Prevalence of risk factors for coronary artery disease among day and shift workers.

JONSSON Prevalence of risk factor s for coronary artery disease among day and shift workers. Scand J Work Environ Health 14 (1988) 317-321. Several recent studies have indicated that shift work is associated with increased risk of coronary artery disease. In this cross sectional study 361 shift workers were examined with respect to some major risk factors for coronary artery disease; 240 day workers constituted the reference group . A higher proportion of shift workers smoked (54 versus 39 %). Shift workers also had significantly higher levels of serum triglycerides (1.61 versus 1.43 mmolll). Body mass index and the blood pressure and total cholesterol levels did not differ between the groups. Multiple regression analyses demonstrated that shift work was significantly related to serum triglyceride levels also when age, smoking, body mass index, and other variables were controlled for. It was concluded that shift work is associated with several risk factors for coronary artery disease.

Shift work has been associated with sleep disturbances and gastrointestinal disease (2,4,11,20). Recent reports have also indicated increased risk of ischemic heart disease in shift workers (1, 3,14 ,15). Very few studies exist, however, of traditional risk factor s among shift wor kers. Although TheIle et al (19) found elevated serum triglyceride levels among shift workers, it was not clear whether the elevated level was related to body con stitution, smoking, age, or other factors. Rosen et al (18) conducted a survey of about 12000 subjects belonging to different occupational groups in Sweden . They found that da y workers smoked less than tho se with " other wor k schedules," but their report lacked information about possible confounders.
The present stud y was aimed at comparing shift and da y workers with respect to risk factor s, eg, total cholesterol, triglycerides, smoking, and high blood pressure , for coronary art ery disease. Particular care was taken to select companies where the work load and socioeconomic background would be similar between th e shift and da y wor kers.

Subjects and methods
The subjects consisted of 601 male blue-collar workers in three plant s in the north of Sweden. Two plant s manufactured paper and one produced carboxymethyl cellulose.
The shift workers (N = 361) worked a rotating threeshift schedule and were process operators with a considerable amount of manual work. The shifts changed at 0600, 1400, and 2200. Table 1 shows the shift schedule at the three plant s. The day workers (N = 240) were predominantl y maintenance men who worked 40 h a week. Th e average income of the shift workers exceeded that of the day workers by approximately 17 0/0. The data were cotlected in 1975 and 1976.
All the subjects filled out a short questionnaire which included question s about the work schedule, the physical work load, psycho social factor s, smoking habits, and curr ent medical symptoms. Smoking habits were recorded in terms of the daily consumption of tobacco. A worker with a daily consumption of one cigarette or more (for pipe smokers one gram of to bacco) was defined as a smoker.
Th e coded an swers for the following six question s were added to form an index (physical index) measuring physical strain at work: Is your work site comfortable? Are machine s and tools good enou gh? Can you change positions durin g work? Is your work too physicall y demanding? Does you r work requ ire heavy liftin g? Is it possible to redu ce the physical demands at your wor k site?
The score sum of the answers for the following nine questions were computed to measur e the psychosocial work environment (psychosocial index): Is your work too mentall y demanding? Is the planning of your work too irregular?   Do you have to do too much overtime work? Is the working speed too dependent on machines? Is the working speed too dependent on fellow workers? Do you have too many job tasks? Are you satisfied with your job? . Are you satisfied with the management? Does the job give enough opportunites to take a break?
The reliability tests yielded Cronbach's coefficient alphas of 0.53 and 0.58 for the physical and psychosocial indices, respectively.
Venous blood samples were drawn from an arm vein in the morning after a night's sleep. The subject fasted from 2400 the night before. Plasma was centrifuged within 2 h, stored in a refrigerator at 4°C, and analyzed within 3 d. The blood plasma was analyzed by an autoanalyzer according to routine methods of a major hospital laboratory. In this paper the results of the total cholesterol, triglycerides, and blood glucose tests are reported.
Body weight (BW) in kilograms and height (H) in meters were measured, and the body mass index (BMI) was calculated (BMI = BW1H Z ) . The diastolic and systolic blood pressures were measured with a sphygmomanometer with the subject supine after 5 min of rest.

Statistical methods
The significance of the differences between the mean group values was tested by Student's t-test and the chisquare test (two-tailed). Stepwise multiple regression 318 analyses and reliability tests were performed by an SPSS program (16). Table 2 shows that the mean age of the shift workers did not differ significantly from that of the day workers. The duration of employment was significantly longer for the day workers. A higher proportion of shift workers smoked. The shift workers also had significantly higher levels of serum triglycerides. The body mass index, blood pressure, and the total cholesterol and blood glucose levels did not differ between the groups. The self-rated physical work conditions did not differ, but the shift workers reported a significantly lower quality of psychosocial work conditions. Serum triglyceride concentrations are usually associated with age and body weight, and, in order to compute the relation between shift work and the serum triglyceride levels, we performed a multiple regression with the serum triglyceride concentrations as the dependent variable. Independent variables were age, body mass index, smoking, shift work, and the physical and psychosocial indices. Table 3 shows that the serum triglyceride levels were significantly predicted by body mass index, smoking, and shift work. We also tested the interaction between smoking and shift work, but no interaction was found.

