DOI: https://doi.org/10.21203/rs.2.24262/v1
Background: Numerous studies conducted over the past 30 years have pointed to the presence of Epstein–Barr virus (EBV) in gastric cancer samples. This study was aimed to provide a meta-analytic review of the prevalence of EBV in gastric cancer patients, and to clarify the relationship between EBV infection and gastric cancer.
Methods: A literature search was performed electronically using online databases for English language publications until July 1, 2019. The pooled EBV prevalence and 95% confidence intervals (CIs) were estimated using a random effects model. To determine the association between EBV and gastric cancer, pooled odds ratio (OR) and its 95% CI were computed for case-control studies with matched pairs design.
Results: The pooled prevalence of EBV in 20411 gastric cancer patients was 8.78% (95% CI: 7.75-9.93%; I 2 =83.0%). The proportion of EBV-associated gastric cancer among male cases was significantly higher than among female cases (10.85%, vs 5.72%) ( P <0.01). EBV was more prevalent in the cardia (12.47%) and in the body (11.68%) compared to the antrum (6.29%) ( P <0.01). There were 20 studies with matched pairs design, including tumor and tumor-adjacent normal tissue pairs from 4116 gastric cancer patients. The pooled OR between EBV infection and gastric cancer risk was 18.56 (95% CI: 15.68–21.97; I 2 = 55.4%).
Conclusion: EBV infection is associated with more than 18 times increase risk of gastric cancer. Although the prevalence of EBV was higher in male patients than in female patients with gastric cancer, women are more likely than men to develop EBV-associated gastric cancer.
According to GLOBOCAN statistics in 2018, gastric cancer is the fifth most frequently diagnosed cancer and the third leading cause of cancer-related mortality in the world, accounted for 8.2% of all cancer deaths. Over 1,000,000 new cases of gastric cancer diagnosed in 2018 around the world, with an estimated 783,000 deaths [1]. Gastric cancer arises from a combination of multiple environmental and genetic risk factors, and infectious agents are one of the critical environmental factors which contribute to an increased risk of developing a number of malignancies [2].
Epstein-Barr virus (EBV), as a member of the Herpesviridae family, is the first described human cancer virus, and is responsible for approximately 1.8% of all human cancers, including Hodgkin lymphoma, Burkitt lymphoma, NK/T cell lymphoma, and nasopharyngeal carcinoma [3]. However, the role of EBV in the development of other malignancies is still under investigation. The first discovery of EBV from gastric cancer tissues made by Burke et al in 1990 using the polymerase chain reaction (PCR) technique [4]. Subsequently, in situ hybridization (ISH) technique developed in laboratories to detect EBV-encoded small RNAs (EBERs) and led to the facilitation of EBV detection in cancer tissues. Since then, numerous studies have been conducted in various countries to investigate the prevalence and the role of EBV infection among gastric cancer patients.
Our meta-analysis aims to determine the association of EBV infection with gastric cancer, and to provide an updated pooled prevalence for EBV infection among gastric cancer patients. It is anticipated that the results of the present study will direct future experimental studies toward the elucidating the role of EBV infection in the carcinogenesis of gastric cancer, and will inform clinicians and policy-makers to improve preventive intervention and control.
The present systematic review and meta-analysis was performed according to the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [5].
Search Strategy
A rigorous literature search was conducted using PubMed, Web of Science, Scopus, EMBASE, and Google scholar to identify all published articles reporting the prevalence of EBV in patients with gastric cancer. Databases were searched from inception to July 1, 2019. The bibliographies of all articles obtained were also reviewed for additional relevant publications. The list of keywords used for this systematic review and meta-analysis is provided in Appendix 1.
Study Selection
All records were imported to EndNote software version X8 (Thomson Reuters, California, USA) and duplicate entries were removed. The screening of the title and abstract of the remaining records was independently conducted by two researchers. The full-texts of the remaining records were then retrieved and reviewed, and any disagreements were resolved through discussion by a third investigator.
