基于倾向性评分匹配法和逆概率处理加权法探索血清镍与口腔癌发病的关联

Association between serum nickel and oral cancer incidence using propensity score matching and inverse probability of treatment weighting

[背景]血清镍(Ni)与口腔癌(OC)发生关联尚不明确,且既往研究多为未控制组间混杂因素的观察性研究。[目的]以倾向性评分匹配法(PSM)和逆概率处理加权法(IPTW)为基础,评估血清镍与口腔癌发病之间的关联。[方法]选取2011年11月—2019年5月期间福建医科大学附属第一医院经组织病理学确诊的新发口腔癌患者456例,选择同期前往医院和社区的健康体检人群作为对照组,共纳入1410例。基于电感耦合等离子体质谱法测定血清样本中Ni质量浓度(简称浓度)。采用1∶1 PSM(卡钳值为0.02)匹配病例-对照,基于IPTW对病例组与对照中的研究对象加权后进行后续分析。研究对象的一般特征通过χ^(2)检验及标准均数差完成匹配前后的均衡性检验,利用限制性立方样条探索血清镍与口腔癌两者非线性的剂量-反应关系,同时应用条件logistic回归和加权logistic回归分析血清镍元素与口腔癌发病的关联。[结果]经PSM和IPTW控制组间协变量后,剂量-反应曲线显示,随着血清Ni浓度的增加,患口腔癌的风险呈先下降后上升的趋势。PSM结果显示,与对照组相比,血清Ni浓度在0.09~16.80μg·L^(−1)组的OC发病风险与其浓度呈负相关(OR=0.36,95%CI:0.24~0.54),Ni浓度>16.80μg·L^(−1)的OC发病风险与其浓度呈正相关(OR=5.43,95%CI:2.76~10.68);IPTW结果显示,血清镍浓度在0.09~20.55μg·L^(−1)时,口腔癌的罹患风险与血清镍浓度呈负相关(OR=0.39,95%CI:0.29~0.52),而Ni浓度>20.55μg·L^(−1)时可升高口腔癌的罹患风险,OR及95%CI为5.54(3.62~8.49)。[结论]血清Ni水平与OC发病风险呈J形关系,高浓度血清Ni浓度(>20.55μg·L^(−1))可能为OC发病的危险因素。

[Background]The association between serum nickel(Ni)and oral cancer incidence is unclear and most of the previous studies were observational studies that did not control for confounding factors between groups.[Objective]To assess the correlation of serum Ni with oral cancer incidence based on propensity score matching(PSM)and inverse probability of treatment weighting(IPTW).[Methods]A cohort of 456 newly diagnosed oral cancer patients was recruited from the First Hospital of Fujian Medical University during November 2011 to May 2019,and residents ordered their health check-up in hospitals or local community health centers over the same period were selected as a control group,which included a total of 1410 participants.Serum Ni was evaluated by inductively coupled plasma mass spectrometry.Case-control pairs were selected using a 1:1 PSM(caliper value of 0.02),and the study subjects in the case group and control group were weighted for subsequent analysis by IPTW.The general characteristics of the study subjects were tested for equilibrium before and after matching by chi-square test and standardized mean difference(SMD).This was followed by exploring the potential nonlinear dose-response relationship between serum Ni and oral cancer using restricted cubic splines as well as analyzing the association between serum Ni and oral cancer incidence by conditional logistic regression and weighted logistic regression.[Results]After controlling for between-group covariates by PSM and IPTW,the dose-response curves demonstrated that the risk of developing oral cancer tended to decline and then increase with the increasing serum Ni level.The outcome of the analysis using PSM demonstrated that as compared to the control group,the risk of developing oral cancer in the 0.09-16.80μg·L^(−1)serum Ni group was negatively correlated with serum Ni level(OR=0.36,95%CI:0.24-0.54),whereas the risk of developing oral cancer in the>16.80μg·L^(−1)serum Ni group was positively correlated with serum Ni level(OR=5.43,95%CI:2.76-10.68).After applying IPTW,a negative association was found between the risk of oral cancer and serum Ni concentration within a serum Ni window ranging from 0.09 to 20.55μg·L^(−1)(OR=0.39,95%CI:0.29-0.52),while a positive association with an OR and 95%CI of 5.54(3.62-8.49)for the Ni concentration>20.55μg·L^(−1).[Conclusion]In this study,a J-shaped relationship between serum Ni concentration and the risk of developing oral cancer is found,which shows that high serum Ni concentration(>20.55μg·L^(−1))may be a risk factor for oral cancer.