These associations were very robust, which did not vary materially when the sensitivity analyses (exclusion the study with controls not in HWE) were performed. The effect of the genotype TT on cancer especially exists in Caucasians and female subjects. Only female specific cancers were included in female subgroup in our meta-analysis, which indicates that the genotype TT is significantly Danusertib nmr associated with an increased risk for female specific cancers. The molecular

basis of gender specific effect of the HIF-1α 1772 C/T polymorphism on cancers is unclear. Studies have shown that estrogen can induce the expression of HIF-1α [28, 29]. The substitution of C to T at positions 1772 of the exon 12 of the HIF-1α gene S63845 ic50 further increase the transactivation capacity of the HIF-1α gene and thus promote the development of female specific cancers. We also observed a marginally significant association between the genotype TT and increased cancer risk in East Asians. However, subjects with mutant homozygotes were only detected in two studies of East Asians. The CI for this subgroup was very wide, and the association could have been caused by chance. More studies based on larger population should be conducted to further examine this association. For the HIF-1α 1790 G/A polymorphism, the meta-analysis on all studies showed no evidence that the HIF-1α 1790 G/A polymorphism was significantly associated with increased

cancer risk. We also performed the stratification analyses by gender, ethnicity, and cancer types. The pooled AMN-107 manufacturer ORs for allelic frequency comparison and dominant model comparison suggested the 1790 G/A polymorphism was significantly associated with an increased cancer risk in Caucasians. However, the sensitivity analysis did not suggest this association. Because the results from the sensitivity analysis were more valid, our meta-analysis also does not strongly suggest the association between the HIF-1α 1790 G/A polymorphism and cancer risk in Caucasians [23]. The pooled effects for allelic frequency comparison and dominant model comparison suggested a significant association between the HIF-1α 1790 G/A polymorphism and a

decreased breast cancer risk. Because the conclusion is inconsistent with the general understanding that the 1790 A alleles enhances HIF-1α transcriptional activity and the presence of the variant allele might be associated with increased cancer susceptibility, we further performed the meta-analysis for the other cancers to detect the specific effects of cancer type [6]. The results suggested a significant association between the A allele and increased cancer risk in other cancers. A marginal association between the 1790 G/A polymorphism and increased cancer risk in other cancers was also detected under dominant model. However, the reanalysis after exclusion the studies with controls not in HWE did not suggest these associations.