Global Burden of Alzheimer’s Disease Attributable to High Fasting Plasma Glucose: Epidemiological Trends and Machine Learning Insights

Yixiao Ma,1,2,* Shuohan Huang,1,* Yahong Dong,1 Qiguan Jin1 1College of Physical Education, Yangzhou University, Yangzhou, Jiangsu, People’s Republic of China; 2GBD Collaborator, Institute for Health Metrics and Evaluation, University of Washington, Seattle, WA, USA*These authors contributed equally to this workCorrespondence: Qiguan Jin, College of Physical Education, Yangzhou University, Huayang West Road, Yangzhou, Jiangsu Province, 225127, People’s Republic of China, Email [email protected]: High fasting plasma glucose (HFPG) is a known risk factor for Alzheimer’s disease (AD). This study aims to analyze global trends in AD death rates and disability-adjusted life years (DALYs) rates attributable to HFPG from 1990 to 2021 and assess the potential of glucose-related biomarkers in predicting cognitive impairment.Methods: Data from the Global Burden of Disease 2021 database were used to analyze AD death rates and DALY rates due to HFPG across 204 countries. All rates were age-standardized. Joinpoint regression, age-period-cohort models, and ARIMA were employed to analyze trends and make future predictions. NHANES data were used to build machine learning models (including logistic regression, SVM, random forests, etc). to evaluate glucose-related biomarkers in predicting cognitive impairment.Results: From 1990 to 2019, global AD death rates attributable to HFPG increased from 2.64 (95% UI: 0.11, 8.38) to 3.73 (95% UI: 0.15, 11.84), with the highest increases in high-income North America, North Africa, and Sub-Saharan Africa. DALY rates also rose globally, from 47.07 (95% UI: 2.72, 126.46) to 66.42 (95% UI: 3.83, 178.85). The greatest impact was observed in females, particularly those aged 80 and above. Joinpoint analysis indicated a significant rise in death rates from 1995 to 2000, followed by a slower increase in recent years. ARIMA model predictions indicate a gradual decline in death rates and DALY rates over the next 15 years. Logistic regression models showed the highest accuracy (90.4%) in predicting cognitive impairment, with 2-hour postprandial glucose and fasting plasma glucose being key predictors.Conclusion: From 1990 to 2021, global AD death rates and DALY rates due to HFPG significantly increased, with a greater burden in females and regions with higher socio-demographic development. Machine learning models are effective tools for identifying individuals at high risk of elevated blood glucose leading to cognitive impairment.Keywords: epidemiology, diabetes, cognitive decline, disability-adjusted life years, machine learning, public health