 246–247). Among those employed part-time, the proportion was higher, at 17.8 percent (CBHSQ, 2016, pp. 246–247).³

Determining whether an association exists between cannabis use and occupational injury is the subject of ongoing research. According to the 1994 National Research Council and IOM report Under the Influence?: Drugs and the American Workforce, evidence on the relationship between employee drug use and accidents in the workplace is mixed (NRC and IOM, 1994, p. 144). This section updates these findings with a review of the current evidence on cannabis use and occupational injury.

Is There an Association Between Cannabis Use and Occupational Injury?

Systematic Review

The committee did not identify a good- or fair-quality systematic review that reported on the association between cannabis use and occupational injury.

___________________

² Cost estimate is in 2010 dollars.

³ These percentages correspond to 17,042,000 and 5,770,000 U.S. adults ages 18 or older with full-time and part-time employment, respectively.

Primary Literature

The committee identified six primary literature articles addressing the association between cannabis use and occupational injury. Case series of occupational fatalities, with or without forensic investigation, were not considered if there was no consideration of risk compared to non-cannabis-exposed groups.

To investigate the potential association between cannabis use and work-related and non-work-related injuries and accidents, Wadsworth et al. (2006) sent questionnaires on drug use, history of accidents and injuries, and problems with memory or attention to 30,000 residents of two communities in Wales. Based on data from 7,979 completed questionnaires, there was no statistically significant association between cannabis use in the previous year and the risk of minor occupational injuries (i.e., work-related injuries not requiring medical attention) (odds ratio [OR], 1.17, 95% CI = 0.74–1.86), work-related accidents at work requiring medical attention (OR, 0.91, 95% CI = 0.43–1.89), or work-related traffic accidents (OR, 3.01, 95% CI = 0.89–10.17) as compared to no illicit drug use and after adjusting for potentially confounding risk factors (e.g., mental and physical health problems, history of risk-taking behavior, limited work experience).

Wadsworth et al. (2006) also stratified the study population into groups with low and high levels of potential risk factors for work-related accidents and injuries, and they determined the association between cannabis use and the risk for occupational injury for each. Compared to participants who did not use illicit drugs in the previous year and who had few other risk factors, those who used cannabis in the previous year had a significantly elevated risk of suffering minor occupational injuries in the past year if they also had several other risk factors (OR, 8.49, 95% CI = 5.37–13.42), but not if they had few other risk factors (OR, 1.10, 95% CI = 0.47–2.57). The risk of suffering a work-related accident requiring medical attention in the previous year was also significantly elevated for participants who used cannabis in the previous year and had several other risk factors (OR, 3.85, 95% CI = 1.89–7.82), but not for participants who used cannabis in the previous year and had few other risk factors (OR, 0.92, 95% CI = 0.22–3.92) when compared to those who reported no illicit drug use in the previous year and who had few other risk factors. When individuals who used no illicit drugs in the previous year and who had few other risk factors were the referent, the risk of work-related traffic accidents in the previous year was significantly increased for individuals who used cannabis in the previous year, whether or not they had high levels (OR, 6.06, 95% CI = 1.37–26.77) or low levels (OR, 3.24, 95% CI = 1.19–8.79) of other risk factors.

Hoffmann and Larison (1999) used data on 9,097 full- and part-time employees ages 18 and older who participated in the 1994 NHSDA to

evaluate the potential association between cannabis use and the risk of work-related accidents (i.e., accidents that occur at work and that result in damage to property or equipment, injury to oneself, and/or injury to others). They found no statistically significant association between any category of former cannabis use (i.e., used 3 or more years ago, used 1–3 years ago) or any category of current use (i.e., used 1–2 days in past year, used 3–51 days in the past year, used at least weekly in past year) and the risk of work-related accidents as compared to never using cannabis.⁴

Shipp et al. (2005) conducted a cross-sectional study to assess the association between self-reported nonfatal occupational injuries and the self-reported use of substances among 3,265 students attending high school in Texas who indicated that they currently (or had previously) worked for pay. Compared to currently employed students who did not smoke cannabis, those who reported using cannabis on one to nine occasions in the previous 30 days reported a significantly increased risk of occupational injury (OR, 1.37, 95% CI = 1.06–1.77) after adjusting for potential confounders, including year in high school, biological sex, and ethnicity. Heavier cannabis use was associated with higher risk: students who reported using cannabis more than 40 times in the past 30 days were more than twice as likely to have suffered a nonfatal occupational injury as those who did not use cannabis (OR, 2.47, 95% CI = 1.64–3.71) during this period. Adjusting for intensity of work (hours of work per week) decreased the strength of the association between cannabis use and occupational injury; nevertheless, that association remained statistically significant for students who had used cannabis one or more times over the course of their lifetimes (1 to 9 times: OR, 1.45, 95% CI = 1.10–1.90; 10 to 39 times: OR, 1.46, 95% CI = 1.01–2.12; 40+ times: OR, 1.87, 95% CI = 1.38–5.34) or 40 or more times in the previous 30 days (OR, 2.23, 95% CI = 1.34–3.71) as compared to students who did not used cannabis during these periods.

To investigate the association between cannabis use and occupational injury, urine samples collected from individuals working in the United States who had experienced an occupational injury were tested for the presence of cannabis metabolites and were compared to samples collected from individuals selected for a random employee drug test (Price, 2014). To control for the potential confounding effect of other substances, individuals with samples containing amphetamines, phencyclidine, or cocaine or opiate metabolites were removed from the analysis. Among the

___________________

⁴ ORs for these variables ranged from 1.51 for “used 1–2 days in past year” to 0.98 for “used 3–51 days in past year,” where the referent was never use of cannabis. Hoffmann and Larison (1999) did not provide confidence intervals for these ORs, though they indicated in the text that none achieved statistical significance at the p <0.05 level.

remaining 961 cases and 2,834 controls, individuals whose urine samples contained detectable levels of cannabis metabolites were not significantly more likely to have suffered an occupational injury than those whose samples did not (OR, 0.814, 95% CI = 0.625–1.060).

