Abstract
Previous studies examining the association between personality and cancer incidence or mortality have reported inconsistent findings. During June through August 1990, 30, 277 residents of Miyagi Prefecture in northern Japan completed the Japanese version of the short-form Eysenck Personality Questionnaire-Revised and another questionnaire on various health habits. During 7 years of follow-up until December 1997, we identified 986 incident cases of cancer. We used Cox regression to estimate the relative risk of total cancer according to the four levels of four personality subscales (extraversion, neuroticism, psychoticism, and lie) , with adjustment for sex, age, education, smoking, drinking, body mass index, and family history of cancer. Multivariate relative risks for the highest verses the lowest level of personality subscales were 0.9 for extraversion (95 percent confidence interval, 0.7-1.1 ; trend p=0.32), 1.1 for psychoticism (95 percent confidence interval, 0.9-1.3 ; trend p=0.96), 0.9 for lie (95 percent confidence interval, 0.7-1.0 ; trend p=0.19), and 1.2 for neuroticism (95 percent confidence interval, 1.0-1.4 ; trend p=0.06). Although neuroticism showed significant, positive, linear associations with 671 cancer cases prevalent at baseline (trend p<0.001) and with 320 cases diagnosed within the first three years of follow-up (trend p=0.03), it showed no association with 666 cases diagnosed during the fourth through the seventh years of follow-up (trend p=0.43). This prospective study does not support hypothesis that personality is a risk factor for cancer incidence. The association between neuroticism and prevalent cases of cancer seems to be a consequence, rather than a cause, of the diagnosis or symptoms of cancer.
