【作者】 邱红燕；

【导师】 袁聚祥； 范红敏；

【作者基本信息】 华北煤炭医学院 ， 流行病与卫生统计学， 2008， 硕士

【摘要】 目的2006年我国煤炭产量已达26亿吨,煤炭从业人员在500万左右,居世界第一。截至2006年底,我国累计检出尘肺患者达69万,其中煤工尘肺占尘肺病例总数的51%左右。煤工尘肺是煤炭行业最严重的职业病,且患病人数在不断增长,并将存在很长一段时间。煤工尘肺的发生是一个连续的过程,其确诊主要依据X线影像学的表现,但X线影像学的变化只有在病变达到一定程度后才能反映出来。患者早期肺功能变化也不明显,可在相当长时间内无明显自觉症状,且影像学表现与症状、体征及肺功能变化也不一致。因此,实现ILO和WHO共同提出的至2030年全球消除矽肺计划的目标,能否找到一些简便迅速的方法用于尘肺筛查或用于X线影像学变化发生前的筛检,是目前的一个新任务。众多的研究都提示,氧化损伤程度与尘肺的发生发展有密切联系,探讨氧化损伤指标是否可作为筛检煤工尘肺的指标是目前研究的方向之一。本研究根据流行病学特征选择接尘工人作为对照进行研究,对选取的氧化损伤指标的筛检价值进行比较和评价,希望探索筛选出可以作为筛检煤工尘肺的敏感氧化损伤生化指标,并对最终筛选确定的指标进行各种联合,构建判别方程,比较其筛检价值,以期从氧化损伤方面探寻筛检煤工尘肺的最佳方法,以利于发现煤矿接尘工人中煤工尘肺的高危个体,为从病因学的角度采取提高高危个体机体抗氧化能力及改善其氧化损伤程度的预防措施,减少或延缓疾病的发生,达到一级预防的目的提供线索和依据。材料和方法严格按照病例和对照排除及纳入标准,从河南平顶山和山东枣庄矿业集团公司某矿选取全部有后前位X线胸片,并由有资质认证的尘肺诊断组严格按照我国《尘肺病诊断标准》（GBZ70-2002）作出诊断的I期尘肺患者314人（其中河南平顶山某矿178例,山东枣庄某矿136例）和接尘工人560人（其中河南平顶山某矿278例,山东枣庄某矿282例）中,根据两矿实际构成随机抽取已确诊的I期尘肺患者112人（其中河南平顶山矿63人,山东枣庄某矿49人）和经过体检的与病例组年龄相近的健康接尘工人144人（其中河南平顶山某矿71人,山东枣庄某矿73人）作为研究对象。抽取空腹静脉血5ml, 2%EDTA抗凝,在室温下分别采用硝酸还原酶法、硫代巴比妥酸显色法、黄嘌呤氧化酶法、二硫代二硝基苯甲酸法、尿酸还原酶法、化学比色法（一氧化氮合成酶分型试剂盒）、对苯二胺盐酸盐法及菲罗啉比色法测定病例组和对照组血浆氧化损伤指标NO、MDA、CuZn-SOD、GSH-Px、UA、iNOS、CP及T-AOC的含量或活力。本研究现场工作的调查员由教师、本人和本专业的研究生同学担任,经过培训,统一调查内容及方法。在资料收集、整理、分析过程中实施了严格的质量控制。实验室工作在可靠预试验基础上进行。所有资料数据经整理后用Excel录入,SPSS 13.0统计软件进行数据分析。计量资料采用t检验进行分析,计数资料采用χ2检验进行分析,采用多因素Logistic回归分析控制主要影响因素,运用ROC曲线、判别分析等比较筛检效力的优劣。结果1基本情况病例组与对照组平均年龄分别为（57.7±6.3）岁和（56.1±7.2）岁,平均累积接尘工龄分别为28.5（11-43）年和28.3（9-48）年。病例组与对照组在年龄、累积接尘工龄、工种、吸烟、饮酒等方面,差异均无统计学意义（P > 0.05）。2病例组与对照组氧化损伤程度比较病例组与对照组氧化损伤指标T-AOC、MDA、iNOS、NO、CuZu-SOD、GSH-Px、CP和UA分别进行比较,发现两组CP含量差异无统计学意义（P > 0.05）,其他各项指标两组比较差异均有统计学意义（P < 0.05）。3煤工尘肺发病主要影响因素的多因素Logistic回归分析多因素Logistic回归分析结果显示,调整年龄、累计接尘工龄、吸烟情况和饮酒情况对煤工尘肺的影响后,T-AOC、CuZu-SOD、GSH-Px、UA、MDA和iNOS仍为煤工尘肺的影响因素（P<0.05）,而CP、NO对煤工尘肺的发生无统计学意义（P>0.05）。4煤工尘肺氧化损伤指标的筛选及评价ROC曲线结果显示,T-AOC、iNOS、MDA和NO的ROC曲线位于基准线之上,其曲线下面积大于基准线下面积,且差异有统计学意义（P<0.05）,其中T-AOC的ROC曲线下面积最大,其他三项氧化损伤指标的ROC曲线下面积依次减小。本次研究将最接近ROC曲线左上角的那一点定为最佳临界值,即取ROC曲线图对角线与四条ROC曲线的交点,得出T-AOC、MDA、iNOS和NO四项氧化损伤指标筛检煤工尘肺的最佳界值分别为39.405 U/ml、5.308 nmol/l、32.478U/ml和7.000μmol/l。分别用四项氧化损伤指标以上临界值对研究对象进行筛检判定分析,结果显示T-AOC用于煤工尘肺筛检时的灵敏度和特异度最高,假阴性和假阳性率较其他三项指标低。本研究四项氧化损伤指标T-AOC、iNOS、NO和MDA筛检煤工尘肺的判定符合率分别为69.2%、62.8%、62.4%和60%,差异有统计学意义（P<0.05）,其中T-AOC筛检煤工尘肺的误判率最低。四项指标筛检煤工尘肺的Kappa值均介于0.21-0.40之间,根据Kappa值判断标准,表明四项指标筛检煤工尘肺的一致性强度均尚好,但正确指数却不理想。四项指标筛检煤工尘肺的阴性预测值均高于阳性预测值,T-AOC筛检煤工尘肺的阳性预测值和阴性预测值均高于其他三项指标。5 T-AOC、iNOS、NO和MDA筛检煤工尘肺的收益本研究根据多因素Logistic回归分析和ROC曲线分析结果,最终筛选确定MDA、T-AOC和iNOS作为可以用于筛检煤工尘肺的指标。由于单项氧化损伤指标筛检煤工尘肺的正确指数及阳性预测值不理想,因此为提高筛检的效率,增加特异度,需采用串联试验对研究对象进行筛检判定,进一步评价其筛检收益。本研究显示,MDA、T-AOC和iNOS三项指标不同串联方式筛检煤工尘肺的灵敏度,均低于单项指标筛检煤工尘肺的灵敏度,其中T-AOC与MDA串联筛检煤工尘肺的灵敏度高于其他串联方式。可见,采用氧化损伤指标的串联试验筛检煤工尘肺的收益较低。