Screening for Obstructive Sleep Apnea in truck drivers

Abstract Professional drivers show a higher prevalence of obstructive sleep apnea (OSA) compared with the general population. OSA has been widely associated with an increased risk of traffic accidents. This article aims to investigate the presence of risk factors for OSA, its prevalence and the value of screening tools in a truck drivers’ cohort. Descriptive and analytical prospective study. Demographic, anthropometric, Epworth Sleepiness Scale, STOP-Bang and Berlin Questionnaire were used to select subjects with suspicion of OSA. Polysomnography (PSG) was performed in individuals with positive screening. Mean age was 44.6±7 years, mean body mass index was 28.7±4 kg/m². Of the 281 truck drivers screened, 88 were positive for potential OSA. Of these, 63 completed PSG study and the diagnosis was confirmed in 85.7% (prevalence of 19.2%). The following variables showed a positive correlation with the apnea-hypopnea index: neck circumference and STOP-Bang. The combination of a predominantly male population, obesity, age distribution and lifestyle could account for the high prevalence of OSA in this specific population. Questionnaires proved to be a valuable screening tool. Screening, treatment, and management of OSA should be a priority as a public safety policy.


introduction
Professional drivers are an understudied and medically underserved population, with high rates of obesity and related comorbidities 1 .Research has shown that drivers with multiple comorbidities, such as cardiovascular and cerebrovascular diseases, psychiatric conditions, diabetes, medication use, uncorrected visual defects and obesity seem to be at increased risk of being involved in a crash 2 .Sleep disturbances and sleepiness are also recognized to be associated with increased accident risk, exceeding the risk associated with many other medical disorders 3,4 .
The most common type of sleep-disordered breathing (SDB) is obstructive sleep apnea (OSA).OSA is a clinical entity characterized by recurrent episodes of apnea and/or hypopnea during sleep, it occurs due to the total or partial collapse of the upper airway tract 5,6 .OSA is associated with an increase in morbidity and mortality related to cardiovascular, cerebrovascular and metabolic disorders and neurocognitive functions 7,8 .Due to the constant sleep disruptions, OSA decreases both sleep quality and quantity and, subsequently, may be associated with daytime sleepiness and impaired psychomotor vigilance.Therefore, it is not surprising that untreated OSA increases the risk of crashes by four to five-fold in comparison with subjects without OSA among commercial drivers 9 .Furthermore, there is additional evidence supporting that adequate treatment of OSA reduces the risk of motor vehicle crashes to control levels 9,10 .
This condition has a high prevalence, and it is thought to be underdiagnosed.A systematic review carried out in 2016 by Senaratna et al. 6 shows a prevalence in the general population of 9% to 38%.Professional drivers show a higher prevalence of OSA, between 28% and 78% 11 , probably because they tend to be obese, male and middle-aged, the most common risk factors for OSA 12 .Despite this risk, OSA remains underdiagnosed and untreated in this work category, mainly due to lack of appropriate screening and sleep study referrals 13,14 .
The American Academy of Sleep Medicine (AASM) has issued a comprehensive and detailed document on screening, and diagnostic and therapeutic approaches for commercial drivers 11 .However, in Portugal and many other countries, there is still no consensus in the OSA screening approach, and a standardized model is needed.
Screening tools such as the Epworth sleepiness scale (ESS), the Berlin questionnaire (BQ) and STOP-Bang questionnaire (SBQ) have been widely used for detecting OSA.The ESS is one of the most used tools to measure sleepiness 15 and requires patients to rate their likeliness of falling asleep in eight different situations; the higher the score, the higher the person's level of daytime sleepiness 16 .The ESS appears to be a convenient, standardized, and cost-effective way to measure sleepiness in patients who suffer from sleep disorders.The BQ categorizes patients as either high or low risk for OSA based on self-reports of snoring, daytime sleepiness, hypertension and obesity 17 .It has modest-high sensitivity and low specificity to detect clinically relevant OSA 18 .The STOP-Bang questionnaire, due to its ease of use, efficiency, and high sensitivity, has been widely adopted and validated in various populations 19 .The questionnaire includes four dichotomous (yes/no) items (STOP: snoring, tiredness, observed apnea and high blood pressure) and four demographic questions [BANG: body mass index (BMI), age, neck circumference and gender] 20 .Patients can be classified for OSA risk based on their respective scores 19 .
Identifying and treating commercial drivers with OSA should decrease crash-related injuries and fatalities and improve drivers' safety and health.Occupational medicine examinations present a unique opportunity for detecting OSA as part of drivers' fitness for their job.This study aimed to investigate the presence of risk factors and symptoms for OSA and its prevalence and the value of screening tools in a truck drivers' cohort.

