[Benchmark dose estimation of polycyclic aromatic hydrocarbons exposure base on Bayesian kernel machine regression].

Objective: To explore benchmark dose (BMD) estimations of polycyclic aromatic hydrocarbons (PAHs) based on Bayesian kernel machine regression (BKMR) . Methods: A total of 155 adult residents of a coking plant in Shanxi Province who were surveyed in summer (June to August) from 2014 to 2019 were selected as the research objects. Fasting elbow vein blood of the subjects was collected in the morning for automatic analysis and detection of blood routine. Morning urine samples were collected for automatic analysis and detection of urine routine and urine creatinine detection. BKMR model combined with BMD method was used to calculate the acceptable doses of PAHs exposure on red blood cell damage in non-occupational population. Results: The concentration of hydroxylpolycyclic aromatic hydrocarbons (OH-PAHs) in the red blood cells abnormal group (n=117) was significantly higher than that in the normal group (n=38) (P<0.01). In the combined effect of OH-PAHs, 2-hydrol-naphthalene contributed the most, and the posterior inclusion probability (PIP) value was 0.9354. When OH-PAHs ≥P(55) concentration, the joint effect on the risk of red blood cell abnormalities increased as the concentration of the OH-PAHs mixture increased. When OH-PAHs were at P(65) and P(75) concentrations, respectively, the risk of red blood cell abnormalities in adults were 3.09 and 4.98 times that of OH-PAHs at P(50) concentrations, respectively. Compared with high concentration, low concentration of OH-PAHs exposure was more sensitive to red blood cell darmage. The acceptable doses of 8 kinds of OH-PAHs were 1.010 μmol/mol Cr (2-hydrol-naphthalene), 0.743 μmol/mol Cr (1-hydrol-naphthalene), 0.901 μmol/mol Cr (2-hydroxy-fluorene) and 0.775 μmol/mol Cr (1-hydroxy-phenanthrene), 0.737 μmol/mol Cr (1-hydroxy-pyrene), 0.607 μmol/mol Cr (9-hydroxy-fluorene), 0.713 μmol/mol Cr (2-hydroxy-phenanthrene) and 0.628 μmol/mol Cr (3-hydroxybenzo[a] pyrene), respectively. Conclusion: OH-PAHs mixture has positive combined effect on red blood cell damage in non-occupational population, and low concentration of OH-PAHs exposure is more sensitive to red blood cell damage. It is recommended that the exposure dose of PAHs should be controlled within 1 μmol/mol Cr.目的： 基于贝叶斯核机回归（Bayesian kernel machine regression，BKMR）探讨多环芳烃混合暴露的基准剂量。 方法： 选取2014至2019年，每年夏季（6至8月）调查的山西省某焦化厂下风向成年居民155人为研究对象。采集研究对象清晨空腹肘静脉血液进行血常规自动分析检测，采集晨尿用于现场尿常规自动分析检测、尿肌酐检测。运用BKMR模型结合基准剂量法计算多环芳烃暴露对非职业人群红细胞损伤的内暴露可接受剂量值。 结果： 红细胞异常组（117人）的多环芳烃羟基代谢混合物（hydroxylpolycyclic aromatic hydrocarbons，OH-PAHs）浓度明显高于红细胞正常组（38人）（P<0.01）。OH-PAHs联合作用中，2羟基萘贡献最大，后验包含概率（PIP）值为0.935 4。当OH-PAHs≥P(55)浓度时，红细胞异常的风险随着OH-PAHs混合物暴露量增加而增大；当OH-PAHs分别处于P(65)和P(75)浓度时，成人发生红细胞异常的风险分别是OH-PAHs处于P(50)的3.09倍和4.98倍。与高浓度相比，低浓度OH-PAHs暴露对红细胞损伤敏感。联合作用下8种OH-PAHs的内暴露可接受剂量分别为1.010 μmol/mol Cr（2羟基萘）、0.743 μmol/mol Cr（1羟基萘）、0.901 μmol/mol Cr（2羟基芴）、0.775 μmol/mol Cr（1羟基菲）、0.737 μmol/mol Cr（1羟基芘）、0.607 μmol/mol Cr（9羟基芴）、0.713 μmol/mol Cr（2羟基菲）和0.628 μmol/mol Cr（3羟基苯芘）。 结论： OH-PAHs混合物对非职业人群红细胞损伤存在正联合效应，低浓度OH-PAHs暴露对血细胞损伤更为敏感。建议多环芳烃的内暴露剂量控制在1 μmol/mol Cr以内。.