| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
Department of Preventive Medicine, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan. The Fukuoka Heart Study Group members are listed in Appendix 1.
Address reprint requests to: Kouichi Yoshimasu, Department of Preventive Medicine, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, Japan.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Study subjects comprised 290 cases (173 male workers and 117 women) of nonfatal acute myocardial infarction (AMI) and 489 controls (303 male workers and 186 women) recruited from among residents of the study area. Participation rates of cases and controls were 86% and 50%, respectively. Type A behavior pattern, job-related psychosocial factors, and other coronary risk factors were ascertained by interview using a questionnaire. Type A behavior pattern was measured by 12 questions, and job strain by the method of Karasek. Logistic regression analysis was used to calculate odds ratio and 95% confidence interval with adjustment for potential confounding variables. The relation to job strain was examined only in men.
RESULTS: Type A behavior pattern was significantly associated with an increased risk of AMI in male workers (adjusted OR = 1.9, 95% CI, 1.2–2.9) and in women (adjusted OR = 2.8, 95% CI, 1.6–4.9). Although job control and job demand showed no material association with AMI, job strain, a combination of low job control and high job demand was associated with an increased risk of AMI (adjusted OR = 2.2, 95% CI, 1.1–4.5). Abbreviated type A score and job demand score were higher among nonparticipant control candidates than participant controls in male workers.
CONCLUSIONS: Type A behavior pattern may be an important risk factor for AMI especially in Japanese women. The present study also revealed persons in stressful psychosocial circumstances were less likely to participate in the study.
Key Words: job strain • type A behavior pattern • myocardial infarction • Japanese
Abbreviations: CHD = coronary heart disease;; MI = myocardial infarction;; AMI = acute myocardial infarction;; JAS = Jenkins Activity Survey;; OR = odds ratio;; CI = confidence interval
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
Most of the previous studies regarding psychosocial factors and CHD have been performed in Western countries. The type A behavior pattern of Western people may differ from that of the Japanese in its context, and the stressful effect of job strain may also differ in different populations. In Japan, job-centered life-style was specifically shown to be positively associated with CHD (15). In another study of Japanese men (16), type A behavior pattern and high job demand were each related to an increased risk of AMI and angina pectoris combined independently. In this study (16), type A behavior pattern was assessed by the Tokai Activity Survey questionnaire which was developed with reference to the JAS, Framingham type A scale, and Bortner Rating Scale with inclusion of a job-related component (17). Interestingly, it was reported that type A behavior pattern assessed by JAS buffered job stress among Japanese male employees (18). The present study investigated the relation of type A behavior pattern assessed by a questionnaire modified to Japanese characteristics and job-related psychosocial factors to the risk of nonfatal AMI among Japanese men and women.
|
METHOD |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
Research nurses checked all admissions with a diagnosis of AMI or suspected AMI, and asked eligible patients to participate in the study. Collaborating cardiologists were responsible for the diagnosis of AMI. A total of 507 potentially eligible patients were identified and 435 patients (311 men and 124 women; 86%) participated in the study; 11 refused interview, 18 were discharged alive before interview, and 43 were too sick to be interviewed.
Two controls for each case, matched by gender, birth year (within 2 years), and proximity in residence, were to be recruited by using the resident registers. Between September 1996 and March 1998, a total of 1429 control candidates were approached by mail. Two reminders were sent, and contact by telephone was done if the telephone number was available. Finally, 664 (477 men and 187 women) participated in the study. The other outcomes were as follows: undelivered mail, 36, dead persons, 10, nonresident, 14, prior history of myocardial infarction, 44, refusal, 506 (including too sick persons), and no response, 155. Thus, the net participation rate was estimated to be 50% (664 of 1325).
Because there were few women with a full-time job, the present study used men having a full-time job for the analysis of job-related psychosocial factors. After the exclusion of men without a full-time job or whose jobs were indeterminable (114 cases and 155 controls), 197 cases and 322 controls with a full-time job remained. We also excluded men whose information on relevant factors were incomplete (24 cases and 19 controls). Thus, 173 cases and 303 controls of male workers remained in the analysis of job-related factors. Likewise, in women, seven cases and one control with incomplete data were deleted, and 117 cases and 186 controls remained in the analysis.
