Obstructive Sleep Apnea and Risk of Motor Vehicle Crash: Systematic Review and Meta-Analysis
ABSTRACT
Study Objectives:
We performed a systematic review of the OSA-related risk of crash in commercial motor vehicle (CMV) drivers. The primary objective involved determining whether individuals with obstructive sleep apnea (OSA) are at an increased risk for a motor vehicle crash when compared to comparable individuals who do not have the disorder. A secondary objective involved determining what factors are associated with an increased motor vehicle crash risk among individuals with OSA.
Design/Setting:
Seven electronic databases (MEDLINE, PubMed (PreMEDLINE), EMBASE, PsycINFO, CINAHL, TRIS, and the Cochrane library) were searched (through May 27, 2009), as well as the reference lists of all obtained articles. We included controlled studies (case-control or cohort) that evaluated crash risk in individuals with OSA. We evaluated the quality of each study and the interplay between the quality, quantity, robustness, and consistency of the body of evidence, and tested for publication bias. Data were extracted by 2 independent analysts. When appropriate, data from different studies were combined in a fixed- or random-effects meta-analysis.
Results:
Individuals with OSA are clearly at increased risk for crash. The mean crash-rate ratio associated with OSA is likely to fall within the range of 1.21 to 4.89. Characteristics that may predict crash in drivers with OSA include BMI, apnea plus hypopnea index, oxygen saturation, and possibly daytime sleepiness.
Conclusions:
Untreated sleep apnea is a significant contributor to motor vehicle crashes.
Citation:
Tregear S; Reston J; Schoelles K; Phillips B. Obstructive sleep apnea and risk of motor vehicle crash: systematic review and meta-analysis. J Clin Sleep Med 2009;5(6):573-581.
INTRODUCTION
Of all occupations in the United States, workers in the trucking industry experience the third highest fatality rate, accounting for 12% of all worker deaths. In 2006, there were 368,000 police-reported large truck crashes, resulting in 4,321 fatalities and 77,000 injuries.1 The Federal Motor Carrier Safety Administration (FMCSA) was established as a separate administration within the U.S. Department of Transportation (DOT) pursuant to the Motor Carrier Safety Improvement Act of 1999. The primary mission of the FMCSA is to reduce crashes, injuries and fatalities involving large trucks and buses. Among the strategies employed by the FMCSA to accomplish this goal are the development and maintenance of medical fitness standards for drivers of commercial vehicles; these standards are applied by medical examiners to commercial drivers, who are required by Federal statute to undergo medical qualification examinations at least every 2 years.
Obstructive sleep apnea (OSA) is a prevalent and potentially dangerous condition among commercial motor vehicle (CMV) drivers. While OSA is conservatively estimated to affect approximately 5% of the general population,2 the condition appears to be much more prevalent in commercial drivers. Howard et al. estimated that 50% of more than 3000 commercial drivers were at risk for sleep apnea.3 Pack et al. found that 28.2% of 406 commercial drivers had at least mild sleep apnea and 4.7% had severe sleep apnea by conventional criteria.4 The majority of research indicates that OSA is a significant cause of motor vehicle crashes.3,5–9 Thus, assessment of the risk of OSA and development of effective methods to identify and treat commercial drivers with OSA is an important part of the mission of the FMCSA. Since the most recent standards for medical examiners regarding OSA are from a Federal Highway Administration (FHWA) sponsored conference in 1991,10 these standards required an evidence-based update.
The current study was designed to provide evidence for updating the standards by conducting a systematic review of the relevant literature concerning OSA and CMV drivers. The literature consists predominantly of cohort and case-control studies. Given that few studies specifically enroll CMV drivers, studies that included non-CMV drivers were also evaluated.
The primary objective of this study was to determine whether individuals with OSA are at an increased risk for a motor vehicle crash when compared to individuals without OSA. If so, a secondary objective was to identify disease-related factors associated with an increased motor vehicle crash risk.
METHODS
Identification of Evidence Bases
Separate evidence bases for each of the objectives of this evidence report were identified using a process consisting of a comprehensive search of the literature; examination of abstracts of identified studies to determine which articles would be retrieved; and the selection of the actual articles that would be included in each evidence base.
A total of 7 electronic databases (MEDLINE, PubMed (PreMEDLINE), EMBASE, PsycINFO, CINAHL, TRIS, and the Cochrane library) were searched (through May 27, 2009). All database searches were conducted by masters-level medical librarians. To supplement the electronic searches, we also examined the bibliographies/reference lists of included studies, recent narrative reviews, and scanned the content of new issues of selected journals and selected relevant gray literature sources. A complete list of the electronic databases searched and the search strategy used to identify relevant studies are available upon request. Admission of an article into an evidence base was determined by the inclusion criteria listed in Table 1.
| Inclusion Criteria (General) |
| Article must have been published in the English language. |
| Article must be a full-length article (no abstracts or letters to the editor). |
| Article must have enrolled 10 or more subjects. |
| Article must have enrolled subjects aged ≥18 y. |
| If the same study is reported in multiple publications, the most complete publication will be the primary reference. Data will be extracted to avoid double-counting individuals. |
| Studies were limited to individuals with OSA only (no central apneas). |
| Studies that evaluated both OSA and other sleep disordered individuals were included as long as data for OSA subjects could be analyzed separately from that of other subject populations. |
| Inclusion Criteria (Primary Objective) |
| Article must describe a study that attempted to directly determine the risk for a motor vehicle crash (risk for a fatal or nonfatal crash) associated with OSA using a direct measure of crash (no indirect measures; e.g., driving simulator data). |
| Article must describe a study that includes a comparison group comprised of comparable subjects who do not have OSA. |
| Article must present motor vehicle crash-risk data in a manner that allows ECRI Institute to calculate (directly or through imputation) effect-size estimates and confidence intervals. |
| Inclusion Criteria (Secondary Objective) |
| Article must describe a study that attempted to determine the disease-related factors associated with an increased risk for a motor vehicle crash (risk for a fatal or nonfatal crash) among individuals with OSA. |
| Article must describe a study that includes a comparison group comprised of comparable subjects with OSA who did not have a motor vehicle crash. |
| Article must present motor vehicle crash-risk data in a manner that will allow ECRI Institute to calculate (directly or through imputation) effect-size estimates and confidence intervals. |
Analytic Methods
Data were extracted by 2 independent analysts. Individual study quality was assessed using quality scales modified from the Newcastle-Ottawa quality instruments for cohort and case-control studies.11 For uncontrolled case series we used a separate instrument developed at ECRI Institute. These quality instruments are available upon request.
