Wang Jianing, Zeng Hao-Long, Hui Ying, Cui Feipeng, Ma Yudiyang, Tang Linxi, Xing Meiqi, Zheng Lei, Chen Ning, Zhao Xinru, Li Dankang, Liu Run, Chen Shuohua, Cheng Liming, Wu Shouling, Wang Zhenchang, Tian Yaohua
Ministry of Education Key Laboratory of Environment and Health, and State Key Laboratory of Environmental Health (Incubating) (J.W., F.C., Y.M., L.T., M.X., L.Z., N.C., X.Z., D.L., R.L., Y.T.), School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China.
Department of Maternal and Child Health (J.W., F.C., Y.M., L.T., M.X., L.Z., N.C., X.Z., D.L., R.L., Y.T.), School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China.
Stroke. 2025 Aug;56(8):2138-2146. doi: 10.1161/STROKEAHA.124.050048. Epub 2025 May 8.
Evidence about the impact of multiple metal exposure on brain neuroimaging metrics remains limited. We aim to investigate the effects of single and mixed metal exposure on white matter hyperintensities (WMHs).
This cross-sectional study included 1183 subjects without stroke history from the META-KLS (Multi-modality Medical Imaging Study Based on Kailuan Study), which is an existing prospective cohort in Tangshan, China. Plasma metal levels, including Mg, Ca, V, Mn, Co, Ni, Cu, Zn, As, Se, Rb, Cs, Tl, Pb, and Cd, were measured using an inductively coupled plasma mass spectrometer. Ordinal and binary logistic regression models were used to examine the effects of metal exposure on the WMH burden, deep white matter hyperintensity, and periventricular white matter hyperintensity. All metal concentrations were naturally log-transformed to reduce skewness and were analyzed as both continuous and tertile forms. Weighted quantile sum regression, quantile-based g-computation model, and Bayesian Kernel Machine Regression were used in the metal mixture analysis.
Compared with the first tertile, the adjusted odds ratios and 95% CIs for the WMH burden in the third tertile were 1.57 (1.05-2.34) for As, 2.01 (1.28-3.18) for Cu, 1.68 (1.14-2.50) for V, 1.61 (1.07-2.44) for Cs, and 1.56 (1.04-2.34) for Tl (all for trend<0.05). Additionally, Pb, Se, and Mg showed significant positive associations with WMH burden exclusively in continuous analysis, with odds ratios of 1.27 (1.02-1.56) for Pb, 1.32 (1.07-1.61) for Se, and 1.27 (1.04-1.55) for Mg for per interquartile range increase in Ln-transformed metal concentrations. The weighted quantile sum index revealed a significant positive correlation with WMH burden risk (each interquartile range increment in the weighted quantile sum index was associated with 60% higher odds for WMH burden [95% CI, 1.09-2.34]). The primary contributors to the weighted quantile sum index were As (39.4%), followed by Pb (12.5%) and Cu (11.3%). The bivariate exposure-response relationships suggested potential interactions between As and Cu, as well as As and Co.
There were positive associations between individual exposures to As, Pb, Cu, V, Se, Cs, Tl, and Mg, and mixed metal exposure with WMH burden among the Chinese population, strengthening the evidence of detrimental effects of specific metals on brain health.
关于多种金属暴露对脑影像指标影响的证据仍然有限。我们旨在研究单一和混合金属暴露对白质高信号（WMH）的影响。
这项横断面研究纳入了来自中国唐山一项现有的前瞻性队列研究——开滦研究衍生的META-KLS（基于开滦研究的多模态医学影像研究）中的1183名无卒中病史的受试者。使用电感耦合等离子体质谱仪测量血浆金属水平，包括镁（Mg）、钙（Ca）、钒（V）、锰（Mn）、钴（Co）、镍（Ni）、铜（Cu）、锌（Zn）、砷（As）、硒（Se）、铷（Rb）、铯（Cs）、铊（Tl）、铅（Pb）和镉（Cd）。采用有序和二元逻辑回归模型来检验金属暴露对WMH负担、深部白质高信号和脑室周围白质高信号的影响。所有金属浓度均进行自然对数转换以减少偏度，并以连续和三分位数形式进行分析。在金属混合物分析中使用加权分位数和回归、基于分位数的g计算模型和贝叶斯核机器回归。
与第一三分位数相比，第三三分位数中，砷（As）导致WMH负担的调整比值比及95%置信区间为1.57（1.05 - 2.34），铜（Cu）为2.01（1.28 - 3.18），钒（V）为1.68（1.14 - 2.50），铯（Cs）为1.61（1.07 - 2.44），铊（Tl）为1.56（1.04 - 2.34）（所有趋势P值<0.05）。此外，铅（Pb）、硒（Se）和镁（Mg）仅在连续分析中显示与WMH负担存在显著正相关，经自然对数转换的金属浓度每增加一个四分位数间距，铅的比值比为1.27（1.02 - 1.56），硒为1.32（1.07 - 1.61），镁为1.27（1.04 - 1.55）。加权分位数和指数显示与WMH负担风险存在显著正相关（加权分位数和指数每增加一个四分位数间距，WMH负担的比值比高60% [95%置信区间，1.09 - 2.34]）。加权分位数和指数的主要贡献者为砷（39.4%），其次是铅（12.5%）和铜（11.3%）。双变量暴露 - 反应关系表明砷与铜以及砷与钴之间可能存在相互作用。
在中国人群中，个体暴露于砷、铅、铜、钒、硒、铯、铊和镁以及混合金属暴露与WMH负担之间存在正相关，这进一步证明了特定金属对脑健康的有害影响。