Eligibility Criteria
Studies were considered eligible for inclusion in the present meta-analysis, if they met the following criteria: (1) Studies using cross-sectional and case-control designs reporting the prevalence of EBV infection in patients with different types of gastric carcinoma; (2) Studies using EBER-ISH technique to detect the presence of EBV transcripts or nucleic acids; (3) Studies using the formalin-fixed paraffin-embedded (FFPE) tissues and biopsies samples; (4) Studies published in peer-reviewed journals in English language.
Studies with following characteristics were excluded from the present meta-analysis: (1) Studies using serological techniques such as enzyme-linked immunosorbent assay (ELISA) to detect circulating antibodies to EBV infection; (2) Studies evaluating the presence of EBV in serum, plasma or peripheral blood mononuclear cell (PBMC) samples; (3) Studies evaluating the presence of EBV in gastric carcinoma patients with underlying disorders; (4) Studies evaluating the presence of EBV by molecular methods such as PCR, nested-PCR and Real-Time PCR; (5) Studies addressing remnant gastric cancer, gastric lymphoma, and other types of gastric malignancies; (6) Studies using techniques other than EBER-ISH, (7) Studies published in languages other than English; (8) Reviews, letters to the editor, abstracts, and case reports.
Data Extraction
Two investigators independently extracted data from all eligible studies in a pre-designed data extraction form using Microsoft Excel 2013 (Microsoft Corporation, Redmond, Washington, USA). The two investigators cross-checked each other's data extraction and any disagreements were resolved by a third investigator. The following characters were extracted from each study: first author’s name, publication date, study location, study design, sample size, sex, type of specimen, histological type, number of EBV-positive samples, tumor anatomical location, depth of invasion, tumor stage, and lymph node invasion.
Statistical analysis
The present meta-analysis had two main purposes; first, providing an updated estimate of the pooled prevalence of EBV among patients with gastric cancer, and secondly, investigating the association between EBV and the development of gastric cancer. A random-effect meta-analysis using the inverse variance method was applied to estimate the pooled prevalence of EBV (DerSimonian-Laird method) [6]. The logit transformation was used for stabilizing the variance and data normalization, and the Clopper-Pearson method was applied to determine the 95% confidence intervals (CIs) for proportions [7].
In order to evaluate the strength of the association between EBV infection and gastric cancer risk, the pooled odds ratios (ORs) with 95% CIs were generated from a random effects model based on the DerSimonian-Laird method. For studies with a zero cell, a continuity correction of 0.5 was applied. We also conducted the subgroup analyses to identify the possible sources of heterogeneity. The heterogeneity among the studies was assessed by means of I2 statistics [8]. To explore potential publication bias and symmetric assumption among the included studies, a Begg's funnel plot was constructed [9]. All the above-mentioned analyses were conducted using the R package Meta (version 3.5.3 [2019–03–11]) [10, 11], and P values less than 0.05 were considered statistically significant. Furthermore, for each case-control study with matched pairs design, we separately computed matched-pairs OR and its corresponding variance using “escalc” function, in the R “metafor” package [12] (version 2.1-0 [2019–05–13]. The obtained results were then used for performing meta-analysis to calculate the matched pairs pooled OR.
Literature selection
The electronic database searches were identified 597 articles, and additional 14 relevant records were found through bibliographic hand searching. Of these 611 articles, 151 duplicates were excluded, so a total of 460 articles was screened according to their title and abstract. A total of 353 articles was eliminated after reading the title and abstract due to obvious irrelevance. The remaining 107 articles were assessed for agreement with the inclusion and exclusion criteria by the full-text review, and finally, 72 papers were included in this systematic review and meta-analysis. Figure 1 shows the process of literature retrieval and screening using a flow chart.
