Life-Style Related Factors and Idiopathic Dilated Cardiomyopathy - A Case-Control Study Using Pooled Controls -

A case-control study was conducted to investigate how basic habits of life including dietary habit, physical activity, cigarette smoking, and drinking, are involved in the development of idiopathic dilated cardiomyopathy (DCM). Collection of cases was entrusted to the clinical research group of DCM, and national pooled controls established by sex and age category by the epidemiological research group of intractable diseases were used to ensure representativeness of the controls. Fifty-eight cases of DCM which developed in and after January 1991 were collected, and 5,912 controls matched with the cases by residential area, sex, and age were selected. Analysis of the results of the study showed that items in the questionnaire suggestive of viral infection, such as "susceptibility to common cold" and "susceptibility to diarrhea", items concerning dietary habit, including "taking no breakfast", "ingestion of salty food", and "ingestion of fatty food", and such items as "cigarette smoking" and "lack of sleep" tended to be observed in the case group at significantly higher frequencies. Since viral infection has been suspected as a causative factor of DCM, further research of this area is thought to be of particular importance for determining the etiology of DCM. J Epidemiol, 1999 ; 9 : 286-296

MATERIALS AND METHODS

Collection of cases
It is essential in collecting cases that specific diagnostic criteria be applied in the diagnosis of cases.Hence, in the present study collection of cases was entrusted to the Ministry of Health and Welfare (MHW) clinical research group of DCM (Yoshio Yazaki, group leader).Accordingly, the diagnosis manual 1.2) prepared by that research group is used in diagnosing the cases, and the diagnosis of the collected cases are considered to be uniform.The form "Questionnaire Concerning Habits of Life and Health" prepared by the MHW epidemiological research group was used to collect necessary information from the cases.The questionnaire, which is filled in by the patient himself, contains a total of 63 items, including usual condition of health, dietary habit, drinking, cigarette smoking, and history of pregnancy and childbirth 3,4).
To collect cases, members of the MHW clinical research group of DCM and 10 collaborating investigators were asked by group leader Yazaki in June 1994 to provide 20 cases each, and 86 cases were thus collected.However, because omissions of entry, overlapping answers, contradictory answers and the like were noted at a rather high frequency, it was decided on review by the preventive factor subgroup of the epidemiological research group to check and confirm the answers, and in December 1994 this was requested of the investigators providing the cases.It was also decided in the review by the preventive factor subgroup to use cases developing in and after January 1991 in the analysis of the case-control study, because the habits of life and others of cases who have had the disease for a long time could have been modified by its presence and pre-onset information relating to the cause of the disease would not necessarily be obtained.However, as it turned out that cases newly occurring in and after January 1991 accounted for only about half of the collected cases and it was feared that some cases of hypertrophic cardiomyopathy might be mixed among them, additional cases were collected and checked to see if any cases of hypertrophic cardiomyopathy were included among them from June to October 1995.

Selection of controls
Securing of representativeness has been a problem in selecting controls.To keep the problem to the minimum , in the present study, so-called controls 1) collected by the MHW epiderniological research group from local and occupational groups of people of 12 regions across Japan were used and all individuals matching with the cases in area of residence , sex and age were used as the control group.When, like the pooled controls, there is already a large number of collected controls, there is no need to re-select a small number of controls, like 5 for l case, and as it is said preferable to use every available control, for this would increase the accuracy of the odds ratio, it was decided to use all eligible individuals.Therefore, each case has a different number of controls.
The form "Questionnaire Concerning Health and Life Habits" prepared by the MHW epidemiological research group was also used to collect necessary information from the controls 4).

