Association between Dietary Vitamin E Intake and Esophageal Cancer Risk: An Updated Meta-Analysis

Epidemiological studies have provided ambiguous evidence on the association between vitamin E and esophageal cancer risk. To resolve this controversy, we performed this meta-analysis. The literature was searched by using Excerpta Medica Database (EMBASE), PubMed, the Web of Science, and the Cochrane Library from the inception to April 2018. A random effect model was utilized to calculate the odds ratio (OR) with the 95% confidence interval (95% CI). Twelve articles reporting 14 studies involving 3013 cases and 11,384 non-cases were included. By comparing the highest category with the lowest category of dietary vitamin E intake, we found that dietary vitamin E intake was inversely related to esophageal cancer risk (OR = 0.47, 95% CI: 0.36–0.60). Subgroup analysis revealed that dietary vitamin E intake had a significantly negative association with both the esophageal squamous cell carcinoma risk (OR = 0.29, 95% CI: 0.18–0.44) and the esophageal adenocarcinoma risk (OR = 0.66, 95% CI: 0.49–0.88). No study significantly affected the findings in the sensitivity analysis. Publication bias was discovered, however, the OR (95% CI) remained unchanged after the trim-and-fill analysis. This meta-analysis showed that the higher dietary vitamin E intake is associated with a lower esophageal cancer risk. However, the association still needs to be upheld by more large-scaled randomized controlled trials and prospective studies.


Introduction
Esophageal cancer is the eighth most common type of cancer worldwide. With an estimated 455,784 new cases and approximately 400, 169 deaths in 2012, esophageal cancer is the sixth most common cause of death from cancer. The highest mortality rates of esophageal cancer are in Eastern Asia (14.1 per 100,000) and Southern Africa (12.8 per 100,000) in men and in Eastern Asia (7.3 per 100,000) and Southern Africa (6.2 per 100,000) in women [1]. There are two main pathological subtypes of esophageal cancer: esophageal adenocarcinoma (EAC) and esophageal squamous cell carcinoma (ESCC). EAC is widespread in the western world, whereas ESCC is the predominant type for esophageal cancer worldwide [2,3]. EAC tends to occur in the distal esophagus from Barrett's esophagus [3]. ESCC, however, has a tendency to occur in the squamous epithelium undergoing inflammatory changes [3]. Many possible factors cause esophageal cancer, including smoking, drinking, hot tea, red meat, poor oral health, low intake of fresh fruit and vegetables, low socioeconomic status, high body mass index (BMI), and Helicobacter pylori (H. pylori) infection [3,4]. Owing to the poor survival rate for esophageal cancer (overall ratio of mortality to incidence was 0.88) [1], nutritional prevention has been increasingly focused by scholars in recent decades.
Vitamin E is a fat-soluble vitamin including eight homologues: four tocopherols (α-tocopherol (T), β-T, γ-T, δ-T) and four tocotrienols (α-tocotrienol (TT), β-TT, γ-TT, δ-TT). A-tocopherol, the major active form of vitamin E with strong antioxidant properties, can eliminate reactive oxygen species, inhibit carcinogenesis and tumor growth, and stimulate cancer cell apoptosis [5][6][7][8]. Accordingly, it was presumed that higher dietary vitamin E intake was inversely associated with the risk of esophageal cancer. Up to now, many epidemiological studies have explored the association between vitamin E intake and the esophageal cancer risk. Some case-control studies reported that dietary vitamin E intake was inversely related with esophageal cancer risk [8][9][10][11][12], however, some studies were unable to find the inverse association [13,14]. To solve the divergence mentioned above, we performed this meta-analysis to systematically and quantitatively estimate the association between dietary vitamin E intake and esophageal cancer risk based on the relevant literature.