Macdonald et al. (2010) conducted a literature review to answer several research questions related to workplace drug testing for cannabis, including whether employees who report using cannabis or who test positive for cannabis are at an increased risk for occupational injuries. Findings from the reviewed studies were mixed, with not all studies showing a statistically significant association between cannabis use and occupational injury. The authors also sought to determine whether chronic cannabis users have cognitive deficits that place them at an increased risk for occupational injuries, and they reported that although some studies suggest an association between cannabis use and reduced cognitive functioning, the impact of any such deficits on the risk of occupational injury has not been determined.

Dong et al. (2015) evaluated longitudinal data on 12,686 participants in the National Longitudinal Survey of Youth in order to identify factors associated with work-related incidents resulting in injury or illness. Among participants ages 14 to 22 years at study baseline and who reported working in construction between 1988 and 2000, there was no statistically significant association between either lifetime cannabis use on 1–10 occasions (OR, 1.04, 95% CI = 0.94–1.15) or lifetime cannabis use on 11 or more occasions (OR, 1.10, 95% CI = 0.99–1.21) and the incidence of occupational injury or illness when never use of cannabis was the referent.

In addition to the articles reviewed above, the committee identified several articles that—while relevant—were published prior to 1999 (Kaestner and Grossman, 1995, 1998; Zwerling et al., 1990) or that considered research questions closely related—but not identical—to the one addressed here (Fransen et al., 2006). Although these articles did not directly inform the committee’s conclusions, they aided the committee in orienting themselves to the broader literature on risk factors for occupational injury.

Discussion of Findings

Although Wadsworth et al. (2006, p. 11) concluded that their findings “suggest a detrimental impact of cannabis use on safety that is apparent both in and out of the workplace,” they also list several limitations of the study and recommend caution in interpreting its results. Data on cannabis use was derived from self-report and did not measure duration or frequency of cannabis use nor the timing of cannabis use in relation to accidents or injuries. Furthermore, the study may not have completely

controlled for the effect of potential confounders, which may work independently of, or interactively with, cannabis use to modify the risk of occupational injuries or accidents. Finally, the risk for occupational injury posed by cannabis use may be attenuated by processes of self-selection in which cannabis users choose on average to work in lower-risk occupations and nonusers choose to work in higher-risk occupations.

Findings from Hoffmann and Larison (1999) also have several limitations. First, the study did not distinguish between work-related accidents resulting in damage to property and those resulting in injury. Second, the study did not determine whether cannabis use took place while at work; consequently, this type of cannabis use could pose a risk for occupational injury, even if current or former cannabis use in general does not. Third, it is not possible to determine from the NHSDA data whether cannabis use occurred proximate to the injury or whether it preceded or followed an occupational accident.

Shipp et al. (2005) note that the scarcity of research on the association between substance abuse and occupational injuries in adolescent populations prevents the comparison of their results with those from other studies. Because the students who were absent from school on the day of the survey may have had a higher or lower risk of injury compared to students who completed the survey, the potential for selection bias exists. Other limitations of the study include the inability to determine whether cannabis use occurred during work hours or at another time, whether cannabis use preceded or followed the injury, or how closely in time the two events occurred.

In Price (2014), urine samples were collected from men and women of different ages living in different states and employed in a variety of industries with unequal levels of safety sensitivity. The analysis did not control for these variables or determine whether they affect the risk of occupational injury. Furthermore, the study results could not be used to distinguish between recent and remote cannabis use or to determine the chronicity of cannabis use or the extent of an individual’s tolerance for cannabis.

Results from Dong et al. (2015) were limited to those participants who reported working in construction and do not address the potential association between cannabis use and the risk of occupational injury in other industries. Participants who stated they had experienced an occupational injury during a specific time period were not asked how many such injuries occurred. As a result, the study may have underestimated the true number and risk of occupational injuries. Finally, the reference period for survey questions were long and changed over the course of the study, creating the possibility for recall bias.

In addition to these limitations, the studies were extremely diverse

in terms of the characteristics of study participants and their occupations, the specificity and scope of data on cannabis use and occupational injuries, and the extent to which the authors effectively controlled or accounted for potential confounders or effect modifiers. In light of the diversity among and limitations of these studies, it was not possible to determine whether general, nonmedical cannabis use is associated with a clearly increased risk of occupational accidents and injuries across a broad range of occupational and industrial settings in the absence of other important risk factors.

MOTOR VEHICLE CRASHES

In 2011, motor vehicle crashes (MVCs) were the leading cause of death among U.S. adolescents and adults ages 16 to 25 years (NHTSA, 2015). Among all age groups, MVCs occurring in 2014 resulted in more than 32,000 fatalities and more than 2 million nonfatal injuries in the United States (CDC, 2016a; NHTSA, 2016).⁵ Nationally, the combined medical and work loss costs associated with these fatal and nonfatal injuries is substantial at $44 and $51.3 billion, respectively (Bergen et al., 2014; CDC, 2015).⁶

In 2014, 3.2 percent of individuals ages 16 to 25 years reported driving while intoxicated by cannabis (Azofeifa et al., 2015), and the prevalence of THC metabolites detected in the blood or oral fluids of weekend nighttime drivers participating in the National Roadside Survey rose from 8.6 percent in 2007 to 12.6 percent in 2013–2014 (Berning et al., 2015). Given the public health burden of MVC-related morbidity and mortality and the

___________________

⁵ NHTSA defines a fatal crash as “a police-reported crash involving a motor vehicle in transport on a trafficway in which at least one person dies within 30 days of the crash.” Total includes drivers and passengers of motor vehicles, motorcyclists, pedestrians, and cyclists (NHTSA, 2016). Data on nonfatal injuries obtained from the Centers for Disease Control and Prevention’s (CDC’s) Web-based Injury Statistics Query and Reporting System (WISQARS). Total includes all unintentional injuries that occurred on a public road or highway and were traffic related and that resulted in an emergency department visit (CDC, 2016a).

⁶ Total lifetime medical and work loss costs associated with fatal injuries consequent to MVC, based on MVCs occurring in 2013, was $44 billion (CDC, 2015). Total lifetime medical ($18.4 billion) and work loss ($32.9 billion) costs associated with nonfatal injuries consequent to MVC, based on MVCs occurring in 2012, was $51 billion (Bergen et al., 2014). Work loss costs are defined as “estimates of how much a person who died in a motor vehicle crash would have earned over the course of their life, had they not died,” and include salary, estimated benefits, and value of household work (CDC, 2015).

presence of cannabis use and intoxication while driving, there is a need for research to understand the effects of cannabis use on the incidence and severity of motor vehicle crashes and the safety and performance of drivers.