Test of significance of questionnaire items
The items in the questionnaire presenting a significant difference in their selection rate and odds ratio are shown in Figures 1-3.The selection rates and odds ratios of all questionnaire items are shown for the cases and controls in Appendix 2 at the end of this report.
(1) Condition of health before onset (one year before diagnosis) The selection rate of "yes" to the item "susceptibility to common cold" was 17.9% for the control group , but it was a high rate of 38.2 for the case group, and the odds ratio of "yes" with "no" as 1 was 2 .88(95% CI: 1.53-5.41,p=0.001) , showing a significantly high frequency of those responding in the affirmative in the case group.(Figure 1, Appendix 2) The selection rate of "yes" to the item "susceptibility to diarrhea" was 20.0% for the control group, but it was a high rate of 32.7% for the case group, and the odds ratio of "yes" with "no" as 1 was 2.30 (95 % CI: 1.18 -4.47, p=0.014), showing a significantly high frequency of those responding in the affirmative in the case group.(Figure 1, Appendix 2) The selection rate of "yes" to the item "cough up phlegm" was 8.6% for the control group, but it was a high rate of 14.5% for the case group, and the odds ratio of "yes" with "no" as 1 was 2.41 (95% CI: 1.00-5.81,p=0.050), showing a tendency for the frequency of those responding in the affirmative to be high in the case group.
Other questions asked concerned "eczema, nettle rash", weekly "exercising time", and daily "walking time", but the selection rates of these items showed no significant difference between the case and control groups.(Appendix 2) (2) Dietary habit before onset The selection rate of the item "took no breakfast" was 9.0% for the control group, but it was a high rate of 20.7% for the case group, and the odds ratio of "took no breakfast" with "took breakfast" as 1 was 2 .88(95% CI: 1.37-6.03,p=0.005), showing a significantly high frequency of those responding in the negative in the case group.(Figure 2, Appendix 2) The selection rate of "liked very much" for the item "salty food" was 11.9% for the control group, but it was a high rate of 31.0%for the case group, and the odds ratio of "liked very much" with "ordinary" as I was 3.16 (95% CI: 1.58-6.34,p=0.001), showing a significantly high frequency of those responding in the affirmative in the case group.(Figure 2, The selection rate of "liked very much" for the item "fatty food" was 4.3% for the control group, but it was a high rate of 15.5% for the case group, and the odds ratio of "liked very much" with "ordinary" as 1 was 6.66 (95% CI: 2.67-16.6,p<0.001), showing a significantly high frequency of those responding in the affirmative in the case group.(Figure 2, Appendix 2).
Other items relating to dietary habit inquired about were coffee, Japanese tea, beef, pork, ham, poultry, liver, eggs, milk, yogurt, cheese, butter, margarine, fries, fried vegetables, fresh fish, boiled fish paste, spinach, carrot, tomato, cabbage, Chinese cabbage, wild plants, mushroom, potato, seaweed, pickled vegetables, food boiled down in soy bean curd, orange, natural fruit juice, and other fruits and cakes, but no significant difference was observed in the selection rate of any of these items.(Appendix 2) (3) Drinking habit before onset The selection rate of "drank" for alcoholic beverages was 66.8% for the control group and 75.4% for the case group.
The odds ratio of "drank" with "did not drink" as I was 1.61 (95% CI: 0.78-3.31,p=0.199), and although the frequency of those who drank seems too high in the case group, no significant difference was observed.(Appendix 2) (4) Smoking habit before onset The selection rate of "smoked" for smoking habit was 45.0% for the control group, but it was a high rate of 58.6% for the case group, and the odds ratio of "smoked" with "did not smoke" as 1 was 2.91 (95% CI: 1.27-6.70,p=0.012), showing a significantly high frequency of those with smoking habit in the case group.(Figure 3, Appendix 2) (5) Usual sleeping time To the question concerning average sleeping time on a usual day, the selection rates of less than 5 hours and less than 5 -6 hours were 0.6% and 3.7%, respectively, in the control group, but the rates were 5.3% and 8.8% and high in the case group, and the odds ratio of less than 5 hours with 7 -8 hours as 1 was 4.91 (95% CI: 0.34-23.9,p=0.037), showing a significantly high frequency of those with short sleeping time in the case group.(Figure 3, Appendix 2) DISCUSSION Idiopathic cardiomyopathy is defined as "myocardial disease of unknown etiology."1.2) Accordingly, the etiology of dilated cardiomyopathy (DCM), which is one of its clinicopathologic types and comprises mainly myocardial contraction insufficiency and ventricular dilatation, is also still unknown 2).And, no systematic study has yet been made concerning how it is related to basic habits of life.
Accordingly, a case-control study was made to explore how basic habits of life such as dietary habit, physical activity, smoking and drinking and usual condition of health are involved in the development of DCM.
(1) Items suggesting involvement of viral infection or immune dysfunction First, study of health condition before onset (one year before onset) showed a significantly high frequency of items suggesting infections such as "susceptibility to common cold" and "susceptibility to diarrhea ," especially viral infection, in the case group.It was also suggested that the frequency of "cough up phlegm" is high in the case group.Odds ratio was 2.88 for "susceptibility to common cold and 2 .30for "susceptibility to diarrhea," and the risk of developing the disease was construed to be 2 -3 times higher in the presence of such health abnormalities.
