Objectives: This study is to evaluate the association between serum folate level and cervical cancer.
Methods: PubMed, Medline, Springer, Elsevier Science Direct, Cochrane Library and Google scholar was searched for relevant trials. Rev.Man5.1 and Stata 11.0 software were applied for the meta-analysis. Odds Ratio (OR) and 95% confidence intervals (CI) were collected and calculated in a fixed-effects model (the Mantel-Haenszel method) or a random-effects model (the DerSimonian and Laird method) when appropriate.
Results: A total of 6 studies including 2383 participants (case group: 873; control group: 1510) were performed in the meta-analysis. The overall meta-analysis showed that there were significant differences (OR=1.94, 95%CI=1.14 to 3.31, P<0.05) between the two groups, suggesting that deficient serum folate level was associated with an increase in cervical cancer risk. The further stratification subgroup analysis indicated that there existed significant differences of the associations between deficient serum folate level and cervical cancer risk in sample size (OR=2.98, 95% CI=1.92 to 4.62, P<0.01) which were less than 500 and in Asian population (OR=3.45, 95% CI=2.06 to 5.76, P<0.01) between the two groups.
Conclusions: Our results indicated that the deficiency of serum folate level was associated with an increase for cervical cancer risk, a larger sample size of studies or meta-analysis is necessary in the future research to verify these results.
Key words: Cervical cancer; folate; Meta-analysis
Nowadays, cervical cancer is the most common cancer among women all over the world which jeopardizes the women’s health seriously due to its high cancer morbidity and mortality 1-7. There are several advances among the development of diagnosis and treatment of cervical cancer, but it’s still indispensable to explore its pathogenesis. Folic acid, a water-soluble B vitamin, must be ingested from diet of mammals and its deficiency has been identified to be associated with the development of several cancers including cervical cancer 8. Folic acid is critical for the biosynthesis of purines and thymidylates which are the key factors for normal DNA replication and repair and it is also involved in the DNA methylation 8. The deficiency of folic acid results in ineffective DNA synthesis, impaired cellular physiology and abnormal cell morphology which are related to the occurrence and development of cancer 9.
Several studies on the association between the serum folate level and cervical cancer were conducted but there were controversial effects. High intake of folate has been suggested as an important factor for prevention of cervical cancer and the positive results showed that serum folate was markedly lower in cervical neoplasia cases compared with the control group 10, 11. However, some studies have shown there was no obvious and direct connection between the serum folic acid level and the appearance and development of cervical cancer 12-15.
The influences of serum folic acid level on the risk of cervical cancer are inconsistent due to the limitation of the sample size. The integration of relevant studies by a meta-analysis could provide statistical power to allow a definite conclusions which offers great benefit for scientific research 16. Accordingly, we conducted an eligible meta-analysis of well-designed systematic studies to evaluate the association between serum folate level and cervical cancer.
Material and methods
We retrieved the relevant trials on the association between serum folate level and cervical cancer from several public databases, mainly including PubMed, Medline, Springer, Elsevier Science Direct, Cochrane Library and Google scholar up to May 2014 following the key words: “folate”, “folic acid”, “cervical cancer”, “cervical carcinoma”, “study” and “trial”. The reference list of retrieved papers were checked for additional studies. We only collected data from the full-published English papers excluding any meeting or conference abstract and the included Publications date was not restricted in our research.
Inclusion and exclusion criteria
This investigation was conducted between the case group (cervical cancer patients) and the control group (non cervical cancer patients) to discuss the association between serum folate level and cervical cancer. Studies only including the data of serum folate level or cervical cancer were included and the reduplicated studies or studies with no comparison between case and control group were excluded
The retrieve of the electronic databases was done by six investigators independently and two persons in pairs were for the retrieve of two databases, respectively. All this retrieve was conducted with the same method. Two investigators from different groups assessed studies for inclusion eligibility through further retrieve of the abstracts. References in this studies were also screened for other potentials by another two investigators from different groups and the discrepancies between them was judged with an additional assessment by a third investigator. The final included literatures should be made according to the three investigators’ decision.
The quality of included studies was analyzed by Newcastle-Ottawa Scale (NOS). Studies with scores ≥7 were considered as high-quality report and conversely were low-quality report. Relevant data were directly extracted from the literatures by two investigators with standard protocols and verification was done by the third investigator. And the further clarification of data was obtained by contacting the authors. The items of extracted data was recorded: first author’s name, year of publication, country, participants’ geographic location (ethnicity), serum folate level, sample size, study design, and age. Any discrepancies were resolved by discussing within our research or contracting with the original investigator .
Estimates of odds Ratio (OR) and its 95% confidence interval (CI) were achieved by using fixed or random effects models. Mantel–Haenszel method used in the fixed effect model 17 or DerSimonian and Laid method used in the random effect model 18were conducted to estimate of ORs. The test for heterogeneity was showed by Cochran’s Q-statistic 19 and I2 statistic 20. The heterogeneity existed when P<0.10 (Q-statistic) or I2>50% (I2-statistic) and data were analyzed by the random effects model which takes into account of within-study and between-study variation in meta-analysis. Otherwise, the fixed effects model was used. Egger’s linear regression test 21 and a funnel plot were showed to examines the effect of publication bias. Sensitivity analysis was performed by exclusion any one of the included study to estimate the reliability of our meta-analysis 22. All the analysis was performed using the software of Review Manager 5.1 (Cochrane Collaboration, http://ims.cochrane.org/revman) and the STATA package v.11.0 (Stata Corporation, College Station, TX, USA). All the P–values were two-sided and Pï¼œ 0.05 was considered statistically signiï¬cant.
