Patterns of Harmful Alcohol Consumption among Truck Drivers: Implications for Occupational Health and Work Safety from a Systematic Review and Meta-Analysis

Alcohol consumption is one of the main causes of productivity losses arising from absenteeism, presenteeism, and workplace injuries. Among occupational categories most affected by the use of this substance, truck drivers are subject to risk factors and risky behaviors that can have a serious impact on their health, their work, and the general road safety. The use of alcohol during truck-driving activities is, indeed, an important risk factor for traffic accidents. The present systematic review and meta-analysis aims at synthesizing the literature regarding harmful alcohol consumption patterns among truck drivers in a rigorous way. A ‘binge drinking’ prevalence of 19.0%, 95% confidence interval or CI (13.1, 26.9) was present. An ‘everyday drinking’ pattern rate of 9.4%, 95% CI (7.0, 12.4) was found, while the rate of alcohol misuse according to the “Alcohol Use Disorders Identification Test” (AUDIT)—“Cut down-Annoyed-Guilty-Eye opener questionnaire” (CAGE) instruments was computed to be of 22.7%, 95% CI (14.8, 33.0). No evidence of publication bias could be found. However, there is the need to improve the quality of published research, utilizing standardized reliable instruments. The knowledge of these epidemiological data can be useful for decision makers in order to develop, design, and implement ad hoc adequate policies.


Introduction
"If you bought it, a truck brought it." A popular saying.
Alcohol consumption is among the main causes of absenteeism, presenteeism, and workplace injuries [1][2][3]. In the USA, alcohol-induced impairment directly affects an estimated 15% of the workforce and causes more than 22% of the deaths as a result of injuries at work [4,5].
Our study adds to the recently published work by Girotto et al. [16] in that it significantly updates and expands the published systematic review and performs a rigorous quantitative synthesis of the available scientific evidences and systematically studies the determinants of alcohol use among truck drivers.

Registration of Protocol with International Prospective Register of Systematic Reviews
The protocol of the present study has been reported according to the "Preferred Reporting Items for Systematic Reviews and Meta-Analysis-Protocols" (PRISMA-P) guidelines [47]. In accordance with these guidelines, the systematic review protocol has been registered with the International Prospective Register of Systematic Reviews (PROSPERO) [48] on 1 April 2016 (registration number CRD42016037077) [49].
The results of the study are reported in line with the PRISMA guidelines [50].

Data Sources and Search Strategy
A systematic literature search has been performed searching different scholarly databases, including nine different bibliographic thesauri (namely, PubMed/MEDLINE (NLM), Scopus, SciVerse ScienceDirect, Science Citation Index Expanded (SCIE) and Social Sciences Citation Index from ISI/Web of Science, ProQuest Research Library, ABI/INFORM, CBCA), via the UNO per TUTTI Primo Central (Ex Libris) platform databases.
All prevalence studies on alcohol use among truck drivers were included in the current study. The search was performed using the following search terms: "(truckers OR truck drivers OR lorry OR commercial vehicles OR large good vehicles OR large vehicles OR heavy vehicles OR long vehicles OR trucking industry OR haul transport) AND (alcohol OR ethanol)". The search strategy was adapted for the other databases. Additionally, we searched reference lists of the chosen studies and prior reviews. We extensively mined different databases and we used a broad keyword string in order to capture the highest number of potentially relevant studies, minimizing the chances of missing pertinent items.
When it was not possible to make a decision on a study's inclusion or exclusion based on the title and/or abstract, the full text of the study was examined (Table 1).

