Introduction

Sleepiness while driving is a major international public health issue. Sleepiness has a relevant impact on road safety, as it substantially contributes to the burden of road-related morbidity and mortality [1,2,3,4,5]. The risk of road death is more than three times higher in low-income countries than in high-income countries, with an average rate of 27.5 deaths per 100,000 population and 8.3 deaths per 100,000 population, respectively. Road injuries are the eighth cause of death for all age groups and are the most important cause of death for children and young adults aged 5–29 years [6]. Road crashes cost most countries 3% of their gross domestic product [7]. A recent European survey about sleepiness while driving reported a 17% prevalence of sleepiness episodes during driving [8]. Over the last two decades, studies have estimated the rate of sleep-related road accidents from 3, 10, 20–22 to 33% according to studies conducted in the USA [9], France [10], UK [11], Italy [12], and New Zealand [4].

A recent meta-analysis showed a strong association between sleepiness while driving and motor vehicle accidents (MVAs). Of the 17 studies included, all but one found that a higher risk of MVAs for drivers reporting sleepiness while driving compared to other drivers. The pooled risk estimate for MVAs due to sleepiness while driving was more than doubled [13]. Sleep-related accidents have multifactorial causes and can be erroneously attributed to other causes such as drug or alcohol consumption, inattention, fatigue, decreased energy, speed, and adverse weather conditions [14, 15].

Sleep deprivation, mainly caused by prolonged wakefulness, shift work, or poor sleep quality, is the most frequent cause of sleepiness. Furthermore, somatic diseases and mental disorders can induce sleepiness. Obstructive sleep apnea syndrome (OSA) should receive close attention, given an estimated prevalence of 23.4% among women and 49.7% among men older than 40 years old [16]. Patients with OSA experience a significantly higher risk of accidents with worse personal injuries [17, 18]. Calculated with the population attributable fraction (PAF), the overall risk for MVAs OSA-related in a general population of male drivers is about 7% [19]. This risk can dramatically decrease with medical treatments with sleep hygiene or continuous positive airway pressure (CPAP) for OSA and changes in behavior. Sleepiness-related MVAs are potentially preventable [18, 20, 21].

The European Union Directive 2014/85/EU [22] included OSA among the clinical conditions of Annex III and established new driver licensing regulatory requirements for patients with OSA. Not treated moderate to severe OSA associated with excessive daytime sleepiness (EDS) represents a medical condition leading to unfitness to drive [23, 24].

The scrupulous evaluation of EDS is a pressing social and medical need for professional and non-professional drivers. Agreement on how to assess EDS in patients with OSA is lacking, and the clinical tools to assess it are not sufficient. This paper provides a brief descriptive review of sleepiness and its implications for safety. It describes the available methods to evaluate EDS in a clinical and non-clinical setting to appraise the fitness to drive in patients with OSA following the introduction of Directive 2014/85/EU in Europe.

Materials and methods

The scholarly literature was examined from inception to the present utilizing PubMed, Medline, and EMBASE. Search terms comprised sleepiness at the wheel, excessive daytime sleepiness, sleepiness measures, sleep-wake cycle, obstructive sleep apnea, driving license, and fitness to drive. The authors also consulted the Legal references for European judicial decisions.

Sleepiness and excessive daytime sleepiness

Sleep is a ubiquitous biological imperative that seems to linger evolutionarily across species [25]. High levels of attention and cognitive performance require both that sleep has sufficient duration, continuity, and intensity, and that has not circadian perturbation. Good quality and quantity of sleep are mandatory to prevent individuals from experiencing physiological changes that may predispose to adverse health outcomes [26, 27]. A lifestyle of chronic partial sleep loss may be partly explained by the fact that individuals frequently carry out other activities (work and free time) sacrificing timing and duration of sleep [28]. In late chronotypes, the restraints of early work schedules lead to an increasing sleep debt during the week that diminishes in the weekends. Many people move their sleep and activity times noticeably between the work week and the weekend or other free days; a circadian schedule referred to as “social jetlag” [29]. In our day, sleep deprivation is the most frequent cause of sleepiness and is linked to adverse cardiometabolic outcomes [30, 31].