Is There an Association Between Cannabis Use and Motor Vehicle Crashes?

Systematic Reviews

The committee identified a total of six systematic reviews of fair or good quality that summarized the association between driving under the influence of cannabis (DUIC) and MVCs (Asbridge et al., 2012; Calabria et al., 2010; Elvik, 2013; Hartman and Huestis, 2013; Li et al., 2012; Rogeberg and Elvik, 2016). Rogeberg and Elvik (2016) was both the most comprehensive and most recently published systematic review. This review pooled studies reviewed in three earlier meta-analyses (Asbridge et al., 2012; Elvik, 2013; Li et al., 2012) and also performed a structured search of online databases. Calabria et al. (2010) evaluated the association between DUIC and fatal MVCs only, but, with the exception of Bedard et al. (2007), all of the studies in this earlier review were also included in Rogeberg and Elvik (2016). Bedard et al. (2007) was excluded by Rogeberg and Elvik (2016) because it was an analysis of cross-sectional data collected by the U.S. Fatal Accident Reporting System registry.

The meta-analysis by Rogeberg and Elvik (2016) summarized evidence from 21 case-control or culpability studies in 13 countries with a combined sample count of 239,739 participants. There were a total of 28 estimates available from these 21 observational studies. The authors of this systematic review limited their analysis to evidence from either case-control studies or culpability studies and did not include evidence from cross-sectional or cohort studies. The primary criterion for inclusion in the review was the quality of information that indicated cannabis use (i.e., laboratory analyses of blood samples, saliva samples, and urine samples; prescriptions; or self-report) and whether cannabis had been used while driving or enough time prior to driving for effects to still persist. The authors included a wide range of recent studies, including non-peer-reviewed data published by Compton and Berning (2015). Rogeberg and Elvik (2016) argued that culpability studies need to be adjusted for baseline culpability rates because the odds of culpable MVCs associated with DUIC are de facto higher than the overall increase in crash risk. Another important strength of this review is the careful adjustment for potential confounders, including alcohol, in the analysis.

Overall, the meta-analysis by Rogeberg and Elvik (2016) found that

DUIC, as indicated by self-reported cannabis use or the presence of THC metabolite in blood, saliva, or urine, was associated with 20 to 30 percent higher odds of an MVC. The authors described the magnitude of this association as low to moderate in range, and the committee agrees with that assessment. Specifically, the estimated ORs were 1.36 (95% CI = 1.15–1.61) for an analysis that used a random-effects approach and 1.22 (95% CI = 1.10–1.36) for a meta-regression analysis using a precision effect estimate with standard errors (PEESE) technique. Subgroup analyses that accounted for alcohol intoxication found that the magnitude of these ORs weakened to 1.11 (95% CI = 1.04–1.18) when using random-effects and to 1.18 (95% CI = 1.07–1.30) when using PEESE; by contrast, an analysis that did not account for alcohol intoxication found that the ORs were 1.79 (95% CI = 1.28–2.51) and 1.69 (95% CI = 1.25–2.28), respectively.

Primary Literature

The committee did not identify any relevant, good-quality primary literature that reported on the association between cannabis use and motor vehicle crashes and were published subsequent to the data collection period of the most recently published good- or fair-quality systematic review addressing the research question. Of the three identified papers with publication dates during or after 2015 that were not included in Rogeberg and Elvik (2016), none contributed new data on the association between DUIC and MVC risk (Allen et al., 2016; Lemos et al., 2015; Meibodi et al., 2015).

Discussion of Findings

Two important methodological limitations of Rogeberg and Elvik (2016) were noted by other researchers (Gjerde and Morland, 2016). First, DUIC may have not just referred to acute intoxication. Indeed, many of the studies considered in this review scored case and control counts as positive using criteria that would also be satisfied by drivers with recent or regular cannabis use but who were neither intoxicated nor impaired while driving (Gjerde and Morland, 2016). Moreover, the association between THC levels in blood and either acute intoxication or driving impairment remains a subject of controversy, and it could represent an important limitation in the interpretation of findings in culpability studies based on blood THC levels (Desrosiers et al., 2014; Khiabani et al., 2006; Logan et al., 2016; Menetrey et al., 2005; Papafotiou et al., 2005). Second, 3 of the 21 studies used different methods to assess cases and controls, which may lead to a non-differential misclassification of exposure. A missing component in this review is a better determination of the dose

at which driving becomes sufficiently unsafe as to increase MVC risk. Finally, Rogeberg and Elvik (2016) did not provide evidence from cohort studies to address DUIC in MVC.

Simulator studies were also not included in Rogeberg and Elvik (2016). Some laboratory and simulator studies that have examined the effects of acute cannabis intoxication on driving performance have found that the psychomotor skills necessary for safe driving become increasingly impaired at higher doses of cannabis (Sewell et al., 2009). While these experiments may have high internal validity regarding dose-related effects on psychomotor performance, they do not necessarily reflect the complex nature of driving ability and MVC risk attributed to DUIC in a real-world scenario. Epidemiological studies of MVC in populations may help to address these limitations and are the only reasonable and ethical alternative to controlled experiments outside the laboratory. However, cannabis smokers have demographic characteristics that are similar to those of other groups with a high crash risk, including youth, males, and those with a high prevalence of drugged and drunk driving (Bergeron and Paquette, 2014; Richer and Bergeron, 2009). In particular, confounding or effect modification with alcohol is an important driver-related factor that needs to be better taken into account. The bulk of the evidence available describing the association between DUIC and MVCs comes from case-control studies that evaluate the odds of a MVC by DUIC status and from culpability studies which evaluate the odds of culpability in drivers involved in collisions by DUIC status.

