中国不同地区人群结直肠癌筛查结肠镜顺应性的综合性分析

Comprehensive analysis of colonoscopy compliance for colorectal cancer screening cross populations in China

目的采用生态学研究方法,从群体水平描述我国结直肠癌筛查的结肠镜顺应性水平,并分析其与结直肠癌粗发病率(CIR)及所选初筛方法阳性预测值(PPV)的相关性。方法系统检索PubMed、Embase、Web of science、Cochrane、中国知识基础设施工程(CNKI)、万方数据资源系统(Wangfang)和中国生物医学文献数据库(CBM)7个数据库,筛选2019-08-25前发表、基于中国平均风险人群结直肠癌筛查的相关文献。提取初筛阳性率、灵敏度、特异度和PPV信息以及结肠镜顺应性。从《中国肿瘤登记年报》、全球健康数据交换库和相关文献中提取不同地区筛查当年全人群结直肠癌CIR。以结肠镜顺应性的对数转换值为应变量,采用LOESS方法描述其与结直肠癌CIR及PPV的关系,并以结直肠癌CIR为自变量进行Joinpoint回归,在变化点处进行分层Spearman相关分析。结果从2460篇文章中最终纳入115篇,提取141项研究。最常用的初筛方法是风险评估与粪便隐血试验(FOBT)并联,其次是单用FOBT。结肠镜顺应性总体上较低,中位数和四分位间距仅为48.3%(34.6%~64.5%),且与筛查当年结直肠癌CIR呈负相关,r=-0.216,P=0.01。LOESS曲线显示,结肠镜顺应性随结直肠癌CIR升高而呈先上升再下降的态势。Joinpoint回归分析显示,结肠镜顺应性变化点位于结直肠癌CIR为47.3/10万处。在CIR≥47.3/10万人群中,结肠镜顺应性与CIR(r_(s)=-0.286,P=0.013)及初筛阳性率(r=-0.307,P=0.009)呈负相关,而在CIR<47.3/10万人群中呈正相关,但未达到有统计学意义水平。大规模筛查人群的结肠镜顺应性较低,但与PPV呈正相关,r=0.468,P=0.058。结论我国结直肠癌筛查结肠镜顺应性较低,且因全人群CIR水平、筛查规模及初筛方法而异。需选择合适的初筛方法,加强筛查过程管理,提高初筛PPV,采用激励性经济方式,提高筛查人群的结肠镜顺应性和筛查效果。

Objective To overview the colonoscopy adherence in cascade screening of colorectal cancer(CRC) in Chinese populations, and evaluate the potential influence of crude incidence of CRC and positive predictive value(PPV) of the initial tests based on an ecological evaluation.Methods We searched PubMed, Embase, Web of science, Cochrane, China National Knowledge Infrastructure(CNKI),Wanfang and CBM for eligible studies which were conducted in average risk populations in China and published before August 25,2019.The sensitivity, specificity, PPV of initial screening tests and adherence to subsequent colonoscopy were extracted.The crude incidence(CIR) of CRC in populations in the year of screening were derived from the Annual Report of Cancer Registry in China,the Global Health Data Exchange(GHDx),and related literature.LOESS was used to outline colonoscopy adherence along with the CIR and the PPV of initial screening tests.Spearman’s correlation analysis was conducted by the change point of CIR for adherence which was identified by Joinpoint regression.Results Of 2 460 articles identified,115 articles(141 studies)were included.Most studies applied cascade screening of CRC with parallel tests of risk assessment and FOBT as an initial tests,followed by FOBT alone.Colonoscopy compliance was generally low,with the median and interquartile range(IQR)being only 48.3%(34.6%-64.5%).The adherence was negatively correlated with the CIR of CRC(r=-0.216,P=0.01).LOESS curve of adherence to colonoscopy(log transformed)along with CIR showed a nonlinear relationship.The change point in adherence was identified at the CIR of 47.3/100 000 by Joinpoint regression.Adherence to colonoscopy was negatively correlated with CIR(r_(s) =-0.286,P=0.013)and positivity rate of initial tests(r=-0.307,P=0.009)in populations with CIR ≥47.3/100 000,but slightly positively in those with CIR <47.3/100 000.In large-scale screening of CRC,adherence to subsequent colonoscopy was lower but positively correlated with PPV(r=0.468,P=0.058).Conclusions Adherence to colonoscopy in CRC screening in China is low and varies by CIR,sample size and type of initial tests.Initial tests with high PPV,standardized screening procedure,and effective economic policies may help improve adherence to subsequent colonoscopy and thus effectiveness of CRC screening in the population.