Bhatti Parveen, Sigurdson Alice J, Wang Sophia S, Chen Jinbo, Rothman Nathaniel, Hartge Patricia, Bergen Andrew W, Landi Maria Teresa
Radiation Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, NIH, Department of Health and Human Services, 6120 Executive Boulevard, EPS 7091, MSC 7238, Bethesda, MD 20892-7238, USA.
Cancer Epidemiol Biomarkers Prev. 2005 Oct;14(10):2449-53. doi: 10.1158/1055-9965.EPI-05-0463.
The differences in common genetic polymorphism frequencies by willingness to participate in epidemiologic studies are unexplored, but the same threats to internal validity operate as for studies with nongenetic information. We analyzed single nucleotide polymorphism genotypes, haplotypes, and short tandem repeats among control groups from three studies with different recruitment designs that included early, late, and never questionnaire responders, one or more participation incentives, and blood or buccal DNA collection. Among 2,955 individuals, we compared 108 genotypes, 8 haplotypes, and 9 to 15 short tandem repeats by respondent type. Among our main comparisons, single nucleotide polymorphism genotype frequencies differed significantly (P < 0.05) between respondent groups in six instances, with 13 expected by chance alone. When comparing the odds of carrying a variant among the various response groups, 19 odds ratios were </=0.70 or >/=1.40, levels that might be notably different. Among the various respondent group comparisons, haplotype and short tandem repeat frequencies were not significantly different by willingness to participate. We observed little evidence to suggest that genotype differences underlie response characteristics in molecular epidemiologic studies, but a greater variety of genes should be examined, including those related to behavioral traits potentially associated with willingness to participate. To the extent possible, investigators should evaluate their own genetic data for bias in response categories.
参与流行病学研究意愿不同导致的常见基因多态性频率差异尚未得到探究，但对内部效度的威胁与非基因信息研究相同。我们分析了三项不同招募设计研究中对照组的单核苷酸多态性基因型、单倍型和短串联重复序列，这些研究包括早期、晚期和从不回复问卷者，有一项或多项参与激励措施，以及血液或口腔DNA采集。在2955名个体中，我们按回复者类型比较了108种基因型、8种单倍型以及9至15个短串联重复序列。在我们的主要比较中，单核苷酸多态性基因型频率在六个实例中回复者组间存在显著差异（P < 0.05），而仅随机预期为13个实例。在比较不同回复组携带变异的几率时，19个比值比≤0.70或≥1.40，这些水平可能存在显著差异。在不同回复组比较中，单倍型和短串联重复序列频率按参与意愿无显著差异。我们几乎没有发现证据表明分子流行病学研究中回复特征的基础是基因型差异，但应检测更多种类的基因，包括那些与可能和参与意愿相关的行为特征有关的基因。在可能的范围内，研究者应评估自身基因数据在回复类别中的偏倚情况。