OVERDOSE INJURIES AND DEATH

According to the American Association of Poison Control Centers (AAPCC), 2,047 calls to position control centers in the United States made in 2014 were in response to single-substance exposures to cannabis, up from 1,548 such exposures in 2013 (Mowry et al., 2014, 2015). Of these exposures, 37 were classified as having major effects, and death was the outcome in 1 (Mowry et al., 2015).⁷ However, these data do not account for overdose injuries or deaths that did not prompt calls to poison con-

___________________

⁷ Major effects are defined as those that are “life-threatening or [that] resulted in significant residual disability or disfigurement” (Mowry et al., 2015, p. 1125). Exposures classified as resulting in death are those where “the patient died as a result of the exposure or as a direct complication of the exposure” (Mowry et al., 2015, p. 1125).

trol centers. Data from the Wide-ranging Online Data for Epidemiologic Research (WONDER) database of the Centers for Disease Control and Prevention indicate that in 2014 there were 16,822 deaths in the United States due to accidental poisoning by and exposure to narcotics and psychodysleptics—a broad category that includes cannabis as well as cocaine, heroin, codeine, morphine, and several other narcotics (CDC, 2016b; WHO, 2016). Due, in part, to the limitations of current surveillance tools and medical record coding systems, there is a limited amount of more comprehensive and precise data on the association between cannabis use and overdose injury or death.

Meanwhile, the increasing availability, diversity, and potency of cannabis products create the potential for an increased risk of adverse health effects related to cannabis use, including overdose injury and death. Accidental ingestion of cannabis by young children can result in respiratory failure and coma, as noted by several case reports (Amirav et al., 2011; Appelboam and Oades, 2006; Carstairs et al., 2011), and the consumption of cannabis edibles has been identified as a contributing factor in the accidental death of at least one adolescent (Hancock-Allen et al., 2015).

Thus, the emerging cannabis products market creates the potential for an increased risk of cannabis-related overdose injury or death, while limitations in the current clinical and public health surveillance system hinder efforts to detect, characterize, and respond to this population health issue. This section reviews the available evidence on the association between cannabis use and overdose injury and death and discusses possible actions to improve the state of research on this health endpoint.

Is There an Association Between Cannabis Use and Overdose Injuries or Death?

Systematic Reviews

The committee did not identify a good- or fair-quality systematic review that reported on the association between cannabis use and overdose injuries or death.

Primary Literature

The committee identified a number of studies that directly or indirectly reported on the association between acute cannabis intoxication and overdose death in either adults or children. An analysis of the National Poison Data Systems database involving more than 2 million human exposure cases in 2012 did not list cannabis among the top causes of death related to pharmaceutical products (Dart et al., 2015). According

to AAPCC annual reports, among all calls to U.S. poison centers involving single-substance exposures to cannabis, death was the outcome in two cases in 2012, no cases in 2013, and one case in 2014 (Mowry et al., 2013, 2014, 2015), although the reports do not indicate whether cannabis exposure was a contributing factor in these outcomes. Cannabis was not found to be the main cause of death in any of the fatal intoxications among drug addicts submitted for medico-legal autopsy and toxicological analysis in Denmark, Finland, Iceland, Norway, or Sweden in either 2007 or 2012 (Simonsen et al., 2011, 2015). Nonetheless, tetrahydrocannabinol was commonly identified (21 percent to 38 percent of cases) in the blood samples of these fatal intoxications.

Case reports on cannabis-related deaths are also uncommon. In Colorado, cannabis intoxication was determined to be a chief contributing factor in the death by trauma of a teenager who jumped from a fourth-floor balcony after ingesting a cookie containing 65 mg of THC (Hancock-Allen et al., 2015). Postmortem analyses revealed no evidence of polysubstance abuse and a delta-9 carboxy-THC whole blood concentration of 49 ng/ml—almost nine times the legal limit for driving in Colorado. Colorado law states that a single-serving edible cannabis product should contain no more than 10 mg of THC; however, currently available edible cannabis products such as cookies and brownies, which are otherwise generally understood as single-serving products, may contain as much as 100 mg (or 10 servings) of THC.⁸ In a study on unintentional pediatric cannabis exposure, Wang et al. (2016) described a case where hospital staff members were unable to resuscitate an unresponsive 11-month-old child who presented with tachycardia and metabolic acidosis and who tested positive for THC in a urine drug screen. The authors noted that any relationship between cannabis exposure and the patient’s symptoms or outcome was unclear. Although presented here for discussion, these case reports did not inform the committee’s conclusions on the association between cannabis use and overdose death.

By comparison with the minimal literature on cannabis-related overdose death in adults or children, several studies reported on potentially serious symptoms associated with cannabis exposure in pediatric populations. Le Garrec et al. (2014) reported that, over a 3.5-year period, seven children ages 11 to 33 months were admitted to a pediatric intensive care unit in Paris with accidental cannabis poisoning. All of the children had central nervous system symptoms, including drowsiness and coma, and three were intubated and placed on mechanical ventilation for less than 24 hours. Between 2010 and 2013, an Arizona poison control center received

___________________

⁸ Colorado Code of Regulations. Department of Revenue. Marijuana Enforcement Division. Retail Marijuana Rules. 1 CCR 212-2 R604 (C5) (2).

49 calls related to unintentional medical marijuana ingestions among children ages 7 years and younger (Lovecchio and Heise, 2015). Among the 39 records with complete information, the most commonly reported symptoms were lethargy (48 percent of cases), an inability to walk (53 percent), coma (10 percent), and vomiting (21 percent). These and other symptoms, including respiratory depression and aspiration pneumonia, underscore the importance of observation in children suspected or known to have unintentionally ingested cannabis. Although presented here for discussion, these case series were published as letters in scientific journals and therefore did not inform the committee’s conclusions on the association between cannabis use and overdose injuries.