Random- and fixed-effects meta-analyses were used to pool data from different studies.12–16 Differences in the findings of studies (heterogeneity) were identified using the Q-statistic and I2.17–19 Heterogeneity was explored with meta-regression of potential explanatory variables (study quality, clinic versus general population study) in Stata 10. Sensitivity analyses aimed at testing the robustness of findings included the use of cumulative fixed- and random-effects meta-analysis.20–22 The presence of publication bias was tested for using the “trim and fill”method.23–25
RESULTS
OSA and Crash Risk
Eighteen articles describing 18 unique studies met the inclusion criteria for the primary objective.3,5,7,8,9,26–38 Study characteristics are presented in Supplemental Table 1 (available online at www.aasmnet.org/JCSM). Only 2 of these studies enrolled distinct populations of CMV drivers.3,26 The remainder of the studies included private motor vehicle license holders, an unknown number of whom may have held commercial driver licenses. The studies used either a cohort design or a case-control design. The most commonly used methodology (16 studies) was to select a cohort of drivers with OSA and compare the incidence of crash over a defined time period with the incidence of crash occurring over a similar time period among comparable individuals without the condition. The less commonly used approach (2 studies) was to compare the prevalence of OSA among individuals who experienced a crash (cases) and those who did not (controls).
A design problem common in many risk-assessment studies is the failure to control adequately for exposure. In this instance, the exposure variables of critical importance are the number of miles driven per unit time and the time frame over which data were collected. If cohorts or cases and controls are not well-matched for exposure to risk, then any observed differences in the risk may simply be the consequence of differences in exposure. A majority of the included studies attempted to control for both of these exposure variables.
Crash rates were determined from data obtained from 2 primary sources: databases and questionnaires. The degree of confidence that one can have in crash rates derived from questionnaires is unclear, primarily because questionnaires depend on reliable reporting by the individual being questioned.
Studies of CMV Drivers
Two studies included only CMV drivers. Howard et al. compared crash risk among drivers with OSA (symptom diagnosis) and drivers not diagnosed with OSA (controls).3 They measured the prevalence of excessive sleepiness and sleep-disordered breathing, and assessed crash-risk factors in 2,342 respondents to a questionnaire distributed to a random sample of 3,268 Australian commercial vehicle drivers and another 161 drivers among 244 invited to undergo formal sleep studies. Howard et al. presented the odds ratio (OR) for having a crash in the past 3 years in drivers with OSA adjusted for age, hours of driving, and alcohol intake. Drivers diagnosed with OSA based on a Multivariable Apnea Prediction Score (MAPS)39 ≥ 0.5 and Epworth Sleepiness Scale (ESS)40 score ≥ 11 were found to be at an increased risk for motor vehicle crash (OR = 1.3, 95% 1.00-1.69). The value of the findings of this study is weakened by the fact that individuals were diagnosed with sleep apnea using questionnaires only. The accuracy of this diagnosis was not confirmed via sleep lab investigations. Because diagnosis based on these questionnaires is subjective, it is unclear whether all individuals had received a correct diagnosis. This, combined with the retrospective outcome measurement and the reliance on self-reported crash data, resulted in this study being judged as low quality.
Stoohs et al. assessed a possible independent effect of OSA on traffic crashes in long-haul commercial truck drivers.26 The study design included integrated analysis of recordings of sleep-related breathing disorders, and self-reported and company-recorded automotive crashes. A cross-sectional population of 90 commercial long-haul truck drivers 20 to 64 years of age was studied. Main outcome measures included presence or absence, as well as severity, of sleep-disordered breathing and frequency of automotive crashes. A “crash” was defined as the collision of the index case's vehicle with a stationary or moving object or as driving off the road in the absence of an obstacle.
The study was performed at the main hub of a long-haul trucking company.26 Two hundred thirteen drivers were scheduled to spend the night at the facilities. Of these, 193 (92%) agreed to undergo monitoring during sleep: 34 had to terminate the monitoring prematurely and their data had to be discarded. Subjects who agreed to be monitored were tested overnight with an ambulatory screening device, the Mesam IV, a microprocessor that continuously monitors 4 variables throughout the night: heart rate, snoring sounds, oxygen saturation (SpO2), and body position/movement. These investigators performed 159 recordings of appropriate duration for analysis. Since some drivers were students with little professional driving experience, the authors decided only to include drivers with a driving history ≥ 2 months. Overnight recordings, completed questionnaires, and crash records were analyzed for 90 truck drivers. Analyses of overnight recordings were used to identify obstructive hypopnea and apnea. The sleep logs were used to calculate total sleep time (TST) and oxygen desaturation index (ODI).
For analysis, Stoohs et al. considered the total number of vehicle crashes.26 They obtained information on mileage both from the trucking company and from the drivers' self-reported usage of private vehicles. All crash rates were adjusted for annual mileage of individual truck drivers. These investigators found that truck drivers identified with sleep disordered breathing (SDB) had a 2-fold higher crash rate per mile than drivers without SDB. Crash frequency was not dependent on the severity of the sleep related breathing disorder. Obese drivers with a body mass index (BMI) ≥ 30 kg/m2 also presented a 2-fold higher crash rate than non-obese drivers. In addition, the authors found that a complaint of EDS was related to a significantly higher automotive crash rate in long-haul commercial truck drivers. SDB with hypoxemia and obesity are risk factors for automotive crashes. This study was assessed as moderate quality because the majority of crash data came from company records rather than self-report.
Studies of All Drivers
Of the 18 included studies, 16 studies reported on the incidence of crashes occurring among populations of individuals with OSA and the incidence of crashes occurring among individuals without the disorder.2,3,5,7,9,26–28,30–34,36–38 These studies were mostly rated as low quality due to retrospective design, lack of adjustment for important potential confounders, and self-reported outcome or lack of independent outcome assessment. Ten of these studies provided enough data to determine the crash relative risk (RR) and 95% confidence intervals (CIs) for individuals who have OSA versus comparable individuals without the disorder.5,7,26–28,31–33,36,38 A test of homogeneity found that the findings of the 10 studies were heterogeneous (Q = 83.9, P < 0.001; I2 = 89%). Exploration of this heterogeneity using meta-regression techniques found that neither study quality nor derivation of study groups (clinic versus general population) explained the heterogeneity. Therefore, we pooled these data using a random-effects meta-analysis (Figure 1). The findings of this meta-analysis provide support for the contention that drivers with OSA are at a significantly increased risk for experiencing a motor vehicle crash when compared to comparable individuals without OSA (crash RR = 2.43, 95% CI: 1.21-4.89: p = 0.013). In other words, if one assumes that the underlying crash risk for a driver is 0.08 crashes per person-year, the crash risk for a driver with OSA can be estimated to be 0.19 (95% CI: 0.10 to 0.39) crashes per person-year. A series of sensitivity analyses (removal of one study at a time, cumulative meta-analysis by publication date) demonstrated our finding that individuals with OSA are at an increased risk for a motor vehicle crash to be robust. The “trim and fill” test did not detect publication bias.
Crash Risk among Individuals with OSA Compared to Controls (Random-effects Meta analysis)
Two of the 18 studies presented data on the odds of an individual who experienced a crash having OSA relative to the odds of a comparable individual who did not crash having OSA.8,29 One study was rated as low quality for reasons described above; the remaining study was rated as moderate quality because crash data was obtained from secure records. The findings of these studies are summarized in Figure 2. One of the 2 studies8 suggested that OSA increased crash risk, and the other29 found no evidence of an increase or a decrease in crash risk.
