INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: snoring; sleepiness; occupational accidents; epidemiology; population-based

Occupational accidents are a major social problem. In the United States alone, more than 10 people per day die from injuries on the job and fatal and nonfatal occupational injuries cost more than $100 billion annually in lost wages and productivity, administrative expenses, health care, and other costs (1). Several variables affecting worker arousal and attentiveness are known to be associated with increased risk of occupational accidents. Alcohol, medications, and shift work have all been implicated as contributors to occupational injuries (2).

Knowledge about the influence of sleep disturbances on occupational accidents is, however, still limited. Snoring and excessive daytime sleepiness (EDS) are the main symptoms of obstructive sleep apnea syndrome (OSAS), which is one of the most frequent sleep disorders affecting 4% of males and 2% of females in a middle-aged, working population (8). Patients with OSAS are involved in traffic accidents two to three times more often than the general population (9, 10). In simulated driving tests some patients with OSAS have been found to perform even worse than control subjects impaired by alcohol (11). Ulfberg and coworkers found that sleepiness at work was reported much more frequently both by patients with OSAS and by heavy snorers compared with referents (12). In a 10-yr retrospective analysis they further reported that involvement in occupational accidents was more frequent among patients referred because of suspected OSAS than subjects from the general population (13).

The aim of this study was to prospectively analyze whether men with reported snoring and EDS run an increased risk of being involved in occupational accidents.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Population

In 1984 a random sample of men aged 30-69 yr and living in Uppsala, Sweden was sent a postal questionnaire developed to analyze the associations between sleep disturbances and medical disorders (14, 15) and to estimate the prevalence of OSAS (16). As Swedish citizens become retired at age 65 yr only men aged 30-64 yr at baseline were included in this analysis. From a random sample of 3,720 men in this age interval, a total of 2,874 men participated (response rate 77.3%) in 1984. Additional 53 subjects had not answered the question on snoring or the one on sleepiness in the 1984 questionnaire and were thereby excluded from the follow-up. Ten years later 150 of the men had died. Of the remaining 2,724 males, 2,451 (90.0%) responded to a new postal questionnaire in 1994. Of those who responded at the follow-up, 2,009 (82.0%) had responded to the work-related questions and were used as the study population. Complete data were hence obtained from 54.0% of the original sample and from 57.5% of those who were still alive.

Questionnaires

The 24 questions used in the 1984 questionnaire concerned topics such as snoring, sleep disturbances, associated daytime symptoms, and medical disorders. In addition, weight and height were asked for and body mass index (BMI) was calculated. For questions on symptoms related to sleep disturbances the subjects were asked how often they experienced specific symptoms and the answers were given on a five-point scale: 1 = never, 2 = seldom, 3 = sometimes, 4 = often, and 5 = very often. To assess snoring, the subjects were asked to state how often they snored loudly and disturbingly. Those who answered 3-5 were regarded as snorers. Daytime sleepiness was assessed with the question "How often do you fall asleep involuntarily for a short period during the day, for example when there is a pause at work?" Excessive daytime sleepiness (EDS) was defined as answering 3-5. The approach applied to categorize the men according to snoring and sleepiness is shown in Figure 1.

View larger version (21K): [in this window] [in a new window]   Figure 1.   Approach applied to categorize the population when analyzing the risk factors for occupational accidents. The proportion of subjects within each circle represents those who reported snoring and EDS, respectively, in the 1984 questionnaire. Subjects without snoring and EDS were used as a reference group.

The questionnaire used in 1994 consisted of 71 questions, with the first 24 identical to those used in 1984. In addition, there were questions on smoking, alcohol consumption, and questions related to work during the past 10 yr.

Smoking habits were assessed by six questions and from the answers it was possible to establish whether the subject had been a smoker in 1984 (17). Alcohol use was investigated using the CAGE questionnaire (18, 19). The subjects who answered "yes" to at least two of the four questions in the CAGE questionnaire were defined as being alcohol dependent. Change in BMI during the 10-yr period (BMI) was calculated as BMI 1994BMI 1984.

The participants were asked how many years they had worked during the past 10 yr and their main employment during that period. They were further asked to state for how many months during the past 10 yr they had had night or shift work, and for how many months they had been occupationally exposed to noise, organic solvents, exhaust fumes, and whole-body vibrations.

In the analyses, data on snoring, EDS, and BMI were calculated from 1984 questionnaire responses. Whether the subject was a smoker in 1984 was calculated from the information given in the 1994 questionnaire, from which data on alcohol consumption and work-related information also were obtained.

Occupations and Occupational Accidents

Based on the questionnaire responses on occupation in 1994, occupations for all subjects were categorized as either blue- or white-collar worker using the Nordic Classification of Occupation (NYK 82), which is based on the International Standard Classification of Occupations (ISCO), with a few modifications (20).

