Truck collisions attributed to falling asleep at the wheel in two commercial drivers prescribed oral appliance therapy for obstructive sleep apnea

Abstract

Falling asleep at the wheel is attributed to sleepiness, and obstructive sleep apnea is a significant cause of sleepiness that increases the risk of motor vehicle collisions due to falling asleep at the wheel. Although continuous positive airway pressure therapy for obstructive sleep apnea reduces the risk of motor vehicle collisions, similar evidence for alternatives such as oral appliance therapy is lacking. We discuss two truck collisions attributed to microsleep confirmed with dashcam video footage of commercial drivers with obstructive sleep apnea. Our results highlight the current situation where there is insufficient evidence for the prevention and reduction of the risk of motor vehicle collisions by oral appliance therapy, objective adherence monitoring of oral appliance therapy, and effectiveness confirmation tests. Therefore, it is suggested that for commercial truck drivers who require a high level of driving safety, careful selection for oral appliance therapy, systematic follow-up, and monitoring of the driver and truck status with dashcam video footage are crucial.

Citation:

Kumagai H, Tsuda H, Kawaguchi K, et al. Truck collisions attributed to falling asleep at the wheel in two commercial drivers prescribed oral appliance therapy for obstructive sleep apnea. J Clin Sleep Med. 2023;19(12):2117–2122.

INTRODUCTION

Falling asleep at the wheel is attributed to sleepiness, and obstructive sleep apnea (OSA) is one of the significant causes of sleepiness that increases the risk of motor vehicle collisions (MVCs) due to falling asleep at the wheel. The estimated prevalence of OSA in commercial truck drivers is 28%–78%, which is higher than that in the general population (3.7%–34%).^1–4 The risk of MVCs involving untreated patients with OSA is 2.4–3.4 times higher than that in those without OSA.^1,2 Unlike noncommercial drivers, commercial truck drivers often experience difficulty in taking brief breaks when they feel sleepy since their duty involves delivering cargo within a set time frame. Consequently, these professional obligations may force them to continue driving while struggling intensely against sleepiness. Moreover, night-shift truck drivers are at a higher risk of MVCs because of the deviation of their driving time from natural human biological rhythms and the accumulated fatigue.^5,6 Since collisions involving trucks are prone to lead to serious accidents, including fatalities,¹ countermeasures for preventing falling asleep at the wheel by commercial truck drivers are an urgent issue.

The treatment options for OSA include continuous positive airway pressure (CPAP) therapy, oral appliance (OA) therapy (OAT), hypoglossal nerve stimulation, and sleep surgery. OAT is frequently used as an alternative to the gold standard, CPAP therapy, and has been suggested to improve OSA by reducing upper airway collapse, mainly by advancing the mandible. OAT is generally well tolerated and preferred over CPAP despite showing inferior results for most objective parameters.⁶ Moreover, no previous study has clarified whether OAT can reduce the risk of MVCs involving commercial truck drivers or the management criteria of OAT for commercial truck drivers with OSA.

We report two truck collisions caused by falling asleep at the wheel attributable to microsleep that was confirmed by dashcam video footage of commercial truck drivers undergoing OAT for OSA. The aim of this study was to use the insights from these cases to clarify the current problems in OAT for commercial truck drivers.

REPORT OF CASES

The drivers involved in both cases were employed by the same truck company. This company conducted OSA screening tests using a type-4 device for all truck drivers when they entered the company, and the drivers were required to undergo additional detailed examinations if OSA was suspected. After the polysomnographic examination, treatment was initiated if necessary, and the physician judged whether the driver could drive. Treatment selection, eg, the choice between CPAP or OA, was decided by the physician based on the patient’s severity of OSA and preferences. Although the workplace had rules to confirm the management of the drivers’ OSA, the company did not have any CPAP- or OA-specific management strategy for drivers receiving treatment for OSA at the time. Both drivers provided written informed consent for this study. This study was conducted in accordance with the principles of the Declaration of Helsinki, and the study protocol was approved by the Ethical Committee for Epidemiology of Hiroshima University (E-2759, Feb. 4, 2022).