6 MDA、T-AOC和iNOS筛检煤工尘肺判别方程的建立和分析由于单一氧化损伤指标筛检煤工尘肺的正确指数及阳性预测值不太理想,串联试验筛检的收益较低,为提高筛检效率,根据判别分析的原理,对研究对象的MDA、T-AOC和iNOS三项氧化损伤指标两两组合,以及分别从中选取一项指标进行三项指标的组合,逐一构建判别方程,对研究对象进行回代判定检验分析,以寻求最佳的筛检煤工尘肺的方法。结果显示,不同组合方式构建的判别方程判定符合率比较,差异有统计学意义（P<0.05）,其中以三指标联合建立的判别方程筛检煤工尘肺的灵敏度、特异度最高,误判率最低,优于单项指标及两项指标串联的筛检效率。MDA、T-AOC和iNOS联合构建的具体判别方程为: Y0=-3.315+0.253X₁+0.007X₂+0.129X₃ Y1=-5.896+0.332X₁+0.019X₂+0.164X₃ （Y0和Y1分别代表非煤工尘肺和煤工尘肺;X₁代表MDA;X₂代表T-AOC;X₃代表iNOS。）从其标准化系数看,其中以T-AOC（0.634）的作用最明显,其次为iNOS（0.555）,再次为MDA（0.382）,这与三项氧化损伤指标的ROC曲线下面积大小、筛检评价的结果均一致。结论1.本研究ROC曲线分析、筛检价值评价结果显示, T-AOC、iNOS和MDA三项氧化损伤相关指标可以作为筛检煤工尘肺的指标,其最佳界值分别为39.405 U/ml、32.478U/ml和5.308 nmol/l。2.本研究T-AOC、iNOS和MDA三项氧化损伤指标联合构建的判别方程筛检煤工尘肺的灵敏度、特异度及判定符合率均高于单个指标和串联试验,为职业病的防治提供新思路和手段。更多还原

【Abstract】 ObjectivesIn 2006, the outcome of coal added up to 26 hundreds million ton, and about 50 million people were engaged in this industry, which occupies the first place in the world. With the number of patients increasing constantly, Coal workers’pneumoconiosis（CWP）is the most serious occupational disease and which will exist for a long time. Once suffered from the disease, it will inreversible and developes fast. The final diagnosis of coal workers’pneumoconiosis depended on X-ray image,While the image only expressed with some extent pathogenetic condition. There is no great change in pulmonary function at the early stage and the patients haven’t any subjective symptoms. And the X-ray image is not coincidence with pulmonary function and symptoms. In order to carry out the target of“eliminate silicosis in 2030”posted by ILO and WHO, to explore some simple and rapid ways to survey and screen CWP is one of our new tasks. Many scholars pointed out that there was relativity between peroxidation and CWP. At present discussing the possibility and necessity of the screening value of oxygenic injury indexes has been becomeing a direction of studying. According to characters of epidemiology, the research selected coal workers as control group, and evaluated their possible roles of oxygenic injury indexes in the screening of CWP, and established Fisher mathematical models of selected oxygenic injury screening indexes. The research desired to explore a best method to screening CWP in order to do help to find high risk individual from coal workers, and take preventive measures to reduce and slow down the occurrence and development of the diease. Subjects and methodsBy strictly eliminated and internalized criterion, 112 patients of I CWP examined with radiography and diagnosed by specialist group according to“Diagnosis Standard of coal workers’pneumoconiosis”（GBZ70-2002） were selected at random from a certain coal mine in Pingdingshan and Zaozhuang coal limited company. At the same time 144 coal workers were selected as the control group diagnosed without pneumoconiosis. 