探索提高化学发光法检测梅毒螺旋体准确性和效率的方法。

回顾性收集华中科技大学同济医学院附属同济医院3个院区2019年1月至2021年12月门诊和住院的496 013例患者的化学发光微粒子免疫分析法（CMIA）检测梅毒抗体和梅毒螺旋体颗粒凝集试验（TPPA）结果，使用SPSS 23.0对不同年份TPPA阳性率差异和TPPA法检测阳性样本的S/CO值分布进行统计学分析，并利用受试者工作特征（ROC）曲线确定CMIA法的最佳S/CO值。

根据样本采集时间分组，TPPA的阳性率呈逐年下降的趋势，但是不同年份TPPA阳性率差异无统计学意义（χ²=4.847、P=0.089）。496 013例样本经CMIA筛查试验检测梅毒抗体阳性样本8696例。经TPPA确证阳性样本6875例，CMIA初筛试验假阳性率为0.37%，阳性预测值为79.06%，特异性为99.63%。根据S/CO值可将CMIA阳性样本分为1≤S/CO<2、2≤S/CO<3、3≤S/CO<4、4≤S/CO<18、S/CO≥18共5组，当S/CO≥18时，TPPA的阳性预测值为100%，当4≤S/CO<18时，TPPA的阳性预测值为95.81%。5组间的阳性预测值差异有统计学意义 （χ²=3 825.005、P<0.05）。当S/CO=4.635时，ROC曲线下面积为0.923 9，约登指数最大为0.729 8，敏感度为80.12%，特异性为92.86%，阳性似然比为11.22，阳性预测值为97.69%。

华中科技大学同济医学院附属同济医院梅毒感染人数已呈现出逐年下降的趋势。CMIA法具有较高的特异性但存在假阳性，对于1≤S/CO<4.635的样本应当用TPPA来确认，降低假阳性率；对于S/CO≥4.635的样本，可以选用胶体金等便捷、快速的梅毒特异性抗体检测方法复核，复核结果不一致，再加做TPPA进行确证，可缩短报告周期，提高检测效率，但是否可应用于临床需要进一步的评估。

To explore methods to improve the accuracy and efficiency of chemiluminescence immunoassay in the detection of treponema pallidum.

A retrospective collection of 496 013 outpatient and inpatient samples from January 2019 to December 2021 in 3 campuses of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology was conducted. Primary screening test for syphilis antibodies used the chemiluminescence microparticle immunoassay (CMIA) and confirmation test used treponema pallidum particle agglutination (TPPA). Statistical analysis of the differences of the TPPA positivity rates in different years and the distribution of S/CO values of samples tested positive by the TPPA was performed by SPSS 23.0 software. The receiver operating characteristic curve (ROC) was used to determine the optimal S/CO value for the CMIA.

When grouped according to sample collection time, the positive rate of TPPA showed a decreasing trend year by year, but there was no statistically significant difference in TPPA positive rates among different years (χ²=4.847, P=0.089). Of the 496 013 cases tested by CMIA, 8696 were positive for syphilis antibodies. Among them, 6875 samples were confirmed as positive by TPPA, resulting in a false-positive rate of 0.37%, a positive predictive value of 79.06%, and a specificity of 99.63% for the CMIA initial screening test. According to the S/CO value, the positive samples tested by CMIA can be divided into 5 groups ranging from 1≤S/CO<2、2≤S/CO<3、3≤S/CO<4、4≤S/CO<18 and S/CO≥18. The positive predictive value of TPPA was 100% when S/CO≥18, and 95.81% when 4≤S/CO<18. The difference in positive predictive values between the 5 groups was statistically significant (χ²=3 825.005, P<0.05). When the S/CO=4.635, the area under the ROC curve was 0.923 9, the maximum Jorden index was 0.729 8, the sensitivity was 80.12%, the specificity was 92.86%, the positive likelihood ratio was 11.22, and the positive predictive value was 97.69%.

The number of syphilis infections has shown a decreasing trend year by year. The CMIA has a high specificity, but there are false positives. For samples 1≤S/CO<4.635 should be confirmed by TPPA to reduce the rate of misdiagnosis, and for samples with S/CO≥4.635, convenient and rapid syphilis-specific antibody assays such as colloidal gold can be used for review, and if the review is inconsistent, TPPA should be added to confirm the test, which can reduce the reporting cycle and improve the efficiency of the test. However, further evaluation is needed to determine whether this conclusion can be applied in clinical practice.