Description of eligible studies
We received 699 potential studies relevant to our search terms through a systemic retrieve (PubMed: 162; Medline: 113; Springer: 125; Elsevier Science Direct: 93; Cochrane Library: 15; Google Scholar: 191). After examining for duplication, relevance and the reference type, 657 studies were excluded. 36 studies of the remaining articles were also removed for 20 only reporting folate data but not for comparison and 16 without providing available data. Only 6 studies were selected to be eligible in this meta-analysis (Fig.1).
A total of 6 studies containing 2383 participants concerning the association between serum folate level and cervical cancer were included in this meta-analysis and the characteristics were presented in Table 1. The included studies were published between 1991 and 2012, the sample size of which were between 165 and 895 and average age was between 17 and 74. All 6 articles were performed by case-control study and the scores of which were above or equal to 7 belonging to high-quality reports (Table 2).
Six separate studies with 873 cases and 1510 controls were analyzed to explore the association between the serum folate level and the risk for cervical cancer (Table 3 and Figure 2). Results showed that there was significant heterogeneity among the included studies and the random effects model was used in this meta-analysis to compare the association between the case and control group. According to the result, there were significant differences (OR=1.94, 95%CI=1.14 to 3.31, P<0.05) between the two groups suggesting that deficient serum folate level was associated with an increase of cervical cancer risk.
We also performed subgroup analysis stratified by sample size and participants’ geographic location. As is shown in Table 3, significant differences (OR=2.98, 95% CI=1.92 to 4.62, P<0.01) were found between deficient serum folate level and cervical cancer risk in sample size which were less than 500. Another result of subgroup analysis showed that there were alsosignificant differences (OR=3.45, 95% CI=2.06 to 5.76, P<0.01) between case and control group in Asian population.
Analysis of publication bias and sensitivity
We performed a sensitivity analysis by removing one study each time and rerunning the model to test the effect on each overall estimate. The result showed that the above meta-analysis estimates changed little after removal of certain studies (OR range 1.57(1.00, 2.49)-1.43(1.19, 4.70)), which implied that our results were reliable.
Cervical cancer remains a burden for women all over the world which causes a number of attention on the research of the prevention and treatment of this tumor. Epidemiological studies showed that an increased risk of cervical canceris related to the deficiency of folic acid 23, 24 . In this meta-analysis, a total of 6 studies containing 2383 participants (case group: 873 and control group: 1510) were combined and analyzed. The results indicated that deficiency of serum folate level could cause an significant increase of the cervical cancer risk (OR=1.94, 95%CI=1.14 to 3.31, P<0.05). Our results is consistent with the previous reports which showed that higher intake of folic acid could provide an obvious protective effect 25, 26 and non-significant protective effect 27, 28 for the cervical cancer. As is known, the persistence of human papillomavirus (HPV) is considered to be a critical factor for the risk of cervical cancer 29. Pillai et al 30 identified that there are connections between folate intake and HPV persistence and Piyathilake et al 31 reported that plasma folate provided a significant protective effect on risk of HPV persistence. Due to the existence of heterogeneity of the overall effects, the further subgroup analysis was conducted in sample size and participants’ geographic location. Significant differences (OR=2.98, 95% CI=1.92 to 4.62, P<0.01) of the associations between low folate intake and cervical cancer risk also existed in sample size which was less than 500 and Asian population (OR=3.45, 95% CI=2.06 to 5.76, P<0.01). There was no publication bias in subgroup analysis according to the results of the Egger’s linear regression test.
A persistent high-risk infection of human papillomavirus (HPV) is considered be main culprit for cervical cancer 32, but the presence of genetic and epigenetic alternations are necessary for the development of carcinogenesis 33. The association between the low folate intake with the risk of tumors should be ascribed to two principal mechanisms. On the one hand, the biosynthesis of purines and thymidylates which is influenced by folic acid, is the necessary segment for DNA replication and cell division 34. A deficiency of folate will directly lead to ineffective DNA synthesis and then lead to the occurrence of cells with impaired physiology and abnormal morphology which are the early indicator for several human tumors. Another mechanism for the intake of folic acid correlating the risk of cervical cancer is DNA methylation. DNA methylation is one mechanism of epigenetic modification and indicated to be involved in tumorigenesis 35, 36. Methylation at specific locations such including constitutive heterochromatin regions and promoter regions in the DNA is always companied with the silence of gene transcription 8, 37. Folic acid is essential for the formation of SAM (S-adenosylmethionine) which is the principal methyl donor in the biochemical reaction in vivo including DNA methylation 9. The deficiency of folate certainly will affect the formation of SAM and lead to an insufficient methyl supply for DNA methylation. Previous studies showed that decreased DNA methylation has been recognized as an early genomic alteration in human tumors since it can lead to the inappropriate activation of proto-oncogenes and then induce malignant transformation 35.
The deficiency of folate intake could increase the risk of cervical cancer according to our study, however, the interpretation of this meta-analysis should be circumspect for there are several limitations. First, signiï¬cant between-study heterogeneities were detected in the current meta-analysis which may distort the results. The heterogeneity 38 related to the validity of the results may be caused by the sample size and participants’ geographic location. Second, the publication bias existed in this mete-analysis which have an influence on the stability of the results due to the small number of studies meeting our inclusion criteria. The further research needs more and high-quality RCTs to test and verify the results of this meta-analysis. Despite these limitations, this meta-analysis findings after preparing a detailed protocol before initiating , conducting a meticulous search for published studies and applying explicit methods for study selection, data extraction and data analysis will provide a reference for future study exploring the association between the folate deficiency and cervical cancer.
Our results indicated that deficient serum folate level was associated with an increase in cervical cancer risk. A larger sample size of studies or meta-analysis is necessary for the future research to verify these results.
Acknowledgments We would like to thank all respondents of the study and all the people who give the help for this study.
Conflict of interest statement The authors have declared that no competing interests exist.