Search Strategy
Two investigators independently searched the electronic databases of Excerpta Medica Database (EMBASE), PubMed, the Web of Science, and the Cochrane Library up to April 2018. The following terms were used for the search: "(Vitamin E or tocopherol or tocopherols or α-Tocopherol or γ-Tocopherol or β-Tocopherol or δ-Tocopherol) and (esophageal or esophagus or oesophageal or oesophagus) and (neoplasm or neoplasms or cancer or cancers or carcinoma or tumor or tumors or tumour or tumours)". In addition, the lists of references and reviews were also searched as potential supplements. The irrelevant articles were excluded by viewing the title, abstract, and full text. If one full-text could not be viewed or downloaded from the aforementioned electronic databases, we attempted to get the full-text by searching the online academic search and full-text delivery system or by contacting the corresponding author.

Study Selection
The inclusion criteria were the following terms: (1) cohort or randomized controlled trial (RCT) or case-control studies; (2) cases were diagnosed with esophageal cancer; (3) the dietary vitamin E intake was assessed by food frequency questionnaire and the vitamin E food composition database; (4) odds ratio (OR), hazard ratio (HR), or relative risk (RR) with their corresponding 95% confidence interval (CI) were reported or calculated; (5) detailed sample sizes were available; (6) articles were published in English; (7) as to dose-response analysis, the OR/HR/RR with 95% CI for esophageal cancer and person-years or the number of cases for each quantitative category of dietary vitamin E intake were available.
The following exclusion criteria were also considered: (1) nonhuman studies, review articles, letters to editor, or case reports; (2) cross-sectional studies; (3) studies where only vitamin E supplement use or serum vitamin E level were reported instead of dietary vitamin E intake; (4) studies where OR/RR/HR with 95% CI were not reported or calculated; (5) studies where detailed sample sizes were not reported.

Data Extraction and Quality Assessment
All data were extracted independently by two researchers and the ambiguities were resolved by a third investigator. The following information was extracted: (1) the last name of the first author; (2) publication year; (3) gender and age of participants; (4) the study period; (5) country; (6) source of control; (7) study design; (8) method of identifying cases; (9) the number of cases and non-cases; (10) assessment method of dietary vitamin E intake; (11) RR, OR, or HR (95% CI) for the maximizing adjusted model for the highest versus the lowest dietary vitamin E intake; (12) the adjusted covariance for multivariate analysis.
We assessed the methodological quality of included articles according to the instruction of Newcastle-Ottawa Scale (NOS) [15], including three quality parameters: selection, comparability, and outcome (for prospective studies) or exposure (for case-control studies). For the NOS score, one star means one point.

Statistical Analysis
OR, HR, and RR were used for analyzing dichotomous data in one meta-analysis. OR was a fair approximation of RR when the incidence of an event was rare. Since esophageal cancer incidence is small, OR and HR are theoretically similar to RR [16]. Therefore, OR was used only for pooling the dichotomous data in this meta-analysis. The heterogeneity among studies was estimated by Cochrane's Q and I 2 statistics. Generally, the 25%, 50%, and 75% of I 2 values refer to low, moderate and high levels of heterogeneity, respectively [17]. If I 2 > 50% and p < 0.05, a random-effect model was applied to pool OR (95% CI), if not, a fixed-effect model was used for pooling OR (95% CI) [18].
We performed meta-regression and subgroup analysis to explore the potential source of heterogeneity based on pathological type (ESCC, EAC, or mixed subtype), geographical distribution (America, Europe, or others), study type (case-control, or cohort study), source of control (population-based or hospital-based), sample size (≥500 or <500), the method for identifying esophageal cancer cases (pathological diagnoses or cancer registries), dietary vitamin E assessment method (validated food frequency questionnaire (FFQ) or unvalidated FFQ), quality score (>5 or ≤5), OR/HR/RR adjusted for age (yes or no), gender (yes or no), smoking (yes or no), drinking (yes or no), and BMI (yes or no), respectively. In addition, sensitivity analyses were also performed to explore the potential study, which greatly affected the robustness of the result. Due to different definitions of exposure categories in separate studies, a fixed-effect dose-response analysis was performed to quantitatively estimate the potential association between the increment per day for dietary vitamin E intake and the esophageal cancer risk. We recalculated the pooled OR of esophageal cancer for every 3 mg/day increment of dietary vitamin E intake through previously described methods [19,20]. Furthermore, a non-linear dose-response analysis was likewise performed by using restricted cubic splines with three knots at fixed percentiles (10%, 30% and 50%) of the distribution of exposure [21,22]. We calculated the overall p-value by testing that the two regression coefficients were simultaneously equal to zero. We also calculated p-values for non-linearity by testing that the coefficient of the second spline was equivalent to zero. The details of this method have been cited previously [23,24]. The aforementioned method involves the following items: (1) the number of cases and person-years or non-cases and the RRs or ORs with the variance estimates for at least three quantitative exposure categories; and (2) the median or mean level of these exposures in each category (if ranges was reported, mean levels were calculated by averaging the lower and upper boundary; if the lower boundary of the lowest category was not reported, we set zero as the lower boundary; if the upper boundary of the highest category was not reported, we added the same amplitude with the closest category to the lower boundary as the upper boundary) [25,26].
Finally, Egger's and Begg's test were performed to assess the potential publication bias. If both of p values for Egger's test and Begg's test were less than 0.05 (p < 0.05), it indicated the publication bias existed [27]. If the publication bias existed, we used the "trim and fill" method to add the potentially missing studies for adjustment for the funnel plot's asymmetry [28,29], and observed the variation of pooled OR (95% CI). If the pooled OR (95% CI) changed weakly or remained unchanged, it indicated that the result was robust. All the statistical analyses were performed by using Stata 11.2 software (Stata Corp Lakeway Drive, College Station, TX, USA), and a two-sided p value < 0.05 was considered statistically significant.

