<p>Objectives: The purpose of this study is to confirm the secular changes in smoking rate in each prefecture and in the whole country of Japan, and to ecologically examine the associations between secular changes in smoking rate and those in death rates from selected major causes among prefectures in Japan.</p><p>Methods: Smoking rate (SR) was calculated from Comprehensive Survey of Living Conditions (the large-scale survey conducted every three years). Smoking data in 2001, 2004, 2007, 2010, 2013, 2016, and 2019 were collected from e-Stat, which is portal site of official statistics of Japan. Cause-specific death rate (CSDR) was calculated from data of Vital Statistics and Population Estimates, etc. from 2001 to 2019, which were collected from e-Stat. The age-standardized SR and CSDR (standardized to the 2015 model population of Japan by the investigation committee for revision) were calculated between 2001-2019. We investigated the average annual percent change (AAPC) of age-standardized SR and CSDR in each prefecture and in the whole country of Japan from 2001 to 2019. Then, the associations between AAPC of the age-standardized SR and the AAPCs of age-standardized CSDR from selected major causes, were examined as an ecological study with all prefectures.</p><p>Results: Age-standardized SR in the whole country significantly decreased after 2001 in both genders, with AAPC of -2.8% (95% confidence interval: -2.9, -2.6) in men and -2.3% (-2.7, -1.9) in women. Age-standardized SR significantly decreased after 2001 in all prefectures in men, however in women it significantly decreased in prefectures excluding Aomori, Tottori, Saga and Kagoshima prefectures. The associations of age-standardized SR and CSDR among the prefectures were different according to the causes of death and genders. We confirmed that the correlations between AAPC of age-standardized SR and the AAPCs of age-standardized CSDR from all-cause (Pearson’s correlation coefficient: men 0.456, women 0.439), malignant neoplasms (men 0.359, women 0.431), and cerebrovascular diseases (men 0.460, women 0.331) were statistically significant in both genders. These correlations mean that the larger the average annual decreasing rates of age-standardized SR, the larger the average annual decreasing rates of age-standardized death rates from all-cause, malignant neoplasms, and cerebrovascular diseases in both genders.</p><p>Conclusion: The AAPC of age-standardized SR and AAPCs of the age-standardized death rates from all-cause, malignant neoplasms, and cerebrovascular diseases were significantly positively correlated among prefectures in both genders. Our analysis method would be one of the methods to confirm the association between secular changes in cause and result.</p>