These findings are supported by retrospective reviews and cohort studies. Wang et al. (2013) retrospectively reviewed cases of unintentional cannabis ingestions among children ages 11 and younger who required medical attention at a children’s hospital in Colorado between 2005 and 2011. Out of 1,378 unintentional ingestions, only 14 were cannabis related, of which 13 were observed in the emergency room or admitted to the hospital. Symptoms included lethargy, ataxia, dizziness, and respiratory insufficiency. The proportion of unintentional ingestions that were cannabis related increased from 0 percent in 2005–2009 to 2.4 percent in 2009–2013, a statistically significant increase coinciding with the October 2009 decision by the U.S. Department of Justice to no longer prosecute users and suppliers of cannabis who act in accordance with state laws. In a subsequent study, Wang et al. (2016) reported the prevalence of unintentional pediatric cannabis exposures occurring between 2009 and 2015 at a children’s hospital and a poison center in Colorado. The average number of cannabis-related calls per 1,000 calls to the poison center increased significantly from 0.9 in 2012–2013 to 2.3 in 2014–2015, periods corresponding to the 2 years before and after legalization of recreational cannabis in Colorado. Between these same periods, the average number of cannabis-related emergency department visits per 1,000 visits also increased, though nonsignificantly, from 4.3 to 6.4. Symptoms reported in the 163 calls received by the poison center included drowsiness and/ or lethargy (49 percent of cases), ataxia and/or dizziness (12 percent), and agitation (8 percent). Out of 81 cases received by the children’s hospital, 40 percent were observed in the emergency department, 22 were admitted to an inpatient ward or the intensive care unit, and 2 required respiratory support. Onders et al. (2016) reviewed data from the National Poison Data System and found that between 2000 and 2013, U.S. poison centers received 1,969 calls related to cannabis exposure among children younger than 6 years old. Most exposures were unintentional (92.2 percent) and occurred as a result of ingesting cannabis or a cannabis product (75.0 percent). Drowsiness and/or lethargy accounted for nearly half of reported

clinical symptoms (45.5 percent), while more serious effects, including coma (0.9 percent), cardiovascular symptoms (4.1 percent), and respiratory depression (0.7 percent), occurred less frequently. The annual rate of exposures increased over time, from a national average of 4.21 per million children in 2006 to 10.42 per million children in 2013, corresponding to a statistically significant increase of 147.5 percent. During the same period, the increase in the annual rate of exposures among states that had legalized medical cannabis prior to 2000 was significant, at 609.6 percent.

Collectively, these findings indicate that state-based legalization of cannabis is associated with a subsequent increase in pediatric cannabis exposures in those states. A similar trend emerges when comparing exposure rates among states where cannabis is legal to exposure rates in states where it is not. Wang et al. (2014) reported that between 2005 and 2011 the rate of calls to poison centers for unintentional pediatric cannabis exposures did not increase in states where cannabis remained illegal as of 2012; increased by 11.5 percent (95% CI = −0.4–24.7) in states where legislation to legalize cannabis was passed between 2005 and 2011; and increased by 30.3 percent (95% CI = 22.5–38.5) in states where cannabis was legalized before 2005. Among children unintentionally exposed to cannabis, those living in states where cannabis was legalized before 2005 were more likely to be evaluated in a health care facility (OR, 1.9, 95% CI = 1.5–2.6), to experience major or moderate effects (OR, 2.1, 95% CI = 1.4–3.1), and to be admitted to critical care units (OR, 3.4, 95% CI = 1.8–6.5) as compared to those living in states where cannabis remained illegal as of 2012. Accounting for 78 percent of all incidents, ingestion was the most common route of unintentional pediatric exposure. Onders et al. (2016) reported that between 2000 and 2013 the annual rate of poison center calls related to cannabis exposures among children younger than 6 was 2.82 times higher in states that had legalized medical cannabis prior to 2000 than in states where medical cannabis remained illegal as of 2013. Another study found that the mean number of calls to poison control centers for unintentional pediatric cannabis exposures increased by 34 percent per year between 2009 and 2015—a significant increase that was also significantly greater than the 19 percent annual increase in cannabis-related calls received by poison control centers throughout the rest of the United States during that same period (Wang et al., 2016). Informed, in part, by these and other findings, a special committee of the Colorado Department of Public Health and Environment found moderate evidence that more unintentional pediatric cannabis exposures have occurred in states with increased legal access to cannabis and that the exposures can lead to significant clinical effects requiring medical attention (CDPHE, 2015).

Discussion of Findings

The committee identified few studies that report on the association between cannabis use and overdose death. Cannabis was not identified as a main cause in the intoxication deaths of drug addicts in five Nordic countries or a top cause of U.S. deaths related to pharmaceutical products. However, studies on the risks to Nordic populations posed by cannabis products available in those countries may not reflect the risks to U.S. populations posed by domestically available cannabis products, and cannabis might still be associated with overdose deaths without also being a top cause among pharmaceutical-related exposure deaths. Data from the National Poison Data System indicate that death was the outcome in a small number of single-substance exposures to cannabis; however, lacking further information, it is not possible to determine whether and to what extent cannabis contributed to these deaths. Case reports implicate acute cannabis intoxication in one accidental death and suggest that cannabis use may pose a risk for sudden cardiac death. However, these individual case reports cannot be used to infer a general association between cannabis use and overdose deaths. Overall, the committee identified no study in which cannabis was determined to be the direct cause of overdose death.

Several studies report that unintentional pediatric cannabis exposure is associated with potentially serious symptoms, including respiratory depression or failure, tachycardia and other cardiovascular symptoms, and temporary coma. Similar symptoms were not reported in adults exposed to cannabis. Most study limitations were related to the origin, quality, and completeness of data. For example, Wang et al. (2013) noted that findings based on data from a single children’s hospital or regional poison centers may not be generalizable to other health care facilities or poison centers, especially those in areas where laws regarding cannabis use are different than in Colorado. Search strategies employed in retrospective reviews of records from hospitals and poison centers may fail to capture all pertinent records, and some records may be incomplete (Wang et al., 2016). Data from poison centers will capture only the subset of cannabis-related overdose injuries or deaths that resulted in a call to a poison center and may overrepresent serious cases or cases from states where cannabis is legal (Wang et al., 2014). Moreover, Onders et al. (2016) observed that cannabis exposures are not identical to poisonings and overdoses; consequently, data on trends in cannabis exposures do not necessarily allow for an estimation of trends in cannabis overdose or poisoning.

CONCLUSION 9-4

9-4(a) There is insufficient evidence to support or refute a statistical association between cannabis use and death due to cannabis overdose.

9-4(b) There is moderate evidence of a statistical association between cannabis use and increased risk of overdose injuries, including respiratory distress, among pediatric populations in U.S. states where cannabis is legal.