material and methods type of study
Prospective, descriptive, and analytical.

Participants and study design
This study was conducted between January 2016 and December 2021 on a sample of truck drivers employed by a Portuguese road freight company.All truck drivers were invited to participate in the study.The participants were informed about the study procedures prior to signing informed consent.The drivers' work schedule was variable, they worked during the day and night, for long hours, with no set times for sleeping or meals.
The truck drivers were invited to complete anthropometric assessments (weight, height, and neck circumference (NC) measurements), per-formed by a health professional, to determine their body mass index (BMI) using the criteria established by the World Health Organization 21 .
Subsequently, participants completed a questionnaire evaluating sociodemographic characteristics, work, and sleep habits.In addition, it included questions concerning personal habits like smoking, consumption of alcohol and use of hypnotic-sedative drugs.Drivers' past medical history was also assessed, with a special attention given to cardiovascular and respiratory disorders.Participants were also asked about the presence of common symptoms of OSA (including night-time and daytime symptoms, presented in Table 1).The presence of hypersomnolence was measured by the Epworth Sleepiness Scale (ESS), a validated eight-item self-rating scale with a maximum of 24 points 16 .A total score ≥11 was considered indicative of excessive daytime sleepiness (EDS).The presence of insomnia was evaluated by SleepMed Insomnia Index Questionnaire (SMIIQ), which is a clinical measure of the characteristics and severity of the insomnia 22 .A total score ≥15 was considered indicative of insomnia.
The Berlin Questionnaire (BQ), which has been validated into the Portuguese language 23 , was also included in this investigation.It is a subjective screening tool for OSA consisting of 10 questions divided into three categories including snoring, sleepiness and fatigue, hypertension, and obesity (defined as a BMI≥30 kg/m²).Drivers who fell into at least two of these categories were classified as being at high risk for having OSA.Furthermore, all the participants were asked to complete the STOP-Bang questionnaire which has been validated into the Portuguese language 24 , and information concerning BMI, age, neck circumference, and gender was collected by a research assistant.The risk score was calculated using the sum of the "yes" answers (8  parameters), with each answer "yes" being equal to 1 point, totalling a minimum score of 0 and maximum of 8; high risk for having OSA was considered when the risk score was ≥3.
A positive screening for OSA was considered when: ESS ≥11 or high risk BQ or STOP-Bang questionnaire.Drivers with suspected OSA were submitted to an overnight portable polysomnogtable 1. Summary of the statistics of the sample studied and the comparison between groups of drivers with negative and positive screening for obstructive sleep apnea (OSA).

All drivers n=281
Screened All subjects who underwent PSG received a report to deliver to their primary care physician to be referred for sleep consultation whether to discuss treatment options or to perform laboratory PSG, to confirm or exclude the diagnosis.

