Introduction
Ovarian cancer is the leading cause of gynecology-related cancer death, with an estimated 22,240 new cases and 14,030 deaths in the United States alone in 2013 [
1]. Ovarian cancer is difficult to diagnose early, so it is crucial to identify risk factors in order to promote prevention. Unfortunately the etiology of ovarian cancer remains unclear. Some epidemiological studies suggest that genetic factors play an important role. Candidate risk factors include the genes encoding insulin receptor substrate 1 [
2], Lysyl oxidase G473A [
3], murine double minute 2 [
4], and progesterone receptor [
5].
Another candidate risk factor that has received a lot of attention is the gene encoding microsomal epoxide hydrolase (mEH). This gene is overexpressed in ovarian tissue [
6], where it plays a role in estrogen production [
7]. The protein is also a phase II metabolic enzyme that plays an important role in the activation and detoxification of exogenous chemicals, as well as in the metabolism of epoxides and endogenous steroids [
8,
9]. Since epoxides are highly reactive oxidative metabolites, mEH is thought to act on the most toxicologically active forms of drugs and environmental chemicals. Despite its important protective function, its net effects on the body can be complex, since it plays a dual role of procarcinogen detoxifier and activator [
10].
Recently, a number of studies have examined the possible association between the Tyr113His mEH (rs1051740) polymorphism and ovarian cancer risk [
11‐
15], but the results have been inconsistent. Since individual case–control studies may fail to detect complicated genetic relationships because of small sample size, we performed a meta-analysis of several published studies to provide a more rigorous test of the effects of mEH polymorphism on ovarian cancer risk.
Materials and methods
Search strategy
All clinical and experimental case–control studies of mEH polymorphism and ovarian cancer risk published through January 31, 2013 were identified through systematic searches in PubMed, EMBASE, Google Scholar, and Chinese National Knowledge Infrastructure databases. No language restrictions were imposed. The search terms used were: mEH; HYL1; EPHX; microsomal epoxide hydrolase; each of these four terms in combination with polymorphism, variation, genotype, genetic or mutation; each of the above terms in combination with ovarian cancer or carcinoma of ovary. Reference lists of relevant articles were also manually searched to identify additional relevant publications.
Inclusion criteria
A study was included in the meta-analysis if it satisfied the following criteria: (a) it assessed the association between ovarian cancer and the Tyr113His mEH gene polymorphism, (b) it used a case–control design, and (c) it provided sufficient published data for estimating an odds ratio (OR) with a 95% confidence interval (95% CI). In the case of multiple studies based on the same population, we selected the study with the largest number of participants.
Two authors (JHZ and LQL) independently searched the literature and identified eligible articles based on the inclusion criteria. The following data were extracted: first author’s family name, year of publication, patient ethnicity or country, numbers and genotypes of cases and controls, and Hardy-Weinberg equilibrium (HWE) of controls. Discrepancies were resolved by consensus. Only those studies that met the predetermined inclusion criteria were included.
Statistical methods
The unadjusted odds ratio (OR) with 95% confidence interval (CI) was used to assess the strength of the association between the Tyr113His mEH polymorphism and ovarian cancer risk based on the genotype frequencies in cases and controls. The potential association of different Tyr113His mEH genotypes with ovarian cancer risk was examined by comparing the C allele with the T allele, comparing homozygous genotypes, and applying recessive and dominant genetic models.
All statistical tests for this meta-analysis were performed using RevMan 5.14 (Cochrane Collaboration) and Stata 11.0 (StataCorp, College Station, USA). Fixed-effect and random-effect models were used to calculate pooled ORs. The statistical significance of the pooled ORs was determined using the Z-test, and P < 0.05 was considered statistically significant. The assumption of heterogeneity was evaluated by applying a chi-squared-based Q-test among the studies. In this approach, the Q value is defined to be identical to the effect size of the chi-squared test. P > 0.10 for the Q-test indicates a lack of heterogeneity and suggests that variability in effect sizes is larger than that expected from chance alone [
16]. In these cases, the fixed-effect model was used to calculate a pooled OR for each study. Otherwise, the random-effect model was used to calculate pooled ORs. HWE in the control group was assessed using the asymptotic test, with P < 0.05 considered significant. Small-study bias was assessed by Harbord’s modified test [
17]. As much as possible, the meta-analyis was carried out according to the PRISMA guidelines [
18].
Discussion
Activity of mEH is critical to the metabolism of xenobiotics and procarcinogens that may be involved in initiating cancers such as ovarian cancer. This meta-analysis aimed to evaluate the hypothesis that mEH polymorphism, by altering xenobiotic metabolism, may affect risk of developing ovarian cancer. The results suggest that in the total population and among Caucasians, although not necessarily among Asians, the CC genotype of Tyr113His mEH (rs1051740) is a risk factor of ovarian cancer according to homozygous and recessive contrast. These results were robust to sensitivity analysis in which one study [
14] using PCR-RFLP instead of allele-specific PCR was excluded. The available data are insufficient to determine whether particular mEH genotypes affect risk of particular histological subtypes of ovarian cancer.
