Grossarth-Maticek and colleagues have shown, in their prospective studies, a strong relationship of their personality types, Types 1 and 2, to cancer and coronary heart disease (CHD), respectively. Relevant information is limited from replication studies, and little is known about psychosocial factors in relation to cancer or CHD in Japan. Subjects included 95 cases of lung cancer (LC), 94 cases of myocardial infarction (MI) and 596 controls. The controls were men and women who visited a clinic for a health checkup. The Grossarth-Maticek personality types, Types 1 to 6, were assessed using the Short Interpersonal Reactions Inventory. The distributions of the 6 personality types were compared between the case and control groups, adjusting for sex and age class. The relation of each of the 6 types to LC and MI were examined in terms of odds ratio, using a logistic regression model controlling for age, sex, job status, education level, and smoking status. As regards the distribution analysis, Types 1 and 2 in the LC and MI groups each were not more prevalent than the controls, respectively. High score of the Type 1 scale was associated with a statistically nonsignificant decrease in LC risk. MI risk was significantly, positively associated with the Type 2 and 5 scales, and unexpectedly, positively related to the Type 3 scale. The present findings partly supported the Grossarth-Maticek theory, but there remain some conflicting issues to be confirmed in future studies. J Epidemiol, 2001 ; 11 : 281-287