Statistical analysis
A descriptive and inferential analysis was performed using the SPPSS® 22.0 software (Statistical Package for the Social Sciences).Categorical variables are presented as absolute frequencies and percentages, and continuous variables as means and standard deviations (SD), or medians and interquartile ranges for variables with skewed distributions.Normal distribution was checked using Shapiro-Wilk or skewness and kurtosis.
Categorical variables were compared with the use of Fisher's exact test or the Chi-squared test, as appropriate.Continuous variables were expressed as means ±SD and compare with use of Student's t-test.Pearson's correlation coefficient was used to assess the correlation between NC and AHI and between STOP-Bang questionnaire and AHI.All reported P values are two-tailed, with a P value of 0.05 indicating statistical significance.
For subjects who underwent confirmatory diagnostic testing, the positive predictive value (PPV) of the screening criteria was estimated as: (Subjects with confirmed OSA diagnosis/Subjects with a positive OSA screen) x 100%.
Of the 281 truck drivers screened, 72 (25.6%) reported usual snoring (defined as snoring three or more nights per week), 25  According to the BQ and STOP-Bang questionnaires, 79 (28.1%) and 80 (28.5%) patients were at high risk for OSA, respectively.Therefore, a total of 88 (31.3%) subjects screened positive for OSA.Table 1 summarizes the study population and compares drivers with positive and negative OSA screens.Drivers meeting the criteria for high-risk for OSA were significantly older, and with higher averages BMI and NC.Moreover, the percentage of drivers with obesity, arterial hypertension and use of hypnotic-sedative drugs was significantly higher in this group.They also reported significantly more daytime and night-time symptoms typical of OSA, considering the percentage results.
Figure 1 illustrates the clinical yields of screening.
From the 88 subjects with suspected OSA, 10 (11.4%) refused to perform PSG, and 15 did not perform PSG because they lost follow-up.A total of 63 (71.6%) drivers completed PSG study and the diagnosis was confirmed in 54, yielding an estimated positive predictive value of 85.7% for the screening criteria.Of these 18 (33.3%)were diagnosed with mild OSA, 24 (44.4%) with moderate OSA, and 12 (22.2%)with severe OSA.This accounts for a prevalence of at least 19.2%.

