Evaluation Of Combined Exposure To Environmental Heavy Metals And Its Impact On Important Health Outcomes

Background: Occupational and environmental exposure to toxic metals has been proved to be an important risk factor for human health,among which the exposure to lead,cadmium,mercury and arsenic is particularly harmful to human health.Extensive studies have shown that exposure to heavy metals may lead to various health problems,including cardiovascular disease,nerve injury,kidney injury,diabetes and even cancer.At present,most studies focus on the marginal effect of single metal exposure on health outcomes.However,considering the possible confounding caused by the complex interaction mechanism between heavy metals,our research focuses on the risk of metal interaction on health outcomes,providing new insights into the relationship between metal exposure and cardiovascular,renal and respiratory diseases.Objective: Using unsupervised machine learning method,the NHANES population was divided into different risk groups based on the concentrations of lead,cadmium,mercury and arsenic in urine and blood,and the association between combined exposure to heavy metals and different types of health outcomes was studied according to the risk stratification information of the population.Methods: We analyzed the data of 9962 subjects in six consecutive cycles（2003-2004 to 2013-2014）of the NHANES cohort study.Based on the concentrations of three heavy metals in urine（total arsenic,lead and cadmium）and three heavy metals in blood（lead,cadmium and mercury）,the population was stratified by two-step analysis method,and the correlation analysis was made with important health outcomes and related indicators.In the first step,the unsupervised machine learning clustering algorithm k-medoids method is applied to the population clustering of NHANES data,and the population is divided into different subgroups based on the metal concentration in urine or blood.Step two,Risk subgroups were compared with different health outcomes and related indicators on the premise of controlling seven possible confounding factors（age,gender,race / ethnicity,education,smoking status,BMI and urinary creatinine）（GGT,systolic blood pressure,diastolic blood pressure,hypertension,total mortality,heart disease,malignancy,chronic lower respiratory tract disease,cerebrovascular disease,Alzheimer’s disease,diabetes,influenza and pneumonia,and nephritis related mortality）were analyzed.result: Based on the levels of heavy metals in blood and urine,k-medoids algorithm divides NHANES population into two subgroups.The group with higher concentration is defined as "high exposure" group,and the group with lower concentration is defined as "low exposure" group.Association analysis with different health outcomes and related indicators showed that based on blood and urine metal concentrations,the total mortality,mortality caused by malignant tumors γ-Glutamyltransferase（GGT）was significantly higher.In addition,the high exposure group based on blood level was also significantly associated with mortality related to systolic blood pressure,hypertension,heart disease and chronic lower respiratory diseases,while the high exposure group based on urine concentration was significantly associated with higher mortality related to nephritis.conclusion: We propose an unsupervised clustering method to divide the population into high exposure group and low exposure group according to the combined exposure of heavy metals.The high exposure group characterized by higher metal concentration had significantly higher γ-Glutamyltransferase（GGT）,systolic blood pressure,diastolic blood pressure and mortality indicate the harmful effects of exposure to these heavy metals.NHANES population stratification based on exposure pattern provides a reference method for studying the impact of metal exposure on health outcomes.