Related Experiment Video
Updated: Sep 1, 2025
Ameliorating Osteoarthritis in Mice Using Silver Nanoparticles
Published on: June 2, 2023
Identification for heavy metals exposure on osteoarthritis among aging people and Machine learning for prediction: A
Fang Xia1, Qingwen Li1, Xin Luo1
1Department of Public Health, Wuhan Fourth Hospital, Wuhan, China.
Objective:
Heavy metals are present in many environmental pollutants, and have cumulative effects on the human body through water or food, which can lead to several diseases, including osteoarthritis (OA). In this research, we aimed to explore the association between heavy metals and OA.
Methods:
We extracted 18 variables including age, gender, race, education level, marital status, smoking status, body mass index (BMI), physical activity, diabetes mellitus, hypertension, poverty level index (PLI), Lead (Pb), cadmium (Cd), mercury (Hg), selenium (Se), manganese (Mn), and OA status from National Health and Nutrition Examination Survey (NHANES) 2011-2020 datasets.
Results:
In the baseline data, the t test and Chi-square test were conducted. For heavy metals, quartile description and limit of detection (LOD) were adopted. To analyze the association between heavy metals and OA among elderly subjects, multivariable logistic regression was conducted and subgroup logistic by gender was also carried out. Furthermore, to make predictions based on heavy metals for OA, we compared eight machine learning algorithms, and XGBoost (AUC of 0.8, accuracy value of 0.773, and kappa value of 0.358) was the best machine learning model for prediction. For interactive use, a shiny application was made (https://alanwu.shinyapps.io/NHANES-OA/).
Conclusion:
The overall and gender subgroup logistic regressions all showed that Pb and Cd promoted the prevalence of OA while Mn could be a protective factor of OA prevalence among the elderly population of the United States. Furthermore, XGBoost model was trained for OA prediction.
