Abstract：
The impact of exposure to metals on chronic kidney disease (CKD) has only been investigated in two-way or single metal interactions in previous studies. We investigated the associations between five single metals in blood and their mixed exposure and CKD by using the machine learning approach. Relevant data were extracted from the National Health and Nutrition Examination Survey (NHANES 2011-2020), and the level of five metals in blood detected by inductively coupled plasma mass spectrometry was considered as exposures, namely, cadmium (Cd), lead (Pb), total mercury (Hg), manganese (Mn), and selenium (Se). The correlations between individual metal and metal mixtures and CKD were then evaluated by survey-multivariable logistic regression (SMLR), generalized weighted quantile sum (WQS), and Bayesian kernel machine regression (BKMR). Altogether, our study included 12,412 participants representing 572.6 million non-institutionalized US adults. Several single metals with the high quartile of exposure showed a positive association with the CKD ratio including Cd [(AOR = 1.873, 95% CI: 1.537, 2.284), Q4], Pb [(AOR = 1.559, 95% CI: 1.295, 1.880), Q4], and total Hg [(AOR = 1.169, 95% CI: 1.018, 1.343), Q2], while Mn [(AOR = 0.796, 95% CI: 0.684, 0.927), Q2] and Se [(AOR = 0.805, 95% CI: 0.664, 0.976), Q4] were negatively associated with the CKD ratio. In light of the positive fit of the WQS regression model, a significantly positive correlation was found between mixed metals and CKD (AOR = 1.373, 95% CI: 1.224, 1.539) after full covariate adjustment, and a similar finding was also detected in the BKMR model. Our study revealed that each single metal including Cd, Pb, and total Hg might have a positive association with CKD while this association was negative for both Mn and Se. The five metals might have a positive joint effect on CKD.
摘要：
在先前的研究中，仅在双向或单金属相互作用中研究了接触金属对慢性肾脏疾病（CKD）的影响。我们使用机器学习方法研究了血液中五种单一金属及其混合暴露与CKD之间的关联。相关数据来自国家健康和营养检查调查(NHANES2011-2020)，通过电感耦合等离子体质谱法检测到的血液中五种金属的水平被认为是暴露量，即,镉(Cd),铅(Pb),总汞(Hg)，锰(Mn),硒（Se）。然后通过调查多变量逻辑回归（SMLR）评估单个金属和金属混合物与CKD之间的相关性，广义加权分位数和(WQS)，和贝叶斯核机回归(BKMR)。总之,我们的研究包括12,412名参与者,代表5.726亿非住院的美国成年人.具有高四分位数暴露的几种单一金属与CKD比率（包括Cd）呈正相关[（AOR=1.873，95％CI：1.537，2.284），Q4],Pb[(AOR=1.559，95%CI：1.295，1.880)，Q4],和总汞[(AOR=1.169，95%CI：1.018，1.343)，Q2],而Mn[(AOR=0.796，95%CI：0.684，0.927)，Q2]和Se[(AOR=0.805，95%CI：0.664，0.976)，Q4]与CKD比率呈负相关。鉴于WQS回归模型的正拟合，经过完全协变量调整后，混合金属与CKD之间存在显着正相关（AOR=1.373，95％CI：1.224，1.539），在BKMR模型中也检测到类似的发现。我们的研究表明，包括Cd在内的每种金属，Pb,总汞可能与CKD呈正相关，而Mn和Se均呈负相关。五种金属可能对CKD具有积极的联合作用。