Dashcam video recordings and analysis

Each truck had two dashcams. A dashcam for monitoring vehicle behavior was located on the upper center of the windshield, and another dashcam for monitoring driver behavior was located on the upper-left side of the windshield. Dashcam video footage of the drive and the collisions recorded on 16-Gigabyte Secure Digital memory cards for both collisions were available (DR-9100 and DRT-7100; HORIBA, Kyoto, Japan). The recorded video footage for 3 minutes after the start of driving and 3 minutes before the collision was analyzed using DR Player EX software (HORIBA). All dashcam video footage was reviewed by all authors. The video footage could be zoomed in and out, and the playback could be paused; the footage was visually analyzed for microsleep-related behaviors.

Definition of microsleep-related behaviors

Microsleep-related behaviors were classified into three categories for video footage analysis: “anti-sleepiness behaviors,” “behavioral signs of microsleep,” and “abnormal vehicle behavior.”⁷ Details of the microsleep-related behaviors are shown in Table 1.

Table 1.

Classification and prevalence of microsleep-related behaviors.

Classification	Specific Behavioral Items	Prevalence*
		3 Minutes After Starting		3 Minutes Before Collision
		Case 1	Case 2	Case 1	Case 2
Anti-sleepiness behaviors	Touching face or body	6	4	8	7
	Stretching upper body	0	0	39	0
	Fidgeting of the legs	0	0	0	0
	Groaning, yelling	0	0	9	0
	Yawning	0	0	22	7
	Rapid blinking of the eyes	0	0	0	0
	Hitting face or body	0	0	3	0
Behavioral signs of microsleep	Absence of body movement	0	0	48	20
	Eyes closed halfway or more	0	0	0	0
	Closed eyes	0	0	0	0
Abnormal vehicle behaviors	Inappropriate line crossing	0	0	13	21
	Speed reduction	0	0	21	16

^*Cumulative time for each behavior within the dashcam video footage analysis time.

Case 1

Details regarding the collision

A 53-year-old man (body mass index 25.9 kg/m²) was a commercial truck driver for a transport company working 5 days a week only on night shift, not shift work. The driver was not taking medications, had no smoking habit, and had a habit of drinking approximately 20–40 g of alcohol per day, 5 days a week. The driver averaged 5.5–6 hours of sleep and felt he was getting enough sleep. The driver was involved in a side-impact collision while driving on the highway in October 2022. The driver showed repeated microsleep-related behaviors while driving at approximately 6 am, after which the truck deviated from its lane and collided with the tunnel’s left wall. The collision occurred 12 hours and 48 minutes after the start of driving, during which time the driver rested for a total of 75 minutes on four separate occasions, according to the truck’s digital tachometer. The driver self-reported on the day of the collision that he slept for a normal 5.5 hours the day before the collision.

Course of treatment for OSA

In April 2020, the driver had undergone home sleep apnea testing using a type-4 device (Pulsox-300i; Konica Minolta, Tokyo, Japan) for OSA screening, and his 3% oxygen desaturation index was 19.4 events/h. The Epworth Sleepiness Scale score at that time was 1 point. Type-1 polysomnography (PSG-1100; Nihon Kohden, Tokyo, Japan) performed for the final diagnosis in June 2020 revealed an apnea-hypopnea index of 33.9 events/h, while the apnea-hypopnea index in the nonsupine position was 0.9 events/h. Since the driver had an Epworth Sleepiness Scale score of 3 points at polysomnography testing and had position-dependent OSA, he consulted his physician and opted for OAT. He showed no difference in perceived daytime sleepiness with or without OA despite an average sleep time of 5.5–6 hours. A follow-up sleep test was not conducted to confirm treatment efficacy before the collision. No abnormal findings other than high body mass index were observed on physical examinations conducted in February and August 2022. The driver self-reported on the day of the collision that he began to neglect OAT on his own initiative based on the results of the physical examination from around April 2022. In October 2022, after the collision, he underwent home sleep apnea testing with a type-3 device (Alice NightOne; Philips, Amsterdam, Netherlands) while using OA, and the 3% oxygen desaturation index was 13.2 events/h, which was lower than that before treatment.

