Self‐reported automobile accidents involving patients with obstructive sleep apnea
Citation Manager Formats
Make Comment
See Comments

Abstract
We developed a routine survey instrument, which included data on self-reported motor vehicle accidents (MVA), among 253 patients who attended the UCLA Sleep Disorders Center to evaluate whether patients with diagnosed sleep apnea syndrome (SAS) have a higher ratio of MVAs. We used unconditional multiple logistic regression to determine the odds ratio between MVA and SAS, adjusted for the following covariates: age, sex, work shift, daytime nap, alcohol and coffee intake, and history of neurologic diseases. Thirty-one percent of patients with SAS compared with 15% of patients without SAS reported at least one MVA (p <0.01). The adjusted odds ratio found through logistic multiple regression analysis was 2.99 (p <0.01). The results demonstrate that patients with SAS had a higher ratio of self-reported MVA than did those without SAS and that the characteristics of falling asleep at inappropriate times and driving past destinations were two good indicators of probability of having an MVA. Nevertheless, sleep apnea is a very treatable condition; once treated the risk factors diminish considerably.
Automobile accidents are the third leading cause of death and injury in the United States, accounting for 40 to 50 thousand deaths and 2 million accidents per year. l Sleep loss and sleep disturbance are thought to play a major role in automobile accidents. 2 Among patients with sleep apnea syndrome (SAS), numerous studies found an excessive tendency to fall asleep while driving, which could increase their risk of having a motor vehicle accident (MVA), especially if they have untreated SAS. 3,4
Several studies demonstrated that patients with SAS compared with non-SAS patients had greater automobile accidents that resulted in serious injuries or deaths; 4 tended to have worse performance on driving simulation tests, particularly on both highway and city or rural driving; and hit a greater number of obstacles during their simulated drive. 5 In another study, patients with SAS compared with non-SAS patients had poorer brake reaction time, lateral position deviation, and off-road episodes when tested with an advanced driving simulator during a 90-minute rural drive at twilight conditions. 6 However, in both studies, the authors found that the driving performance of patients improved once they were treated with continuous positive airway pressure (CPAP) 5 or uvulopalatopharyngoplasty. 6
Other studies also demonstrated that individuals with SAS may not compensate for their condition by driving more carefully. Haraldsson et a1. 8 conducted two studies to determine the impact of SAS on driving. In the first study, they surveyed 140 patients and 142 control subjects without symptoms associated with SAS. They found that 73 patients had a complete triad of SAS-associated symptoms (heavy snoring, sleep disturbances, and daytime somnolence) and that 52% of patients reported habitual sleeping spells while driving compared with less than 1% of the control subjects. 7 In the second study, the authors used data from a national random sample of 1,214 persons in Sweden and found that diurnal hypersomnia was only prevalent among 2.2% of respondents. 8
Furthermore, studies conducted in Canada, 9 the University of Virginia Sleep Disorders Center, 10 and the Sleep Disorder Center in Michiganll demonstrated that patients with diagnosed SAS compared with control subjects experienced a greater number of accidents (ranging from two times up to a seven-fold increase) and also had a greater rate of more than one accident compared with control subjects. 9–11
In the present cross-sectional study, we surveyed 253 subjects to assess whether patients with SAS have a higher ratio of self-reported MVAs than patients without SAS to determine what characteristics distinguish those who have MVAs from those who do not have MVAs.
Methods
Selection of subjects
All subjects who had a nocturnal polysomnography test (NPSG) at the Sleep Disorders Center at the UCLA Medical Center in 1993 for complaints of having sleep problems, loud snoring, daytime tiredness, or breathing problems were selected. From this initial sample of 319 patients, we excluded 24 patients who did not drive, resulting in 295 eligible patients.