RESEARCH GAPS

To address the research gaps relevant to injury and death, the committee suggests the following:

• There is a need for long-term, well-designed cohort studies to determine the association between cannabis use and all-cause and cause-specific mortality among large, representative populations. These studies will need to assess the effects of the various characteristics of cannabis use (e.g., frequency, duration, cumulative exposure) on mortality among demographic and clinical subgroups of interest, to use credible measures of cannabis exposure, and to control for known confounders.
• The association between cannabis use and occupational injury needs to be explored across a broad range of regions, populations, workplace settings, workplace practices (e.g., drug use prevention programs, safety standards), worker characteristics (e.g., medical history, history of drug and alcohol use), work patterns, and occupations.
• There is a need for research to evaluate whether and how the form of cannabis (e.g., edibles, flower, concentrates) affects the risk of overdose and to characterize the incidence and prevalence of overdose deaths in children and adults due to accidental or intentional exposure to edible cannabis.
• There is a need for well-designed surveillance studies to determine the prevalence of acute cannabis use and intoxication among U.S. drivers. Research is also needed to explore how patterns of cannabis use, the degree of acute cannabis intoxication, and geographic and demographic variables affect MVC incidence, driver and passenger outcomes, and driver safety and performance. Finally, research is needed to identify the causal channels through

• which cannabis use may adversely or therapeutically affect MVC risk.
• There is a need for research on the association between cannabis use and injury and mortality among unstudied and understudied demographic groups, such as minority groups, working adolescents, and employed older populations.

SUMMARY

This chapter discussed the associations between cannabis use and all-cause mortality, occupational injury, motor vehicle crash, and death and injury due to overdose. Box 9-1 provides a summary of the conclusions from this chapter. Notably, the committee found substantial evidence of a statistical association between cannabis use and motor vehicle crashes. These findings suggest the need for research to further specify the strength of this association and to identify any mediating factors, as well as the need for broader surveillance efforts to track patterns of cannabis use, especially where cannabis use may pose risks to personal and public health.

Apart from illuminating potential research objectives, these findings also suggest enacting policies such as making DUIC a direct target for both policy and policing. Such efforts could include checkpoints for DUIC in conjunction with those for sobriety, the development of point-of-care kits for DUIC testing, and a consideration of zero tolerance laws. These proposals find parallels in policies that restrict or prohibit the use of alcohol while driving, and there is both domestic and international precedent for policing the use of cannabis while operating motor vehicles. In Colorado and Washington, an individual whose blood contains 5 ng/ml or more of THC while driving is considered to be under the influence and is guilty of DUIC.⁹ In Australia, it is illegal to drive with any level of THC in oral fluid or blood samples (Boorman and Owens, 2009).¹⁰ Some research suggests that policies that legalize cannabis for medical use have been associated with a decrease in the incidence of MVC. For example, an ecological study found a net reduction in traffic crashes associated with the introduction of laws for medical cannabis use (Anderson et al., 2013).

The committee also found moderate evidence of a statistical association between cannabis use and an increased risk of overdose injuries among pediatric populations in states in where cannabis is legal. The potential risks associated with the use of highly potent cannabis products suggest a need for public health policies, such as regulations that require packaging for cannabis products to include child-focused safety features, warnings that ingested cannabis can have different effects from smoked cannabis, and guidance on how to respond to potential emergencies. Again, precedents for such policies exist. For example, Colorado regulations require that medical and retail cannabis products be sold in packages that are child-resistant, that list the potency of the product in mg of THC and cannabidiol, and that contain several warning statements, including the direction to keep the product out of the reach of children.^11,12

The available evidence was insufficient to draw any conclusions regarding the association between cannabis use and occupational injury or all-cause mortality. The high economic and social costs associated with occupational injuries in this country suggest the need for further research to determine whether these injuries are associated with cannabis use. In pursuing this research, it will be important to determine which individual and work-related factors protect against, or expose workers to, the risk of injury. Emerging evidence suggests that access to legal cannabis can

___________________

⁹ Wash. Rev. Code Ann. § 46.61.502 (1) (b). Colo. Rev. Stat. Ann. § 42-4-1301 (6) (a) (IV).

¹⁰ Road Traffic Act 1974, Part V, Division 2, Section 64AC (1).

¹¹ Colorado Code of Regulations. Department of Revenue. Marijuana Enforcement Division. Medical Marijuana Rules. 1 CCR 212-1 M1004.5 (B) and M1005 (B).

¹² Colorado Code of Regulations. Department of Revenue. Marijuana Enforcement Division. Retail Marijuana Rules. 1 CCR 212-2 R1006 (A–B).

increase the incidence of accidental cannabis ingestion among pediatric populations and that such ingestion can lead to depressed respiratory function and other symptoms of overdose. If state-level changes in cannabis policy continue to make cannabis more accessible, there will be an increased need for research to assess the prevalence of injuries and death due to cannabis overdose, especially among children and other vulnerable populations.

REFERENCES

Boorman, M., and K. Owens. 2009. The Victorian legislative framework for the random testing drivers at the roadside for the presence of illicit drugs: An evaluation of the characteristics of drivers detected from 2004 to 2006. Traffic Injury Prevention 10(1):16–22.

Calabria, B., L. Degenhardt, W. Hall, and M. Lynskey. 2010. Does cannabis use increase the risk of death? Systematic review of epidemiological evidence on adverse effects of cannabis use. Drug and Alcohol Review 29(3):318–330.

Carstairs, S. D., M. K. Fujinaka, G. E. Keeney, and B. T. Ly. 2011. Prolonged coma in a child due to hashish ingestion with quantitation of thc metabolites in urine. Journal of Emergency Medicine 41(3):e69–e71.

CBHSQ (Center for Behavioral Health Statistics and Quality). 2016. 2015 national survey on drug use and health: Detailed tables. http://www.samhsa.gov/data/sites/default/files/NSDUH-DetTabs-2015/NSDUH-DetTabs-2015/NSDUH-DetTabs-2015.pdf (accessed December 27, 2016).

CDC (Centers for Disease Control and Prevention). 2015. State-specific costs of motor vehicle crash deaths. https://www.cdc.gov/motorvehiclesafety/statecosts/index.html (accessed October 18, 2016).

CDC. 2016a. WISQARS: Nonfatal injury reports, 2001–2014. http://webappa.cdc.gov/sasweb/ncipc/nfirates2001.html (accessed October 18, 2016).