Data on occupational accidents during the period (January 1, 1985-December 31, 1994) were obtained from the Occupational Injury Statistics Division of the Swedish National Board of Occupational Safety and Health. This is a national register where the Injury Statistic Division compiles statistics on all occupational accidents and diseases in Sweden, including injuries not requiring sick leave. Employer reporting of occupational accidents requiring one or more day's sick leave is mandatory. The injury report form includes information on occupation, the enterprise, working conditions, circumstances of the injury, and number of days of absence from work. Injuries resulting from falls, equipment and machine-related injuries, injuries due to falling or flying objects, etc. were included in the analysis. Injuries due to overexertion and accidents occurring during travel to or from work were excluded in the present analysis.

The informed consent of all participants was obtained and the study was approved by the Ethics Committee of the Medical Faculty at Uppsala University. Potential participants were informed in the letter accompanying the mailed questionnaire that the study would make use of their personal occupational injury data available in the Swedish national register.

Statistical Analysis

The computations were performed using the StatView 5.0 software package (SAS Institute Inc, Cary, NC). The chi-square test was used to test for differences between proportions and the Mann-Whitney U test was used when the comparison involved continuous variables. To study the influence of several possible explanatory variables on occupational accidents, multiple logistic regression was performed and the results are presented as adjusted odds ratios (OR) with 95% confidence intervals (95% CI). The null hypothesis was rejected at a level of p < 0.05. The Kaplan-Meier survivorship function was used to study the cumulative risk of being involved in an occupational accident during the follow-up period. Adjustment for confounders was then performed using the Cox proportional hazards model. Subjects who had been working for less than 10 yr and who had not been involved in any accidents were treated as censored at the time they quit the job.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Responders, Nonresponders

Of the 2,009 subjects who responded to the questionnaires both in 1984 and in 1994, 766 (38.1%) were snorers without EDS in 1984 according to the criteria described above. EDS without snoring was reported by 85 (4.2%) while 108 (5.4%) suffered from the combination of snoring and EDS (Figure 1).

In the questionnaire used in 1994, 96.1% of the population answered the question on smoking while the response rate for the questions on alcohol was 96.9%.

Compared with the responders in 1994, the 715 nonresponders were older (45.1 ± 11.2 versus 42.7 ± 9.1 yr, p < 0.001) and had a somewhat higher BMI in 1984 (24.7 ± 3.1 versus 24.3 ± 2.8 kg/m², p < 0.05). Nonresponders had further reported the combination of snoring and EDS in 1984 more often (8.8 versus 5.4%, p < 0.01) while the prevalence of snoring without EDS and EDS without snoring did not significantly differ between the groups.

Occupational Accidents

During the 10-yr period, a total of 345 occupational accidents had been reported by 247 of the men (12.3%). Forty of the men had reported two accidents, 15 had reported three, and 7 subjects had reported four or more accidents. The most frequently reported accident was falling either on the same level (n = 45) or to a lower level (n = 36). Sick leave was a consequence in 164 of the accidents for a mean time of 44 ± 103 d (median 11, range 1-938 d).

Characteristics of the total population and differences between those who had or had not been involved in at least one occupational accident are presented in Table 1. The prevalence of blue-collar workers was strikingly higher in the group involved in accidents. Exposures to noise, organic solvents, exhaust fumes, and whole-body vibration were more common in those involved in accidents while the prevalence of night and shift work did not significantly differ between the groups. Furthermore, a higher BMI was correlated with occupational accidents, while smoking, weight gain, and alcohol dependence was not.

Of the 286 men involved in occupational accidents, 52.4% reported snoring in 1984 compared with 42.8% of those with no occupational accident (p < 0.01). However, when subdividing the snorers into sleepy and nonsleepy, only snorers with sleepiness were significantly overrepresented in the accident group (Table 1).

Blue- and White-collar Workers

Occupational accidents were more common in blue-collar than in white-collar workers. Compared with the 1,272 white-collar workers, the 737 blue-collar workers were somewhat older (mean age 43.5 ± 9.2 versus 42.2 ± 9.0, p < 0.01) and had a higher BMI in 1984 (24.9 ± 2.8 versus 24.0 ± 2.7, p < 0.0001). The blue-collar workers were more often smokers (41 versus 30%, p < 0.0001) and alcohol dependent (13 versus 10%, p = 0.055). Shift work was more prevalent among the blue-collar workers (10 ± 29 versus 5 ± 21 mo, p < 0.001), and they also reported longer periods with exposure to noise (30 ± 44 versus 5 ± 20, p < 0.0001), organic solvents (10 ± 29 versus 2.2 ± 13 mo, p < 0.0001), exhaust fumes (25 ± 42 versus 4 ± 19 mo, p < 0.0001), and whole-body vibration (10 ± 28 versus 1.4 ± 11, p < 0.0001). In 1984, 7.7% of the blue-collar workers and 4.0% of the white-collar workers (p < 0.05) reported the combination of snoring and EDS. The prevalence of snoring without EDS was also more common among blue-collar workers (42 versus 36%, p < 0.05) while the prevalence of EDS without snoring was similar in both groups (4.3% versus 4.2%, n.s.).