The Tyr113His mEH polymorphism has been extensively investigated for its potential involvement in various types of cancers, such as colorectal cancer [
19], lung and upper digestive tract cancer [
20], esophageal carcinoma [
21], and hepatocellular carcinoma [
22]. Most of these studies, however, have shown that the polymorphism is not a risk factor, except in the case of hepatocellular carcinoma [
22]. The present study finds compelling evidence that the homozygous mEH CC genotype may increase ovarian cancer risk. Thus, Tyr113His mEH polymorphism may play an important role in the development of at least two common cancers.
Patients with different histological subtypes of ovarian cancer may have different prognoses [
23,
24], making it important to understand whether genetic risk applies differently to particular subtypes. Unfortunately this question could not be adequately addressed because most included studies did not report detailed genotype data for different subtypes. Nevertheless, four [
11‐
13,
15] of the five included studies reported genotype frequences for some histological subtypes. Among these four studies, two demonstrated that the TT genotype was associated with a decreased risk of endometrioid [
15] and serous [
11] ovarian cancer. Our meta-analyses based on three studies [
11,
13,
15] involving 922 cases and 718 controls found that the C allele significantly decreased endometrioid/clear cell ovarian cancer risk (Table
3). At the same time, a single study with 1571 cases and 2046 controls [
12] showed that the C allele significantly increased serous ovarian cancer risk. Unfortunately the single large study [
12] did not describe histological subtype data in detail, so it could not be combined in a meta-analysis with the other three [
11,
13,
15]. Small sample size may explain the discrepancy among the studies. Large, well-designed cohort studies examining the genetic susceptibility of different histological subtypes of ovarian cancer are needed.
The causes of ovarian cancer appear to be complex, involving multiple inherited [
25‐
27], environmental [
28,
29] and acquired factors [
30,
31]. While mEH polymorphism may participate directly in influencing risk of ovarian cancer, it may also do so through interactions with other genes or with the environment. For example, one study [
20] found cigarette smoking to be associated not only with increased risk of lung cancer but also with increased likelihood of mEH polymorphism. Arguing against this possibility, two studies [
12,
15] in the present meta-analysis failed to find evidence of synergistic interactions between alcohol or tobacco use and mEH polymorphism on risk of ovarian cancer. The remaining three included studies [
11,
13,
14] did not explore gene-environment or gene-gene interactions. Future studies should examine these interactions in detail, especially given the complexity of ovarian cancer risk factors.
The 113His codon variant is relatively common in the mEH gene, but allele frequency differs significantly among ethnicities. For example, the frequency of the homozygous CC genotype (His113His) is approximately 12% in Caucasians but 40% in Chinese (Table
1). These data are in line with other studies [
32,
33] showing that the frequency of the His113His variant is greatest in the Asian population (18-42%) and intermediate in the European population (about 10%). Surprisingly, despite its higher frequency in Asians, we did not find any association between the CC genotype or C allele and risk of ovarian cancer in our Chinese ethnic subgroup analysis. In contrast, meta-analysis of all five included studies, with a total participant population that was 95% Caucasian, showed the CC genotype to be a risk factor, and the largest included study [
12], involving Caucasians, indicated an association between the C allele and increased risk of serous ovarian cancer. Future work should examine carefully whether the mEH polymorphism affects risk of ovarian cancer differently in different ethnicities. It may be, for example, that the C allele normally increases cancer risk, as observed here in Caucasians, but no longer in Asians because of compensatory mutations with the population could not support such a high frequency of the risk-bearing allele.
The strengths of this meta-analysis include the large number of subjects investigated and the attempt to provide a complete picture taking in account ethnicity, genotyping method and histological subtype of ovarian cancer. However, our review also has several limitations. First, although we included clinical and experimental studies in both English and Chinese in order to avoid local literature bias [
34], the number of included studies was quite small. We were unable to increase this number even after systematically searching the databases four months after the original searches. Thus selective publication bias may exist. Second, the results may be affected by additional confounding factors, such as tumor status, age or gender, but most studies either did not report these baseline data or aggregated them in different ways, making it impossible to include them in the meta-analysis. In addition, the vast majority (95.2%) of data may came from Caucasian populations. The numbers of Chinese were relatively small (4.8%). Therefore, the results with Chinese subjects should be interpreted with caution, and future studies on Chinese and African populations should evaluate the ethnicity-specific effects of the mEH polymorphism observed here. In fact, future studies should be more careful about including ethnicity as a confounder variable, since some of the included studies from the US and UK did not report this in detail, potentially leading us to overestimate the proportion of Caucasians in our study population. Finally, although the CC genotype was significantly associated with an increased risk of ovarian cancer compared to the TT genotype or TT + TC genotypes, the 95% CI is near 1, suggesting that the association borders on nonsignificance. Given that the meta-analysis involved more than 5,000 cases and controls altogether, this finding suggests that the effects of mEH risk are likely to be small and modulated by interactions with other genes or the environment.
Implications for future practice and study: This meta-analysis suggests that the mutant CC genotype of Tyr113His mEH may be associated with increased risk of ovarian cancer. However, since this meta-analysis included few studies from non-Caucasian populations, large, well-designed studies in Asian and African populations are warranted to re-evaluate these associations. Our findings also highlight the need for larger, well-designed studies that take into account histological subtype as well as clinically relevant outcomes like overall survival and recurrence rate.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
ZJH contributed to study conception and design. ZJH and ZZM searched the databases and extracted data. ZJH performed all statistical analysis. ZJH and LLQ drafted the manuscript. All authors read and approved the final manuscript.