OSA and Crash Risk (OR)
Disease-Related Factors and Crash Risk
Our assessment of the evidence pertaining to crash risk found that drivers with OSA (both commercial and noncommercial) are at a significantly increased risk for a motor vehicle crash when compared with comparable drivers who do not have the disorder. Not all individuals with OSA, however, appear to be at increased risk. A secondary objective was to determine whether there are specific risk factors that are predictive of which individuals with OSA are at the greatest risk for a crash. The identification of such risk factors is important, because it will enable medical examiners to differentiate high-risk individuals from low-risk individuals when making decisions about fitness-to-drive certification.
Thirteen articles describing 13 unique studies met the inclusion criteria for this objective, and are summarized in Supplemental Table 2 (available online at www.aasmnet.org/JCSM).5,9,26,27,33,37,41–46 Three of these studies were graded as being of moderate quality, while the remaining 10 studies were graded as being of low quality due to retrospective design, lack of adjustment for important potential confounders, and self-reported outcome or lack of independent outcome assessment. One of the studies assessed the factors predictive of crash among CMV drivers with OSA. All of these studies examined several factors caused by OSA that are thought to be associated with an increase in an individual's risk for a motor vehicle crash (Supplemental Tables 2 and 3, available online at www.aasmnet.org/JCSM). These factors—all of which serve as surrogate indicators of disease severity—included the presence and degree of daytime sleepiness,7,9,27,37,42–46 the severity of disordered respiration during sleep,5,7,26,27,33,37,42–46 and nighttime oxygen saturation (SpO2).37,43–46 In addition to these three factors, some included studies also examined the relationship between BMI and the risk of a motor vehicle crash.7,26,27,43 Since a high BMI is a risk factor for OSA, it may also be considered to be a surrogate marker for OSA severity because it is strongly correlated with the severity of the disorder.47–50 In addition, 3 studies examined the relationship between cognitive and psychomotor functioning and the risk of a motor vehicle crash.41,42,44
Sleepiness
Eight included studies reported on the relationship between sleepiness and crash risk among populations of individuals with OSA,7,27,37,42–45 and 3 of these 8 studies (rated as low quality for reasons mentioned above) provided data sufficient to calculate effect-size estimates (and 95% confidence intervals) which could be pooled using meta-analysis.7,43,45 All 3 measured daytime sleepiness subjectively using the ESS.40 A test of homogeneity found that the findings from these 3 studies for which an effect-size estimate could be calculated were heterogeneous (Q = 6.46, p = 0.040; I2 = 69.05%). Consequently, we pooled the data from the three studies using a random-effects meta-analysis. The result suggested a trend toward an increased crash risk in individuals with higher ESS scores, but the finding was not quite statistically significant (SMD = 0.64, 95% CI: −0.03 to 1.30; P = 0.061). Also, these 3 studies did not adjust for the effect of other factors (such as age, gender, and alcohol use). Likewise, 4 additional studies could not confirm an increased crash risk based on higher ESS scores,9,27,42,44 (although one of these studies found a significantly higher risk of near-miss accidents among drivers with higher ESS scores). However, daytime sleepiness from any cause has been associated with increased crash risk.51–53
Two studies used the multiple sleep latency test (MSLT)54 to objectively measure sleepiness. In one report Aldrich found that there were no significant differences in mean sleep latency between individuals with crashes and those without (males, 8.2 min vs. 7.8 min; females, 7.3 min vs. 7.6 min).37 Young et al. also found no significant differences on the MSLT, although there was a trend toward lower scores (indicating greater sleepiness) among men with more than one crash (4.5 ± 2.7) compared to men who did not crash (8.8 ± 0.3).9
Disease Severity
Eleven included studies reported on the relationship between disease severity and crash risk among populations of individuals with OSA.5,7,26,27,33,37,42–46 Three of these 11 studies (judged to be of low quality for reasons mentioned above) provided data sufficient to calculate an effect-size estimate.7,43,46 All 3 employed the apnea plus hypopnea index (AHI) to quantify the severity of disordered respiration during sleep. A test of homogeneity found that the findings of the 3 studies for which an effect-size estimate could be calculated did not differ substantially (Q = 1.6, p = 0.45; I2 = 0.0%). Consequently, we pooled the data from the 3 studies using a fixed-effects meta-analysis. The results suggested a trend toward greater severity of disordered respiration during sleep (measured using the AHI) among individuals with OSA who crashed, but again the findings did not quite reach statistical significance (SMD = 0.27, 95% CI: −0.006 to 0.54; p = 0.055). The findings of the 8 studies not included in the meta-analysis were mixed. Three studies found that severity of disordered breathing during sleep was associated with an increased risk for a motor vehicle crash.5,33,37 The remaining 5 studies found that severity of disordered breathing during sleep was not associated with an increased risk for a motor vehicle crash.26,27,42,43,45 Though the findings suggest the possibility that the severity of disordered breathing may be related to crash risk, a definitive conclusion cannot be drawn at this time.
Oxygen Saturation
Five included studies reported on the relationship between a measure of SpO2 and crash risk among populations of individuals with OSA.37,43–46 Data from these 5 studies (rated as low quality for reasons mentioned earlier) were reported using several different methods, and as a result we were precluded from pooling their findings in a meta-analysis. Two studies found that total oxygen desaturation time (time during which oxygen saturation was decreased < 90%)45 or nocturnal hypoxemia46 correlated with crash score45 or sleep-related near-miss crashes.46 One study determined that there was no statistical difference between individuals who experienced a crash and individuals who did not experience a crash with regards to mean SpO2 and lowest SpO2.43 However, the data did indicate that individuals who experienced a crash were more likely to have lower SpO2 levels. A separate study found that males who experienced a crash had a significantly lower minimum SpO2 compared to males who did not experience a crash.37 Finally, one study found that neither mean SpO2 nor time below 90% SpO2 was related to number of crashes.44 Taking all of this information into account, it appears that hypoxemia may be a risk factor for a motor vehicle crash in individuals with OSA, but the level at which this occurs, and how best to measure it are unclear from these studies.