Study Screening and Selection
Once retrieved via the UNO per TUTTI Primo Central (Ex Libris) platform databases, duplicate studies were automatically removed. The list of non-redundant items was handled with the open source Review Manager 5 (RevMan 5) software.
The studies have been independently screened by two authors (NLB, GD) looking at study titles and/or abstracts for potential eligibility. Screening questions were developed and pilot-tested with a subset of records before implementation (Table 2). Disagreement was assessed using Cohen's κ statistics and has been resolved through discussion; a third reviewer (AM) has been involved if necessary. We have provided tables with characteristics of included studies and of excluded studies with reasons for their exclusion (Tables 3-5).
Studies meeting the following PICOS/PECOS criteria were considered for inclusion:

Alcohol Consumption Pattern
In this systematic review and meta-analysis, we followed the WHO terminology related to alcohol consumption: "health-wise" (non-hazardous), "hazardous", and "harmful" alcohol use. The following paragraphs provide the readers with an overview of these definitions. We focused on alcohol pattern (frequency of drinking and number of drinks per occasion/event) rather than the mean alcohol intake, as the latter is an incomplete risk predictor of alcohol-related harm.

"Harmful Alcohol Use"
The tenth edition of International Classification of Diseases (ICD-10) developed by the WHO defines "harmful alcohol use" as a pattern of substance use that causes damage to physical or mental health. This definition was closely similar to the concept of "alcohol abuse" developed by the American Psychiatric Association (APA) in the Diagnostic and Statistical Manual of Mental Disorders-Fourth edition (DSM-IV). Both of these concepts were introduced in order to gather clinically important problems associated with alcohol consumption that nonetheless could not be characterized as "alcohol dependence".

"Binge Drinking"
One pattern of harmful use is called "binge drinking", defined by the National Institute on Alcohol Abuse and Alcoholism (NIAAA) as a pattern of drinking four or more drinks for women, five or more drinks for men, in a two hour timeframe, which typically brings blood alcohol concentration (BAC) levels to 0.8 g/L.
This pattern is similar to another definition called "heavy episodic drinking" (HED), defined as drinking at least 60 grams or more of pure alcohol on at least one occasion in the past 30 days. Although the two terms are often used synonymously, we decided to use the former because it is more widely used by researchers.

"Everyday Drinking"
"Everyday drinking" among professional drivers can be considered as a pattern of consumption that increases the risk of accidents, which can therefore be harmful. Although this pattern has been referred to with several denominations, such as "daily drinking", "continuous drinking", and "steady drinking", we preferred "everyday drinking" because it has been correlated to problematic use and because it is the most unambiguous term.

"AUDIT/CAGE"
The Diagnostic and Statistical Manual of Mental Disorders -Fifth edition (DSM-V) introduced another concept, the "Alcohol Use Disorder" (AUD), which both overcomes the distinction between "abuse/harmful use" and "dependence" and includes them in a single category. The WHO has developed a tool to identify this pattern of use called Alcohol Use Disorder Identification Tool (AUDIT). A shorter three item version, called AUDIT-C identifies harmful patterns of consumption, such as frequency of drinking and "binge drinking" episodes, while the full 10-item AUDIT investigates symptoms of alcohol dependence as well. The CAGE (Cut down-Annoyed-Guilty-Eye) questionnaire is another tool previously developed to identify at risk alcohol users, an acronym for typical symptoms such as feeling the need to cut-down the amount of alcohol, being annoyed by criticism, feeling guilt, and using ethanol as an eye-opener in the morning.

"Truck Drivers"
In this report, "truck-drivers" are defined as any person whose activities involve driving lorry or commercial vehicles with the following characteristics: large, heavy or long vehicles.

Appraisal of Study Quality
Two reviewers are content experts (AM, GD) and one reviewer (NLB) is an experienced biostatistician/epidemiologist. The contents experts have only assessed potential publications with respect to the appropriateness of the research questions tested. The biostatistician has only evaluated the appropriateness of methods employed. Disagreement has been resolved by consensus.
The "Joanna Briggs Institute Critical Appraisal tools for use in JBI Systematic Reviews-Checklist for Prevalence Studies" has been used to assess the quality of studies included in the current systematic review and meta-analysis. This tool explores different domains of quality: namely, (i) the appropriateness of the sample frame to address the target population; (ii) the participants sampling technique; (iii) the adequateness of the sample size; (iv) the completeness of the description and details concerning the study subjects and the setting; (v) the coverage of the sample; (vi) the validity of the methods and (vii) their reliability; (viii) the appropriateness of the statistical analyses; and (ix) the adequateness of the response rate.
Concerning the third domain, the adequateness of the sample size was computed using the formula: where n is the sample size, Z is the Z statistic for a given level of confidence (1.96), P is the expected prevalence or proportion (in proportion of one; if, for instance, 20%, P is 0.2), and d is the precision (in proportion of one; if 5%, d = 0.05). Expecting a prevalence of alcohol consumption rate in the range 9-19%, an adequate sample size should comprise a minimum of 131-236 subjects.