Results
In addition smoking, the other predictor of coronary artery disease that differed between the groups, was subjected to a regression analysis. The independent variables were shift work, age, and the physical '1.'=5.29, P<0.15 (not significant). Table 3. Last step in the step-wise regression analysis with serum triglyceride level as the dependent variable. or alcohol consumption, neither of which had been included in the present study. We do not think that the strong relation between shift work and the increased levels of triglycerides was caused by a phase-shift of the circadian rhythm. First, the blood specimen was drawn from both the shift and day workers in the morning after a night's sleep. Second, the analyses did not yield any association between shift work and other blood variables that have the same diurnal rhythm as triglycerides.
Elevated levels of serum triglycerides in shift workers have been reported also by TheIle et al (19) and Orth-Gomer (17). In an experimental study the latter demonstrated that the type of shift schedule used might affect the serum triglyceride levels. Thus a group of policemen showed lower levels of serum triglycerides during counterclockwise shift rotation than in clockwise rotation. and psychosocial indices. This analysis yielded a significant association only between smoking and shift work (standardized regression coefficient = 0.15, P<O.OOI). Table 4 shows the proportion of subjects with increased levels of serum lipids. In all, 55 070 of the shift workers and 57 % of the day workers were obese (defined as BMI > 25). Twenty percent of both the day and the shift workers had a diastolic blood pressure exceeding 94 mm Hg (12.5 kPa).
In table 5 the prevalence of the three major risk factors for coronary artery disease, ie, increased serum lipids, high blood pressure, and smoking, is presented. "Increased serum lipids" was defined as a total cholesterol level higher than 5.1 mmo!!l or triglyceride levels higher than 2.2 mmo!!!. "High blood pressure" was defined as a diastolic blood pressure higher than 94 mm Hg (12.5 kPa).

Discussion
The present cross-sectional study showed that smoking and high serum triglyceride levels were more prevalent among the shift workers than the day workers.
The three leading predictors of coronary artery disease are hypercholesterolemia, hypertension, and cigarette smoking. The role of elevated serum triglycerides as a separate risk factor for atherosclerosis and coronary artery disease is unclear, however. Thus Hulley et al (13) have argued that the Western Collaborative Study does not demonstrate that serum triglyceride levels are an independent risk factor. On the other hand Carlsson & Bottiger (7) have concluded that serum triglycerides are a risk factor for "hard" end points of coronary artery disease. This argument is supported by the findings of the Stockholm Prospective Study (6,7,8).
Hypertriglyceridemia has been noted to be a common finding in men with coronary artery disease (5). Moreover, it is frequently accompanied by obesity and abnormal glucose tolerance. The difficulty in interpreting triglyceride values, as well as their frequent association with hypercholesterolemia has prevented hypertriglyceridemia from emerging as a separate risk factor.
In the present study body mass index was the most powerful predictor of the serum triglyceride levels. Interestingly, however, this variable did not differ between the two groups studied and cannot be a contributor to the observed difference in triglyceride levels between the groups.
Smoking, however, is known to affect lipids, and the groups differed significantly with respect to this variable. Thus part of the observed difference between the groups may have been due to smoking.
Yet there still remains a significant, independent contribution from shift work. The reason for this finding is not clear, but one explanation may be dietary habits In the present study smo king was more common among the shift workers. There have been ver y few studies which ha ve compared shift and day workers with respect to smoking behavior. In the Tromso heart study project (19) a higher prevalence of smoking was reported for shift workers than for day workers (64.9 versus 54.8 alo). The level of significance was not reported. In a recent Swedish population survey day workers smoked less than those with "other work schedules" (18). However, no info rmation on work schedules or possible confounders wa s provided. The investigations by Angersbach et al (4) indicated that smoking substantially contributes to gastrointestinal disea se among shift workers. The percentage of smokers was very high (75 0/0) , but no figures were given for day workers.
Why, then, do shift workers smoke more than day workers? One explanation may be that there is a selection of smokers into shift work. Such a selection could operate through, eg, socioeconomic or marital status. Previous investigations have shown that people from lower social strata, blue-collar workers, and people with low education smoke more than other socioeconomic groups (9,10,12,18). However , in the present study, we had excluded white-collar workers and foremen . In addition , tho se shift and da y workers which were included in the study had very similar living and economic conditions. Actually, the shift workers earned somewhat more than the day workers, a fact which should have sugge sted a lower prevalence of smoking. Unfortunately we did not collect an y data on marital status in the present study, but in a current sur vey of blue-collar workers in one o f the three plant s, we did not find that marital status was a significant confounder with respect to smo king among the da y and shift workers (unpublished data).
Another possible cause of smo king might be the work environment or the work task. However, the work environment was similar for the day and shift workers in 1975-1976. When the study was carried out, there were very few restrictions on smoking in the three plants studied. Both the day and shift workers could very easily take a break and smoke. In addition very similar ratings were obtained for the physical work conditions, and the total number of workhours was 10 % lower for the shift workers. Furthermore, the mental load of the work task did not seem to explain the smo king behavior. Another, more likely explanation for the difference in smo king behavior is that smoking is influenced by the workhours, perhaps as a stim ulant or as a wa y to pas s time during the long night shift. The data do not, however, allow us to draw an y definite conclusions.
The results of the presen t study, to gether with previous findings, support the hypothesis that shift work is related to risk for coronary heart disease. This greater risk seems, at least partiall y, to be due to wellknown risk factors , for example , smoking. How ever, in a recent study reporting association between shift work and coronary artery disease, a higher proportion of shift workers smoked, but the association between shift work and coronary artery disease remained significant also after age and smoking habits were controlled for (14). Thus it is too earl y to rule out the po ssibility of a direct association between shift wo rk and coronary artery disease independent of mediation through established risk factors.
As cardiovascular disease is the most common cau se of death in industrialized countries, any links with shift work would mean that the harmful effects of shift work on health would have to be regarded as more seriou s than ha s pre viousl y been thought the case.