Body Mass Index (BMI)
Four included studies reported on the relationship between BMI and crash risk among populations of individuals with OSA.7,26,27,43 Differences in measures of effect precluded us from pooling their findings in a meta-analysis. Stoohs et al. examined the relationship between BMI and automobile crashes in 90 commercial long-haul truck drivers (moderate quality study).26 Individuals were classified into 4 categories: BMI < 25 kg/m2, BMI ≥ 25 < 28 kg/m2, BMI ≥ 28 < 30 kg/m2, and BMI ≥ 32 kg/m2. Drivers whose BMI exceeded ≥ 30 kg/m2 were classified as obese. The authors found that automobile crash rate increased with increasing BMI: 0.031 crashes/10,000 miles (BMI < 25 kg/m2); 0.041 crashes/10,000 miles (BMI ≥ 25 < 28 kg/m2); 0.079 crashes/10,000 miles (BMI ≥ 28 < 30 kg/m2); and 0.101 crashes/10,000 miles (BMI ≥ 32 kg/m2) (p < 0.05). In addition, Stoohs et al. reported that non-obese drivers had a mean of 0.045 crashes/10,000 miles within the last 5 years compared to a mean of 0.1 crashes/10,000 miles (p < 0.03) within the last 5 years in obese truck drivers. Using the scores for obesity ( ≥ 30 kg/m2) as a predictor for crashes, they found that this predictor had a sensitivity of 49% and a specificity of 71%. Horstmann et al. examined the relationship between BMI and automobile crashes in 130 individuals who were diagnosed as having sleep apnea syndrome (SAS).7 Individuals were categorized on the basis of whether or not they had experienced a crash during the previous 3 years. Mean BMI was then compared between the 2 groups. The authors found that individuals who experienced a crash during the previous 3 years had a mean BMI of 35.1 kg/m2, whereas individuals who did not experience a crash had a mean BMI of 30.9 kg/m2 (p = 0.02). Yamamoto et al. examined the relationship between BMI and automobile crashes in 39 individuals who were diagnosed with OSA.43 Individuals were categorized on the basis of whether or not they had experienced a crash during the previous 2 years. Mean BMI was then compared between the 2 groups. The authors found that individuals who experienced a crash during the previous two years had a mean BMI of 32.4 kg/m2, whereas individuals who did not experience a crash had a mean BMI of 28.0 kg/m2 (p < 0.05). Mulgrew et al. found a small but statistically significant association between higher BMI and crash risk in a multivariable model (rate ratio 1.01, 95% CI: 1.00 to 1.02, p = 0.008).27 In summary, all 4 studies reporting on BMI and crash risk found that higher BMI is a risk factor for a motor vehicle crash in individuals with OSA.
Cognition and Psychomotor Function
Three included studies reported on the relationship between cognitive/psychomotor function and crash risk among populations of individuals with OSA.41,42,44 The low quality of these studies and differences in measures of effect precluded us from pooling their findings in a meta-analysis. Turkington et al.. performed a multivariable analysis and found a relationship between the number of simulated off-road events and crashes in the previous year (OR 1.004, 95% CI 1.0004–1.008, p < 0.03), but the effect size is very small.42 In contrast, a multivariable analysis by Barbe et al. found that mean reaction time, reaction fatigue, and percentage of simulated crashes on a driving simulator were not related to the number of crashes in drivers with SAS.44 Although Pizza et al. reported that a history of car crashes was associated with poor simulated driving performance, they did not adjust for other factors in their analysis.41
In summary, 4 factors may be associated with crash risk among the general driver population with OSA. These factors include BMI, severity of disordered respiration during sleep (as measured by the AHI), and hypoxemia. The presence and degree of daytime sleepiness also may be associated with crash risk, but the available instruments for measurement of daytime sleepiness (ESS and MSLT) appear to insufficiently distinguish crash risk within the group of drivers with OSA.
DISCUSSION
The primary findings of this study of OSA and crash risk are that individuals with OSA are clearly at increased risk for crash. The findings are somewhat stronger for private than for commercial drivers because so few studies specifically enrolled commercial drivers. The mean crash-rate ratio associated with OSA is likely to fall within the range of 1.21 to 4.89 (95% CI of random-effects summary effect-size estimate). Thus, if the underlying crash risk for a driver is 0.08 crashes per person-year, the crash risk for a driver with OSA can be expected to be in the range of 0.10 to 0.39 crashes per person-year. Our analysis indicates that the characteristics which may predict crash in drivers with OSA include AHI, hypoxemia, BMI, and possibly daytime sleepiness.
This report confirms and expands the systematic review of Ellen.55 In addition to confirming that drivers with OSA have roughly twice the risk of crash as comparable drivers who do not, we were able to identify some factors that may increase crash risk. In particular, BMI alone (in the absence of documented OSA) is associated with crash. While confirming the finding of Ellen et al. that AHI may increase risk for crash, we also found that severity of hypoxemia is associated with crash risk. On the other hand, evidence to support use of the ESS, MSLT, or maintenance of wakefulness test (MWT) to measure daytime sleepiness is very weak.
The lack of substantial association between ESS scores and crash risk in studies of drivers with diagnosed OSA may be due in part to low statistical power and that there may be less variability in ESS scores among diagnosed OSA patients than in a general driver population. Also, the ESS depends upon driver honesty in answering questions and upon driver awareness of sleepiness. Drivers may misrepresent sleepiness levels to prevent license revocation and those aware of sleepiness may develop compensatory strategies. Furthermore, drivers unaware of sleepiness may be at greater risk of crash.27
There are significant implications of these findings for both commercial and noncommercial drivers. OSA is prevalent and becoming more so, as the population ages and becomes more obese. OSA clearly poses a significant driving risk; however, most individuals with OSA do not crash. Identifying those drivers with OSA who are at greatest risk for crash is a critical challenge, and is an important part of the FMCSA's mission. The results of this analysis suggest that it may be possible to predict crash risk by use of simple measures such as BMI or oximetry. Subjective sleepiness is difficult to measure and is probably not applicable to the evaluation of commercial vehicle drivers for at least 2 reasons. First, many individuals with sleep apnea may fail to recognize or acknowledge that they are sleepy.56–58 Second, reliance on self-reported or subjective sleepiness is unlikely to be reliable in individuals whose livelihood or mobility is at stake, since they may have a vested interest in denying such symptoms.59 AHI, as determined by in-laboratory or portable testing, also weakly predicts crash risk, as may SpO2, which can be determined by oximetry alone. Finally, measured BMI is associated with crash, independent of any measured effect of BMI on OSA. This may be because obesity is a risk factor for OSA, but could also be because obesity is an independent cause of sleepiness.60–62
Although the goal of this review was to assess the risks of OSA in CMV drivers, the majority of data available pertains to private drivers. Thus, the generalizability of the findings to CMV drivers is unclear. Exposure to risk is far lower among noncommercial vehicle drivers than commercial drivers, because their driving exposure is lower. In addition, studies of general driver populations tend to include a greater proportion of women than represented in the CMV population. However, in the studies of OSA and driving risk, the number of males included in the studies of private motor vehicle license holders ranged from 79% to 98%, suggesting that gender may not be an issue when considering generalizability to CMV drivers. The ages of the private motor vehicle license holders included in these studies are similar to those of CMV drivers. Because data from studies of CMV drivers with OSA is scarce, relevant data from studies that investigated crash risk associated with OSA among the general driver populations is currently the primary source of evidence about OSA and crash risk. While the generalizability of the findings of these studies to CMV drivers may not be clear, such findings do at the very least allow one the opportunity to draw evidence-based conclusions about the relationship between OSA and motor vehicle crash risk in general.