Discussion
Our results are consistent with previous findings regarding the assessed risk factors for OSA.Obesity is well recognized as one the most important risk factors for OSA 25 .In our study, 31.3% and 50.2% of subjects were obese and overweight, respectively.Additionally, we found that the average BMI and NC of drivers with high risk for OSA was significantly higher than negative OSA screening drivers (p<0.001) and that there was a positive correlation between NC and AHI (r=0.42;p=0.001).These findings are in line with some authors who stated that the neck circumference correlates well with obesity and an increased NC is associated with a higher risk of OSA 26,27 .The reasons for the high incidence of obesity in professional drivers is speculated to be related to their sedentary life style, with prolonged working hours, inappropriate dietary habits, physical inactivity and occupational stress 10,28,29 .The percentage of drivers with arterial hypertension was also higher in the group with positive screen for OSA.Untreated OSA is associated with an elevated risk of hypertension and cardiovascular disease 30 .Observational studies have reported a prevalence of OSA over 30% among hypertensive patients and nearly 80% among resistant hypertensive patients 31,32 .Several pathophysiologic factors contribute to the relationship between OSA and cardiovascular risk, including endothelial dysfunction, neurohormonal dysregulation and inflammation 30 .Patients in that group were also older, reinforcing the data from current literature confirming the high prevalence of OSA in older individuals 33 .Patients who screened positive for OSA reported significantly more daytime and night-time symptoms typical of OSA presented in Table 1, which is consistent with the current literature 34,35 .Although 11% reported daytime sleepiness, only 3.9% presented excessive daytime sleepiness when evaluated through ESS (ESS≥11), which can be explained by the fact that in some patients the ESS may not be a meaningful measure of EDS 35 .ESS only measures the tendency to fall asleep in specific situations which might differ in different groups of individuals and, therefore, may affect their answer regardless of how sleepy they are, for example depending on gender and age 36,37 .Given that ESS and self-reported EDS appear to measure distinct aspects of sleepiness, many subjects with EDS will go undiagnosed.We found a prevalence of insomnia of 8.5%.OSA and insomnia are highly prevalent in the general population and can often coexist.They share some clinical features and may aggravate each other 38 .Insomnia may reduce compliance to the treatment of OSA with positive airway pressure therapy, and OSA can make insomnia more refractory to cognitive-behavioural therapy 39,40 .Therefore, it is highly recommended for clinicians to investigate the potential co-occurrence of these disorders 38 .The COVID-19 pandemic could have also contributed to this result because many individuals reported sleep disturbances, including insomnia during this period 41 .A larger number of patients in the group who screened positive for OSA were under hypnotic-sedative drugs.Given their central nervous system depressant properties and potential myorelaxant effects, early studies found that these agents can prolong respiratory events and worsen overnight hypoxemia in OSA 42 .However, according to more recent studies, there is no evidence that hypnotics impair pharyngeal muscle activity during sleep and standard doses of common hypnotics do not systematically worsen OSA severity as measured via the AHI 43 .This difference between groups can be explained by the possible coexistence of insomnia and thus, be exposed to hypnotic use 44 .Also, patients undiagnosed for OSA are usually under hypnotic medication due to complains of sleep disruption.In our study, 60.1% of patients had smoking habits, however, we did not find a higher prevalence of smokers in the group who screened positive for OSA.Many authors have studied the potential association between smoking and OSA and it is reported that its impact on OSA is due to various mechanisms including changes of sleep architecture, of upper airway neuromuscular function, of arousal mechanisms and of enhanced upper airway inflammation 45 .It has also been found that untreated OSA is related to smoking addiction 46 .Nevertheless, data remain controversial and it is still unclear whether smoking represents a risk factor for OSA.Our findings are in line with other studies that failed to conclusively establish a significant association between cigarette smoking and OSA [46][47][48] .
Regarding the screening tools we used in our study, the Epworth sleepiness scale (ESS), the Berlin questionnaire (BQ) and the STOP-Bang questionnaire (SBQ) are widely used for OSA 49 .Three different studies [49][50][51] observed that the sensitivity and diagnostic odds ratio (DOR) of the SBQ were higher than those of the BQ and ESS for detecting mild, moderate, and severe OSA.By contrast, compared with the ESS, the SBQ had limited value in screening out patients without OSA.In order to enhance our screening accuracy, we used three validated and widely used questionnaires.In fact, sleep quality should be evaluated using a combination of the different tools, in order to obtain a complete picture of both sleep and daytime impairments 15 .We found a positive correlation between STOP-Bang and AHI (r=0.37;p=0.006), which suggests that the STOP-Bang test is useful for predicting OSA severity.These results are corroborated by Farney et al. 52 and Chung et al. 53 , who also demonstrated that as the STOP-Bang score increased, the probability of having more severe OSA also increased.
In this sample of Portuguese truck drivers, we found 88 (31.3%) cases that screened positive for OSA.These results are comparable with another sample of 714 Portuguese truck drivers in which 29% were at high risk for having OSA 54 .Of the 63 drivers who underwent PSG, 54 had confirmation of the diagnosis of OSA, yielding an estimated PPV of 85.7%.Previous studies that used screening criteria similar to this study found PPV that varied between and 78.5 and 100% 12,14,55,56 .In our study, the prevalence of OSA obtained by a combination of questionnaires and PSG was at least 19.2%.This is higher than the prevalence observed in a similar study 14 but lower than the estimated in previous studies which identified sleep-disordered breathing in between 25 and 78% of commercial drivers 11 .