Procedures
We gave a questionnaire to 295 eligible subjects before their NPSG test that had three questions related to their driving: Have you had driving accidents or near-accidents due to sleepiness? Have you driven miles past your destination with little or no awareness? and Have you fallen asleep at inappropriate times, particularly while driving? In addition, we asked questions about work shift, alcoholic beverage intake (more than one beverage per day), history of episodes of severe dizziness, Parkinson's disease, seizures, epilepsy, or loss of consciousness, which are related to increased risk of having a car accident. We also included questions on regular daytime naps and coffee intake (more than one cup per day and/or drinking coffee to maintain wakefulness) because some studies found that some SAS patients can maintain wakefulness under motivating conditions. 11 All variables were adjusted in the analyses conducted to determine the impact of SAS on MVAs.
In addition to the questionnaire, we conducted the NPSG test to diagnose SAS in these patients. The NPSG test was monitored with EEGs, electroocculograms, and submental EMG using standard techniques. Total sleeping time, sleep stages, number and duration of obstructive, central or mixed apnea, and total arousal and awakening time were calculated for each patient. The respiratory disturbance index (RDI) (i.e., the number of respiratory disturbances producing arousals per hour of sleep) and the apnea or hypopnea index (AHI) (i.e., the number of apnea or hypopnea episodes lasting ≥ 10 seconds causing 4% oxygen desaturations) were considered for the estimation of sleep apnea. Scores of >5 were considered normal for both RDI and AHI. Subjects with an RDI or AHI >5 were classified as having SAS; the others were classified as non-SAS patients. Additionally, we relied on clinical evaluation by physicians in the Sleep Disorders Center of the UCLA Medical Center for the final diagnosis of SAS.
Statistical analyses
We used numerous tests to compare the groups and to assess the impact of the numerous variables on MVAs. In this cross-sectional study, the independent variable was the diagnosis of SAS and the dependent variable was either having or almost having an MVA due to sleepiness. We used chi-square tests to compare both groups on the various variables. We calculated the odds ratio (OR) of having or almost having an MVA in SAS patients relative to non-SAS patients by controlling for the following mediators: age, sex, night work shift, daytime nap, coffee intake, and alcoholic beverage intake. We used EPI INFO 612 for these analyses. To examine whether diagnosis of SAS was associated with WAS, we calculated the Mantel Haenszel OR (ORM-H) to estimate the effect of SAS (SAS patient or non-SAS patient) adjusted for various predictors. We conducted unconditional multiple logistic regression analyses to evaluate the adjusted ORs of having an MVA after controlling for the presence of multiple covariates simultaneously. All logistic regression procedures were tested using the computer program Proc Logistic of the Statistical Analysis Systems. 13 We used 95% CIS and two-tailed p values.
Results
From the original 295 that were eligible for the study, we received 253 surveys for a response rate of 85.8%. Among these 253 subjects, most were diagnosed with sleep apnea (68%), were men (71%), and were mainly between the ages of 20 to 64 years of age (83%). As noticed in table 1, when the subjects were stratified according to diagnosis of SAS, we found that both the SAS patients and the non-SAS patients were very similar, although SAS patients were mostly men and had a significantly higher proportion of accidents than the control subjects. In addition, SAS patients evidenced a higher proportion of snoring and reports of falling asleep at inappropriate times.
Table 1 Demographic and other behavioral characteristics of SAS and non-SAS patients
We analyzed groups based on whether they self-reported having an accident. Sixty-six subjects (26%) reported having had at least one accident. Among those who reported at least one car accident, 52 (82%) were diagnosed with sleep apnea. Among individuals who reported either having or almost having an accident, sleep apnea, history of severe dizziness (D.), Parkinson's disease (P.), seizures/epilepsy (S.), and loss of consciousness (L.), report of passing destination with little awareness, and falling asleep at inappropriate times were significantly higher among this group than the nonaccident group. The crude OR of having an accident among those with sleep apnea was 2.57 (95% CI = 1.22 to 5.51, p <0.0063).
We determined the ORM-H of having a motor vehicle accident in SAS patients relative to those without SAS adjusted for the following covariates: sex, age, shift worker, alcohol intake, coffee intake, daytime nap, history of D.P.S.L., history of passing destination, and falling asleep at inappropriate times. There was an increased ratio of self-reported MVA among patients with polysomnographically confirmed SAS, even after controlling for the different covariates. All variables with the exception of falling asleep at inappropriate times put patients compared with control subjects at relatively increased risk of having an MVA. A history of these variables increased the risk of MVAs among patients compared with control subjects from 2.41 to 2.73 (CI from 1.18 to 5.59).