The findings of this systematic review highlight the limitations of published studies and how little is known about the factors contributing to crash risk. For example, 75% of the studies evaluated in this systematic review enrolled patients derived from clinical, rather than general populations, which carries inherent bias. Individuals who have visited clinics are more likely to have experienced symptoms that affect driving performance, and may not be representative of all individuals with OSA in the general population. In addition, other established road crash risk factors such as age, alcohol, sleep duration, circadian/shiftwork, drugs, and other medical conditions are rarely assessed in studies of crash risk. Further, our analysis indicates that tests of simulated driving do not reliably predict crash. Research is urgently needed to identify tools to identify those drivers or characteristics of drivers that increase crash risk.
The current FMCSA medical standard applicable to OSA is 49 CFR 391.41 (b)(5) which indicates that the driver; “Has no established medical history or clinical diagnosis of a respiratory dysfunction likely to interfere with his ability to control and drive a motor vehicle safely” Two previous conference reports have addressed the issue of OSA and commercial drivers.10,63 The 1991 Respiratory/Pulmonary report for the Federal Highway Administration of the US Department of Transportation recommended screening drivers by asking if they snore and frequently fall asleep during the day.10 The report recommended that those with suspected or diagnosed but untreated OSA sleep apnea should not return to work for one month, and should not be medically qualified to drive until the diagnosis was eliminated or the condition was successfully treated. The report also stated that prior to returning to safety-sensitive work, the driver should have either a repeat sleep study, showing resolution of the apneas or a normal MSLT. Yearly sleep studies or MSLTs were recommended for follow up. The 1988 Neurologic Disorders report, also prepared for the Federal Highway Administration, recommended that CMV operators with sleep apnea and excessive daytime sleepiness not be permitted to operate in interstate commerce.63 The latter report only addressed surgical treatment, and a 3-month wait and laboratory studies (MSLT or polysomnogram) were recommended prior to resuming commercial driving. Based on this review of the literature, it is difficult to justify the expensive, cumbersome and time-consuming recommendations of the 1988 report. Evidence about the ability of the MWT or MSLT to predict crash is lacking. Data suggest that simulated driving performance may improve after 2 nights of CPAP treatment.64,65 Based on the evidence report and their expertise, a Medical Expert Panel made specific recommendations for changes in guidance related to CMV drivers with OSA.66 Among their recommendations were:
Annual recertification is required for an individual with untreated OSA and an AHI ≤ 20, who is not sleepy or who is effectively treated;
Disqualification for drivers who report sleepiness while driving, have experienced a crash with falling asleep, or have an AHI > 20 until CPAP is instituted, or have had surgery for OSA but not had repeat PSG, or have a BMI > 33 kg/m2 (until evaluated).
Conditional certification for one month after initiation of CPAP, with continued certification for 3 months if compliance ( > 4 hours of use for at least 70% of days) is documented during that month.
Required evaluation for the presence of OSA in individuals who have a high risk score on the Berlin Questionnaire, a BMI ≥ 33 kg/m2, or high risk based on clinical evaluation.
Use of overnight PSG as the preferred method of diagnosis, but acceptance of portable testing that includes at least 5 hours of oxygen saturation, nasal pressure, and sleep wake time.
DISCLOSURE STATEMENT
This study was funded by the Federal Motor Carrier Safety Administration of the Department of Transportation. The work was conducted by MANILA Consulting Group, McLean, VA and its subcontractor, the ECRI Institute, Plymouth Meeting, PA. None of the investigators has any affiliations or financial involvement that conflicts with the material presented in this manuscript. Dr. Philips has received honoraria from The American College of Chest Physicians, Boehringer Ingelheim, the Federal Motor Carrier Safety Administration, GlaxoSmithKline, and Ventus Corporation. The other authors have indicated no financial conflicts of interest.
REFERENCES
1
National Highway Traffic Safety Administration (NHTSA), Fatality Analysis Reporting System (FARS) and General Estimates System (GES) Accessed Feb 23, 2008 http//www.fmcsa.dot.gov/facts-research/facts-figures/analysis-statistics/cmvfacts.htm.2
Epidemiology of obstructive sleep apnea: a population health perspectiveAm J Respir Crit Care Med2002165121739, 11991871 3
Sleepiness, sleep-disordered breathing, and accident risk factors in commercial vehicle driversAm J Respir Crit Care Med2004170101421, 15317672 4
Impaired performance in commercial drivers: role of sleep apnea and short sleep durationAm J Respir Crit Care Med200617444654Epub 2006 May 11, 16690976 5
Falling asleep while driving and automobile accidents among patients with obstructive sleep apnea-hypopnea syndromePsychiatry Clinical Neurosci2002563334 6
Reduction in motor vehicle collisions following treatment of sleep apnoea with nasal CPAPThorax20015650812, 11413347 7
Sleepiness-related accidents in sleep apnea patientsSleep20002338389, 10811382 8
The association between sleep apnea and the risk of traffic accidents. Cooperative Group Burgos-SantanderNew Engl J Med199934084751, 10080847 9
Sleep-disordered breathing and motor vehicle accidents in a population-based sample of employed adultsSleep19972060813, 9351127 10
U.S. Department of Transportation, Federal Highway Administration, Office of Motor CarriersConference on Respiratory/Pulmonary Disorders and Commercial Drivers. Publication No. FHWA-MC-91-0041991 Accessed Feb 23, 2008 WashingtonUSDOT http//www.fmcsa.dot.gov/documents/pulmonary1.pdf.11
The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. [internet] accessed 2006 May 11 Ottawa (ON)Ottawa Health Research Institute (OHRI)[2 p]. Available: http://www.ohri.ca/programs/clinical_epidemiology/oxford.htm.12
Combining estimates of effect sizeThe handbook of research synthesis1994New York (NY)Russell Sage Foundation26177 13
Extracting summary statistics to perform meta-analyses of the published literature for survival endpointsStat Med199817281534, 9921604 14
Fixed effects modelsThe handbook of research synthesis1994New York (NY)Russell Sage Foundation28599 15
