Bradley D T, Badger S A, McFarland M, Hughes A E
Centre for Public Health, School of Medicine, Dentistry and Biomedical Sciences, Queen's University Belfast, Institute of Clinical Sciences, Block B, Royal Victoria Hospital, Belfast BT12 6BA, UK.
Mater Misericordiae University Hospital, Eccles Street, Dublin, Ireland.
Eur J Vasc Endovasc Surg. 2016 Jan;51(1):64-75. doi: 10.1016/j.ejvs.2015.09.006. Epub 2015 Oct 12.
OBJECTIVE/BACKGROUND: Many associations between abdominal aortic aneurysm (AAA) and genetic polymorphisms have been reported. It is unclear which are genuine and which may be caused by type 1 errors, biases, and flexible study design. The objectives of the study were to identify associations supported by current evidence and to investigate the effect of study design on reporting associations.
Data sources were MEDLINE, Embase, and Web of Science. Reports were dual-reviewed for relevance and inclusion against predefined criteria (studies of genetic polymorphisms and AAA risk). Study characteristics and data were extracted using an agreed tool and reports assessed for quality. Heterogeneity was assessed using I(2) and fixed- and random-effects meta-analyses were conducted for variants that were reported at least twice, if any had reported an association. Strength of evidence was assessed using a standard guideline.
Searches identified 467 unique articles, of which 97 were included. Of 97 studies, 63 reported at least one association. Of 92 studies that conducted multiple tests, only 27% corrected their analyses. In total, 263 genes were investigated, and associations were reported in polymorphisms in 87 genes. Associations in CDKN2BAS, SORT1, LRP1, IL6R, MMP3, AGTR1, ACE, and APOA1 were supported by meta-analyses.
Uncorrected multiple testing and flexible study design (particularly testing many inheritance models and subgroups, and failure to check for Hardy-Weinberg equilibrium) contributed to apparently false associations being reported. Heterogeneity, possibly due to the case mix, geographical, temporal, and environmental variation between different studies, was evident. Polymorphisms in nine genes had strong or moderate support on the basis of the literature at this time. Suggestions are made for improving AAA genetics study design and conduct.
目的/背景：已有许多关于腹主动脉瘤（AAA）与基因多态性之间关联的报道。目前尚不清楚哪些关联是真实存在的，哪些可能是由一类错误、偏差和灵活的研究设计所导致的。本研究的目的是确定当前证据支持的关联，并调查研究设计对关联报道的影响。
数据来源为MEDLINE、Embase和科学网。根据预定义标准（基因多态性与AAA风险的研究）对报告进行双重审查以确定其相关性和纳入情况。使用商定的工具提取研究特征和数据，并对报告进行质量评估。使用I²评估异质性，对于至少有两项研究报告且有任何关联报道的变体进行固定效应和随机效应荟萃分析。使用标准指南评估证据强度。
检索到467篇独特文章，其中97篇被纳入。在97项研究中，63项报告了至少一种关联。在进行多次检验的92项研究中，只有27%对其分析进行了校正。总共研究了263个基因，87个基因的多态性存在关联报道。荟萃分析支持CDKN2BAS、SORT1、LRP1、IL6R、MMP3、AGTR1、ACE和APOA1中的关联。
未校正的多次检验和灵活的研究设计（特别是检验多种遗传模型和亚组，以及未检查哈迪-温伯格平衡）导致了明显错误关联的报道。异质性明显，这可能是由于不同研究之间的病例组合、地理、时间和环境差异所致。目前基于文献，九个基因的多态性有较强或中等程度的支持。对改进AAA遗传学研究设计和实施提出了建议。
