Scholarlink

<p>Biological age is an important measure of aging that reflects an individual’s physical health and is linked to various diseases. Current prediction models are still limited in precision, and the risk factors for accelerated aging remain underexplored. Therefore, we aimed to develop a precise biological age and assess the impact of socio-demographic and behavioral patterns on the aging process.We utilized Deep Neural Networks (DNN) to construct biological age from participants with physical examinations, blood samples, and questionnaires data from the China Kadoorie Biobank (CKB) between June 2004 and December 2016. △age, calculated as the residuals between biological age and chronological age, was used to investigate the associations of age acceleration with diseases. Socio-demographics (gender, education attainment, marital status, household income) and lifestyle&#xa0;characteristics (body mass index [BMI], smoking, drinking, physical activity, and sleep) were also assessed to explore their impact on age acceleration. 18,261 participants aged 57 ± 10&#xa0;years were included in this study. The DNN-based biological age model&#xa0;has demonstrated accurate predictive performance, achieving a mean absolute error of 3.655&#xa0;years. △age was associated with increased risks of various morbidity and mortality, with the highest associations found for circulatory and respiratory diseases, with hazard ratios of 1.033 (95% CI: 1.023, 1.042) and 1.078 (95% CI: 1.027, 1.130), respectively. Socio-demographics, including being female, lower education, widowed or divorced, and low household income, along with behavioral patterns, such as being underweight, insufficient physical activity, and poor sleep, were associated with accelerated&#xa0;aging. Our DNN model is capable of constructing a&#xa0;precise biological age using commonly collected data. Socio-demographics and lifestyle factors were associated with accelerated aging, highlighting that addressing modifiable risk factors can effectively slow age acceleration and reduce disease risk, providing valuable insights for interventions to promote healthy aging.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Construction and validation of a DNN-based biological age and its influencing factors in the China Kadoorie Biobank

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年3月7日

查看全文
  • Yushu Huang,
  • Lijuan Da,
  • Yue Dong,
  • Zihan Li,
  • Yuan Liu,
  • Zilin Li,
  • Xifeng Wu,
  • Wenyuan Li

摘要

Biological age is an important measure of aging that reflects an individual’s physical health and is linked to various diseases. Current prediction models are still limited in precision, and the risk factors for accelerated aging remain underexplored. Therefore, we aimed to develop a precise biological age and assess the impact of socio-demographic and behavioral patterns on the aging process.We utilized Deep Neural Networks (DNN) to construct biological age from participants with physical examinations, blood samples, and questionnaires data from the China Kadoorie Biobank (CKB) between June 2004 and December 2016. △age, calculated as the residuals between biological age and chronological age, was used to investigate the associations of age acceleration with diseases. Socio-demographics (gender, education attainment, marital status, household income) and lifestyle characteristics (body mass index [BMI], smoking, drinking, physical activity, and sleep) were also assessed to explore their impact on age acceleration. 18,261 participants aged 57 ± 10 years were included in this study. The DNN-based biological age model has demonstrated accurate predictive performance, achieving a mean absolute error of 3.655 years. △age was associated with increased risks of various morbidity and mortality, with the highest associations found for circulatory and respiratory diseases, with hazard ratios of 1.033 (95% CI: 1.023, 1.042) and 1.078 (95% CI: 1.027, 1.130), respectively. Socio-demographics, including being female, lower education, widowed or divorced, and low household income, along with behavioral patterns, such as being underweight, insufficient physical activity, and poor sleep, were associated with accelerated aging. Our DNN model is capable of constructing a precise biological age using commonly collected data. Socio-demographics and lifestyle factors were associated with accelerated aging, highlighting that addressing modifiable risk factors can effectively slow age acceleration and reduce disease risk, providing valuable insights for interventions to promote healthy aging.