Association of the selenoprotein 15-kDa (SEP15) polymorphisms with cancer risk: a meta-analysis

Selenoproteins are involved in antioxidant defense, the redox signaling pathway and cell homeostasis. Primary studies have shown that single-nucleotide polymorphisms in the selenoprotein gene (SEP15) are associated with cancer risk. However, conflicting outcomes warrant a meta-analysis to obtain more precise estimates. Literature search yielded 18 case-control studies from 12 articles. We calculated pooled odds ratios (OR) and 95% confidence intervals (CI) of two SEP15 polymorphisms (rs5845 and rs5859) using standard genetic models (homozygous, recessive, dominant and codominant). Subgroup analysis was based on statistical power (80% cutoff) and cancer type (breast/respiratory/genitourinary/colorectal). Heterogeneity of the outcomes necessitated examining their sources (outlier treatment). Multiple comparison outcomes were corrected with the False Discovery Rate (P^aF). Our core findings lay in the post-outlier recessive subgroup outcomes, where risks in the powered study (≥ 80%) was increased (OR 1.26, 95% CI 1.02-1.57, P^aF = 0.047) while that in genitourinary cancer was protective (OR 0.29, 95% CI 0.20-0.43, P^aF < 10^-4). The potency of outlier treatment in unmasking significant associations and generating homogeneity provides good evidence of SEP15's role in cancer. In the clinical sense, selenium chemo-intervention may be of benefit among persons with particular SEP15 genotypes.AbbreviationsAnumber of unduplicated articles that contributed to instabilityAManalysis modelBnumber of robust comparisonsBCbreast cancerBLCbladder cancercDNAcomplementary deoxyribonucleic acidCIconfidence intervalCIDconfidence interval differenceCRCcolorectal cancerDdecreased riskEHeliminated heterogeneityFfixed-effectsFDRFalse Discovery RateGUCgenitourinary cancersGSgained significanceHBhospital-basedHWEHardy-Weinberg EquilibriumIincreased riskI²measure of heterogeneitykDakiloDaltonLAClaryngeal cancerLUClung cancermafminor allele frequencynnumber of studiesNnumber of comparisonsNMnot mentionedNOSNewcastle-Ottawa ScaleORodds ratioP^aP value for associationP^aδP value for association (pre-FDR)P^aFP value for association FDR-correctedP^bP value for heterogeneityPBpopulation-basedPCprostate cancerPRISMAPreferred Reporting Items for Systematic Reviews and Meta-AnalysesPROpre-outlierPSOpost-outlierRrandom-effects[R]referenceRCrespiratory cancersRNSretained non-significanceROSreactive oxygen speciesSEPselenoproteinsSEP15selenoprotein geneSNPsingle nucleotide polymorphismSWShapiro-Wilk testUSAUnited States of Americavvvariantwvheterozygouswwwild-type.