Study Characteristics
Table 1 shows the characteristics of eligible studies included in the systematic review and meta-analysis. Out of 72 studies, 30 were case-control and 42 were cross-sectional in design. Publication dates ranged from 1993–2019, and over half of the studies (59.7%) described specimens recruited prior to 2005. Among the studies included in this meta-analysis, four were from Africa, 16 were from America, 36 were from Asia, and 17 were from Europe. Of the 73 studies included, 47 provided information on patients’ sex, 41 studies provided information on histological type and 35 had information on tumor anatomical location. The largest study included 2226 gastric cancer cases and the smallest included 19 cases. Most studies were from Japan (n = 15).
Author [Ref.] | Year | Location | Study design | Type of sample | No. of case | No. of control | No. of case positive | No. of control positive |
---|---|---|---|---|---|---|---|---|
Rowlands [13] | 1993 | UK and Japan | Cross-sectional | FFPE | 174 | 9 | ||
Shibata [14] | 1993 | USA | Cross-sectional | FFPE | 187 | 19 | ||
Tokunaga [15] | 1993 | Japan | Cross-sectional | FFPE | 1848 | 122 | ||
Tokunaga [16] | 1993 | Japan | Cross-sectional | FFPE | 999 | 69 | ||
Imai [17] | 1994 | Japan | Case-control | FFPE | 1000 | 1000 | 70 | 0 |
Ott [18] | 1994 | Germany | Case-control | FFPE | 39 | 39 | 7 | 0 |
Shousha [19] | 1994 | UK | Case-control | FFPE | 19 | 9 | 1 | 5 |
Yuen [20] | 1994 | China | Case-control | FFPE | 74 | 36 | 7 | 0 |
Harn [21] | 1995 | Taiwan | Case-control | FFPE | 55 | 49 | 6 | 0 |
Gulley [22] | 1996 | USA | Case-control | FFPE | 95 | 95 | 11 | 0 |
Moritani [23] | 1996 | Japan | Case-control | FFPE | 132 | 132 | 15 | 0 |
Selves [24] | 1996 | France | Case-control | FFPE | 59 | 59 | 5 | 0 |
Shin [25] | 1996 | South Korea | Case-control | FFPE | 89 | 37 | 12 | 0 |
Galetsky [26] | 1997 | Russia | Case-control | FFPE | 206 | 206 | 18 | 0 |
Clark [27] | 1997 | Singapore | Cross-sectional | FFPE | 137 | 6 | ||
Ojima [28] | 1997 | Japan | Cross-sectional | FFPE | 412 | 83 | ||
Yanai [29] | 1997 | Japan | Cross-sectional | FFPE | 124 | 12 | ||
Herrera-Goepfert [30] | 1999 | Mexico | Cross-sectional | FFPE | 135 | 11 | ||
Kume [31] | 1999 | Japan | Case-control | FFPE | 344 | 344 | 40 | 0 |
Takano [32] | 1999 | Japan | Cross-sectional | FFPE | 513 | 33 | ||
Wan [33] | 1999 | China | Case-control | FFPE | 58 | 58 | 6 | 0 |
Chapel [34] | 2000 | France | Case-control | FFPE | 56 | 56 | 7 | 0 |
Wu [35] | 2000 | Taiwan | Cross-sectional | Biopsy | 150 | 30 | ||
Corvalan [36] | 2001 | Chile | Case-control | FFPE | 185 | 185 | 31 | 0 |
Kijima [37] | 2001 | Japan | Cross-sectional | FFPE | 313 | 23 | ||
Ishii [38] | 2001 | Japan | Cross-sectional | FFPE | 119 | 23 | ||
Koriyama [39] | 2001 | Brazil | Cross-sectional | FFPE | 300 | 24 | ||
Luqmani [40] | 2001 | UK | Case-control | FFPE | 20 | 79 | 1 | 9 |
Burgess [41] | 2002 | UK | Cross-sectional | FFPE | 534 | 9 | ||
Kang [42] | 2002 | South Korea | Cross-sectional | FFPE | 233 | 21 | ||