Some of the deleterious effects of sleep deprivation are due to the disruption of the synergy between homeostatic process depending on sleep and wake (Process S) and the circadian process (Process C) controlled by the circadian pacemaker. Process S increases during wakefulness and declines during sleep, within a value range that changes with a periodicity that is usually trapped to day and night by the Process C. When S reaches the range’s lower boundary, it triggers awakening, while close to the upper boundary it triggers sleep [32]. Internal and external stimulation of various kinds (proprioceptive, exteroceptive, enteroceptive, cognitive, and emotional) can contribute to initiating and maintaining waking state [33, 34]. These stimuli induce impulses which travel along the central nervous system to activate the ascending arousal system (AAS), composed of several brain structures within the brainstem and cerebral hemispheres as the thalamus, hypothalamus, and basal forebrain. AAS releases multiple neurotransmitters (monoaminergic and cholinergic) co-driving wakefulness [35] and promoting arousal within the thalamus and cerebral cortex [36]. One proposed term for this arousal drive is Process A since it is driven by the afferent nervous system [33]. Since this process includes emotional and cognitive inputs [34] as well as sensory inputs and is also mostly voluntary, the term psycho-sensory (PS) was proposed [36]. Unlike the input to the C wake drive, input to the PS wake drive can be both voluntary and reactive to its surroundings. The PS output is transient and cannot be predictable. Given the complex regulation of this triple-drive model (Processes C, S, and PS) for sleep and wake, it is not hard to regard sleepiness as a complex, multifactorial, and multi-dimensional condition.

Sleepiness appears to be characterized by subjective feelings and self-perceptions that encompass different situations, with various degrees of sleep propensity, alertness, or vigilance (i.e., the capacity to remain awake together with an adequate, sustained, and selective attention over prolonged periods) affecting human performance [37,38,39,40].

Whereas sleepiness is a physiological condition, resulting from the non-linear interplay between triple-drive model [41], EDS represents a pathological condition characterized by an abnormally increased likelihood of falling asleep or lapsing into drowsiness when sleep is not sought or desired. The American Academy of Sleep Medicine (AASM) defines EDS as “the inability to maintain wakefulness and alertness during the major waking episodes of the day, with sleep occurring unintentionally or at inappropriate times almost daily for at least three months” [42]. EDS represents a severe, pathological, disabling condition with essential effects related to gender, age, and comorbidities, such as depression and other disorders commonly linked with fatigue. Fatigue is, however, a different concept which should not be confused with sleepiness [43]. Epidemiological studies of EDS are complicated by varying definitions and by the use of a broad array of subjective and objective instruments. Prevalence rates of EDS consequently range from 4 to 21%, depending on the definition and the method of appraisal [44]. The National Sleep Foundation has performed several population-based surveys, according to which about 30% of respondents suffer from EDS that interferes with quality of life [45]. Severe EDS affects behavioral, physiological, and cognitive operating, as well as the quality of life. EDS results in a reduction of reaction time, vigilance, alertness, and concentration, and impairs the ability to sustain attention-based activities [46]. Sleep loss enhances emotional susceptibility and tolerance to pleasant stimuli and also impairs prefrontal cortical functioning and connectivity involved in inhibitory control [47]. Furthermore, significant sleep deprivation increases sensitivity to rewards and risk-taking [48]; sleep loss can also increase risk-taking behavior. Sleepiness uniquely predicts risk-taking, following the adjustment for the influence of recent sleep, chronotype, chronic sleepiness, and trait sensation seeking [49].

EDS and OSA

The most common medical cause of EDS is OSA. OSA is a widespread breathing disorder related to sleep, which is characterized by repetitive, partial (hypopnea) or complete (apnea) episodes of pharyngeal and upper way collapse, resulting in weak gaseous exchange, oxygen desaturation, and bouts of concomitant hypoxia and hypercapnia. Anatomical factors determining upper airway narrowing together with physiological, non-anatomic factors contribute to cause apnea [50, 51]. The clinical manifestations of OSA differ considerably among individuals. Clustering of symptoms and comorbidities such as arterial hypertension, cerebral-cardio-metabolic complications, neurocognitive disorders, or severe obesity allows discrimination between clinical phenotypes [52, 53].

EDS is a distinctive feature of clinically significant OSA. EDS leads to impaired function, primarily executive functions, reduced quality of life, and poorer cardiovascular outcomes [54,55,56,57]. OSA imposes a considerable clinical and epidemiological burden [16]. However, despite its importance, OSA still represents a relatively overlooked and undiagnosed public health condition [58, 59].

