Leandro F M Rezende1, Gerson Ferrari2, Dong Hoon Lee3, Dagfinn Aune4,5,6,7, Bing Liao8, Wentao Huang8, Jing Nie9, Yafeng Wang10, Edward Giovannucci3,11. 1. Departamento de Medicina Preventiva, Escola Paulista de Medicina, Universidade Federal de São Paulo, São Paulo, Brazil. 2. Universidad de Santiago de Chile (USACH), Escuela de Ciencias de la Actividad Física, el Deporte y la Salud, Santiago, Chile. 3. Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA, USA. 4. Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London, London, UK. 5. Department of Nutrition, Bjørknes University College, Oslo, Norway. 6. Department of Endocrinology, Morbid Obesity and Preventive Medicine, Oslo University Hospital, Oslo, Norway. 7. Unit of Cardiovascular and Nutritional Epidemiology, Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden. 8. School of Nursing, Guangdong Pharmaceutical University, Haizhu District, Guangzhou, China. 9. Department of Sociology & Institute for Empirical Social Science Research, School of Humanities and Social Sciences, Xi'an Jiaotong University, 28 Xianning West Road, Xi'an, 710049, Shaanxi, China. njamm.5818@stu.xjtu.edu.cn. 10. Department of Epidemiology and Biostatistics, School of Health Sciences, Wuhan University, Wuhan, China. 11. Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA, USA.
Abstract
BACKGROUND: Lifestyle risk factors have been associated with increased all-cause and cause-specific mortality, but the influence of reverse causation has been underappreciated as a limitation in epidemiological studies. METHODS: Prospective cohort study including 457,021 US adults from the National Health Interview Survey 1997-2013 linked to the National Death Index records through December 31, 2015. Multivariable Cox models were performed to examine the association of lifestyle risk factors with all-cause and cause-specific mortality. Participants with prevalent diseases and the first 2, 5, 10, and 15 years of follow-up were excluded to account for reverse causation. RESULTS: During 4,441,609 person-years, we identified 60,323 total deaths. Heavy alcohol drinking (HR 1.12; 95% CI 1.08 to 1.16), smoking (HR 1.78; 95% CI 1.74 to 1.83) and lack of physical activity (HR 1.51; 95% CI 1.47 to 1.54) were associated with increased all-cause mortality. Overweight was associated with lower all-cause mortality (HR 0.88; 95% CI 0.86 to 0.90). After exclusion of participants with diseases and first 10 years of follow-up, associations changed to: heavy alcohol drinking (HR 1.31; 95% CI 1.20 to 1.43), smoking (HR 1.99; 95% CI 1.87 to 2.11), lack of physical activity (HR 1.21; 95% CI 1.15 to 1.27), and overweight (HR 1.05; 95% CI 1.00 to 1.10). CONCLUSIONS: Methods to account for reverse causation suggest different effects of reverse causation on the associations between lifestyle risk factors and mortality. Exclusion of participants with diseases at baseline, and exclusion of 5-10 years of follow-up may be necessary to mitigate reverse causation.
BACKGROUND: Lifestyle risk factors have been associated with increased all-cause and cause-specific mortality, but the influence of reverse causation has been underappreciated as a limitation in epidemiological studies. METHODS: Prospective cohort study including 457,021 US adults from the National Health Interview Survey 1997-2013 linked to the National Death Index records through December 31, 2015. Multivariable Cox models were performed to examine the association of lifestyle risk factors with all-cause and cause-specific mortality. Participants with prevalent diseases and the first 2, 5, 10, and 15 years of follow-up were excluded to account for reverse causation. RESULTS: During 4,441,609 person-years, we identified 60,323 total deaths. Heavy alcohol drinking (HR 1.12; 95% CI 1.08 to 1.16), smoking (HR 1.78; 95% CI 1.74 to 1.83) and lack of physical activity (HR 1.51; 95% CI 1.47 to 1.54) were associated with increased all-cause mortality. Overweight was associated with lower all-cause mortality (HR 0.88; 95% CI 0.86 to 0.90). After exclusion of participants with diseases and first 10 years of follow-up, associations changed to: heavy alcohol drinking (HR 1.31; 95% CI 1.20 to 1.43), smoking (HR 1.99; 95% CI 1.87 to 2.11), lack of physical activity (HR 1.21; 95% CI 1.15 to 1.27), and overweight (HR 1.05; 95% CI 1.00 to 1.10). CONCLUSIONS: Methods to account for reverse causation suggest different effects of reverse causation on the associations between lifestyle risk factors and mortality. Exclusion of participants with diseases at baseline, and exclusion of 5-10 years of follow-up may be necessary to mitigate reverse causation.