Kattoor [43] | 2002 | India and Japan | Cross-sectional | FFPE | 2226 | 135 | ||
Vo [44] | 2002 | USA | Cross-sectional | FFPE | 107 | 11 | ||
Czopek [45] | 2003 | Poland | Cross-sectional | FFPE | 40 | 5 | ||
Karim [46] | 2003 | Malaysia | Cross-sectional | FFPE | 50 | 5 | ||
Oda [47] | 2003 | Japan | Case-control | FFPE | 97 | 97 | 5 | 0 |
Ishii [48] | 2004 | Japan | Case-control | FFPE | 133 | 133 | 19 | 0 |
Lee [49] | 2004 | South Korea | Cross-sectional | FFPE | 1127 | 63 | ||
Lopes [50] | 2004 | Brazil | Case-control | FFPE | 53 | 53 | 6 | 0 |
van Beek [51] | 2004 | Netherlands | Cross-sectional | FFPE | 566 | 41 | ||
Alipov [52] | 2005 | Kazakhstan | Case-control | FFPE | 139 | 139 | 14 | 0 |
Herrera-Goepfert [53] | 2005 | Mexico | Case-control | FFPE | 330 | 330 | 24 | 2 |
Luo [54] | 2005 | China | Case-control | FFPE | 172 | 172 | 11 | 0 |
Yoshiwara [55] | 2005 | Peru | Cross-sectional | FFPE | 254 | 10 | ||
Campos [56] | 2006 | Colombia | Cross-sectional | FFPE | 368 | 42 | ||
Szkaradkiewicz [57] | 2006 | Poland | Cross-sectional | FFPE | 32 | 14 | ||
Luo [58] | 2006 | China | Cross-sectional | FFPE | 185 | 13 | ||
von Rahden [59] | 2006 | Germany | Case-control | FFPE | 82 | 82 | 5 | 0 |
Abdirad [60] | 2007 | Iran | Cross-sectional | FFPE | 273 | 9 | ||
Jung [61] | 2007 | South Korea | Cross-sectional | FFPE | 111 | 7 | ||
Lima [62] | 2008 | Brazil | Cross-sectional | FFPE | 71 | 6 | ||
Ryan [63] | 2009 | USA | Cross-sectional | FFPE | 113 | 11 | ||
Trimeche [64] | 2009 | Tunisia | Cross-sectional | FFPE | 96 | 4 | ||
Truong [65] | 2009 | USA | Case-control | FFPE | 235 | 72 | 12 | 0 |
Ferrasi [66] | 2010 | Brazil | Case-control | FFPE | 54 | 54 | 5 | 0 |
Koriyama [67] | 2010 | Japan | Cross-sectional | FFPE | 156 | 21 | ||
Chen [68] | 2010 | China | Case-control | FFPE | 676 | 676 | 45 | 3 |
Boysen [69] | 2011 | Denmark | Cross-sectional | FFPE | 131 | 10 | ||
BenAyed-Guerfali [2] | 2011 | Tunisia | Cross-sectional | FFPE | 81 | 12 | ||
de Lima [70] | 2012 | Brazil | Cross-sectional | FFPE | 160 | 11 | ||
Ksiaa [71] | 2014 | Tunisia | Cross-sectional | FFPE | 43 | 4 | ||
Aslane [72] | 2016 | Algeria | Case-control | FFPE | 97 | 10 | 22 | 0 |
Tsai [73] | 2016 | Taiwan | Cross-sectional | FFPE | 1039 | 52 | ||
Zhang [74] | 2016 | China | Cross-sectional | FFPE | 600 | 30 | ||
Liu [75] | 2016 | China | Case-control | FFPE | 206 | 206 | 15 | 0 |
Na [76] | 2017 | South Korea | Cross-sectional | FFPE | 205 | 15 | ||
Boger [77] | 2017 | Germany | Cross-sectional | FFPE | 484 | 22 | ||
Kim [78] | 2017 | South Korea | Case-control | FFPE | 207 | 56 | 13 | 0 |
Nogueira [79] | 2017 | Portugal | Case-control | FFPE | 82 | 33 | 9 | 1 |
Ribeiro [3] | 2017 | Portugal | Cross-sectional | FFPE | 179 | 15 | ||
de Souza [80] | 2018 | Brazil | Cross-sectional | Biopsy | 302 | 62 | ||
Wanvimonsuk [81] | 2018 | Thailand | Case-control | FFPE | 33 | 55 | 4 | 0 |
Martinez-Ciarpaglini [82] | 2019 | Spain | Cross-sectional | FFPE | 209 | 13 |