Meta-Analysis of Dietary Vitamin E Intake and the Esophageal Cancer Risk
Twelve articles involving 14 studies on dietary vitamin E intake and the esophageal cancer risk were included in this meta-analysis. The pooled OR of the esophageal cancer risk for the highest vs. lowest intake of dietary vitamin E was calculated. In a comprehensive analysis, we found that dietary vitamin E intake was inversely associated with the esophageal cancer risk (OR = 0.47, 95% CI: 0.36-0.60) with significant heterogeneity (I 2 = 67.5%; p < 0.001) (Figure 2, Table 2).

Meta-Regression and Subgroup Analysis
Since significant heterogeneity was found in this meta-analysis; meta-regression and subgroup analysis were performed to explore the source of the between-study heterogeneity. Meta-regression analysis found adjustment for variables like age (p = 0.044), gender (p = 0.017), smoking (p = 0.044), and BMI (p = 0.017) was the possible source of the heterogeneity. The following covariance such as geographical distribution (p = 0.475), study type (p = 0.403), source of control (p = 0.433), sample size (p = 0.821), pathological type (p = 0.072), the method of identifying esophageal cancer cases (p = 0.463), dietary vitamin E assessment method (p = 0.394), quality score (p = 0.355), and adjustment for drinking (p = 0.261) did not explain the between-study heterogeneity.

Sensitivity Analysis, Publication Bias Analysis, and Trim and Fill Analysis
A total of four articles reporting five studies involving 970 cases and 2225 non-cases [2,8,14,33] were included in the dose-response analyses based on the relevant criteria mentioned in the method of statistical analysis. Among the four articles [2,8,14,33], the dietary vitamin E intake per day was defined mg/day, except for Murphy's article [14], which was defined intake in µg/day. Therefore, we turned the µg/day to mg/day by dividing 1000. Overall analysis, no heterogeneity was found (p = 0.987), no significant non-linear dose-response relationship was found (pnon-linearity = 0.864), but a marginally linear dose-response relationship was found for 3 mg/day increment of dietary vitamin E intake and the risk of esophageal cancer (OR = 0.78, 95% CI: 0.57-1.06, plinearity = 0.114) ( Figure 3A). For the sensitivity analysis, no one specific study significantly affected the pooled OR (95% CI) (Supplementary Figure S1). Both Egger's test and Begg's test found publication bias existed (p = 0.008, and p = 0.005, respectively). Then, trim and fill analysis was performed for the adjustment of funnel plot asymmetry; no missing studies were added and the adjusted OR was 0.47 (95% CI: 0.36-0.60), which showed the pooled OR (95% CI) remained unchanged. The funnel plot of trim and fill analysis was shown in Figure 3B.