Nonetheless, our study does have a number of limitations and the true prevalence of OSA may be underestimated for several reasons.First, the subjective nature of some items of the questionnaires that were used, may have been subject to under-reporting by the drivers.Regarding the study design, there is lack of PSG data on subjects with negative OSA screens, precluding precise estimates of disease prevalence.In addition, 28.4% of drivers who had positive screening did not undergo PSG because they refused or lost follow-up; this rate is consistent with observations from previous publications 55 .Furthermore, OSA was defined using a home unattended level III PSG.Although portable monitors have been proposed as a cheaper and more accessible technology for OSA detection in specific population, such as individuals with high pre-test suspicion and without complicated comorbidities, the gold standard test remains in-laboratory, technician-monitored PSG.In fact, portable monitors tend to underestimate the severity of OSA, and an important proportion of professional drivers who tested negative will likely require a follow-up with in-laboratory PSG due to invalid results, the relatively low sensitivity and specificity of this method, or potential barriers from human factors 57 .However, overnight PSG performed in accredited sleep units is inappropriate for systematic screening because it is expensive, labour-intensive, not feasible in all subjects and time consuming for professional drivers who are regularly on the road.Therefore, simple, and inexpensive selection of patients with the potentially highest risk of OSA is needed.Another possible limitation would be that the night of the study did not represent a typical night for the patient, for example, due to the "first-night effect", in which there is an alteration of sleep architecture on the first night of PSG studies.This effect can often be observed in studies conducted in sleep laboratories, because of the new environment, discomfort caused by electrodes, being kept under watch by experimenters, or reactions to a novel or stressful situation 58 .To overcome this possible limitation, our studies were carried out at the patient's home, in a familiar environment, and if patients reported difficulties in falling asleep or any other changes in their usual sleep pattern during the night of the study, which could underestimate the result obtained, the study was repeated.Also, a report was sent to all subjects to deliver to their primary care physician to be referred for sleep consultation whether to discuss treatment options or to perform laboratory PSG, to confirm or exclude the diagnosis.It is also important to mention that our study was conducted in a 6-years period and comprised the COVID-19 (disease caused by SARS-CoV-2) pandemic.The pandemic has had a negative impact on the general population, with many individuals reporting sleep disturbances, including difficulties in sleep initiation or maintenance, poor sleep quality, and other sleep impairments 41,59 , which could have affected the individuals reports on the applied questionnaires.The explanation for these changes is likely multifactorial, including social restrictions, fear of infection and financial loss 41 .Regarding the association between OSA and COVID-19, many of the risk factors for OSA are associated with poor COVID-19 outcomes 60 .Patients with OSA often have upregulation of the angiotensin-converter enzyme 2 (ACE2) receptors, which facilitates SARS-CoV-2 virus entry into cells 61 .Also, chronic inflammation status in OSA patients, contributes to the worsening of the cytokine storm in patients with severe forms of COVID-19 62 .Additionally, patients who suffered from COVID-19, particularly severe cases, may be under risk for OSA due to pulmonary fibrosis 60 .Another limitation of our study is that, although our sample was obtained from one of the largest Portuguese road freight companies, the number of participants might not be sufficient to generalize the results.Other studies with a larger cohort are required.
Despite several limitations, our study has important strengths.We used a combination of 3 validated questionnaires, that demonstrated a high PPV, proving to be valuable and simple tools for OSA screening.Our study raises awareness for the significant prevalence of OSA in truck drivers, reinforcing that this pathology should not be neglected, given the associated health and economic burden 11 .Therefore, measures should be taken in order to implement screening tools as part of health policies related to these workers.

Conclusion
The combination of a predominantly male population, obesity, age distribution, and lifestyle could account for the high prevalence of sleep-disordered breathing in truck drivers.Screening, treatment, and management of OSA should be a priority as a public safety policy, along with the benefits for the driver itself.The present study found a significant relationship between indicators of high risk of OSA, obtained by clinical examination, underlining neck circumference evaluation, and questionnaires on OSA symptoms.Most of parameters were obtained after clinical evaluation, which could be performed by occupational medicine.The diagnosis can be confirmed by PSG in workers with high suspicion of OSA.The use of questionnaires proved to be a valuable and simple tool for OSA screening, and in our opinion, this screening should be implemented as an occupational requirement for professional drivers.

Collaborations
M Argel worked on the conceptualization, research, methodology and writing.Â Cunha worked on the methodology and formal analysis.M Conceição worked on the research and data curation.T Abrantes worked on the design and final editing.AS Torres on the conceptualization, supervision and final editing.