Case 2

Details regarding the collision

A 69-year-old man (body mass index 26.8 kg/m²), who worked for the same transport company as the driver in case 1, was a commercial truck driver working only daytime shifts 5 days a week. The driver was a nonsmoker, was taking no medication, and had a habit of drinking approximately 20 g of alcohol per day, 3 days a week. The driver averaged 7.5 hours of sleep and felt he was getting enough sleep. The driver was involved in a side-impact collision while driving on the highway in December 2022. Although the driver had no previous collisions, he drove drowsy and deviated from his lane at approximately 2:30 pm, after which the vehicle collided with the median poles on the right side of the truck. The collision occurred 7 hours and 38 minutes after the start of driving, during which time the driver took a 1-hour break, as confirmed by the truck’s digital tachometer. The driver self-reported on the day of the collision that he slept for 8.5 hours the day before the collision.

Course of treatment for OSA

In January 2019, the driver underwent home sleep apnea testing using a type-4 device (Pulsox-300i; Konica Minolta), and this test showed a 3% oxygen desaturation index of 8.8 events/h. The Epworth Sleepiness Scale score at that time was 1 point and was not reexamined thereafter. Subsequently, type-1 polysomnography (SOMNOscreen BT-plus, Fukuda, Tokyo, Japan) was conducted using the same polysomnography indicators as in case 1 in April 2019, and the result was an apnea-hypopnea index of 20.9 events/h. Based on these findings, OAT was initiated. However, as in case 1, a follow-up sleep study was not performed. The driver’s average sleep time was approximately 7.5 hours, and he was aware that he had a good night’s sleep after starting OAT. When interviewed on the day of the collision, the driver self-reported that he used OA every night when sleeping, including the night before the collision.

Validation of treatment steps for OA in cases 1 and 2

The American Academy of Sleep Medicine guidelines list six important steps in the management of OAT: (1) at diagnosis, (2) at the start of OAT, (3) OA adjustment, (4) OAT monitoring, (5) follow-up sleep testing, and (6) periodic hospital visits.⁸ Considering each of these six steps in the two cases, we found that three and four steps out of six steps had not been followed in cases 1 and 2, respectively. At the time of diagnosis, OAT was selected after thorough consultation between the treating physician and patient, based on the indications for CPAP in both cases. At the start of OAT, noncustom OAs were selected for both patients because custom OAs were not covered by medical insurance in Japan. OA adjustments were made as appropriate at the dentist’s discretion and were not made in case 1 but were made once in case 2. Neither titratable OA nor oversight by qualified dentists was selected for OAT monitoring. Follow-up sleep testing was done after collision only in case 1. Periodic hospital visits were initiated after collision in case 1 but not in case 2 (see Table S1 ^{(73.7KB, pdf)} in the supplemental material).

Dashcam video footage analysis of microsleep-related behaviors in cases 1 and 2

Dashcam video footage analysis revealed that both collisions were attributable to microsleep. In both cases, dashcam video footage immediately before the collision showed microsleep-related behaviors characterized as behavioral signs of microsleep such as the absence of body movements and as abnormal vehicle behaviors such as lane departure (Table 1). Moreover, the frequency of microsleep-related behaviors was confirmed to be higher in the 3-minute period before the collision than those in the 3-minute period after the start of driving (Figure 1 and Figure 2).

Figure 1. Dashcam video images and frequency of microsleep-related behaviors in case 1.

Case 1: (A) Dashcam video image immediately before the collision. (1) Dashcam video image for outside the vehicle showed that the vehicle was approaching the left wall of the tunnel with the left lane departure. (2) Dashcam video image for the inside of the vehicle showed that the driver’s upper body muscle tone had mildly decreased and body movement was absent. (B) Appearance of microsleep-related behaviors by analyzing every second of interior dashcam video footage for 3 minutes after the start of driving. (C) Appearance of microsleep-related behaviors by analyzing every second of exterior dashcam video footage for 3 minutes before the collision. The frequency of microsleep-related behaviors for 3 minutes before the collision was clearly greater than the frequency observed immediately after the start of driving. The horizontal axes of (B) and (C) represent the time in seconds (s) from 0 to 180 seconds immediately after the start of driving in (B) and from 180 to 0 seconds before the collision in (C). The vertical axes of (B) and (C) represent the number of occurrences (n) of each behavior.