Clinical and Behavioral Risk Factors
A questionnaire-based interview elicited details of medical and family history, life-style factors, job titles, the presence of shift work, and psychosocial factors. The history of diseases that were related to AMI (hypertension, hyperlipidemia, angina pectoris, and diabetes mellitus) was defined as positive if subjects had ever been medicated or under dietary or exercise treatment for these conditions. Height and body weight were recorded, and a body mass index (kg/m2) of 25 or greater was defined as overweight. Job titles were coded according to the classification used in the national census. They were grouped into white-collar jobs (professional, managerial, clerical, and sales work) and blue-collar jobs (craft, laboring, and service work).
Psychosocial Factors
Type A behavior pattern was assessed by an abbreviated set of 12 questions which was developed by Maeda (19). This questionnaire is thought to be useful for Japanese regardless of age, sex, or occupation (19). Each question was listed in Appendix 2. The concordance for the behavior pattern between the Jenkins Activity Survey (JAS) and this 12-item questionnaire was 75%. Each question had three responses. Points 2, 1, and 0 were given to the answers of "always," "occasionally," and "hardly" for nine questions, and the points were doubled for three questions. A total score of 17 or greater was defined as type A. Cronbach 
of the type A scale was 0.81 for men and 0.77 for women.
Job stress was assessed in accordance with the Karasek model (20). Questions used in the present study were those developed by Courtney et al. (21) to capture aspects of job demand, job control, and on-the-job social support in an abbreviated form. Each question was listed in Appendix 3. Patients rated each of the 10 statements about their jobs on a four-point scale (strongly agree, agree, disagree, and strongly disagree). Job demand and job control scores were calculated on the basis of responses to two and five questions, respectively, in the original version. However, the Cronbach of the five job control questions in the present subjects was poor ( = 0.34). Kawakami et al. (22) and Kawakami and Fujigaki (23) reported that the Cronbach of the job control scales was in the range from 0.68 to 0.80 in Japanese workers using the Japanese version of the Job Content Questionnaire. Therefore, to confirm the applicability of the abbreviated questionnaire to the Japanese population, we performed a factor analysis on the data from the 303 control subjects, using squared multiple correlations as prior communality estimates and the maximum likelihood method to extract factors. A screen test suggested two meaningful factors, and only two factors were retained in the final step of varimax rotation to produce the rotated factor pattern (Table 1). Because loadings on the second factor were negligibly small for questions regarding "nonrepetitious work" and "freedom as to how to work," which were originally used to obtain the job control score, these questions were excluded in obtaining the job control score in the main analysis. The Cronbach of the job demand score was 0.87, and that of the job control score was 0.63. Job demand and job control scores were divided into tertiles as nearly as possible. In the classification of job strain, job control and job demand scores were dichotomized at the median value. High job strain was defined as a combination of low control and high demand, low job strain was defined as a combination of high control and low demand, and the remainings were combined into the middle strain group. When the tertiles in the job demand and control scores were used in the classification of job strain, the number of subjects in the low job strain was extremely small. Job support score was calculated on the basis of responses to three questions, and subjects were classified into approximate thirds with respect to this score. The analysis for job support was confined to 278 male subjects (103 cases and 175 controls) who had colleagues or superiors in the job environment. The Cronbach of the job support score was 0.78. Reproducibility of the psychosocial measures was assessed in a separate group of 25 patients undergoing coronary angiography. Intraclass correlation coefficients were 0.64 for type A score, 0.51 for job demand score, and 0.71 for job control score, and 0.33 for job support score.
|
To examine the characteristics of control candidates who did not participate in the present study, we mailed an abbreviated questionnaire containing 13 questions to 456 such persons accumulated until December 1997. Of these, 217 responded to the questionnaire, and unanswered questions were completed by telephone interview. Subjects who reported a history of myocardial infarction (N = 4) were excluded in comparison with participating control subjects. Among remaining 213 subjects, 69 men who had a full-time job and 88 wopmen (21% had a full-time or a part-time job) were used in the analysis. In this survey, four questions of type A, one question of job control and one question of job demand were chosen. The four questions of type A represented four components (time urgency, punctiliousness, stiffness, and competition) described by Doubeta (24). These were questions 1, 5, 7, and 12 in the Appendix 2. The question of job control and that of job demand were questions 3 and 4 in the Appendix 3, respectively. The questions of job support were questions 8 and 10 in the Appendix 3.