CDC. 2016b. WONDER: About underlying cause of death, 1999-2014. https://wonder.cdc.gov/ucd-icd10.html (accessed October 18, 2016).

CDPHE (Colorado Department of Public Health and Environment). 2015. Monitoring health concerns related to marijuana use in Colorado: 2014. http://www2.cde.state.co.us/artemis/hemonos/he1282m332015internet/he1282m332015internet01.pdf (accesssed December 27, 2016).

Compton, R. P., and A. Berning. 2015. Drug and alcohol crash risk. Traffic Safety Facts Research Note. DOT HS 812 117. Washington, DC: National Highway Traffic Safety Administration. http://www.nhtsa.gov/staticfiles/nti/pdf/812117-Drug_and_Alcohol_Crash_Risk.pdf (accessed December 20, 2016).

Dart, R. C., A. C. Bronstein, D. A. Spyker, L. R. Cantilena, S. A. Seifert, S. E. Heard, and E. P. Krenzelok. 2015. Poisoning in the United States: 2012 Emergency Medicine Report of the National Poison Data System. Annals of Emergency Medicine 65(4):416–422.

Degenhardt, L., H. A. Whiteford, A. J. Ferrari, A. J. Baxter, F. J. Charlson, W. D. Hall, G. Freedman, R. Burstein, N. Johns, R. E. Engell, A. Flaxman, C. J. Murray, and T. Vos. 2013. Global burden of disease attributable to illicit drug use and dependence: Findings from the global burden of disease study 2010. Lancet 382(9904):1564–1574.

Desrosiers, N. A., S. K. Himes, K. B. Scheidweiler, M. Concheiro-Guisan, D. A. Gorelick, and M. A. Huestis. 2014. Phase I and II cannabinoid disposition in blood and plasma of occasional and frequent smokers following controlled smoked cannabis. Clinical Chemistry 60(4):631–643.

Dong, X. S., X. Wang, and J. A. Largay. 2015. Occupational and non-occupational factors associated with work-related injuries among construction workers in the USA. International Journal of Occupational and Environmental Health 21(2):142–150.

Elvik, R. 2013. Risk of road accident associated with the use of drugs: A systematic review and meta-analysis of evidence from epidemiological studies. Accident Analysis & Prevention 60:254–267.

Fransen, M., B. Wilsmore, J. Winstanley, M. Woodward, R. Grunstein, S. Ameratunga, and R. Norton. 2006. Shift work and work injury in the New Zealand blood donors’ health study. Occupational and Environmental Medicine 63(5):352–358.

Gjerde, H., and J. Morland. 2016. Risk for involvement in road traffic crash during acute cannabis intoxication. Addiction 111(8):1492–1495.

Hancock-Allen, J. B., L. Barker, M. VanDyke, and D. B. Holmes. 2015. Notes from the field: Death following ingestion of an edible marijuana product—Colorado, March 2014. Morbidity and Mortality Weekly Report 64(28):771–772.

Hartman, R. L., and M. A. Huestis. 2013. Cannabis effects on driving skills. Clinical Chemistry 59(3):478–492.

Hoffmann, J., and C. Larison. 1999. Drug use, workplace accidents and employee turnover. Journal of Drug Issues 29(2):341–364.

Imtiaz, S., K. D. Shield, M. Roerecke, J. Cheng, S. Popova, P. Kurdyak, B. Fischer, and J. Rehm. 2016. The burden of disease attributable to cannabis use in Canada in 2012. Addiction 111(4):653–662.

IOM (Institute of Medicine). 1999. Marijuana and medicine: Assessing the science base. Washington, DC: National Academy Press.

Kaestner, R., and M. Grossman. 1995. Wages, workers’ compensation benefits, and drug use: Indirect evidence of the effect of drugs on workplace accidents. American Economic Review 85(2):55–60.

Kaestner, R., and M. Grossman. 1998. The effect of drug use on workplace accidents. Labour Economics 5(3):267–294.

Khiabani, H. Z., J. G. Bramness, A. Bjorneboe, and J. Morland. 2006. Relationship between THC concentration in blood and impairment in apprehended drivers. Traffic Injury Prevention 7(2):111–116.

Le Garrec, S., S. Dauger, and P. Sachs. 2014. Cannabis poisoning in children. Intensive Care Medicine 40(9):1394–1395.

Leigh, J. P. 2011. Economic burden of occupational injury and illness in the United States. Milbank Quarterly 89(4):728–772.

Lemos, N. P., A. C. San Nicolas, J. A. Volk, E. A. Ingle, and C. M. Williams. 2015. Driving under the influence of marijuana versus driving and dying under the influence of marijuana: A comparison of blood concentrations of delta9-tetrahydrocannabinol, 11-hydroxy-delta9-tetrahydrocannabinol, 11-nor-9-carboxy-delta9-tetrahydrocannabinol and other cannabinoids in arrested drivers versus deceased drivers. Journal of Analytical Toxicology 39(8):588–601.

Li, M. C., J. E. Brady, C. J. DiMaggio, A. R. Lusardi, K. Y. Tzong, and G. Li. 2012. Marijuana use and motor vehicle crashes. Epidemiologic Reviews 34:65–72.

Logan, B., S. L. Kacinko, and D. J. Beirness. 2016. An Evaluation of Data from Drivers Arrested for Driving Under the Influence in Relation to Per Se Limits for Cannabis. AAA Foundation for Traffic Safety: Washington, DC. https://www.aaafoundation.org/sites/default/files/EvaluationOfDriversInRelationToPerSeReport.pdf (accessed December 27, 2016).

Lovecchio, F., and C. W. Heise. 2015. Accidental pediatric ingestions of medical marijuana: A 4-year poison center experience. American Journal of Emergency Medicine 33(6):844–845.

Macdonald, S., W. Hall, P. Roman, T. Stockwell, M. Coghlan, and S. Nesvaag. 2010. Testing for cannabis in the work-place: A review of the evidence. Addiction 105(3):408–416.

Manrique-Garcia, E., A. Ponce de Leon, C. Dalman, S. Andreasson, and P. Allebeck. 2016. Cannabis, psychosis, and mortality: A cohort study of 50,373 Swedish men. American Journal of Psychiatry 173(8):790–798.