Figure 2. Dashcam video images and frequency of microsleep-related behaviors in case 2.

Case 2: (A) Dashcam video footage recorded immediately before the collision. (1) Dashcam video footage from outside the vehicle showed that the vehicle was approaching the median poles with the right lane departure. (2) Dashcam video footage from inside the vehicle showed that the driver’s eyes were closed and body movements were absent. (B) Appearance of microsleep-related behaviors by analyzing every second of interior dashcam video footage for 3 minutes after the start of driving. (C) Appearance of microsleep-related behaviors by analyzing every second of exterior dashcam video footage for 3 minutes before the collision. The frequency of microsleep-related behaviors for 3 minutes before the collision was clearly greater than that immediately after the start of driving. The horizontal axes of (B) and (C) represent the time in seconds (s) from 0 to 180 seconds immediately after the start of driving in (B) and from 180 to 0 seconds before the collision in (C). The vertical axes of (B) and (C) represent the number of occurrences (n) of each behavior.

DISCUSSION

Through an analysis of dashcam video footage, two truck collisions involving commercial drivers in a transport company who were receiving OAT were objectively confirmed to be attributable to falling asleep at the wheel caused by microsleep. To prevent truck collisions caused by falling asleep at the wheel in commercial drivers, we would like to emphasize the importance of careful consideration before the administration of OAT for commercial truck drivers because of the insufficient evidence for the prevention of MVCs by OAT, the difficulties in objective monitoring of OAT adherence, and the high risk of residual obstructive events with OA in comparison with CPAP. Furthermore, the findings suggested that dashcam video footage-based monitoring of driver status is also crucial for objective assessment of collision causes.

OSA is usually treated with CPAP or OA. However, unlike the gold-standard CPAP therapy, which allows objective monitoring of effectiveness and adherence, OAT shows many limitations.⁹ CPAP therapy has been shown to effectively reduce the risk of MVCs and is the only treatment option recommended for commercial truck drivers with OSA.^1,2 In contrast, data on the effectiveness of OAT in ensuring safe driving are insufficient; therefore, OAT is not recommended for patients with an apnea-hypopnea index ≥ 20 events/h or those with severe clinical symptoms.¹ Future studies should aim to evaluate the ability of OAT combined with or without CPAP to reduce MVC risk in drivers with OSA.² Therefore, OAT should be used with caution in commercial truck drivers at high risk for MVCs since there is insufficient evidence for the effectiveness of OAT in reducing the risk of MVCs.

In addition to the lack of clear evidence for the efficacy of OAT in reducing the risk of MVCs, two more challenges are associated with this treatment modification. First, a method for confirming OA adherence is not used in all OAT cases, especially in Japan, even though adherence monitoring is possible, and second, sleep testing under OAT is required to confirm treatment efficacy. As an example of the first problem, nonadherence to OAT could not be objectively identified despite the presence of severe OSA in case 1. The fact that the driver had been essentially driving for 6 months without receiving treatment for OSA may have been responsible for the collision attributed to falling asleep at the wheel. Although objective confirmation of OA adherence with a small temperature sensor is recommended by the American Academy of Sleep Medicine,¹⁰ it is not currently used in clinical settings in Japan as it is not covered by medical insurance. The establishment of a method to confirm OA adherence regularly and objectively, similar to the approach in CPAP therapy, is a problem that needs to be resolved urgently for commercial truck drivers. The second problem is that, in both cases, the effects of OAT in improving the severity of OSA were not confirmed. OSA improvement status cannot be confirmed by OAT itself. In case 1, OAT was selected because the patient had position-dependent OSA. However, for such cases, one option would be to introduce CPAP therapy and then switch to OAT if the driver is intolerant, while another would be to continue OAT after confirming that the severity of OSA had been reduced to an acceptable level after initiating OAT. In case 2, a collision attributable to falling asleep at the wheel occurred despite the driver sleeping while wearing an OA every day. This problem can be addressed by follow-up sleep testing of patients prescribed an OA to improve and confirm the effectiveness of the treatment.^1,8 We believe that sleep testing to confirm OAT effectiveness is more important for commercial truck drivers than for the general population. There are OAs for which adherence monitoring is possible, and objective and systematic OAT monitoring, including follow-up by physicians and dentists, is feasible. Therefore, it is desirable that in countries such as Japan, where objective adherence monitoring for OAT is not covered by medical insurance, it should be covered as soon as possible and treatment guidelines for systematic follow-up, including regular follow-up by physicians and dentists, should be established. We believe that an objective adherence monitoring system for OAT should be established as a standard clinical practice worldwide, including confirmation of residual excessive daytime sleepiness, which is a common clinical problem with both CPAP and OAT. In addition, because drivers experience difficulty in self-awareness of the fact that they have slept despite the presence of microsleep, objective confirmation of collision causes by dashcam video footage may be necessary regardless of the cause of dozing.

