Search strategy
The search strategy used was developed for the Pubmed and Medline, EMBASE, Scopus, Proquest, and Web of Science, databases. A librarian was involved in developing the search strategy. Medical Subject Heading Terms (MeSH) and keywords were used in various combinations to develop the search strategies. The search strategies used were reviewed by all the authors. A comprehensive electronic search was performed from the starting date of the databases up to the 30th of September 2022 according to the prepared search strategies without limitations related to language or publication status. Studies will be retrieved from the above mentioned databases using the following search strategy and the complete search strategy is attached as Annexure-1;
(Selenium) AND ((((((((Epithelial Ovarian Cancer) OR (Ovarian Cancer)) OR (Gynecological Malignancies)) OR (Ovarian Carcinoma)) OR (Ovarian tumor)) OR (Ovarian tumor)) OR (ovarian malignancy)) OR (female malignancies)) Sort by: Publication Date
In addition, the references of review articles, systematic reviews, meta-analyses, commentaries, editorials, meeting abstracts, and references of the included studies were evaluated for additional articles. Furthermore, books related to gynecological malignancies and manual searches of journals were done. Moreover, grey literature such as conference abstracts/proceedings, published lists of theses and dissertations, and other literature outside of the main journal literature, were searched. We searched for unpublished outcomes and studies by searching informal sources, including meeting abstracts and Ph.D. theses, and contacting the authors of the included studies.
Inclusion and Exclusion Criteria:
All observational studies (prospective/retrospective, case-control, and nested case-control) with adjusted risk estimates or provided data allowing the calculation of the risk estimates and 95% confidence interval (CI) for the association between Se and EOC, ecological studies, cross-sectional analytical studies, randomized control studies (RCTs) and nonrandomized clinical trials were selected. Studies published across all dates, times, and countries were included in the review. Studies published in other languages were translated into English by Google Translate.
Descriptive studies (i.e. case reports, case series, editorials, and opinion pieces) and animal studies were excluded from the review.
Study selection
Two review authors independently screened studies using COVIDENCE software for systematic reviews under the University of Queensland multiple systematic review license to identify eligible studies according to the following procedure (Fig. 1);
-
Studies were retrieved with COVIDENCE software and duplications were removed
-
The title and abstract of all the studies were assessed, and irrelevant studies were removed
-
The full texts of all the studies identified as possibly relevant were assessed
-
Selected studies for the systematic review and meta-analysis
Any inconsistencies between the two review authors regarding the eligibility of particular studies were resolved through discussion with a third reviewer.
Quality assessment
The quality of two review authors will independently assess the New Castle Ottawa scale for observational studies, the ROBINS-1 tool for quality assessment for the nonclinical randomized studies and the Cochrane risk of bias tool for randomized controlled trials. Disagreements between the two review authors over the risk of bias in particular studies will be resolved by discussion, with the involvement of the third review author where necessary.
Data extraction
The following general study information, where available, was independently retrieved from all included studies: author, publication year, study design, study setting; study population, participant demographics, baseline characteristics, details of the intervention and control conditions, recruitment and study completion rates, outcomes with risk estimates and 95% CIs, adjusted/matched factors for individual studies and times of outcome measurement, and information for the assessment of the risk of bias. Discrepancies were identified and resolved through discussion with the third author where necessary. As per the 2020 PRISMA-P expanded checklist, we will cite studies that appeared to meet the inclusion criteria, but were excluded, and explain why they were excluded.
Synthesis of the findings
One review author abstracted the data into standard evidence tables; the second review author checked the data for accuracy. The findings will be synthesized via a narrative description in the first place. Quantitative synthesis for the pooling of data will be performed by meta-analysis assuming that all of the studies are estimating the same underlying effect and that variation between their results is due to chance..
A subgroup analysis will be performed to examine any sources of significant heterogeneity according to the different types of studies (cohort, case-control, and cross-sectional analytical), study quality, and exposure (highest Se intake [food sources and supplements]) and controls. Heterogeneity between the studies in effect measures will be assessed by using the I2 statistic and Cochran’s Q test. The random effect inverse-variance model was used in the meta-analysis with STATA-17 with the DerSimonian-Laird estimate of tau² .
The results are displayed via forest plots of log ORs against log lower and log upper CIs. The existence of publication bias was also considered using funnel plot asymmetry, Egger”s p-value and Begg”s p-value. All analyses will be conducted using STATA-17.
Commentaries
Ovarian cancer is the most fatal of all gynecologic cancers (11). It is usually detected at advanced stages, and treatment is unlikely to be curative, therefore survival is poor. The majority of ovarian cancers are diagnosed with metastasis as the symptoms of ovarian cancers are often nonspecific (15). Due to the lack of recommended screening tools for ovarian cancer by the United States Preventive Service Task Force (USPTF), the identification of modifiable risk factors and preventive tools is essential for reducing ovarian cancer mortality and morbidity (2).
Selenium is a micronutrient and powerful antioxidant, therefore it reduces oxidative stress and ultimately prevents cell damage (16). There is some evidence that Se has some protective effect against several cancers (i.e. colon, prostate, lung stomach, and esophagus) (8) but epidemiological studies have yielded inconsistent results on the association between Se and EOC risk. Some studies have shown an inverse relationship between Se intake from food sources and EOC (6, 8), while, some other studies have given no association (9, 12). Some were given an inverse relationship between Se from supplements (9, 10) and EOC. In contrast, some other studies showed no associations (11, 20). No systematic reviews or meta-analyses of the association between Se and EOC risk have been conducted.
The search strategy for electronic databases was created with the assistance of an expert librarian at the University of Queensland. To minimize bias two independent reviewers screened the title and abstracts of the uploaded articles, screened the full texts of the selected articles, extracted the data, assessed the risk of bias in the selected studies, synthesized data. Conflicts will be resolved by the third review author. These are strengths of this study. There were no limitations related to the language or publication status of the article.
Although blood/serum levels of Se (14, 17) total toenail levels of Se (18) were considered as proxy measures of Se intake from food sources and supplements when we assessed the association of Se with EOC, blood/serum levels or tissue levels of Se in patients diagnosed with ovarian cancer in case-control studies were not suitable for inclusion in the systematic review and meta-analysis; therefore these studies were excluded from the review.
Confounding bias plays an important role in the true relationship between Se and EOC, and if the confounding bias is uniformly controlled in studies selected for the review, then the relationship between Se and EOC will be more ideal.
If there is a true inverse relationship between Se and EOC according to the meta-analysis this systematic review could provide an excellent basis for additional expanded and technical research on ovarian cancer prevention strategies.