Significant interindividual variability in EDS and personal susceptibility to the systemic effects of OSA exist in subjects with comparable severity of OSA, and reasons for this variability remain slightly understood [60, 61]. Measures of OSA severity, including apnea-hypopnoea index (AHI), scarcely correlate with subjective and objective EDS [62,63,64,65,66]. The explanation for this inconsistency may be the individual susceptibility differences in pathogenicity of main underlying mechanisms of OSA, such as intermittent hypoxia, intrathoracic pressure swings, variations in sympathetic and muscular tone, and hemodynamic instability [43].

Positive predicting factors of EDS in OSA comprise African-American race, younger age, obesity, higher sleep efficiency, short sleep duration, increase in respiratory arousals, and nocturnal hypoxemia with activation of the sympathetic nervous system [53, 67, 68] and inflammatory cytokines [69,70,71,72]. The correlation of chronotype with OSA is unclear and conditioned by obesity and mood, even if morning chronotype seems to have a protective effect on EDS in OSA [62]. In OSA, sleep deprivation enhances oxidative stress, inflammation, and risk for hypertension [73, 74].

If not properly treated, patients with OSA are at higher risk of experiencing MVAs [8, 12, 75]. However, successful treatment of OSA can re-establish fitness to drive along with its consequent EDS [23]. It is substantial that sleep specialists experienced in the management of OSA and EDS evaluate the effectiveness of treatment [24]. OSA intensity and reduction of EDS require careful consideration before attesting the fitness to drive. Currently, there are only few published national guidelines and evidence-based protocols informing and steering the procedure [76]. Residual EDS in patients successfully treated for OSA may still lead to a significant socioeconomic burden, including road and work-related accidents, reduced quality of life and overall health, neurocognitive impairment, as well as all-cause and cardiovascular morbidity and mortality [77].

Assessing EDS for fitness to drive in patients with OSA

Epidemiological studies on EDS are hampered by the varying definitions of the condition and the variety of subjective and objective instruments used to measure EDS. Consequently, the classification of sleepy subjects with OSA remains a controversial issue. An agreement on how to assess EDS in patients with OSA is lacking. Clinical tools for appraisal of EDS are insufficient, particularly regarding the assessment of fitness to drive.

Subjective measures

Various psychometric tools exist to measure EDS quantitatively [78]. Subjective measures are a fast and cost-effective way of estimating EDS. The following paragraphs describe the main subjective measures used.

The “Epworth Sleepiness Scale” (ESS) is the most widely adopted clinical tool to assess subjective sleepiness [79]. The ESS is a questionnaire asking the likelihood to doze off or fall asleep in eight different situations most people encounter in their daily lives. Responses range on a Likert scale from 0 to 3, with a total score from 0 to 24. A score higher than 10 is usually used to define EDS: the higher the score, the higher the subject’s risk of falling asleep during the day. Even though the ESS has good reliability (0.73–0.86), it suffers from several drawbacks. ESS was found to poorly correlate with AHI and OSA-related variables, with correlation coefficients from 0.11 to 0.23 [80]. ESS presents a natural tendency to underestimate EDS when patients complain of insomnia or have difficulties falling asleep. ESS overestimates EDS during some sleep-inducing conditions, which does not necessarily indicate problematic EDS depending on activities of daily life for different genders and ages [81, 82]. Furthermore, it should be emphasized that the ESS usually applies to clinical settings and populations, with a dearth of data about its psychometric properties in non-clinical samples, in particular among drivers. Notable driver studies include those by Baiardi and collaborators in a cohort of 221 commercial drivers [83], and those by Garbarino et al. in a sample of 283 truck drivers of dangerous goods [58], in a cohort of 526 truck drivers [84] and in a sample of 949 truck drivers [85, 86].

Questionnaires such as the ESS may provide more accurate results when compared with additional details from a partner [43]. Focused questions by the clinician on critical features of EDS may be more reliable than the ESS, such as investigating whether EDS is present when the individual is alone and non-active, or when mentally, physically, and socially engaged, and when performing high-risk activities, such as driving.

A systematic review of the literature has indicated higher correlations of the ESS score with other measures of EDS (the Maintenance of Wakefulness Test (MWT), r = − 0.43, and the Multiple Sleep Latency Test (MSLT), r = − 0.27) than with less strictly related constructs (severity of OSA and general health measures, pooled r ranging from 0.11 to 0.23), though all correlation coefficients were lower than expected [87]. The internal consistency of the ESS (Cronbach’s alphas ranged from 0.7 to 0.9) suggests that this tool is trustworthy for group-level comparisons, but it should be used with caution for individual-level comparisons [87].

