ABSTRACT
BACKGROUND: This prospective cohort study was carried out to investigate the relation of lifestyle variables to total mortality in residents aged 60-64 of a rural community in Korea. METHODS: A total of 1,042 residents was interviewed by face-to-face survey for baseline data collection in 1996 and 955 residents were successfully followed up until April 2002, among which 91 died. RESULTS: In univariate logistic regression, consumption of cigarette per day, duration of smoking, status of smoking, frequency of drinking per month, status of drinking, average duration of TV watching per day and average duration of sleeping were statistically significant, predicting total mortality. None of the variables reflecting leisure-time physical activities were statistically significant. Adjusting for sex, age, education, current as well as past chronic diseases history, and limitation of daily living with multiple logistic regression, those smoking more than 40 years and current smokers showed reduced, but statistically significant RRs, 1.89 (95% CI 1.05-3.41) and 1.82 (95% CI 1.01-3.25), respectively. Sleeping more than 10 hours a day showed RR, 2.41 (95% CI 1.11-5.22). CONCLUSION: These results suggested that some lifestyle variables, smoking and sleeping were predictive of total mortality while drinking, TV watching and leisure-time physical activities were not.