Marucci-Wellman, H. R., T. K. Courtney, H. L. Corns, G. S. Sorock, B. S. Webster, R. Wasiak, Y. I. Noy, S. Matz, and T. B. Leamon. 2015. The direct cost burden of 13 years of disabling workplace injuries in the U.S. (1998–2010): Findings from the Liberty Mutual workplace safety index. Journal of Safety Research 55:53–62.

Meibodi, M. K., S. Esfandyari, V. Siyabi, and S. Roosta. 2015. Illicit drug abuse in drivers of motor vehicle collisions. Galen Medical Journal 4(1):39–46.

Menetrey, A., M. Augsburger, B. Favrat, M. A. Pin, L. E. Rothuizen, M. Appenzeller, T. Buclin, P. Mangin, and C. Giroud. 2005. Assessment of driving capability through the use of clinical and psychomotor tests in relation to blood cannabinoids levels following oral administration of 20 mg dronabinol or of a cannabis decoction made with 20 or 60 mg delta9-THC. Journal of Analytical Toxicology 29(5):327–338.

Mowry, J. B., D. A. Spyker, L. R. Cantilena, Jr., J. E. Bailey, and M. Ford. 2013. 2012 annual report of the American Association of Poison Control Centers’ National Poison Data System (NPDS): 30th annual report. Clinical Toxicology 51(10):949–1229.

Mowry, J. B., D. A. Spyker, L. R. Cantilena, Jr., N. McMillan, and M. Ford. 2014. 2013 annual report of the American Association of Poison Control Centers’ National Poison Data System (NPDS): 31st annual report. Clinical Toxicology 52(10):1032–1283.

Mowry, J. B., D. A. Spyker, D. E. Brooks, N. McMillan, and J. L. Schauben. 2015. 2014 annual report of the American Association of Poison Control Centers’ National Poison Data System (NPDS): 32nd annual report. Clinical Toxicology 53(10):962–1147.

Muhuri, P. K., and J. C. Gfroerer. 2011. Mortality associated with illegal drug use among adults in the United States. American Journal of Drug and Alcohol Abuse 37(3):155–164.

NHTSA (National Highway Traffic Safety Administration). 2015. Motor vehicle traffic crashes as a leading cause of death in the United States, 2010 and 2011. https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/812203 (accessed December, 27, 2016).

NHTSA. 2016. Fatality Analysis Reporting System (FARS) encyclopedia. http://www-fars.nhtsa.dot.gov/Main/index.aspx (accessed December 27, 2016).

NRC and IOM (National Research Council and Institute of Medicine). 1994. Under the influence?: Drugs and the American work force. Washington, DC: National Academy Press.

Onders, B., M. J. Casavant, H. A. Spiller, T. Chounthirath, and G. A. Smith. 2016. Marijuana exposure among children younger than six years in the United States. Clinical Pediatrics 55(5):428–436.

Papafotiou, K., J. D. Carter, and C. Stough. 2005. The relationship between performance on the standardised field sobriety tests, driving performance and the level of delta9tetrahydrocannabinol (THC) in blood. Forensic Science International 155(2–3):172–178.

Price, J. W. 2014. Marijuana and workplace safety: An examination of urine drug tests. Journal of Addictive Diseases 33(1):24–27.

Richer, I., and J. Bergeron. 2009. Driving under the influence of cannabis: Links with dangerous driving, psychological predictors, and accident involvement. Accident Analysis & Prevention 41(2):299–307.

Rogeberg, O., and R. Elvik. 2016. The effects of cannabis intoxication on motor vehicle collision revisited and revised. Addiction 111(8):1348–1359.

Sewell, R. A., J. Poling, and M. Sofuoglu. 2009. The effect of cannabis compared with alcohol on driving. American Journal on Addictions 18(3):185–193.

Shipp, E. M., S. R. Tortolero, S. P. Cooper, E. G. Baumler, and N. F. Weller. 2005. Substance use and occupational injuries among high school students in South Texas. American Journal of Drug and Alcohol Abuse 31(2):253–265.

Sidney, S., J. E. Beck, I. S. Tekawa, C. P. Quesenberry, and G. D. Friedman. 1997. Marijuana use and mortality. American Journal of Public Health 87(4):585–590.

Simonsen, K. W., P. T. Normann, G. Ceder, E. Vuori, S. Thordardottir, G. Thelander, A. C. Hansen, B. Teige, and D. Rollmann. 2011. Fatal poisoning in drug addicts in the Nordic countries in 2007. Forensic Science International 207(1-3):170–176.

Simonsen, K. W., H. M. Edvardsen, G. Thelander, I. Ojanpera, S. Thordardottir, L. V. Andersen, P. Kriikku, V. Vindenes, D. Christoffersen, G. J. Delaveris, and J. Frost. 2015. Fatal poisoning in drug addicts in the Nordic countries in 2012. Forensic Science International 248:172–180.

Wadsworth, E. J., S. C. Moss, S. A. Simpson, and A. P. Smith. 2006. A community based investigation of the association between cannabis use, injuries and accidents. Journal of Psychopharmacology 20(1):5–13.

Wang, G. S., G. Roosevelt, and K. Heard. 2013. Pediatric marijuana exposures in a medical marijuana state. JAMA Pediatrics 167(7):630–633.

Wang, G. S., G. Roosevelt, M. C. Le Lait, E. M. Martinez, B. Bucher-Bartelson, A. C. Bronstein, and K. Heard. 2014. Association of unintentional pediatric exposures with decriminalization of marijuana in the United States. Annals of Emergency Medicine 63(6):684–689.

Wang, G. S., M. C. Le Lait, S. J. Deakyne, A. C. Bronstein, L. Bajaj, and G. Roosevelt. 2016. Unintentional pediatric exposures to marijuana in Colorado, 2009–2015. JAMA Pediatrics 170(9):e160971.

WHO (World Health Organization). 2016. Accidental poisoning by and exposure to noxious substances (X40-X49). IDC-10 Version: 2015. http://apps.who.int/classifications/icd10/browse/2015/en#!/X40-X49 (accessed November 30, 2016).

Zwerling, C., J. Ryan, and E. J. Orav. 1990. The efficacy of preemployment drug screening for marijuana and cocaine in predicting employment outcome. JAMA 264(20):2639–2643.

This page intentionally left blank.