The prevalence of EBV among gastric cancer patients
The first aim of the current study was to determine the pooled prevalence of EBV in 20411 gastric cancer patients from 27 countries, and the range was from 1.69–43.75% of the selected individual studies. The pooled prevalence of EBV among gastric cancer patients was 8.78% (95% CI: 7.75–9.93%; I2 = 83.0%). The highest and lowest prevalence of EBV were found in gastric cancer patients from the Poland and the United Kingdom, respectively (25.57%, 95%CI: 6.13–64.36% vs 2.78%, 95%CI: 1.51–5.06%). Table 2 presents more detailed information on the prevalence of EBV infection in gastric cancer patients for subgroups.
Characteristics | Categories | No. of Studies | Pooled prevalence (%) (95% CI) | Heterogeneity test I2 %, p-value | Differences between subgroups; χ2 test (p-value) |
---|---|---|---|---|---|
Overall | - | 72 | 8.78 (7.75–9.93) | 83.0%, P < 0.01 | - |
Study design | Cross-sectional | 42 | 8.25 (6.93–9.78) | 87.9%, P < 0.01 | P = 0.17 |
Case-control | 30 | 9.71 (8.32–11.30) | 59.4%, P < 0.01 | ||
Publication date | ≤ 2005 | 43 | 8.93 (7.68–10.36) | 82.1%, P < 0.01 | P = 0.76 |
> 2005 | 29 | 8.56 (6.81–10.71) | 84.6%, P < 0.01 | ||
Sex | Male | 47 | 10.85 (9.47–12.41) | 72.6%, P < 0.01 | P < 0.01† |
Female | 47 | 5.72 (4.28–7.61) | 74.3%, P < 0.01 | ||
Study location | Africa | 4 | 11.93 (5.97–22.44) | 76.8%, P < 0.01 | P = 0.65 |
America | 16 | 9.51 (7.45–12.07) | 76.8%, P < 0.01 | ||
Asia | 36 | 8.41 (7.19–9.81) | 83.6%, P < 0.01 | ||
Europe | 17 | 8.21 (5.82–11.46) | 80.4%, P < 0.01 | ||
Development status | Developed countries | 34 | 8.42 (7.11–9.94) | 82.1%, P < 0.01 | P = 0.63 |
Developing countries | 40 | 8.94 (7.44–10.72) | 83.1%, P < 0.01 | ||
Sample type | FFPE | 70 | 8.51 (7.56–9.56) | 79.7%, P < 0.01 | P < 0.01† |
Biopsy | 2 | 20.36 (16.89–24.32) | 0%, P = 0.9 | ||
Lauren's histological type | Intestinal type | 41 | 8.15 (6.71–9.87) | 68.5%, P < 0.01 | P = 0.34 |
Diffuse type | 41 | 9.40 (7.55–11.65) | 76.4%, P < 0.01 | ||
Tumor anatomical location | Cardia | 32 | 12.47 (10.39–14.89) | 24.8%, P = 0.1 | P < 0.01† |
Body | 32 | 11.68 (9.96–13.65) | 32.0%, P = 0.04 | ||
Antrum | 35 | 6.29 (4.67–8.42) | 76.8%, P < 0.01 | ||
Depth of invasion | Early | 7 | 13.00 (9.20-18.06) | 0%, P = 0.71 | P = 0.45 |
Advanced | 7 | 10.80 (7.64–15.06) | 58.1%, P = 0.03 | ||
Tumor stage | I + II | 14 | 7.39 (5.79–9.39) | 29.5%, P = 0.14 | P = 0.36 |
III + IV | 14 | 8.80 (6.57–11.68) | 64.4%, P < 0.01 | ||
Lymph node invasion | Absent | 14 | 8.75 (6.02–12.55) | 57.9%, P < 0.01 | P = 0.91 |
Present | 14 | 9.00 (6.33–12.65) | 77.4%, P < 0.01 | ||
FFPE: Formalin-Fixed Paraffin-Embedded; † p-value less than < 0.05 |
The Association Between EBV And Gastric Cancer
