Advance Publication by J-STAGE J Epidemiol 2011 doi:10.2188/jea.JE20100081

Young Investigator Award Winner’s Special Article

Attributable Fractions of Risk Factors for Cardiovascular Diseases
Atsushi Hozawa1,2
1 2

Department of Public Health, Yamagata University, Graduate School of Medical Science, Yamagata, Japan Division of Epidemiology, Tohoku University, Graduate School of Medicine, Sendai, Japan

Received May 21, 2010; accepted November 2, 2010; released online January 29, 2011

ABSTRACT
Background: Cardiovascular disease (CVD) is a leading cause of death in Japan. To reduce the threat of CVD, it is important to identify its major risk factors. The population attributable fraction (PAF) is calculated from the prevalence and relative risk of risk factors and can be used to estimate the burden of these factors with respect to CVD. We analyzed the findings from several prospective studies to determine the PAFs of CVD. Methods: PAF was calculated as pd × (multiadjusted relative risk − 1)/multiadjusted relative risk, where pd is the proportion of patients exposed to that risk factor category, according to data from the Ohsaki Cohort Study, EPOCHJAPAN, NIPPON DATA80, Miyagi Cohort Study, CARDIA Study, and ARIC Study. Results: Nonoptimal blood pressure explained 47% and 26% of CVD mortality in middle-aged and elderly Japanese, respectively. Cigarette smoking explained 34% of all-cause mortality in middle-aged men. The combination of hypertension and cigarette smoking explained 57% and 44% of CVD mortality in younger men and women, respectively. Furthermore, the presence of at least 1 nonoptimal risk factor explained most CVD deaths and all-cause deaths. Conclusions: Established CVD risk factors, especially high blood pressure and cigarette smoking, explained a large proportion of CVD mortality and all-cause mortality. Prevention, early detection, and treatment of these conventional risk factors are required to reduce mortality risk. Key words: cohort study; cardiovascular diseases; population attributable fraction

INTRODUCTION
Cardiovascular diseases (CVDs), namely, heart disease and stroke, are leading causes of death in Japan.1 Furthermore, because stroke is a major cause of certification for long-term care insurance in Japan,1 risk factors for stroke also contribute to a decline in activities of daily living (ADL). Therefore, the prominent risk factors for CVD must be identified if we are to lower the risks for mortality and ADL decline. The population attributable fraction (PAF) is an estimate of the burden of a disease.2 My colleagues and I estimated the PAFs of allcause death, CVD death, CVD incidence, ADL decline, and smoking-related diseases due to established CVD risk factors,3–10 and the results are described herein.

METHODS
Cohort studies Ohsaki Cohort Study The setting and design of the Ohsaki Cohort Study have been

reported in detail elsewhere.11 In brief, this prospective cohort study started in 1994. A self-administered questionnaire requesting information on various health-related lifestyles was delivered to all National Health Insurance (NHI) beneficiaries aged 40 to 79 years living in the catchment area of the Ohsaki Public Health Center, Miyagi Prefecture, Japan. In Japan, the NHI is used by farmers, the selfemployed, pensioners, and their dependents. The Ohsaki Public Health Center, which is a local government agency, provides preventive health services for the residents of 14 municipalities. The questionnaires were delivered to and collected from the subjects’ residences by public health officials in each municipality. This procedure yielded a high response rate of 94.6% (n = 52 029). A total of 776 subjects were excluded from the study because they had withdrawn from the NHI before 1 January 1995, when the prospective collection of NHI claim files began. Thus, 51 253 subjects formed the study cohort. Among the participants in the Ohsaki NHI Cohort Study, 16 515 (32.2%) underwent an