Discussion
Up to now, epidemiology studies have provided ambiguous evidences for the relationship between dietary vitamin E intake and the esophageal cancer risk. To solve the divergence, we performed a meta-analysis on the association between dietary vitamin E intake and esophageal cancer risk (including EAC and ESCC). We found that the higher dietary vitamin E intake was related to a lower esophageal cancer risk, especially for ESCC. The aforementioned findings were in line with many other studies of dietary vitamin E intake and various cancer risks. One published dose-response meta-analysis has suggested that bladder cancer risk decreased by 17% for every 10 mg/day increase of dietary vitamin E intake (RR = 0.83, 95% CI: 0.72-0.95) [35]. Another meta-analysis focused on the association between dietary antioxidant vitamins intake and gastric cancer, which showed the higher dietary vitamin E intake decreased the risk of gastric cancer by 35% (OR = 0.65, 95% CI: 0.57-0.74), and the gastric cancer risk decreased by 21% for every 15 mg/day increase of dietary vitamin E intake (OR = 0.79, 95% CI: 0.66-0.94) [36]. The findings of one latest meta-analysis containing 11 studies (10 prospective studies and one case-control study) also showed that the dietary vitamin E intake was inversely associated with the risk of lung cancer (RR = 0.86, 95% CI: 0.74-0.99) [37]. The findings of one recent meta-analysis containing 11 studies (10 prospective studies and one case-control study) also showed that the dietary vitamin E intake was inversely associated with the risk of lung cancer (RR = 0.86, 95% CI: 0.74-0.99) [37]. Therefore, the combination of the meta-analyses mentioned above and our study further suggested that dietary vitamin E intake was inversely associated with many cancer risks, which might provide relevant population-based evidence for vitamin E preventing tumorigenesis, including in esophageal cancer.
In 2007, Kubo's meta-analysis explored the association between antioxidant intake (including vitamin E) and the risk of esophageal and gastric cardia adenocarcinoma, and showed a marginally inverse association between vitamin E intake and EAC risk (OR = 0.80, 95% CI: 0.63-1.03), but did not report the relationship between vitamin E intake and ESCC risk [38]. In Kubo's research, a total of three case-control studies on the association between vitamin E intake and the EAC risk were included, two of which were on dietary vitamin E intake and the EAC risk and one of which was on dietary α-tocopherol and EAC risk. α-tocopherol is one of vitamin E's homologues, but we mainly take in total vitamin E not just α-tocopherol from various foods such as cereal grains, nuts, particularly vegetable oils [39]. The dietary vitamin E intake was calculated by using the following formulas that came from the literature of Yang Yue et al. [40]: "α-tocopherol equivalent (α-TE, mg) = 1.0 × α-tocopherol (mg) + 0.5 × β-tocopherol (mg) + 0.1 × γ-tocopherol (mg) + 0.02 × δ-tocopherol (mg) + 0.3 × α-tocotrienols", or " vitamin E (mg) = α-tocopherol (mg) + (β + γ)-tocopherol (mg) + δ-tocopherol (mg)" [41]. Importantly, dietary α-tocopherol intake is not equivalent to dietary vitamin E intake. Therefore, Carman's and Terry's studies were excluded in our meta-analysis because of only reporting dietary α-tocopherol intake [40,42].
The present meta-analysis showed that higher vitamin E intake is negatively associated with the risk of esophageal cancer. This may be explained by the strong oxidative properties of vitamin E [43,44]. One previous study showed that vitamin E could inhibit DNA oxidative damage and lipid peroxidation by reducing the level of endogenous reactive oxygen species, and thereby prevented gene mutations as well as tumor formation [45]. Additionally, in N-nitrosomethylbenzylamine (NMBZA)-treated rats, 8OH-dG was significantly increased in various esophageal lesions, whereas early continuous supplementation of vitamin E reduced 8OH-dG in each category of lesions [46]. In vivo, vitamin E may also suppress NMBZA-induced carcinogenesis in rat esophagus by blocking activation of nuclear factor-kappa B (NFκB) and abnormal metabolic pathway of arachidonic acid [47].