Statistical Analysis
For the meta-analysis, data have been extracted from the studies using a standardized documentation form (Table 2). Prevalence ratios were calculated as effect size (ES) estimates. The 95% confidence intervals (CIs) were also generated. More in detail, the logit transformation (l) approach was utilized in the current meta-analysis, being one of the possible approaches for pooling together raw prevalence data. The following equation was used to compute l: Variance was computed using the equation: where N is the population size.
The pooled l was subsequently back-transformed to a proportion using the equation: Additional analyses were performed after stratification considering all the variables listed in Table 2. Meta-analyses were carried out using the commercial software MedCalc Statistical Software version 16.8.4 (MedCalc Software bvba, Ostend, Belgium; https://www.medcalc.org; 2016) and Comprehensive Meta-Analysis CMA v3.

Heterogeneity and Sensitivity Analysis
Statistical heterogeneity has been assessed using the I 2 statistic. I 2 more than 50% was regarded as substantial heterogeneity [51,52]. To identify sources of variation, further stratification was performed relative to study quality and to performance of confirmatory tests. In addition, for the sensitivity analyses, the stability of the pooled estimate with respect to each study was investigated by excluding individual studies from the analysis.

Publication Bias
Potential publication bias has been extensively investigated in the current systematic review and meta-analysis. First, we have visually inspected the funnel plot, looking at asymmetry of the graph. The funnel plot chosen in the current meta-analysis is the funnel plot of precision by logit event rate.
If asymmetry was present based on visual assessment, we performed exploratory analyses to investigate and adjust this using the Duval and Tweedie's trim-and-fill analysis [53]. In addition, the probability of publication bias has been tested using the Egger's linear regression test [54].
In conclusion, it should be emphasized that, in presence of statistically significant heterogeneity and with less than ten studies included, the findings of these tests should be interpreted with caution.
Of these studies, seven articles reported an unspecified alcohol consumption pattern, whereas two studies reported a generic consumption in the last year, meaning that the truck driver had consumed any alcoholic drink during the last year. While this could be useful to identify drinkers and non-drinkers, it is not-in the authors' opinion-useful for evaluating harmful consumption patterns, which is the main objective of the present study. Finally, one study reported alcohol consumption in terms of one per day, another study in terms of grams per kilogram, whilst one study counted the number of drinks per week.
Seven studies conducted biological monitoring of alcohol consumption (two using breath analysis, two utilizing urine samples, two saliva samples, and one blood).
Seventeen studies were retained in the qualitative synthesis of the literature and included in the quantitative synthesis (meta-analysis), as pictorially shown in Figure 1. Table 3 reports studies excluded with reason, whereas the main characteristics of studies included in the meta-analysis are described in Table 4. In Table 5, the critical appraisal of the methodological quality of the retained studies is reported.

Systematic Review and Meta-Analysis of Alcohol Consumption Rates among Truck Drivers
The total study population comprised 11,574 truck drivers, with sample sizes ranging from 91 to 2945 participants. Mean age ranged from 33.8 to 44.0 years old. Male percentage went from 89.0% to 100.0%. Ten studies were performed in Brazil, whereas two studies were carried out in the USA. The remaining studies were conducted in Morocco (one study), the Netherlands (one study), Italy (one study), Nigeria (one study) and Pakistan (one study). Mean work-load ranged from 11.1 to 12.7 h per day and was reported only in four studies. Night-shifters (reported in four studies) varied from 12.5% to 33.0% of participants. Schooling level was described in ten studies: primary education was achieved by 30.8-81.0% of participants, according to the study. Experience years went from 10.0 to 18.1 years (reported in eight studies). Only three studies reported the percentage of truck drivers working for companies (range 40.5-60.9%).