Numerous reports have been made 2,6-9) concerning the possibility of viral infections, especially viral myocarditis, being involved as cause of DCM.The enterovirus and the coxsackie virus are said to be important as viruses involved in viral myocarditis, and about half of the cases develop common coldlike symptoms as precursor symptoms 6).The viral myocarditis is known to develop and persist latently without presenting any typical clinical picture in its acute stage 6), so it is fully possible that DCM is in fact viral myocarditis that has advanced latently.In fact, attempts have come to be made to detect antibody to the coxsackie virus in DCM patients 10), and recently studies have also come to be made to substantiate the viral infection theory by detecting viral genomes of the enterovirus and the coxsackie virus in the myocardial biopsy specimens of DCM patients [11][12][13].The results of our study also supports the possibility of viral infection being involved in the development of DCM, but it is well known that resistance to infection is decreased in heart failyre patients, and it is fully possible that infection is not the cause really, but the result of heart failyre .Therefore, it is difficult to form a conclusion at this stage.
Besides viral infection, impairment of immune function is suggested as the cause of DCM 2), but susceptibility to common cold and susceptibility to diarrhea, which have been frequently observed in the cases of the present study, can be construed as resulting from impairment of immune function among other possibilities.
(2) Dietary habit Study of dietary habit show that the frequency of cases of DCM responding "taking no breakfast" to be higher than the controls, and the odds ratio to be about 3. The frequency of those responding "like very much" for salty food was also high, the odds ratio likewise being about 3, and the frequency of those responding "like very much" for fatty food was also high, the odds ratio being about 7. For other foods, no difference of frequency was observed between the case group and the control group.
Nutritional disorder and deficiency of nutrients or trace ele-ments are also considered as possible causative factors of DCM 8).The significantly high frequency of "taking no breakfast," "like salty food", and "like fatty food" in the case group observed in the present study can be taken to mean that many in the group do not have balanced eating habits, and the possibility remains that nutritional disorder and deficiency of nutrients are a cause.And, it is reported that hypertension is related in some way to DCM 2).High intake of salt is closely related to hypertension, and from the results of the present study, involvement of hypertension in the development of the disease is thought to be possible.
(3) Drinking and smoking The toxication theory is also advanced as the cause of DCM, and there are reports of alcohol, anti-cancer drugs, psychopharmaceuticals, cobalt, and others being possible causes 8) In particular, alcoholic cardiomyopathy presenting the same pathologic image as DCM, which develops by excessive drinking, has been known to exist from long ago.Investigation and analysis were therefore made concerning drinking habits, but no statistically significant differences were observed in this analysis, although a tendency was observed for the frequency of drinking to be slightly higher in the case group.
Tabacco, needless to say, contains numerous harmful substances, but it also is considered to precipitate spasm of micro vessels, which is espoused as a cause of DCM, as well as have a toxic effect on the myocardium8).In the present analysis, the frequency of those who "had smoked" was higher in the case group than in the control group, with an odds ratio of about 3, and those who had smoking habit were observed at a significantly high frequency in the case group.The possibility of smoking being involved in some way in the development of the disease is considered to be great.(4) Sleeping time and others Study of average sleeping time on a usual day revealed that the frequency of those who slept less than 5 hours was higher in the case group than in the control group and the odds ratio was as high as about 5. Short sleeping time cannot in itself be considered to be a cause of DCM, but resistance can he considered to be low when one is exposed to a causative agent in a state where he is wanting in sleep.
Analysis was also made concerning physical activity , but no significant results were obtained.CONCLUSION A case-control study was conducted to investigate whether basic habits of life, including dietary habit , physical activity, cigarette smoking, and drinking , and usual condition of health are involved in the development of idiopathic dilated cardiomyopathy (DCM).As a result , infection, excessive intake of salt and fat, cigarette smoking and lack of sleep were suggested as possible factors related to development of the dis-ease.Physical activity and drinking habit showed no relationship.However, judgment is difficult at this time as to whether these factors are working as causes, whether the finding is the result of modification of the habits of life owing to the existence of the disease, or whether the factors are involved as contributing factors of the disease.In order to confirm how these various factors are involved in the development of DCM further research needs to be made in the fields of basic medicine, clinical medicine and epidemiology.
whose representativeness is different to secure, has been facilitated through use of national pooled controls established by sex and age category by the epidemiological research group of intractable diseases.

Table 1 .
Cases of idiopathic dilated cardiomyopathy and their controls.