This study has certain strengths and limitations. The behavior of the driver and the truck that caused the collision could be analyzed simultaneously using the dashcam video footage, and it was possible to clarify that the collision occurred due to microsleep. In our previous study, microsleep-related behaviors occurred more frequently in the 1 minute before a collision,⁷ and in the present study, the frequency of microsleep-related behaviors was clearly increased in the 3 minutes before the collision, compared to the 3 minutes after the start of driving. This suggests that microsleep-related behaviors may be a predictor of collisions attributed to falling asleep at the wheel. However, future validation studies should be conducted to determine which of the three categories of microsleep-related behaviors—anti-sleepiness behaviors, behavioral signs of microsleep, and abnormal vehicle behavior—is the best predictor of collisions attributed to falling asleep at the wheel and whether anti-sleepiness behavior is a specific behavior before the collision. It may be necessary to modify and evaluate the existing concept of sleepiness-related behaviors at the wheel. Moreover, it is difficult to conclude that OSA was the only cause of the collision in these two cases. However, in both cases, since the drivers reported that they usually slept well in the interview conducted immediately after the collision, and the Epworth Sleepiness Scale scores taken before the collision were both 1 point, the most likely cause was considered to be the unconscious accumulation of mild sleep deprivation due to moderate OSA. Generally, truck drivers tend to work long hours and often find it difficult to take breaks according to their convenience. Therefore, not only OSA but also chronic sleep insufficiency and chronic fatigue can be considered causes of falling asleep at the wheel, and this may be a common problem for all commercial truck drivers, not just the two cases described in this report. Although the potentially limited generalizability of the findings of these case reports cannot be ruled out, some drivers may still be receiving an OA with inadequate OA management, as in these cases. Thus, follow-up or large-scale studies of the clinical issues in OAT demonstrated in this study will need to be conducted in the future.

In summary, we emphasize that, for commercial truck drivers, who require a high level of driving safety, systematic objective monitoring such as confirmation of the effectiveness of OAT for OSA and regular monitoring of OA adherence are essential compared to general drivers. Furthermore, as truck collisions cause far greater damage than passenger vehicle collisions, it is important that physicians and dentists who treat OSA in commercial truck drivers are made aware of the points to note when choosing OAT. Considering the current situation, where the effect of OAT in reducing the risk of MVC is unclear, we propose that careful selection of patients who receive OA, close follow-up, and monitoring of the driver and truck status with dashcam video footage are crucial for commercial truck drivers.

DISCLOSURE STATEMENT

All authors have read and approved the manuscript. This study was conducted at the Department of Sleep Medicine, Graduate School of Biomedical and Health Sciences, Hiroshima University, and the Sleep Disorders Center, Hiroshima University Hospital, Hiroshima. Hajime Kumagai, Hiroko Tsuda, Kengo Kawaguchi, Yuka Kiyohara, Noriyuki Konishi, and Toshiaki Shiomi are affiliated with the Department of Sleep Medicine established by the donation of Fukuyama Express Co., Ltd. The authors report no conflicts of interest.

ABBREVIATIONS

CPAP

continuous positive airway pressure

MVC

motor vehicle collision

OA

oral appliance

OAT

oral appliance therapy

OSA

obstructive sleep apnea