Other questionnaires evaluate the overall feeling of sleepiness, such as the Basic Nordic Sleepiness Questionnaire (BNSQ), by asking how often per week subjects complain EDS [88]. This question correlates with different factors that contribute to EDS, including snoring [89], nasal congestion [90], and OSA severity categories [91]. EDS is usually defined when subjects experience daytime sleepiness at least 3 days per week. The prevalence of EDS in the general population ranges from 17 to 28.5% by using BNSQ [92,93,94]. In sleep research, the BNSQ has been used together with the ESS to identify individuals with EDS at the wheel [1].

Another subjective tool used routinely is the “Stanford Sleepiness Scale” (SSS), initially developed as a 24-item true/false format and administered to 340 undergraduate students. The principal component analysis identified two main dimensions, each one explaining 24.2 and 20.6% of the total variance, respectively [95]. The SSS uses a self-rated seven-point scale of the same range to quantify symptoms [96]. The modified SSS is comprised of only one item on seven-point scale measuring levels of sleepiness throughout the day. A score of more than 3 points at any time when the respondent should be feeling alert indicates serious EDS [97]. The comparison between ESS and SSS in patients attending sleep clinics who receive information about sleepiness and fitness-to-drive [98] provide largely reproducible information when they report their symptoms.

The “Barcelona Sleepiness Index” (BSI) [99] was developed using focus groups in a sample of 53 consecutive patients and their bed partners. The patients were suffering from sleep-disordered breathing and complained of EDS. The tool was subsequently validated against objective measures in another sample of 98 consecutive patients suffering or not from EDS. The objective measures included the “Sustained Attention to Response Task” (SART), MWT, and MSLT. Two items achieved the highest predictive value for the MWT and were sensitive to treatment changes with continuous positive airway pressure, with an overall sensitivity and specificity of 64.9 and 72.1%, respectively.

The “Karolinska Sleepiness Scale” (KSS) gauges the subjective level of sleepiness at a given time of day, a measure of situational sleepiness. On this scale, subjects indicate which level best reflects the psychophysical state experienced in the last 10 min. This is a nine-point scale (1 = extremely alert, 3 = alert, 5 = neither alert nor sleepy, 7 = sleepy—but no difficulty remaining awake, and 9 = extremely sleepy—fighting sleep). A modified KSS exists and contains one other item, 10 = extremely sleepy, falls asleep all the time [100]. Since the KSS measures situational sleepiness rather than trait sleepiness, it has not been widely used for fitness to drive.

Objective measures

Both the MSLT and the MWT are the main objective test tools for EDS, and they are considered the gold standards for the detection of EDS in the clinical practice of sleep medicine [101]. It is important to note, however, that the reported broad variabilities of the MSLT and MWT sleep latencies in the healthy population overlap with findings in medical or sleep disorders characterized by EDS [102, 103].

The MSLT measures the underlying physiological tendency for a subject to fall asleep in lack of alerting factors [104]. The test takes place in the sleep laboratory after polysomnography of the previous night and consists of four or five nap opportunities performed at 2-h intervals. Being used worldwide, the MSLT is the most widely accepted objective measure of physiologic sleep tendency [102, 103].

The MWT, which is a laboratory-based daytime polysomnographic test, gauges the subject’s ability to stay awake in a context with low levels of stimulation [102]. Mitler et al. [105] designed the test to measure daytime wakefulness when there are clinical issues that concern treatment efficacy or patient safety. Recording conditions during the MWT are similar to but not equal to those of the MSLT. The MWT is performed in a sequence of four tests at 2-h intervals beginning 1.5 to 3 h after the individual’s standard waking time [103].

The task force of the AASM [101] underlined that the MSLT is not targeted to assess the effects of EDS in patients dealing with potentially dangerous circumstances (i.e., fitness to drive). On the contrary, the MWT, which requests patients to fight against sleepiness, is better adapted to evaluate fitness to drive [106]. During the 40-min MWT, abnormal sleep latency correlates with impaired driving as measured both on a driving simulator [107] and in real driving conditions [108] in untreated patients with OSA. Another study has confirmed that individuals with shorter sleep latencies on the MWT exhibit impaired driving performance compared to subjects with intermediate and long sleep latencies, regardless of their sleep diagnosis and whether being in treatment or not. This study supports the usefulness of the MWT as a reliable measure of EDS related to impairment for driving in patients with sleep disorders [106]. However, since both the MSLT and the MWT are rather expensive and somewhat impractical, and since they both require a sleep laboratory, there is a need to develop more straightforward, less expensive tests to evaluate objective EDS in patients with OSA to estimate their fitness to drive.