Table 2 presents unconditional multiple regression analyses that evaluate the adjusted ORs of having an MVA after controlling for the presence of multiple covariates simultaneously. We note that three factors are significantly related to an individual's history of self-reported MVAs, which include history of passing destination, falling asleep at inappropriate times, and diagnosis of SAS. We also note that alcohol intake and history of other medical conditions increase the risk, whereas coffee intake diminishes the risk of WAS in these individuals. Additional multiple logistic regression analyses were conducted using various models to ascertain the impact that different factors would have on the overall OR. The results of those analyses were similar to those demonstrated in table 2.
Table 2 Multiple logistic regression analysis of the association between MVA and patients' characteristics
Discussion
We found a positive relationship between diagnosed SAS and self-reported MVAs among our population of sleep apnea patients. Among SAS patients, there was a greater proportion of men, individuals reporting greater frequency of snoring, and falling asleep at inappropriate times. The literature has found that men generally have a higher risk of car accidents than do women. Even though in our study there was a higher proportion of men than women in the patient population, the OR of having an MVA among SAS patients, after controlling for sex, was still significantly high (ORM-H = 2.58, 95% CI = 1.29 to 5.17). Patients with SAS also had a higher proportion of reports of falling asleep at inappropriate times than did the non-SAS patients, which has been found to be associated with sleep apnea in other studies. 3,14 Furthermore, similar to other studies, 7 we found that a greater proportion of our patient population reported snoring (OR = 4.17, 95% CI = 1.98 to 8.86). This is important considering that snoring is one of the symptoms of SAS.
The OR between answering affirmatively that they have passed a destination with little awareness and having an MVA was 7.16, whereas the OR for answering positively that they fell asleep at inappropriate times and that they had or almost had an MVA was 5.28. Using multiple logistic regression analysis, we found that even when we adjusted these variables, the adjusted ORs were 5.72 and 4.69 (table 2). This suggests to us that even in the absence of a clear diagnosis of SAS, the characteristics of falling asleep at inappropriate times and passing destinations with little awareness of doing so are good indicators of increased risk of probably having an MVA.
This study demonstrated that patients with polysomnographically confirmed SAS had an increased ratio of self-reported MVAs even after controlling for numerous covariates. As noted in table 2, even after controlling for various factors simultaneously, the risk of having an MVA was 2.58 among those with a diagnosis of SAS. These results suggest an association between sleep apnea and self-reported MVAs and confirms results found previously by others. For example, Findley et al. 10 reported that the percentage of car accidents among 29 patients with confirmed SAS was greater than that of 35 patients without confirmed SAS (31% versus 6%, p <0.01). In addition, a study conducted in Canada found that 93% of 27 patients with SAS and 53% of control subjects matched by age and sex reported MVAS. 9
The logistic regression analysis also demonstrated that alcohol intake enhances the possibility of MVAs for drivers suffering from SAS. This could be because ethanol could reduce sleep latency and attention performance. 15 Thus, patients with SAS should be more particular about avoiding alcohol before driving and should allow a greater interval between alcohol consumption and driving. We also found that drinking coffee had a reverse effect, allowing patients to stay more alert while driving (OR = 0.43, CI 0.21 to 0.85, p <0.01).
As with any self-reported surveys, there are problems with using a self-administered questionnaire to determine MVAs. Subjects may not recall almost having a car accident or may fail to report an accident. In addition, verifying the information with the Department of Motor Vehicles (DMV) presents a bias because not all accidents are reported to the DMV and also because the DMV does not record who is at fault. Other potentially confounding variables such as driving mileage, duration of driving, weather conditions? and so on should also be controlled when analyzing the relationship between SAS and MVAs. We, however, only used self-reported data on driving accidents or near accidents due to sleepiness so that we could more reasonably analyze the association between SAS and MVA.
