Mitsuaki Sawano1, Shun Kohsaka2, Tomonori Okamura3, Taku Inohara1, Daisuke Sugiyama3, Makoto Watanabe4, Yasuyuki Nakamura5, Aya Higashiyama6, Aya Kadota7, Nagako Okud8, Yoshitaka Murakami9, Takayoshi Ohkubo10, Akira Fujiyoshi11, Katsuyuki Miura7, Akira Okayama12, Hirotsugu Ueshima7. 1. Department of Cardiology, Keio University School of Medicine, Tokyo, Japan. 2. Department of Cardiology, Keio University School of Medicine, Tokyo, Japan. Electronic address: kohsaka@cpnet.med.keio.jp. 3. Department of Preventive Medicine and Public Health, Keio University School of Medicine, Tokyo, Japan. 4. Department of Preventive Cardiology, National Cerebral and Cardiovascular Center, Suita, Japan. 5. Department of Food Science and Human Nutrition, Faculty of Agriculture, Ryukoku University, Otsu, Japan. 6. Department of Preventive Medicine and Epidemiologic Informatics, National Cerebral and Cardiovascular Center, Osaka, Japan. 7. Center for Epidemiologic Research in Asia, Shiga University of Medical Science, Shiga, Japan; Department of Public Health, Shiga University of Medical Science, Otsu, Japan. 8. Department of Cardiology, Keio University School of Medicine, Tokyo, Japan; Department of Health and Nutrition, University of Human Arts and Sciences, Saitama, Japan. 9. Department of Medical Statistics, Toho University School of Medicine, Tokyo, Japan. 10. Department of Hygiene and Public Health, Teikyo University School of Medicine, Tokyo, Japan. 11. Department of Public Health, Shiga University of Medical Science, Otsu, Japan. 12. Research Institute for Lifestyle-Related Disease Prevention, Tokyo, Japan.
Abstract
BACKGROUND AND AIMS: The European Society of Cardiology developed prediction models (SCORE) for low- and high-risk populations in the European countries. However, whether or not these models are valid in different ethnicities is unknown. We aimed to evaluate the performance of the low-risk SCORE model in the general Japanese population. METHODS: Healthy middle-aged Japanese participating in the NIPPON DATA80 cohort had been observed. The predicted 10-year cardiovascular death risk was calculated using the low-risk SCORE model for the overall population as well as for each gender individually. The model performance of the low-risk SCORE model was evaluated with the Harrel's c-statistics for discrimination and the Grønnesby and Borgan goodness-of-fit test for calibration. RESULTS: A total of 4842 participants aged 40-64 years old and 47,606 person-years were evaluated in our study. 203 (4.19%) died within the ten-years of follow-up and 44 (0.91%) CV deaths were observed. The low-risk SCORE model in the overall population had reasonable discrimination (c statistics 0.72, 95% CI 0.71-0.73) but poor calibration (R(2), 0.67, Chi-square value 6.15, p = 0.01). Discrimination was reasonable in both men (c statistics 0.71, 95% CI 0.69-0.73) and women (c statistics 0.71, 95% CI 0.70-0.73). However, calibration was poor in men (R(2), 0.22, Chi-square value 0.749, p = 0.38) compared to women (R(2), 0.96, Chi-square value 1.39, p = 0.24). CONCLUSIONS: Although the low-risk SCORE model performs reasonably well in women, the SCORE models generally overestimated the risk of cardiovascular death risk in the Japanese general population.
BACKGROUND AND AIMS: The European Society of Cardiology developed prediction models (SCORE) for low- and high-risk populations in the European countries. However, whether or not these models are valid in different ethnicities is unknown. We aimed to evaluate the performance of the low-risk SCORE model in the general Japanese population. METHODS: Healthy middle-aged Japanese participating in the NIPPON DATA80 cohort had been observed. The predicted 10-year cardiovascular death risk was calculated using the low-risk SCORE model for the overall population as well as for each gender individually. The model performance of the low-risk SCORE model was evaluated with the Harrel's c-statistics for discrimination and the Grønnesby and Borgan goodness-of-fit test for calibration. RESULTS: A total of 4842 participants aged 40-64 years old and 47,606 person-years were evaluated in our study. 203 (4.19%) died within the ten-years of follow-up and 44 (0.91%) CV deaths were observed. The low-risk SCORE model in the overall population had reasonable discrimination (c statistics 0.72, 95% CI 0.71-0.73) but poor calibration (R(2), 0.67, Chi-square value 6.15, p = 0.01). Discrimination was reasonable in both men (c statistics 0.71, 95% CI 0.69-0.73) and women (c statistics 0.71, 95% CI 0.70-0.73). However, calibration was poor in men (R(2), 0.22, Chi-square value 0.749, p = 0.38) compared to women (R(2), 0.96, Chi-square value 1.39, p = 0.24). CONCLUSIONS: Although the low-risk SCORE model performs reasonably well in women, the SCORE models generally overestimated the risk of cardiovascular death risk in the Japanese general population.