Both among blue- and white-collar workers, snorers with EDS were involved in accidents more often than their nonsnoring, nonsleepy counterparts (Figure 2), although the difference in number of accidents did not reach significance within the white-collar group.

View larger version (27K): [in this window] [in a new window]   Figure 2.   Mean number of occupational accidents per year of employment (± SE) during the 10-yr period in white- and blue-collar workers as a function of snoring and EDS at baseline in 1984.

Multiple Logistic Regression

To assess the contribution of various independent factors to the risk of being involved in at least one occupational accident during the 10-yr period, a multiple logistic regression analysis was performed. Subjects who were sleepy snorers at baseline had an adjusted OR of 2.2 (95% CI 1.3-3.8) for being involved in an occupational accident after adjusting for age, BMI, weight gain, years at work, snoring status, smoking, alcohol dependence, type of work, and exposures, while snoring without sleepiness did not significantly influence the risk (Table 2). Being a blue-collar worker was the most important independent risk factor for occupational accidents. Also exposure to noise was independently related to accidents.

The multiple logistic regression analysis was also performed for blue- and white-collar workers separately. In this model sleepy snorers of both groups had an increased adjusted odds ratio of occupational accidents, although the association was significant only in blue-collar workers. Exposure to noise was an independent risk factor in blue-collar workers while shift work was an independent risk factor in white-collar workers.

Sleepy Snorers and Occupational Accidents

As the influence on occupational accidents of snoring and EDS was significant only in men who reported the combination of both snoring and EDS in 1984, this group was studied more carefully. To study the time that passed between the 1984 survey and the date of report of the first occupational accident, men with the combination of snoring and EDS were compared with all the remaining in a survival analysis. As shown in Figure 3, the cumulative injury rate was significantly higher among men who were sleepy snorers at baseline. The difference between the groups remained significant (p = 0.006) after adjusting for all the possible confounders that are used as independent variables in Table 2.

View larger version (10K): [in this window] [in a new window]   Figure 3.   Cumulative chance of not being involved in any occupational accidents during the 10-yr period. Analysis of men who reported both snoring and EDS in 1984 compared with all the other men, that is, nonsnorers without EDS, snorers without EDS, and nonsnorers with EDS.

Of the 108 men who reported the combination of snoring and EDS in the 1984 questionnaire, 45 still reported both snoring and EDS in 1994. Of the remaining, 38 reported snoring but no EDS while nine were nonsnorers but with persistent EDS at the 10-yr follow-up. Of the subjects who were sleepy snorers in 1984 but not in 1994 six men had been treated with uvulopalatopharyngeoplasty (UPPP), three had undergone surgery to the nose, and one had started treatment with continuous positive airway pressure (CPAP).

Of the subjects without "snoring and EDS" in 1984, 112 reported both these symptoms in 1994. To study changes in "snoring and EDS" over time in relation to accidents, a new multiple logistic regression was performed were the subjects were subdivided into (1) no "snoring and EDS" in 1984 or in 1994, (2) "snoring and EDS" in 1984 but not in 1994, (3) "snoring and EDS" in 1984 and in 1994, and (4) "snoring and EDS" in 1994 but not in 1984. After adjusting for the same independent variables as in Table 2 only men with persistent "snoring and EDS" were at an increased risk with an adjusted OR of 3.1 (95% CI 1.5-6.4). Among the blue-collar workers also men with "snoring and EDS" in 1984 but not in 1994 were at an increased injury risk (Table 3).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main finding of this study is that in a population-based sample of middle-aged men, snorers who also suffer from EDS are at an increased risk of occupational accidents. Furthermore, among the sleepy snorers only those who had persistent symptoms with the combination of snoring and EDS both at baseline and at the 10-yr follow-up were at an increased risk for injury.

As far as we know this is the first population-based prospective survey performed to investigate the influence of snoring and EDS on occupational injury risk. The results of our prospective investigation are in accordance with the retrospective study by Ulfberg and coworkers in which snorers referred to a sleep laboratory because of symptoms suggestive of OSAS reported occupational accidents significantly more often than referents from the general population (13). In that study, all subjects underwent whole-night sleep recordings and, interestingly, the risk of occupational accidents was at least as high in heavy snorers without apneas as in patients with OSAS.