Random effects modelsThe handbook of research synthesis1994New York (NY)Russell Sage Foundation30121 16
Fixed- and random-effects models in meta-analysisPsychol Methods19983486504 17
An evaluation of homogeneity tests in meta-analyses in pain using simulations of individual patient dataPain20008541524, 10781914 18
Evaluation of old and new tests of heterogeneity in epidemiologic meta-analysisAm J Epidemiol199915020615, 10412966 19
Quantifying heterogeneity in a meta-analysisStat Med200221153958, 12111919 20
Clinical decision making using cumulative meta-analysis [editorial]Clin Cardiol1993161678, 8095197 21
Assessing and enhancing reliabilityDecubitus19925424 22
sbe22: Cumulative meta-analysisStata Technical Bulletin199842136 23
Empirical assessment of effect of publication bias on meta-analysesBMJ200032015747, 10845965 24
Practical estimates of the effect of publication bias in meta-analysisAustralasian Epidemiologist199851417 25
A non-parametric ‘trim and fill’ method of assessing publication bias in meta-analysisJ Am Stat Assoc2000958998 26
Traffic accidents in commercial long-haul truck drivers: the influence of sleep-disordered breathing and obesitySleep19941761923, 7846460 27
Risk and severity of motor vehicle crashes in patients with obstructive sleep apnoea/hypopnoeaThorax20086353641, 18234904 28
Effect of continuous positive airway pressure on the risk of road accidents in sleep apnea patientsRespiration200774449, 16804289 29
The role of sleep-disordered breathing, daytime sleepiness, and impaired performance in motor vehicle crashes-a case control studySleep Breath200486172, 15211390 30
Obstructive sleep apnoea: a case-controlled studyNeurol India2003514979, 14742931 31
Treatment with nasal CPAP decreases automobile accidents in patients with sleep apneaAm J Respir Crit Care Med20001618579, 10712333 32
Self-reported sleepiness while driving as a risk factor for traffic accidents in patients with obstructive sleep apnoea syndrome and in non-apnoeic snorersRespir Med2000949716, 11059950 33
Sleep apnea and automobile crashesSleep1999227905, 10505825 34
Risk of traffic accidents in patients with sleep-disordered breathing: reduction with nasal CPAPEur Respir J19969260611, 8980976 35
Self-reported automobile accidents involving patients with obstructive sleep apneaNeurology19964612547, 8628462 36
Clinical symptoms of sleep apnea syndrome and automobile accidentsORL J Otorhinolaryngol Relat Spec1990525762, 2304760 37
Automobile accidents in patients with sleep disordersSleep198912487494, 2595172 38
Automobile accidents involving patients with obstructive sleep apneaAm Rev Respir Dis198813833740, 3195832 39
A survey screen for prediction of apneaSleep199541815866, 7610311 40
A new method for measuring daytime sleepiness: the Epworth sleepiness scaleSleep1991145405, 1798888 41
Daytime sleepiness and driving performance in patients with obstructive sleep apnea: comparison of the MSLT, the MWT, and a simulated driving taskSleep20093238291, 19294958 42
Relationship between obstructive sleep apnoea, driving simulator performance, and risk of road traffic accidentsThorax2001568005, 11562521 43
Long-term effects nasal continuous positive airway pressure on daytime sleepiness, mood and traffic accidents in patients with obstructive sleep apnoeaRespir Med2000948790, 10714485 44
Automobile accidents in patients with sleep apnea syndrome. An epidemiological and mechanistic studyAm J Respir Crit Care Med19981581822, 9655701 45
Daytime sleepiness and automobile accidents in patients with obstructive sleep apnea syndromePsychiatry Clin Neurosci1998522212, 9628162 46
Self-reported use of CPAP and benefits of CPAP therapy: a patient surveyChest199610914706, 8769496 47
Determinants of daytime sleepiness in obstructive sleep apneaChest198894327, 3383654 48
Excess weight and sleep-disordered breathingJ Appl Physiol20059915929, 16160020 49
Obesity and weight loss in obstructive sleep apnea: a critical reviewSleep19961910415, 8855032 50
Objective evidence that bariatric surgery improves obesity-related obstructive sleep apneaSurgery20071413548, 17349847 51
Driver sleepiness and risk of serious injury to car occupants: population based case control studyBMJ200232411259, 12003884 52
Comparison between two quasi-induced exposure methods for studying risk factors for road crashesAm J Epidemiol200616318895, 16306310 53
A systematic review of epidemiological studies investigating risk factors for work-related road traffic crashes and injuriesInj Prev200814518, 18245316 54
the Standards Practice Committee of the American Academy of Sleep MedicinePractice parameters for clinical use of the multiple sleep latency test and the maintenance of wakefulness testSleep20052811321, 15700727 55
Systematic review of motor vehicle crash risk in persons with sleep apneaJ Clin Sleep Med20062193200, 17557495 56
Epworth Sleepiness Scale and sleep studies in patients with obstructive sleep apnea syndromePsychiatry Clin Neurosci1999533012, 10459718 57
Under reporting of sleepiness and driving impairment in patients with sleep apnoea/hypopnoea syndromeJ Sleep Res199762725, 9493528 58
Self assessment of daytime sleepiness: patient versus partnerThorax19955099495, 8539684 59
Sleep apnea and commercial motor vehicle operators: Statement from the joint task force of the American College of Chest Physicians, American College of Occupational and Environmental Medicine and the National Sleep FoundationChest20061309025, 16963693 60
Daytime sleepiness in the obese: not as simple as obstructive sleep apneaObesity (Silver Spring)200715250411, 17925477 61
Obesity-related sleepiness and fatigue: the role of the stress system and cytokinesAnn N Y Acad Sci2006108332944, 17148748 62
Low sleep quality and daytime sleepiness in obese patients without obstructive sleep apnoea syndromeJ Intern Med200325353643, 12702031 63
U.S. Department of Transportation, Federal Highway AdministrationConference on Neurologic Disorders and Commercial Drivers. Publication No. FHWA-MC-88-0421988 Accessed Feb 23, 20009 WashingtonU.S. DOT, Federal Highway Administration, Office of Motor Carriers http://www.fmcsa.dot.gov/documents/neuro.pdf64
Time course of changes in driving simulator performance with and without treatment in patients with sleep apnoea hypopnoea syndromeThorax200459569, 14694250 65
Driving simulator and neuropsychological [corrected] testing in OSAS before and under CPAP therapyEur Respir J200526898903, 16264053 66
U.S. Department of Transportation, Federal Highway AdministrationOSA and CMV Driver Safety: Recommendations of the MEP2008114 Accessed June 26, 2009 http//www.fmcsa.dot.gov/rules-regulations/TOPICS/mep/report/Sleep-MEP-Panel-Recommendations-508.pdf.