5ml vein blood of each subject was phlebotomized and plasma lipid peroxidation of total antioxygen capability（T-AOC）、inducible nitric oxide（iNOS）、malonydialdehyde （MDA）、Nitric Oxide（NO）、glutathione peroxidase （ GSH-PX）、uric acid （UA）、Copper and Zinc superoxide dismutase（CuZn-SOD） and ceruloplasmin（CP） from the two groups were measured respectively.The investigators were trained commonly and composed of teachers、postgraduates and me. The course of collection, sorting and analysis was implemented strictly and the most adaptive experiment condition was set up.The database was established by Excel. According to its distribution, the data was analyzed by t test,χ2 test, ROC curve, Logistic regression analysis and Discriminant analysis with SPSS （13.0） software.Results1 General informationThe average age of case group and control group was （57.7±6.3） and （56.1±7.2） years old respectively. Cumulative exposure length of service was 28.45（11-43）and 28.32（9-48）years respectively. There was no statistical difference of age, cumulative exposure length of service、type of work、condition of smoking and drinking between two groups（P>0.05）.2 The level of plasma lipid peroxidation in case and control groups There was statistical significant between two groups with the content or density of MDA、CuZu-SOD、GSH-Px、T-AOC、UA、NO and iNOS except CP. 3 Mainly affect factors of multiple factors logistic regression analysis of CWPBy the multiple factors logistic regression analysis, there was statistical significant between the indexes of T-AOC、MDA、iNOS、CuZu-SOD、GSH-Px、UA and CWP when the factors of circumstance such as age、cumulative exposure length of service、smoking and drinking were adjusted. And there was no statistical significant between NO、CP and CWP.4 Valuation and selection of peroxidation indexesWith their ROC curve above the reference line, there was statistical significant in area under the curve（AUC） of MDA、T-AOC、NO and iNOS between two groups. Combined the sensitivity and specificity together, regarded the point of intersection of ROC curve and its diagonal line as the cut off point, and the best critical value of MDA、T-AOC、NO and iNOS was 5.308nmol/l、39.405U/ml、7.000μmol/l and 32.478U/ml respectively. With their cut off points, the screening value of T-AOC was better than other indexes, the second was iNOS, then MDA by comparing their validity and predictive value with using sensitivity、specificity、false negative and positive rate、likelihood ratio and et al. The consistency rate of T-AOC、iNOS、MDA and NO was 69.2%、62.8%、62.4% and 60% respectively and there was