Binge Drinking among Truck Drivers
Five studies reported data concerning binge drinking among truck drivers. Based on the I 2 value (95.33), a random-effects model was performed. A rate of 19.0%, 95% CI (13.1, 26.9) was found (  (Table 6).

Systematic Review and Meta-Analysis of Alcohol Consumption Rates among Truck Drivers
The total study population comprised 11,574 truck drivers, with sample sizes ranging from 91 to 2945 participants. Mean age ranged from 33.8 to 44.0 years old. Male percentage went from 89.0% to 100.0%. Ten studies were performed in Brazil, whereas two studies were carried out in the USA. The remaining studies were conducted in Morocco (one study), the Netherlands (one study), Italy (one study), Nigeria (one study) and Pakistan (one study). Mean work-load ranged from 11.1 to 12.7 hours per day and was reported only in four studies. Night-shifters (reported in four studies) varied from 12.5% to 33.0% of participants. Schooling level was described in ten studies: primary education was achieved by 30.8-81.0% of participants, according to the study. Experience years went from 10.0 to 18.1 years (reported in eight studies). Only three studies reported the percentage of truck drivers working for companies (range 40.5-60.9%).

Binge Drinking among Truck Drivers
Five studies reported data concerning binge drinking among truck drivers. Based on the I 2 value (95.33), a random-effects model was performed. A rate of 19.0%, 95% CI (13.1, 26.9) was found (  (Table 6).

"Everyday Drinking" Pattern among Truck Drivers
Seven studies reported data concerning "everyday drinking" consumption rate among truck drivers (total population 4314 subjects, ranging from 91 to 2134 participants). Based on the I 2 value (84.40), random-effects model was utilized. A rate of 9.4%, 95% CI (7.0, 12.4) was found ( Figure 6). Sensitivity and cumulative analyses confirmed the stability of the findings. No statistically significant moderators were computed. Concerning the funnel plot (Figure 7), no evidence of publication bias could be detected, both visually inspecting the graph and performing the Duval and Tweedie's trim-and-fill analysis and the Egger's linear regression test (intercept = −1.98, standard error = 1.76, 95% CI (−6.51, 2.54), t-value = 1.13, p = 0.31111) ( Table 7).

"Everyday Drinking" Pattern among Truck Drivers
Seven studies reported data concerning "everyday drinking" consumption rate among truck drivers (total population 4314 subjects, ranging from 91 to 2134 participants). Based on the I 2 value (84.40), random-effects model was utilized. A rate of 9.4%, 95% CI (7.0, 12.4) was found ( Figure 6). Sensitivity and cumulative analyses confirmed the stability of the findings. No statistically significant moderators were computed. Concerning the funnel plot (Figure 7), no evidence of publication bias could be detected, both visually inspecting the graph and performing the Duval and Tweedie's trimand-fill analysis and the Egger's linear regression test (intercept = −1.98, standard error = 1.76, 95% CI (−6.51, 2.54), t-value = 1.13, p = 0.31111) ( Table 7).

Alcohol Consumption Rate among Truck Drivers Based on AUDIT-CAGE Instruments
Five studies reported data concerning alcohol consumption rate based on AUDIT-CAGE instruments. Based on the I 2 value (97.29), random-effects model was carried out. A consumption rate of 22.7%, 95% CI (14.8, 33.0) was found (Figure 8). Concerning the funnel plot (Figure 9), no evidence of publication bias could be found, both visually inspecting the graph and carrying out the Duval and Tweedie's trim-and-fill analysis and the Egger's linear regression test (intercept = −1.84, standard error = 5.72, 95% CI (−20.04, 16.36), t-value = 0.32, p = 0.76889) ( Table 8).