The interpretation of the data is also limited by the cross-sectional nature of the survey and also by our inclusion of a clinical population. The findings, however, demonstrate that patients with SAS have a significantly higher ratio of self-reported MVAs than did control subjects, that alcohol intake enhances the possibility of having an MVA among these patients, and that falling asleep at inappropriate times and driving past destination with little awareness in patients with SAS were good indicators of increased risk of having a motor vehicle accident. Nevertheless, sleep apnea is a very treatable condition, and once treated, the risk factors diminish considerably.
Acknowledgments
We thank Dr. Jess F. Kraus for providing guidance to the first author on the statistical analyses and to Ms. Rosa Elena Garcia for clerical assistance.
Footnotes
-
Received June 7, 1995. Accepted in final form November 1, 1995.
- Copyright 1996 by the American Academy of Neurology
References
- 1.↵
U.S. Bureau of the Census Statistical Abstract of the United States: 1990, 114th ed. Washington, DC: U.S. Department of Commerce, 1994.
- 2.↵
- 3.↵
Parsons M. Fits and other causes of loss of consciousness while driving. Q J Med 1986;227:295–303.
- 4.↵
Findley W, Weiss JW, Jabour ER. Drivers with untreated sleep apnea. A cause of death and serious injury. Arch Intern Med 1991;157:1451–1452.
- 5.↵
- 6.↵
Haraldsson PO, Carenfelt C, Peason HE, et al. Simulated long-term driving performance before and after uvulopalatopharyngoplasty. J Otorhinolaryngol Relat Spec 1991;53:106–110.
- 7.↵
Haraldsson PO, Carenfelt C, Tingvall C. Sleep apnea syndrome symptoms and automobile driving in a general population. J Clin Epidemiol 1992;45:821–825.
- 8.↵
Haraldsson PO, Carenfelt C, Diderichsen F, et al. Clinical symptoms of sleep apnea syndrome and automobile accidents. J Otorhinolaryngol Relat Spec 1990;52:57–62.
- 9.↵
George CF, Nickerson PW, Hanly PJ, et al. Sleep apnea patients have more automobile accidents [letter]. Lancet 1987;2(8556):447.
- 10.↵
Findley W, Unverzagt ME, Suratt PM. Automobile accidents involving patients with obstructive sleep apnea. Am Rev Respir Dis 1988;138(2):337–340.
- 11.↵
- 12.↵
Dean AG, Dean JA, Burton AH, Dickes RC. Epi Info Version 6: a word processing, database and statistics program for epidemiology on microcomputers. Stone Mountain, Georgia: USD Incorporated, 1990.
- 13.↵
SASISTAT user's guide, version 6. Ed. Cary, NC: SAS Institute Inc., 1990.
- 14.
Leger D. The cost of sleep-related accidents: a report of the National Commission on Sleep Disorders Research. Sleep 1994;17:84–93.
- 15.↵
Roehrs T, Zwyghuizen-Doorenbos A, Timms V, et al. Sleep extension, enhanced alertness and the sedating effect of ethanol. Pharmacol Biochem Behav 1989;34:321–324.
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Dr. Nicole Sur and Dr. Mausaminben Hathidara
► Watch
Related Articles
- No related articles found.
Alert Me
Recommended articles
-
Articles
Sleep-disordered breathing in multiple sclerosisTiffany J. Braley, Benjamin M. Segal, Ronald D. Chervin et al.Neurology, August 15, 2012 -
Review
Treatment options for obstructive sleep apneaPatrick Pavwoski, Anita Valanju Shelgikar et al.Neurology: Clinical Practice, November 10, 2016 -
Articles
Platelet activation, epinephrine, and blood pressure in obstructive sleep apnea syndromeI. Eisensehr, B. L. Ehrenberg, S. Noachtar et al.Neurology, July 01, 1998 -
Clinical and Ethical Challenges
Early diagnosis and treatment of obstructive sleep apnea after strokeAre we neglecting a modifiable stroke risk factor?Arielle P. Davis, Martha E. Billings, W.T. Longstreth, Jr et al.Neurology: Clinical Practice, June 10, 2013