statistical significant among them. The Kappa value of these indexes was between 0.21 and 0.40 that meaned the consistency stability of them was well. But the Youden’s index of them was not at best.5 The screening yield of T-AOC、iNOS、MDA and NO of CWPAccording to the results of multivariate logistic regression and ROC curve analysis,MDA、T-AOC and iNOS was finally seclected as the screening indexes of CWP. We evaluated their screening effect mainly from the aspect of yield. The sensitivity and specificity of compounding trial in series with MDA and T-AOC was 50.9% and 80.4% respectively, the sebsitivity was bigger than other forms of compounding trial in series of two and three indexes. We can see that specificity was improved and sensitivity was decreased in compounding trial in series. The screening yield of compounding trial with Oxygenic injury indexes was low. So we should utilize them carefully according to the circumstance and destination of performance in practical application.6 Establishment and analysis of Fisher mathematical modelsIn order to decrease inconsistency rate and improve screening effect of T-AOC、iNOS and MDA, we established different discriminant mathematical models of them to explore the best screening mothod. By back substitution assessment analysis, there was statistical difference of consistency rate among all the possible models founded by these three indexes（P<0.05）. The discriminant mathematical model composed of MDA、T-AOC and iNOS together was the best one of them. It was: Y-0=-3.315+0.253X₁+0.007X₂+0.129X₃ Y₁=-5.896+0.332X₁+0.019X₂+0.164X₃ （Y₁ and Y0 was represent CWP and not CWP respectively; X₁、X₂ and X₃ was represent MDA、T-AOC and iNOS respectively.）The result of standardized canonical discriminant function coefficients of the discriminant mathematical model was consistent with results of AUC and evaluation. The consistency rate、sensitivity and specificity of the discriminant mathematical model was 84.0%、77.7% and 89.1% respectively, which was increased than single index and compounding trial in series（P<0.05）.Conclusions1. In our research, MDA、T-AOC and iNOS were finally selected as screening indexes of coal workers’pneumoconiosis by multiple factors logistic regression analysis、ROC curve analysis and evaluation. And the best critical value of them was 5.308nmol/l、39.405U/ml and 32.478U/ml respectively. And the screening value of T-AOC was the best one, the second was iNOS, then MDA.2. The Fisher discriminant mathematical model composed of MDA、 T-AOC and iNOS together was a good method to screening coal workers’pneumoconiosis, which provided a new idea and way to prevent occupational diseases such as CWP.更多还原