Among 30 case-control studies, 20 had matched pairs design, including tumor and tumor-adjacent normal tissue pairs from 4116 gastric cancer patients. The remaining 10 non-matched case-control studies included 911 cases of gastric cancer and 436 controls. Using data obtained from studies with non-matched pairs design, the pooled OR of EBV infection was 3.31 (95% CI: 0.95–11.54; I2 = 55.0%), whereas the pooled OR for studies with matched pairs design was 18.56 (95% CI: 15.68–21.97; I2 = 55.4%), indicating a very strong significant positive relationship between EBV infection and gastric cancer (Fig. 2). So we further performed subgroup analysis for studies with matched pairs design. Table 3 presents details on the association between EBV infection and gastric cancer risk for subgroups. Finally, the analysis of the funnel plot did not show evidence of asymmetry (Fig. 3), and Begg's test indicated an absence of publication bias among all the studies included in this meta-analysis (P = 0.18).
Characteristics | Categories | No. of Studies | Pooled ORs (95% CI) | Heterogeneity test I2 %, p-value | Differences between subgroups; χ2 test (p-value) |
---|---|---|---|---|---|
Overall | - | 20 | 18.56 (15.68–21.97) | 55.4%, P < 0.01 | - |
Sex | Male | 8 | 14.07 (10.46–18.93) | 49.0%, P = 0.06 | P = 0.06 |
Female | 8 | 21.47 (15.55–29.63) | 0%, P = 0.55 | ||
Study location | America | 5 | 15.69 (10.82–22.74) | 57.4%, P = 0.05 | P = 0.33 |
Asia | 9 | 21.00 (16.77–26.30) | 59.1%, P = 0.01 | ||
Europe | 6 | 17.23 (13.19–22.51) | 5.6%, P = 0.38 | ||
Development status | Developed countries | 10 | 17.31 (13.38–22.40) | 58.7%, P < 0.01 | P = 0.46 |
Developing countries | 10 | 19.73 (15.56–25.03) | 56.2%, P = 0.01 | ||
Lauren's histological type | Intestinal type | 10 | 15.07 (9.55–23.78) | 62.0%, P < 0.01 | P = 0.27 |
Diffuse type | 10 | 10.69 (7.14-16.00) | 79.0%, P < 0.01 | ||
Tumor anatomical location | Cardia | 10 | 6.65 (5.18–8.52) | 21.8%, P = 0.24 | P = 0.46 |
Body | 10 | 6.31 (2.38–16.69) | 97.0%, P < 0.01 | ||
Antrum | 11 | 15.55 (4.12–58.62) | 98.2%, P < 0.01 | ||
Depth of invasion | Early | 3 | 5.87 (2.78–12.40) | 45.8%, P = 0.16 | P < 0.01† |
Advanced | 3 | 19.94 (13.31–29.85) | 22.9%, P = 0.27 | ||
Tumor stage | I + II | 2 | 33.50 (10.85-103.46) | 73.8%, P = 0.05 | P = 0.52 |
III + IV | 2 | 22.26 (13.05–37.96) | 24.6%, P = 0.25 | ||
Lymph node invasion | Absent | 3 | 16.98 (9.02–31.95) | 1.3%, P = 0.36 | P = 0.58 |
Present | 3 | 23.21 (9.44–57.03) | 80.6%, P < 0.01 | ||
† p-value less than < 0.05 |
Our meta-analysis showed that the pooled prevalence of EBV among gastric cancer patients from 27 countries is 8.78% (95% CI: 7.75–9.93%; I2 = 83.0%). We chose strict inclusion and exclusion criteria to obtain pertinent studies and to increase the chance of finding a valid conclusion. The pooled prevalence and OR obtained in this meta-analysis were calculated from studies which detected EBV infection with ISH method. All studies that investigated the presence of EBV by other methods, including different types of PCR assays, and even immunohistochemistry (IHC), did not consider in our analysis. The reason for this stems from the fact that the sensitivity and specificity of each detection method are different, and it is not reliable to draw a conclusion using the pooled data.