Discussion
The goal of this study was to assess the harmful alcohol consumption in the occupational category of truck drivers, a group of workers of the utmost importance for road safety and the global economy. In particular, we investigated three patterns of alcohol use considered hazardous or harmful: namely, "binge drinking", "everyday drinking", and positivity to AUDIT or CAGE tests.
The findings of the current meta-analysis showed a relevant harmful alcohol use prevalence among truck drivers. The results regarding "binge drinking" (prevalence of 19.0%) are extremely important because this mode of consumption has been linked to impairment in several executive cognitive functions [84,85] necessary for the complex task of driving. Frontal executive functioning of the brain is part of a system that controls the hierarchical order of brain processing, thus permitting control over cognition and behaviors [86,87]. Some authors have suggested that this executive impairment could impact on the predisposition towards the development of harmful habits including alcohol use disorders (AUD) [88,89] and use of illicit drugs. Moreover, this pattern has been associated with relevant road safety issues: "binge drinking" at least once a month increases the chance of being involved in crash accidents by ten-fold [90].
Interestingly, the results of the present study show a positive association between binge drinkers and nationality: the study showed that Brazilian drivers have a higher prevalence of "binge drinking" compared to their North American counterparts. This is comprehensible since alcohol consumption is, and has been, part of human culture, with different historical and social significance, as well as characterized by diverse patterns of consumption and different legislation in every country. Nevertheless, this observation needs further data: most studies that evaluated "binge drinking" were performed in Brazil, while only one was performed in the USA.
Another association was found between "binge drinking" and marriage. Since most truck drivers spend significant amounts of time away from home, days and even weeks at a time, this result can be interpreted-as some authors have suggested-that being distant from families removes a valuable support system that acts as a barrier against stress [28].
Regarding the "everyday drinking" pattern, our results showed a prevalence of 9.4% in the studied population. This issue is particularly relevant for the specific occupational category of truck drivers, a group of workers that most likely drives every working day. Among drivers drinking everyday could be indicative of another alcohol use pattern: AUDs [57]. Indeed, "everyday drinking" puts the user on the threshold of at risk consumption even considering the AUDIT test.
Another important finding of the present study is the prevalence of positives to AUDIT-CAGE tests (22.7%). Although this significant prevalence, and despite the fact that the AUDIT instrument is a reliable and easy to use test, the real epidemiological figures could be even higher, in light of under-reporting during the Occupational Health Surveillance, as suggested by some authors [61].
Moreover, an association between testing positive and being overweight or obese has been reported in the literature [52]. It is well known that excess bodyweight is one of the major risk factors for OSA. Therefore it is not surprising that OSA is more prevalent in truck drivers than the general population [91]. Statistically significant rise in sleep apnea severity and cardiac frequency are induced by 0.5 g alcohol/kg body weight, a level regarded as the safe upper limit by health authorities, on sleep apnea in otherwise healthy habitual snorers with mild-to-moderate OSA [92]. Similar results were obtained in a truck driver sample [72,73].
Even if heterogeneity exists between countries worldwide, the results discussed above show various prevalence rates among truck drivers, approximately in line with the prevalence in the general population, as reported in Table 9. Many countries and international organizations have implemented policies aimed at reducing alcohol use among the general population. One of the guiding principles of the global strategy to reduce alcohol consumption adopted by the WHO is the "protection of populations at high risk of alcohol-attributable harm and those exposed to the effects of harmful drinking by others should be an integral part of policies addressing the harmful use of alcohol" [96]. The "European action plan to reduce the harmful use of alcohol 2012-2020" suggests a development of community and workplace resources for alcohol programs, and an enhanced enforcement of road alcohol tests as well as a reduction of the blood-alcohol content [97].
In Italy, the Prevention National Plan 2014-2018 [98] regards the prevention of substance use, the prevention of traffic accidents, and the prevention of occupational injuries and disorders as three of its ten macro-goals. It is thus clear that the issue is of importance both from a public health as well as from an occupational health perspective. However, there are several factors among truck drivers that must be accounted for: truck drivers are one of the occupational categories identified by the State-Regions Conference in 2006 as at high risk of injuries and harm to others, to which the sale and use of alcohol is prohibited. Furthermore, Italian driving law requires a total absence of blood alcohol concentration (0 g/L) for professional drivers (the so-called "zero effective tolerance" policy), while the threshold for the general population is 0.5 g/L. Not all countries have adopted such policy and, as such, the alcohol level at which a person is considered legally impaired differs among countries.
A combined effect of a low quantity of alcohol with moderate sleep restriction results in significant decrements to subjective alertness and driving performance [99]. The detrimental effect of alcohol is also observed in relation to circadian sleep propensity. Even low consumption of alcohol could be too high when driving under a condition of increased sleep pressure such as during the night hours, in association to the effect of sleep deprivation. The highest values of hourly circadian sleep propensity are during the night, with a secondary maximum in the afternoon. Due to the significant interaction, even low BAC levels strongly increased road accident risk when associated with high sleep propensity [100].
Managing the issue of alcohol use among truck drivers is therefore bidirectional: public health policies can help reduce the consumption rate also among truck drivers, while occupational health prevention and health promotion will contribute to the reduction in the general population consumption. The first necessary step is to further study the epidemiology, especially in Europe, in order to better grasp the current state of the problem, with the aim of filling the current gap of knowledge supporting policy makers in implementing effective measures to contrast it.

