Carcinogenesis of glioma is considered to be a process affected by complicated factors. Environmental factors had been a focus in previous studies, however, except therapeutic ionizing radiation; no other environmental factors had been shown to be associated with glioma susceptibility . Compared with other cancer types, knowledge about the role of genetic SNP in glioma is relatively limited.
EGF, a 6,045-Da single chain polypeptide which activates several signaling pathways such as ras/raf/MAPK and phosphatidylinositol-3-kinase (PI3K) , is of special importance in regulating cell proliferation, migration, adhesion, and inflammatory processes . High expression of EGF had been detected in some cancer tissues, such as gallbladder cancer  and glioblastoma . As an important receptor of EGF with high affinity, EGFR was often overexpressed in glioma cells. Amplification of EGFR gene was shown in half of the glioblastomas, as well as overexpression of EGFR mRNA and protein . The overexpression related to a poorer prognosis in glioma patients . These previous studies suggested that overexpression of EGF and EGFR might affect cell activities in brain tissues, abnormity of which may contribute to the glioma.
How does EGF +61 polymorphism affect the characteristic of EGF? This polymorphism had been proved to be functionally significant in individual variability of EGF expression. Shahbazi et al.  identified that cells from 61 A/A homozygous individuals produced significantly less EGF than cells from G/G (p = 0.0004) or G/A (p = 0.001) individuals. Tanabe et al.  showed that with EGF +61G, transcription product had significant longer (more than 2-fold) half-life than 61A allele. Serum EGF levels were 1.8-fold higher in G/G hepatocellular carcinoma patients with cirrhosis than A/A patients, and liver EGF levels were 2.4-fold higher in G/G patients than A/A patients. EGF+61G enhanced risk of hepatocellular carcinoma.
However, frequencies of EGF +61G allele were diverse among different subpopulation. Comparing to MAF of G allele in European population, the Chinese population occupy 0.633 of G allele (data shown in NCBI SNP database http://www.ncbi.nlm.nih.gov/SNP/snp_ref.cgi?rs=4444903) (Table 2). We also compared the frequencies of the genotypes and alleles between control group in the present study and previous study on this SNP. Table 2 summarized the distribution of EGF +61 G/A polymorphism in different healthy control populations which fit in HWE except the study of Lim et al. . The frequency of the EGF +61A allele in the present study was 27.3%, which was similar to the frequency observed in Chinese controls (30%-31%). The frequency of EGF +61A allele was about 23.9%-32.8% in Asians, which was different from that of Caucasians (53.7%-61.6%) (Table 2). Therefore, constitutive EGF expression may be different in different ethnic groups, which might lead to discrepancy results of association studies in different ethnic groups. Lim et al.  reported somewhat different distribution of EGF genotype in Korea healthy controls, the frequency of +61A allele was higher than that of other Asian populations, with a MAF = 0.4735. Although the discrepancy might be explained by their relatively small control sample size (n = 132), further studies were still needed to address the difference .
Previous study evaluated the influence of EGF+61 G/A to glioma risk, but no consensus had been reached [20, 32]. The effects of EGF +61 G/A allele in various cancers were conflicting too. Shahbazi et al.  showed that G/G was significantly associated with increased malignant melanoma by examining 99 controls and 135 cases. Vishnoi et al.  found that G/G genotype of EGF +61A/G was significantly associated with increased gallbladder cancer risk by examining 126 gallbladder patients and 190 healthy controls; Goto et al.  found no association between EGF +61 G/A and gastric cancer but that EGF +61 A/A genotype showed a trend of protection by research on 202 patients and 454 healthy controls. Kang et al.  examined 432 lung cancer patients and 432 healthy age-and gender-matched control subjects and found that +61 A/A and +61 G/A genotypes were not significantly associated with the risk of lung cancer compared with the +61 G/G genotype. Tanabe et al.  examined 50 patients with cirrhosis and hepatocellular carcinoma and 148 patients with only cirrhosis. Logistic regression analysis demonstrated that the number of copies of G was significantly associated with increased hepatocellular carcinoma. However, the study of Gao et al.  didn't find association between the EGF +61 polymorphism and nasopharyngeal carcinoma in Chinese population by studying 173 patients and 206 controls, but A/A showed a trend to increase risk. By examining 383 breast cancer patients and 500 controls, Araujo et al.  found that carriers of G homozygous genotype had a lower risk for developing breast cancer (OR = 0.68, 95CI: 0.46-1.01). A lower risk for breast cancer in G/G carriers might be explained through EGF receptor internalization promoted by EGF.