Several studies confirmed the negative effect of EDS on driving performance, as assessed by driving simulation (DS) or in real driving conditions [107, 109, 110]. Unfortunately, DS testing cannot be translated into fitness for real driving [84]. However, studies on DS in patients with OSA have assessed the relationship between objective EDS measurements and driving performance. These studies confirmed that tracking error on a divided-attention driving task correlated with MSLT sleep latencies in baseline conditions and improved after the use of CPAP therapy [111, 112]. The monotonous DS testing confirmed the correlation between MSLT sleep latencies and lane-position variability in patients with OSA [109]. DS performance correlated closer with MWT, compared with MSLT for the propensity to fall asleep, with ESS score > 11, and reported the previous history of road accident [109]. As a group, patients with OSA show lower steering performance and exhibit a higher number of road accidents than healthy age-matched controls. However, the variability in driving performance among patients is wide [113].

In a recent systematic review [114] comparing subjective and objective measures of driving performance and road safety, only two studies were examined considering sleep, drowsiness, and monotonous driving in DS. There was no significant time of the day compared to subjective measures of EDS, but there were significant increases in the occurrence of micro-sleeps at the end of the day [115]. Discrepancies between the self-report and objective measures of EDS (electroencephalography) suggest an impairment in the drivers’ ability to evaluate their sleepiness [116].

As both subjective (ESS) and objective (MSLT) measures of EDS may lead to different results in clinical populations with OSA, previous studies have applied a twofold approach of identifying “sleepy” patients based on pairing subjective and objective sleepiness [117, 118]. Sforza et al. [119] compared different subjective (SSS) and objective (polysomnography or PSG, and MWT) measures of EDS in subjects with breathing disorders related to sleep using psychomotor vigilance testing (PVT). The PVT is a computer-based test with a chronometric measure of an individual’s reaction time to specified small visual stimulation in a labile environment. The small changes are assessed at pseudo-random intervals ranging from 2 to 10 s [120, 121]. Individuals are trained to react quickly to a digital signal on a computer terminal by pressing a key. Errors of omission and commission are collected. Reaction times exceeding 500 ms or failure to react are defined lapses. The cumulative count of lapses is the principal outcome measure of PVT performance believed to represent perceptual, processing, or executive failures in the central nervous system. Count of lapses strictly correlates with the effects of sleep deprivation, sleep restriction, on attention, and vigilance [122,123,124,125]. Patients with OSA, compared to controls, show impairments in performance affecting both speed and accuracy. Lapses and false responses were outnumbered in patients with more severe EDS in the MWT, and with higher AHI, while analysis of omissions seems to be more sensitive to the assessment of diurnal impairment [119]. As a consequence of chronic sleep restriction or total sleep deprivation, EDS is associated with impaired PVT performance in a dose-response manner. In patients with OSA, objectively measured EDS via the MWT is associated with slower reaction time in PVT, whereas subjectively assessed EDS measured by ESS is associated with increased numbers of lapses [126].

Lapses in attention on the PVT reflect state instability linked with sleep deprivation. Lapses are the most common PVT metric used in the peer-reviewed literature [127]. Two lapses were reported to be the lowest threshold for discriminating between sleep-deprived and non-sleep-deprived individuals. Albeit the PVT measures the neurobehavioral alertness, it is sensitive to EDS induced through sleep deprivation and associated with OSA and is much less expensive and time-consuming compared with MSLT [54, 119, 126, 128].

Certification of fitness to drive in OSA-EDS patients: we have got something that might work but no guarantees

New EU legislation has been adopted on fitness-to-drive tasks for the detection, treatment, and management of drivers with an AHI of 15 or more in the presence of EDS. Overnight PSG is deemed the “gold standard” for the diagnosis of OSA. However, PSG is not widely available, and the expenses of time and money constitute a burden to public health resources.

The OSA Working Group under European Commission Directive 2006/126 proposed a questionnaire to screen for OSA (Q-OSAS) in order to identify driving license applicants at risk for OSA with EDS. The questionnaire is not mandatory, but member states are strongly encouraged to use it [129, 130]. Five factual elements are easily collected: gender, age, height, weight, and the history of recent MVAs involving injuries or property damage according to police reports. The OSA Working Group investigated the presence of snoring (frequency and intensity), of witnessed pauses in breathing as identified by bed partners, of EDS primarily while driving associated with the restorative quality of sleep, and the diagnosis of arterial hypertension. Most questions required YES/NO/DON’T KNOW answers. EDS can be explored using the ESS [129]. Unfortunately, the ESS shows only moderate correlation with objective measures of EDS in patients with OSA [131].