Strengths and Weaknesses
To the best of the authors' knowledge, this is the first systematic review and meta-analysis studying alcohol consumption in truck drivers. The strengths of this study include comprehensive coverage of the literature, careful appraisal of study quality, risk of bias, consideration of possible subgroup effects, and a focus on relevant endpoints to this specific occupational category. However, this study also presented a few shortcomings.
One of the major limitations of this meta-analysis was the paucity of studies, especially from Europe. This finding is even more relevant if we consider that this region has the highest per capita alcohol consumption and alcohol-attributable disease burden reported in the world. Moreover, there is a high level of "binge drinking" among Europeans, as it is apparent from Table 9. There is a dearth of studies on truck drivers in European countries, for which no/little current epidemiological data are available, which means de facto ignoring the potential risk represented by harmful drinking pattern in this occupational category.
Such a geographical disparity might have influenced the results of the meta-analysis because of different socio-economic contexts and cultural backgrounds. Another important limitation in this study was the qualitative differences concerning the way of collecting data on alcohol use and the demographics of the population between studies.
Several studies had to be excluded from the meta-analysis because data on alcohol consumption in frequency/dose were missing, or because data were related to any consumption pattern in the past 12 months. Another limitation regarded the quantity and quality of the studies assessing the consumption during working hours using biological sampling: although seven studies evaluated alcohol content in bodily fluids, each one used different detection methods and cut-offs. In our assessment, only blood samples or breath correlates are indicative of recent alcohol use. Positive results in urine and saliva tests might be detected even after 48-72 h after alcohol consumption. Finally, as already mentioned in the material and methods section, caution should be taken in interpreting the findings of the statistical tests used for assessing the publication bias, given the high heterogeneity and the small number of included studies.
An improvement over these limitations is needed in order to better understand and evaluate the issue of alcohol use among truck drivers with a higher degree of accuracy. More studies using standardized questionnaires-thoroughly investigating the demographics, psycho-social determinants of the population, and well-defined patterns of alcohol use-need to be performed. Also, studies using comparable biological sampling methodologies are necessary.

Conclusions
This systematic review with meta-analysis provides the first rigorous analytical synthesis of updated epidemiological data regarding truck drivers. Our findings show that the prevalence of alcohol use among this occupational category can be considered harmful and put the light on some existing gaps, including the dearth of studies and data for many countries, but at the same time provide useful insights. This can be useful for decision and policy-makers in order to develop, design, and implement adequate surveillance and preventive policies.
However, in order to better assess the magnitude of the risk, more specified and defined modes of consumption must be investigated.