According to Rosenthal's criteria , the link between EGF +61 G/A polymorphism and cancer is somehow plausible. It is essential to show that the change in the gene under study causes a relevant alteration in the function or level of the gene product, and EGF +61 G/A had been shown to be functional [14, 30, 39]. The EGF-EGFR signaling pathway performed an important role in regulating cell proliferation, migration, adhesion, and inflammatory processes , the correlation between this particular polymorphism and cancer have practical value. However, inconsistency existed in different studies, which may arise from many aspects. Cancer is a complex disease, EGF +61 G/A might play different roles in different kinds of cancers; Ethnic difference may also contribute to different disease susceptibility; Sample size must be large enough to be convincing. Larger studies will have greater power to detect an effect. Increasing the sample size results in an increased frequency of detecting marker/phenotype associations . The number of patients in most of former studies about the association of EGF +61 G/A to cancers were less than 300, which might lead to a different result from our study involving 677 glioma patients and 698 healthy controls. In our paper reviewing process, a study in Northern-Chinese population failed to detect association of EGF+61 A/G and glioma , which might due to the small sample size (168 patients). However, stratification analysis in our study might have the decreased sample size to certain phenotypes. When stratified by WHO, the significant association of genotype A/A was only found in grade II and IV. The insignificant result in grade I and III might arise from the small sample size. Interestingly, stratification for different smoking status suggested A allele only increased the glioma risk in never-smokers. This might be explained of small sample size in ever- and current-smokers.
Our analysis showed the EGF +61 A allele may increase the risk of glioma, the stratification analysis indicated that A allele had no association with astrocytomas except for glioblastoma. Under different histology types of astrocytomas, glioblastoma, and other gliomas, the statistical power was 0.954, 0.952, and 0.920, respectively. For different grade of glioma, the power over 80% was only found in WHO IV, as well as in low grade and high grade gliomas. When adjusted for sex, age, fmc, smoking status, the glioma grade I risk significantly increased under additive genetic model(the power was only 0.743). However, Costa et al.  reported a significant association of the G variant with an increased risk of not only gliomas but also glioblastomas and oligodendroglial tumors in Portugal. Lanuti et al.  found that the association between EGF +61A/G and esophageal adenocarcinoma risk in earlier-stage patients might suggest that the G/G genotype was associated with a less aggressive phenotype, paralleling lung cancer where individuals with EGF pathway driven tumors seem to have fewer molecular alterations [42, 43]; Consider the different allele frequency between Asians and Caucasians, further studies are needed to explore the role of EGF +61 G/A polymorphism in cancer development in different populations.
Cigarette smoking is a plausible behavioral exposure that might modulate glioma risk, but no overall association between glioma risk and cigarette smoking among either men or women [44, 45]. The number of research on the role of smoking in glioma susceptibility was relatively small compared to other cancers such as lung cancer. Excision repair cross-complementing group 1 (ERCC1) is the lead enzyme in the nucleotide excision repair process. Stratified analyses revealed that the A/A genotype of ERCC1 8092C > A polymorphism was a risk factor in nonsmokers, but a protective factor in heavy smokers when compared with the C/C genotype . No study had investigated the association between EGF polymorphism and the risk of glioma under different smoking status. In our study, the genotype-smoking interaction was insignificant. We also calculated the power of the stratification analysis by smoking status. The result showed that the power was 0.964 in never-smokers, which was of strong ability to test the association between EGF +61 G/A and glioma. However, in ever- and current-smokers, the power was 0.421 and 0.550, respectively. It is possible that the insignificant results in ever- and current-smokers might attributable to relatively small number of subjects. Additional studies with more subjects will be needed.
Our study included a relatively large sample size from a homogeneous population of the same ethnicity. However, several limitations in our study need to be addressed. The association between EGF +61 SNP (rs4444903) and risk of gliomas is biologically plausible, since EGF performs a very important role in the proliferation, migration, and differentiation . However, only one SNP in EGF was examined in our study. It remains uncertain whether the rs4444903 is in strong LD with a causative variant located inside or near the EGF locus. Secondly, the association between EGF +61A/G and glioma risk was controversial in our study and Caucasions , Thus, studies with ethnically diverse populations are warranted to confirm our findings and to further elucidate the significance of the polymorphism in the development of gliomas. Finally, further functional experiments are therefore necessary to test the hypotheses.