Statistical Analysis
Chi-square test and unpaired t test were used for the univariate analysis. Logistic regression analysis was used to make adjustment for potential confounding variables. Covariates included in the models were age, hypertension, diabetes mellitus, hyperlipidemia, angina pectoris, overweight, cigarettes smoking, alcohol intake, and parental CHD (angina pectoris and myocardial infarction). In addition, shift work was adjusted for in the analysis of male workers. Regarding women, the subjects were classified into two categories with and without a full-time or a part-time job. Age was categorized into six strata by 5-year class. In the analysis of men, exposure to cigarette smoking was expressed as cigarette-years, the number of cigarettes smoked per day multiplied by years of smoking, which was categorized into 0, 1 to 399, 400 to 799, and 800+ cigarette-years. The average amount of alcohol consumed per day was calculated for current alcohol drinkers, and alcohol use was categorized into never, past, and current use with a consumption of <50 or >50 ml of alcohol per day. In the analysis of women, the subjects were regarded as smokers if they had ever smoked even one cigarette everyday over a period of 1 year or longer. Likewise, alcohol use was defined as drinking alcohol at least once per week over a period of 1 year or longer. Indicator variables were created for categories of the covariates, and adjusted OR and 95% CI were obtained from the corresponding logistic regression coefficients and their standard errors. P values (two-sided) less than .05 were considered statistically significant. All computations were performed using the SAS software package version 6.04 (SAS Institute, Inc., Cary, NC).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
Table 3 shows clinical and life-style factors as well as job-related factors according to type A and job strain each in male control subject. Angina pectoris, heavy smoking, and alcohol use were more frequent among men with type A. Hyperlipidemia and parental CHD were more frequent among men with job strain. In women, clinical and life-style factors did not vary between the controls with and without type A behavior pattern, except that women with type A were younger than those without type A (p = .03). The frequency of type A behavior pattern was statistically significantly higher among women with a full-time or a part-time job than housewives (p = .001).
|
Table 4 presents the relation between type A and the risk of AMI in male workers and women. There was a statistically significant association between type A and AMI in each group. The association was slightly stronger in women than in men.
|
Table 5 summarizes the relation of job control, job demand, job strain, and job support to the risk of AMI. Low job control did not show a measurable association with AMI, and high job demand was related to a small, statistically nonsignificant increase in the risk of AMI. High job strain was statistically significantly associated with an increased risk of AMI. There was no material association between job support and the risk of AMI. The interaction between type A and job strain was far from statistical significance (likelihood ratio test, p = .45). Additional adjustment for job type did not change the results (data not shown).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
In working men, time urgency associated with type A or high job demand may have prevented them from participating in the study. Japanese male workers with type A may regard activities other than their jobs as a nuisance. However, the absence of a bias among type A patients with AMI to avoid participating in the study casts some doubt on this interpretation.
In the present study, both type A behavior patterns and job-related psychosocial factors were assessed retrospectively by a self-administered questionnaire. The history of AMI may have influenced patient’s perception or recall of the behavioral pattern and job-related psychosocial factors before the onset of AMI. There may also be a bias of selective survival, if survival among the AMI patients differs by the behavior pattern. If AMI was more likely to be fatal among nontype A patients, the effect of type A would be overestimated. On the other hand, if type A patient’s attacks were more often fatal, the bias could be reversed. However, in the Western Collaborative Group Study (3), new AMI were equally likely to be fatal among type A and nontype A patients. Although a high correlation between type A assessed in the present study and JAS was reported (19), the present measure of type A may not have represented type A assessed by the structured interview. In fact, type A assessed by JAS was not much correlated with a Japanese version of the structured interview in men and women (25). In this regard, the results need to be interpreted with caution.