If patients know that admission of EDS may have implications regarding their driving license, they may give biased answers to questions using tools like the ESS. This behavior is mainly present in the case that patients’ work depends upon a license to drive [98]. Patients attending sleep clinics do provide largely reproducible information on the ESS or the SSS; only 5.7% of them reconsiders and switches their score from sleepy to non-sleepy. Clinicians should inform patients about potential implications before compiling the ESS. Factors that play a role in the reliability of self-reported data include the setting of individuals’ testing and whether or not results remain private [98].

Another limitation of the ESS was recently reported: feeling sleepy better correlates with many variables that contribute to EDS, such as restless leg syndrome and insomnia, rather than with the risk of dozing [94]. Quite often, sleepy subjects complain of not feeling restful during the day rather than referring sleepiness; they also complain of having a decidedly modest quality of life. Subjects at risk of dozing off during the day but not complaining EDS have comparable symptoms as the non-sleepy population, except for snoring and self-reported apnea [94]. Therefore, since it measures just one aspect of sleepiness, the ESS alone is not adequate to assess subjects for fitness to drive. [81, 132, 133].

Limitations

Several factors may condition the generalizability of the findings of this review. Comparing reported results is challenging due to the varying definitions of EDS, and the many different tools used to measure EDS, both subjectively and objectively. EDS and OSA severity are loosely correlated. Most subjects with moderate to severe OSA are not pathologically sleepy (50–80%), and approximately half of such subjects deny EDS [61, 134]. Fifty percent of the drivers who are habitually sleepy state that sleepiness occurs predominantly at the wheel while denying EDS during any other activities [21]. Residual EDS may remain even with optimal treatment of OSA with CPAP, suggesting that EDS may be related to other coexisting factors and comorbidities [135].

Considering sleepiness as a complex phenomenon with different subjective symptoms and objective correlates, each tool used to measure EDS explores different aspects with variable degrees of accuracy. While reasonably reliable in clinical settings, these tools are not accurate enough to characterize fitness to drive. There is no consensus on the method to use for assessing EDS in patients with OSA.

Future perspectives

The ideal tool for measuring sleepiness should be reliable, reproducible, widely available, inexpensive, and easy to perform. Since objective measures of EDS are unwieldy and expensive, and subjective measures of EDS are unreliable, the search for a predictive biomarker is now a target of the research in progress. Despite the publication of data on biomarkers in OSA, including the effect of treatment, research in this area is still at the beginning [136,137,138,139]. One promising biomarker is the proinflammatory cytokine IL-6, but there are conflicting data on its effects in OSA [140]. Bravo et al. [71] identified levels of IL-6 in a group of men with severe OSA compared with matched controls. Objective and subjective sleepiness was assessed using ESS and MSLT. OSA patients had higher levels of IL-6 compared with controls, but no differences emerged in IL-6 levels between OSA with and without sleepiness [71]. A significant relationship between objective sleepiness by MSLT and elevated IL-6 levels has been reported, but this relationship was not found for severe subjective sleepiness and IL-6 in OSA [72]. The researchers demonstrated a similar association of IL-6 levels with EDS identified by PVT and ESS and with subjective sleepiness, but they did not find an association of IL-6 with PVT measured sleepiness alone [62]. The potential usefulness of IL-6 as a biomarker of both physiological and residual sleepiness in treated OSA, regardless of other confounding factors such as obesity and sleep duration, needs further exploration 140,141,142,143,144,145,146]. A double approach to identify “sleepy” patients, based on combined subjective and objective sleepiness tests and physiological and biochemical biomarkers, may be more reliable than any single measure of EDS [147].

Conclusions

Sleepiness while driving has a significant impact on road safety, contributing to the burden of traffic-related morbidity and mortality. OSA is the most frequent medical cause of EDS. Further investigation is needed to clarify which patients with OSA are vulnerable to EDS while driving and to enhance predictive accuracy for driving risk in OSA.

Road safety programs should be strengthened to inform drivers with OSA about the risk involved with EDS at the wheel and improve their awareness of risk for EDS-related MVAs. Due to the relevant social and economic impact of EDS in professional and non-professional drivers, we recommend the implementation of screening programs for EDS.

Inexpensive, less time-consuming new tools are urgently needed for the objective evaluation of EDS in OSA patients for certification of driving license. Useful screening tools will most likely result from a combination of physiological and biochemical biomarkers.