| Year | No. of Individuals with OSA included (n = ) | Diagnosis (e.g., PSG, questionnaire) | Severity of OSA and How was Assessed (n = ) | Sleepiness and How was Assessed (n = ) | Age Distribution | % Male | % CMV Drivers | Driving Exposure | Definition of Crash | Outcome self-reported? | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Commercial Motor Vehicle Drivers | |||||||||||||
| Howard et al.3 | 2004 | 161 | PSG, ESS, and MAP questionnaire | RDI | ESS | 48 ±9 | 99 | 100 | Minimum 10 hrs/wk for work | Any single- or multiple-motor vehicle crash where enrollee was driver | Yes (questionnaire) | ||
| 5.0-14.9 (27.5-42.7)* | 34.8% | 7.69 ±4.34 | |||||||||||
| 15.0-29.9 (9.3-20.7)* | 14.3% | >11: 24.1% (17.6-31.5)* | |||||||||||
| ≥30.0 (6.3-16.4)* | 10.6% | ||||||||||||
| SAS | |||||||||||||
| (RDI ≥5, ESS ≥11) (10.5-22.5)* | 15.8% | ||||||||||||
| Stoohs et al.26 | 1994 | 90 | Questionnaire and Ambulatory screening device: the Mesam IV® | ODI | NR | 36 ±9 | 93 | 100 | NR | Any collision or as driving off the road in the absence of an obstacle. | Yes (questionnaire)No (employer records) | ||
| <10 | n = 44 | ||||||||||||
| ≥10 <20 | n = 26 | ||||||||||||
| ≥20 <30 | n = 10 | ||||||||||||
| >30 | n = 10 | ||||||||||||
| Noncommercial Motor Vehicle Drivers | |||||||||||||
| Mulgrew et al.27 | 2008 | 643 | PSG | AHI >5, ≤15 (mild) AHI >15, <30 (moderate) AHI >30 (severe) | ESS Mild: 9.9 (5.3) Moderate: 9.3 (5.1) Severe: 10.9 (5.2) Other (AHI 0-5): 10.1 (5.2) | 49.9 (11.6) | 71.1 | NR | Km/week driven: 234.3 (208.3) | Crashes involving minor property damage (<$1000) or major property damage (>$1000), with or without personal injury | No (insurance company records) | ||
| Barbe et al.28 | 2006 | 76 | PSG | Severity ranged from AHI = 21 to AHI = 122 AHI mean = 60 (SD = 2) | ESS Cases: 12 ±1† Controls: 3 ±2† | Cases: 49 ±1† Controls: 46 ±1† | NR | NR | km driven, 1,000 km/year Cases: 25 ±2† Controls: 21 ±2† | A crash resulting in property damage >USD 500 and/or personal injury | No (insurance company records) | ||
| Kingshott et al.29 | 2004 | 60 | PSG | AHI Mean (SD) = 8 (9) AHI >5 48% AHI >15 15% AHI >5 + ESS ≥10 20% AHI >15 + ESS ≥10 7% | ESS Cases: 8 ±4 Controls: 8 ±4 MWT Cases: 17 ±4 minutes Controls: 18 ±3 minutes | Cases: 49 ±11 Controls: 49 ±11 | Cases: 48 Controls: 48 | NR | km driven/year Cases: 15,410 ±12,301 Controls: 15,253 ±21,007 | Driver in a single-vehicle crash or causative driver in a multiple-vehicle crash | Yes (questionnaire) | ||
| Pradeep Kumar et al.30 | 2003 | 20 | Questionnaire | NR | ESS Cases: 13.6 ±6.1 Controls: 4.2 ±4.1 | Cases: 41 ±6 Controls: 41 ±8 | Cases: 100 Controls: 100 | NR | NR | Any motor vehicle crash where enrollee was driver | Yes (questionnaire) | ||
| Shiomi et al.5 | 2002 | 448 | PSG | AHI | ESS | 49 ±14 | 89 | NR | NR | Any motor vehicle crash where enrollee was driver | Yes (questionnaire) | ||
| 5-15 | n = 155 | >11 | n = 93 | ||||||||||
| 15-30 | n = 111 | ||||||||||||
| >30 | n = 182 | ||||||||||||
| Findley et al.31 | 2000 | 50 | PSG | AHI 37 (3.8)† | NR | 56 (2)† | 86 | NR | NR | A crash resulting in property damage >$500 and/or personal injury for which the driver was convicted of a traffic violation. | No (State Records) | ||
| Horstmann et al.7 | 2000 | 156 | PSG | AHI | ESS Cases: 12.9 ±5 .5 Controls: 7.2 ±4.7 | Cases: 56 ±10 Controls: 56 ±12 | Cases: 92 Controls: 90 | NR | Cases: # of drivers (%):130 (83) Mean = 19,416 km /driver /year Controls: # of drivers (%):140 (87)Mean = 14,160 km /driver /year | All reported crashes were subdivided into those with property damage <$600 and into those with property damage >$600 or personal injury. | Yes (questionnaire) | ||
| 10-34 | n = 78 | ||||||||||||
| >35 | n = 78 | ||||||||||||
| Median = 20 | |||||||||||||
| Lloberes et al.32 | 2000 | 122 | PSG | AHI mean = 42.5 (SD = 2) | NR | Cases: 51 ±9Controls: 50 ±9 | Cases: 95 Controls: 84 | NR | NR | Any motor vehicle crash or incidence of driving off the road where enrollee was driver | Yes (questionnaire) | ||
| George and Smiley33 | 1999 | 460 | PSG | AHI | NR | Cases: 51 ±12Controls: 52 ±12 | Cases: 88 Controls: 90 | NR | NR | Any motor vehicle crash where enrollee was driver | No (State Records) | ||
| 10-25 | n = 182 | ||||||||||||
| 26-40 | n = 85 | ||||||||||||
| >40 | n = 193 | ||||||||||||
| Teran-Santos et al.8 | 1999 | 102 | PSG and nocturnal respiratory polygraphy (at home) | AHI | ESS Cases: 5.9 Controls: 5.7 | Cases: 44 ±9 Controls: 43 ±9 | 77 | NR | Cases: 24,011 ±22,359 Km driven/yr 20 ±10 yrs of driving Controls: 16,978 ±18,760 km driven/yr 19 ±8 yrs of driving | Any motor vehicle crash where enrollee was driver | Yes (questionnaire) | ||
| ≥5 | n = 29 | ||||||||||||
| ≥10 | n = 21 | ||||||||||||
| ≥15 | n = 17 | ||||||||||||
| Young et al.9 | 1997 | 221 | PSG | AHI | ESS MSLT | 45 ±8 (range, 30-60) | 59 | NR | NR | A crash resulting in property damage ≥$500 and/or personal injury, or if police or other law enforcement personnel were at the crash scene and filed a report. | No (State Records) | ||
| 5-15 | n = 133 | ||||||||||||
| >15 | n = 88 | ||||||||||||
| Cassel et al.34 | 1996 | 59 | PSG | AHI | Questionnaire MSLT | 49 ±1† | 100 | NR | 29,860 ±2,886 km driven/yr† | Any motor vehicle crash where enrollee was driver | Yes (questionnaire) | ||
| Wu and Yan-Go35 | 1996 | 173 | PSG | AHI >5 RDI >5 | NR | SAS 17-24: 3% 25-44: 31% 45-64: 49% >64: 18% Non-SAS 17-24: 9% 25-44: 36% 45-64: 43% >64: 13% | SAS 79 Non-SAS 53 | NR | NR | Any motor vehicle crash or near-miss where enrollee was driver | Yes (questionnaire) | ||
| Haraldsson et al.36 | 1990 | 140 | Questionnaire | SAS | NR | Cases: 48 ±9 (range: 30-69)Controls: 46 ±11 (range: 30-69) | Cases: 100 Controls: 100 | NR | Cases: 24 ±2 103 km 1,800 103 km (accumulated) Controls: 20 ±3 103 km2,900 103 km (accumulated) | Crashes were categorized as single-car (driving off the road) or combined-car (two or more vehicles) crashes. | Yes (questionnaire) | ||
| Incomplete | N = 67 | ||||||||||||
| No sleep spell | N = 35 | ||||||||||||
| Sleep spell | N = 38 | ||||||||||||
| Complete | N = 73 | ||||||||||||