Although aggression, especially expressive hostility, has been implicated as being an important component of the coronary prone behavior pattern in the United States (26, 27), job-centered life-style or workaholism rather than hostility has been shown to be a coronary prone behavior in Japanese society (15, 17). The present finding corroborates that the Japanese version of type A is predictive of increased risk of AMI. The association in women was stronger than that of men. Booth-Kewley and Friedman (28), in a meta-analysis of coronary prone behavior and disease outcomes, found the connection between type A and cardiovascular outcomes to be "as strong or stronger for women as it is for men." In Japan, there was no study that examined the relation between type A and CHD among women. In a study of working mothers (29), women with high type A scores had greater job stress than those with low type A scores. It was also reported that type A was strongly associated with nonworking mothers’ child-rearing stress (30). Another study (31) showed that women with type A tended to be unaware of their poor stress coping and suggested that the unawareness of poor stress coping might lead to manifestation of coronary heart disease and psychiatric symptoms. Haynes and Feinleib (32) showed that women, regardless of employment status, reported significantly more symptoms of emotional distress than men. In that study (32), it was also reported that working women experienced more daily stress and aging worries and were less likely to display overt anger than either housewives or men. Because of this characteristic, Japanese women also may have greater emotional instability than men and the stressful effect of type A behavior pattern may be greater in women.
The present study first addressed the association between the Karasek model of job strain and CHD risk in Japan. Previously, a case-control study of Japanese men reported that job demand was greater among patients with AMI or angina pectoris than healthy controls (16). Another study suggested that job strain of the Karasek model was significantly associated with hypertension and hyperglycemia in Japanese men (33). In the present study, although high job demand was associated with only a minimal increase in the risk of AMI and low job control was unrelated to AMI, there was a statistically significant increase in the risk associated with high job strain in comparison with low job strain. However, methodological problems cause difficulties in interpreting the present findings. As mentioned above, job strain may have been underestimated differentially in the control group. Although neither low job control nor low job support showed any material association with AMI in the present study, it has been reported that these factors have greater effects on psychological distress than quantitative job demand in Japanese (34).
There is a possibility that the subject’s response to job-stress questions may be affected by the cultural characteristics specific to the Japanese society (34). Iwata et al. (35) noted that Japanese men tended to suppress the expression of positive feelings and the questions including positively worded items, such as job control questions, would not adequately measure job-related stress in Japanese workers. The role of a group or cooperation has been emphasized at the expense of individual initiatives in the traditional work environment in Japan (17). In such a situation, Japanese workers naturally tend to be less critical of their colleagues or superiors, thereby obscuring the association between job support and the risk of AMI.
The interaction between the type A behavior pattern and job strain was of particular interest in Japanese men because the Japanese coronary-prone behavior pattern characterized by job-centered life-style or workaholism may modify or moderate the perception of job strain (18). Furthermore, job strain may be perceived as job satisfaction in Japanese men with job-centered life-style. It was reported that working more than 10 hours per day was positively associated with job satisfaction among type A, but not nontype A, Japanese men (36). In the present study, however, no measurable interaction was found between type A and job strain.
In summary, the present case-control study showed that type A behavior pattern was associated with an increased risk of AMI in Japanese men and women. Job strain was also associated with a statistically significant increase in the risk of AMI in men. The present study also showed that persons with stressful psychosocial circumstances were less likely to participate in this epidemiologic study. Careful consideration is needed in epidemiologic study for psychosocial risk factors as well as biomedical risk factors.
|
APPENDIX 1 |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
APPENDIX 2 |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
APPENDIX 3 |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
Responses were assigned values of 4 for "strongly agree," 3 for "agree," 2 for "disagree," and 1 for "strongly disagree." The job control score was calculated by totaling the responses to questions 2, 3, and 6. The job demand score was calculated by totaling the responses to questions 4 and 5. Social support on the job was measured by the sum of questions 8, 9, and 10.Questions 1 and 7 after reversing the scales were to be used for job control according to the original version, but they were omitted in the present study because of little contribution to the scale.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
2 Points were doubled for question 5, 6, and 9.
Received for publication August 15, 2001.
Revision received February 2, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONAPPENDIX 1APPENDIX 2APPENDIX 3NOTESACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. Gallacher Commentary: Type A behaviour and heart disease: no less inscrutable in Japan Int. J. Epidemiol., December 1, 2008; 37(6): 1406 - 1407. [Full Text] [PDF]
|
|
|
|
 |
|
 K. S. Thomas, R. A. Nelesen, M. G. Ziegler, W. A. Bardwell, and J. E. Dimsdale Job Strain, Ethnicity, and Sympathetic Nervous System Activity Hypertension, December 1, 2004; 44(6): 891 - 896. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y Liu and H Tanaka Overtime work, insufficient sleep, and risk of non-fatal acute myocardial infarction in Japanese men Occup. Environ. Med., July 1, 2002; 59(7): 447 - 451. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