It seems reasonable to assume that a high proportion of the sleepy snorers in our study population suffered from sleep-disordered breathing. Polysomnography was, however, performed only on a small proportion of the sleepy snorers at baseline and at that time we analyzed only whether the sleepy snorers fulfilled the diagnostic criteria of OSAS (16). In the recent report with recommendations for diagnostic criteria presented by the American Academy of Sleep Medicine Task Force, respiratory effort-related arousals (RERA), defined as a progressively more negative esophageal pressure lasting > 10 s, terminated by a sudden change to less negative pressure, and an arousal are given the same value as apneas and hypopneas in the definition of the obstructive sleep apnea-hypopnea syndrome (21). In the cited study by Ulfberg and coworkers occurrence of RERAs was not analyzed, but a high prevalence of RERA among the nonapneic snorers might be an explanation for their daytime sleepiness and increased risk for accident. Furthermore, at baseline of this study in 1984, the importance of RERA was not recognized and a whole-night polysomnography would therefore have only partly answered the question about prevalence of sleep-disordered breathing among the sleepy snorers.

In the present study men suffering from EDS but who were not snorers were not at an increased risk for injury and we can only speculate about possible explanations. In sleep-disordered breathing, sleep is repeatedly interrupted by short arousals with daytime sleepiness as a consequence. These patients are rarely aware of their disturbed sleep and often describe themselves as experiencing sound sleep. On the other hand, people suffering from excessive daytime sleepiness from other causes such as medical disorders, insomnia, or children disturbing their sleep might be more aware of their sleep deprivation and consequently avoid demanding and risky situations.

Night job was not significantly related to occupational accidents and shift job was an independent risk factor in white-collar workers only. However, in studies in which accident risks have been compared between different work shifts, significantly more accidents have been reported during the night shift (4, 7). The lack of significant associations might be explained by the population-based design of the study, which entailed a low prevalence of night and shift workers. Furthermore, night and shift jobs are actually generic terms describing a multitude of work schedules. The participants in our study, drawn from the general population and not a given employer, can be expected to have been exposed to shift work schedules of great variety.

One disadvantage of this study is that the questions on smoking and alcohol dependence were included only in the 1994 questionnaire. Even though 96.1% of the subjects were able to give detailed information on their previous smoking habits, there might have been some recall bias when it comes to their smoking 10 yr ago. The questions relating to alcohol consumption are formulated as "Have you ever...," and, therefore, also involve the past. The CAGE questionnaire used in this study has a high validity in measuring alcohol dependence. It is possible, however, that it does not correlate well with the amount of alcohol consumed.

In 1994, a majority of the nonresponders were excluded because their response to the question on employment during the past 10 yr was not exact enough to permit categorization into blue- or white-collar worker. Compared with the 2,009 responders, the 442 subjects who responded to the follow-up questionnaire but could not be categorized as blue- or white-collar workers were older (47.5 ± 11.5 versus 42.7 ± 9.1 yr, p < 0.001), had been at work for a shorter period (5.8 ± 4.4 versus 8.6 ± 2.2 yr, p < 0.001), and had been involved in occupational accidents less often (8.8% versus 12.3%, p < 0.05). The combination of snoring and EDS was also more common among those not responding to the work-related question (9.0% versus 5.4%, p < 0.01). To analyze the influence of this uncategorized group, a multiple logistic regression analysis as presented in Table 2 but including the 442 "uncategorized" men was performed. Subjects were hence categorized into white-collar workers (reference), blue-collar workers, and "uncategorized." In this model, subjects reporting the combination of snoring and EDS at baseline had an adjusted OR of 1.9 (95% CI 1.2-3.2). The adjusted OR (95% CI) for blue-collar workers was 4.3 (3.1-6.0) and for "uncategorized" was 1.7 (1.1-2.7). No increased risk was found for snorers without EDS or for men reporting EDS but no snoring.

An advantage of the present study is that all information on accidents was obtained from a national register. The problem with recall bias associated with reported data on accidents over long reference periods (22) was thereby avoided. The reporting frequency to the register has been examined (23). Of occupational accidents occurring in a wide range of companies 92-95% had been reported, while of the accidents handled at a Swedish University hospital during 1 yr and categorized as occupational accidents only 77% were later found in the register (23). However, as the employers report the accidents there seems to be no reason to assume that reporting frequency would differ between sleepy snorers and nonsleepy nonsnorers.

We conclude that in a middle-aged male population, sleepy snorers have an increased risk of involvement in occupational accidents. The data indicate that early identification and treatment of sleep-disordered breathing are important when attempting to reduce the number of job injuries in the community.