Association about dietary vitamin C intake on the risk of ovarian cancer: a meta-analysis

Abstract Changes in dietary vitamin C intake have been related to the risks of various cancers. However, the association between dietary vitamin C intake and the risk of ovarian cancer has not been fully determined. A meta-analysis was performed to evaluate the relationship between vitamin C intake and ovarian cancer risk. Observational studies that evaluated the association between vitamin C intake and ovarian cancer risk were identified via systematic search of PubMed and Embase databases. A random-effect model was used to combine relative risk (RR) with corresponding 95% confidence intervals (CIs). As a result, 16 studies (5 cohort studies and 11 case–control studies) with 4553 cases and 439,741 participants were included. Pooled results showed that dietary vitamin C intake had non-significant association on the risk of ovarian cancer (RR = 0.95, 95%CI = 0.81–1.11, I2 = 52.1%, Pfor heterogeneity = 0.008). Subgroup analyses according to characteristics including geographic location and study design showed consistent results with the overall result. In summary, findings from the present study indicated that dietary vitamin C intake is not associated with the risk of ovarian cancer.


Introduction
According to Globocan's estimate in 2018, cancer is the second leading cause of death worldwide, with an estimated 9.6 million deaths [1]. Ovarian cancer is still the most deadly gynecologic malignancy [2]. Meanwhile, it is also the leading cause of cancer-related death in women [2,3]. Previous paper estimated that there were 22,440 new cases and 14,080 deaths of ovarian cancer in 2017 [2]. Therefore, primary prevention of ovarian cancer is necessary. Although ovarian cancer is confirmed to be associated with many genetic factors [4,5], some dietary factors may also affect the development the risk of ovarian cancer. Dietary vitamin C intake has been linked to many cancers, such as pancreatic cancer [6,7], cervical neoplasia [8], renal cell carcinoma [9], esophageal cancer [10], prostate cancer [11], and so on. However, no comprehensive meta-analysis was performed to explore the relationship about vitamin C intake on the risk of ovarian cancer recently. Up to now, several studies have investigated the effectiveness of dietary vitamin C intake on the risk of ovarian cancer, and these results should be re-evaluated to provide robust pooled results. Therefore, the current meta-analysis of available observational studies was conducted to determine the role of vitamin C intake on the risk of ovarian cancer.

Data sources, search strategy, and selection criteria
The present study was performed and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Statement issued in 2009 [12]. Electronic searches for relevant studies about vitamin C intake and the risk of ovarian cancer were conducted of PubMed and Embase from their inception to May 31, 2019. The search terms included 'vitamin C' OR 'vitamin*' combined with 'ovarian Two authors independently conducted the literature search and selected the studies by reading the titles, abstracts, and full-text articles, and any disagreement was resolved by an additional author until consensus was reached.
Studies were included if they met the following criteria: (1) Patients: patients diagnosed with ovarian cancer and ≥18 years of age; (2) Study design: all the observational studies were acceptable; (3) Interested and outcomes: the studies should assess the association about dietary vitamin C intake on the risk of ovarian cancer; (4) Data: the study should provide the available data of relative risk (RR) and 95% confidence intervals (CI). Furthermore, we only included studies that explore the relationship about dietary vitamin C intake only, but not vitamin C supplement, on the risk of ovarian cancer.

Data collection
One author conducted the data collection according to a standard flowchart, while another author checked it. If any disagreement was detected, they discussed the issue until consensus was reached. The data collected included the family name of the first author, publication year, country, cases and participants, age, category of vitamin C intake, the value of RR and 95%CI, adjustment for factors.

Statistical analysis
The combined RR and 95%CI was pooled using STATA software (version 10.0, College Station, TX, U.S.A.) with a random-effects model [13]. Heterogeneity among the included studies was calculated using I-square and P values for Q statistic, and significant heterogeneity was defined as an I-square >50.0% or P < 0.10 [14,15]. The robustness of the pooled results was measured using a sensitivity analysis by sequential exclusion of individual trials. Funnel plot [16] and Egger test [17] were used to evaluate potential publication bias. The inspection levels for all pooled results were 0.05.

Literature search
The initial electronic searches produced 243 articles and one article was identified from the reference of reviews; of them, 205 were excluded due to irrelevant topics and duplication. The remaining 39 full articles were reviewed; of them, 16 articles [18-32] involving 4553 cases and 437,689 participants were included in the final analysis. Fourteen of the included studies come from North America, one from Europe and one from Asia. Five of the 16 articles were cohort design and the remaining 11 articles were case-control design. Table 1 summarizes the general characteristics of the patients and studies.

Publication bias and sensitivity analysis
The funnel plots were symmetry on visual inspection (Figure 3). Results of Egger's regression tests also did not indicate significant publication biases (P = 0.790). Sensitivity analysis showed that no single study had a potential influence on the pooled result ( Figure 4).

Discussion
In the current meta-analysis of 16 studies with 4553 cases and 437,689 participants, we found that the highest category compared with the lowest category of dietary vitamin C intake had no significant association on the risk of ovarian cancer. Moreover, by pooling the subgroup results of geographic locations and study design, we got consistent results with the overall result. Significant heterogeneity (I 2 = 52.1%, P for heterogeneity = 0.008) was found in the overall result about vitamin C intake on the risk of ovarian cancer. As far as we know, between-study heterogeneity is common in a meta-analysis, and it is an essential part to explore the sources of heterogeneity. We used meta-regression to explore the causes of heterogeneity for covariates of publication year, study design, geographic locations, and number of cases. We found that geographic locations (P = 0.017) may be a covariate that could influence this high heterogeneity. As seen in Figure 2, when we did the hierarchical analysis by geographic locations, the heterogeneity in North America was very low (I 2 = 2.0%, P for heterogeneity = 0.427). The I 2 in Europe and Asia was not detected due to only one study in each group. Even though, the result in North American populations was consistent with the overall result.
Although dietary vitamin C intake that is one of antioxidants had some potential role on preventing of cancers [6,8,10] due to inactivating free radicals and reducing oxidative DNA damage, we did not obtain an inverse association between dietary vitamin C intake and ovarian cancer. In our included studies, almost all researches got a non-significant relationship about vitamin C intake on the risk of ovarian cancer. The study by Chang et al. [30] indicated that dietary vitamin C intake (>665 mg/day vs. ≤75 mg/day) could significantly increase the risk of ovarian cancer. The value of highest category (>665 mg/day) was more than that in any other included studies. Otherwise, Zhang et al. [27] suggested that dietary vitamin C intake (≥140.25 mg/day vs. ≤66.50 mg/day) had a lower development on ovarian cancer risk. To our attention, the value of highest category (≥140.25 mg/day) was almost the lowest among all studies. Therefore, the current evidence showed that large amount of dietary vitamin C could not reduce the risk of ovarian cancer, and there may be harm.
Our study has some limitations which should be considered in interpreting the results. First, significant heterogeneity was detected among all the included studies, but it can be successfully explained by a covariate of geographic location. The association was not changed in North America populations. Second, only the subgroup analyses by geographic locations and study design were performed due to the limitation information provided in each individual study. Third, as a meta-analysis of observational studies, although all the included studies were adjusted for age, some related factors such as body mass index (BMI), total energy intake, duration of oral contraception use, and so on were not fully adjusted in every study. Fourth, almost all the included studies come from North America; therefore, more studies conducted in other populations are warranted to further explore the association between geographic locations and ovarian cancer risk. Finally, since we did not get a positive association between dietary vitamin C intake and the risk of ovarian cancer, the dose-response analysis between them was not performed.