| Aldrich37 | 1989 | 181 | PSG | RDI | MSLT | Cases: 50 Controls: 43 | NR | NR | NR | Any motor vehicle crash or near-miss where enrollee was driver | Yes (questionnaire) | ||
| Findley et al.38 | 1988 | 29 | PSG | Desaturation per hour of sleep (at least 4%) | NR | Cases: 47 ±12 Controls: 45 ±12 | NR | NR | Cases: 13,150 ±7,350 miles driven/yr Controls: 11,290 ±7,780 miles driven/yr | A crash resulting in property damage >$500 and/or personal injury. A driver was at fault if he was convicted of a traffic violation that contributed to the crash. | No (State Records) | ||
| Reference | Year | Number of individuals with OSA | Risk Factors Assessed (Method) | Primary Outcome | Definition of Crash | Factors adjusted for in analysis |
|---|---|---|---|---|---|---|
| Commercial Motor Vehicle Drivers | ||||||
| Stoohs et al.26 | 1994 | 46 commercial drivers with sleep-disordered breathing (SDB) | Oxygen Saturation (ODI) BMI | Crashes/10,000 miles | A motor vehicle crash was defined as the collision of the index case's vehicle with a stationary or moving object or as driving off the road in the absence of an obstacle. | Annual mileage |
| Noncommercial Motor Vehicle Drivers | ||||||
| Pizza et al.41 | 2009 | 24 men with severe OSA | Sleepiness | Simulated driving performance | NR | None |
| Mulgrew et al.27 | 2008 | 643 individuals with AHI >5 | Sleepiness (ESS) Disease Severity (AHI) BMI | Crash rate | Crashes involving minor property damage (<$1000), major property damage (>$1000) or personal injury | Kilometers driven, age, gender, alcohol use, sedative use |
| Shiomi et al.5 | 2002 | 448 individuals with OSA-hypopnea syndrome (OSAHS) | Disease Severity (AHI) | Crash rate | Any motor vehicle crash where enrollee was driver. | None |
| Turkington et al.42 | 2001 | 150 individuals with OSA | Sleepiness (ESS) Disease Severity (RDI) | Odds ratio for crashes in the previous year Odds ratio for near-miss crashes in the previous 3 years | NR | Age, gender, alcohol use |
| Horstmann et al.7 | 2000 | 16 individuals with SAS reporting at least one crash compared to 114 individuals with SAS reporting no crashes | Sleepiness (ESS) Disease Severity (AHI) BMI | ESS score AHI BMI | All reported crashes were subdivided into those with property damage <$600 and into those with property damage >$600 or personal injury. | Kilometers driven |
| Yamamoto et al.43 | 2000 | 13 individuals with OSA reporting at least one crash compared to 26 individuals with OSA reporting no crashes | Sleepiness (ESS) Disease Severity (AHI) Oxygen Saturation (SaO2) BMI | ESS score AHI SaO2 Minimum SaO2 BMI | NR | None |
| George and Smiley33 | 1999 | 460 individuals with OSA | Disease Severity (AHI) | Crash rate | Any motor vehicle crash where enrollee was driver | None |
| Barbe et al.44 | 1998 | 60 individuals with SAS | Sleepiness (ESS) Disease Severity (AHI) Oxygen Saturation (SaO2) | Number of crashes | A motor vehicle crash was defined as a crash resulting in property damage >USD 500 and/or personal injury | Annual kilometers driven, age, alcohol use |
| Noda et al.45 | 1998 | 44 individuals with OSA syndrome (OSAS) | Sleepiness (ESS) Disease Severity (AHI) Oxygen Saturation (SaO2) | Correlation between crash score* and ESS score, AHI, and total oxygen desaturation time | NR | None |
| Young et al.9 | 1997 | 221 individuals with OSA (AHI ≥5) | Sleepiness (ESS and MSLT) | Odds ratios of motor vehicle crash in the past 5 years | Any crash with property damage ≥$500 and/or personal injury, or if police or other law enforcement personnel were at the crash scene and filed a report. | Miles driven, age, gender, alcohol use |
| Engleman et al.46 | 1996 | 204 individuals with sleep apnea/hypopnea syndrome (SAHS) | Disease Severity (AHI) Oxygen Saturation (SaO2) | Correlation between AHI and minimum SaO2 | Crashes were divided into near-misses, casualty-free (“minor” crashes), and crashes causing injury (“major” crashes) | Miles driven |
| Aldrich37 | 1989 | 41 individuals with sleep apnea reporting at least one crash compared to 187 individuals with sleep apnea reporting no crashes | Sleepiness (MSLT) Disease Severity (RDI) Oxygen Saturation (SaO2) | MSLT score RDI Minimum SaO2 | Any motor vehicle crash or near-miss where enrollee was driver | None |
| Study | Year | Unit | Risk Factor | |||||
|---|---|---|---|---|---|---|---|---|
| Sleepiness | AHI or RDI | Oxygen Saturation | Body Mass Index (BMI) | Cognitive/ Psychomotor Function | ||||
| Commercial Motor Vehicle Drivers | ||||||||
| Stoohs et al.26 | 1994 | Crashes/10,000 miles [mean (SEM)] | BMI <25: 0.031 (0.012)b BMI ≥25 <28: 0.041 (0.024)b BMI ≥28 <30: 0.079 (0.039)b BMI ≥32: 0.101 (0.026)b | |||||
| Noncommercial Motor Vehicle Drivers | ||||||||
| Pizza et al.41 | 2009 | Drivers with OSA reporting car crash | ||||||
| Drivers with OSA reporting no car crash | ||||||||
| Mulgrew et al.27 | 2008 | Drivers with OSA reporting crash | ||||||
| Drivers with OSA reporting no crash | ||||||||
| Shiomi et al.5 | 2002 | Crashes per driver per year | ||||||
| Turkington et al.42 | 2001 | Odds Ratio for crash in the previous year | ||||||
| Odds Ratio for near-miss crash in the previous 3 years | ||||||||
| Horstmann et al.7 | 2000 | Drivers with OSA reporting ≥1 crash | ||||||
| Drivers with OSA reporting no crash | ||||||||
| Yamamoto et al.43 | 2000 | Drivers with OSA who had a crash | ||||||
| Drivers with OSA who did not crash | ||||||||
| George and Smiley33 | 1999 | Crashes per year (mean ±SD) | AHI 10–25: 0.08 ±0.12 AHI 26–40: 0.06 ±0.14 AHI >40: 0.11 ±0.15 | |||||
| Barbe et al.44 | 1998 | Mean (SEM) number of crashes | ||||||
| Noda et al.45 | 1998 | The crash score† was significantly correlated with the total oxygen desaturation time (r = 0.46, p <0.05). | ||||||
| Young et al.9 | 1997 | MSLT and ESS scores for drivers with or without crash | MSLT and ESS scores were not significantly associated with crash history | |||||
| Engleman et al.46 | 1996 | Correlation | ||||||
| Aldrich37 | 1989 | |||||||
| Drivers with OSA who had a crash | Males: 7.8 minutes Females: 7.6 minutes | Males: 49‡ Females: 48 | Males: 68‡ Females: 75 | |||||
| Drivers with OSA who did not crash | Males: 8.2 minutes Females: 7.3 minutes | Males: 40 Females: 35 | Males: 75 Females: 77 | |||||