Based on the results of meta-regression and subgroup analysis, we found that the possible sources of between-study heterogeneity for this meta-analysis were adjustment for variables like age, gender, smoking, and BMI. As we know, the incidence of esophageal cancer increases with age, and the incidence of esophageal cancer worldwide in males is more than double that in females (2.4:1). Furthermore, one meta-analysis of 12 studies showed that about 68% esophageal cancer cases were attributed to smoking (RR = 3.01, 95% CI: 2.30-3.94) [48]. As for BMI, it was widely used for determining obesity and overweight. When the dose increased by per 5 kg/m 2 for BMI, the esophageal cancer risk was enhanced by approximately 10% (RR = 1.11, 95% CI: 1.09-1.14) [49,50]. The aforementioned findings may respectively explain the reasons for adjustment for variables like age, gender, smoking, and BMI as the potential sources of heterogeneity for this meta-analysis. Besides, when stratified by pathological type, we found the heterogeneity was reduced, which revealed that pathological type was also the potential source of heterogeneity. The higher dietary vitamin E intake has a negative correlation with the risk of ESCC, which were more than the negative correlation double those for the EAC risk.
To our knowledge, the present meta-analysis had several advantages Firstly, the study was a quantitative and comprehensive dose-response analysis to evaluate the association between dietary vitamin E intake and the risk of esophageal cancer. Secondly, meta-regression and subgroup analysis were simultaneously performed to explore the potential sources of heterogeneity. Thirdly, sensitive analysis found no one specific study significantly affected the pooled OR (95% CI). Fourthly, trim and fill analysis did not find potentially missing articles, and the pooled OR (95% CI) was maintained after the adjustment of the asymmetry of the funnel plot, which revealed that our findings might be robust and sound.
However, several limitations should also be taken into consideration. Firstly, most of the eligible studies were case-control studies, which were prone to cause recall bias and selection bias. Secondly, dietary vitamin E intakes were assessed by validated FFQs [8][9][10][11][12][13][14]30], and dietary vitamin E intakes were assessed by unvalidated FFQs [31][32][33][34]. The former was more applicable than the latter, and 33% articles used not-validated FFQ, which might affect the estimation of the dietary vitamin E intake. Thirdly, only a few studies were included in the dose-response analysis, and the corresponding sample size was small. Therefore, the association between the increase of per day dietary vitamin E intake and the esophageal cancer risk should be explored by using more usable data in the future investigation. Last but not least, there was a higher mortality rate for esophageal cancer in Eastern Asia according to the data from GLOBOCAN 2012, but most studies were conducted in Europe and America, and only few studies were performed in Asia. This may induce regional restrictions to the findings, and they may not be generalized to Asian population very well. Therefore, studies on dietary vitamin E and esophageal cancer research should be expanded, especially in areas with high incidence of esophageal cancer, such as in the Asian region. The association between dietary vitamin E intake and esophageal cancer risk should be researched in more areas worldwide.

Conclusions
In summary, our study suggests that dietary vitamin E intake is inversely associated with the esophageal cancer risk. Considering the limitations of the included studies, more large-scale prospective studies and randomized controlled trials are still needed to confirm our results, though our findings could provide certain evidence for esophageal cancer prevention.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6643/10/7/801/s1. Figure S1: The sensitive analyses for studies on the association between dietary vitamin E intake and the esophageal cancer risk. Table S1: Methodological quality of case-control studies included in the meta-analysis, Table S2: Methodological quality of prospective studies included in the meta-analysis.
Author Contributions: L.C. designed the study; L.C., L.L and Y.T. assessed studies for inclusion, extracted data, and assessed validity; L.C. and L.L. and conducted meta-analyses; F.X. and T.Q. tabulated data; L.C. wrote the first draft of the manuscript; L.L. and F.X. provided critical review of the manuscript.