The gold standard technique for the detection of EBV in tissues is ISH with EBV EBERs (EBER-ISH) due to its high sensitivity and specificity to determine the precise intranuclear localization of the EBV-infected cells. The diagnosis of EBV-associated gastric cancer is confirmed by the presence of EBER within the tumor cells and its absence in the normal tissue adjacent to the tumor [3]. Many studies have reported the higher prevalence of EBV among gastric cancer patients by PCR assay than EBER-ISH technique [83]. However, PCR is unable to discriminate between cancer cells and lymphocytes infiltrating in tumor stromal, and thus it is impossible to know from where the EBV genome is amplified. It should be noted that the great majority of people (nearly 90%) are EBV carriers, and their lymphocytes probably contain EBV genomes [84]. Regarding the aforementioned statements, our meta-analysis exclusively focused on the positivity of the EBV-associated gastric cancers by ISH only.
One of the major strong points in this meta-analysis is that the pooled estimate of ORs was calculated from studies with matched pairs and non-matched pairs designs, separately, with different statistical methods. The detailed descriptions about the analysis of data for matched pairs and non-matched pairs studies are available in several previous studies [85]. It has been recommended that a matched-pairs analysis should be used to assess effect sizes for studies with matched pairs design. Accordingly, the pooled OR determined for studies with non-matched pairs and matched pairs designs were 3.31 (95% CI: 0.95–11.54; I2 = 55.0%) and 18.56 (95% CI: 15.68–21.97; I2 = 55.4%), respectively.
To date, several studies have attempted to discover the role of EBV infection in gastric cancer progression. EBV enters B lymphocytes in oropharyngeal lymphoid tissues. The virus then enters the gastric epithelial cells, either by the cell-to-cell contact between B lymphocytes and gastric epithelial cells or by direct entry into the gastric epithelia [86]. It has been reported that EBV entry into the gastric epithelial cells is facilitated by the previous mucosal damage [53]. After the virus enters the cell, EBV establishes type I latency in which a limited set of latent gene is expressed [64]. A recent systematic review study showed that the most of the EBV latent proteins expressed in gastric cancer cases were EBNA1 (98.1%) and LMP2A (53.8%), whereas LMP1 and LMP2B were detected in only 10% of EBV-associated gastric cancer cases. Some of lytic proteins such as BARF1 were also reported to be present in almost half of EBV-associated gastric cancer cases [87]. It is shown that the EBV-encoded BARF1 acts as oncogene and promotes cell proliferation in gastric cancer through upregulation of NF-κB signaling and reduction of the cell cycle inhibitor p21 [88]. It is well known that DNA methylation plays a crucial role in gastric cancer development and progression [89]. Methylation of both viral and cellular genome is one of the important mechanisms involved in the development and maintenance of EBV-associated gastric cancer. It is well documented that EBV latent membrane protein 2A (LMP2A) plays a variety of key roles in the epigenetic abnormalities such as aberrant DNA methylation in host stomach cells, and in the development and maintenance of EBV-associated gastric cancer [90].
Another interesting finding of our meta-analysis is that the prevalence of EBV was 1.9-fold higher in male patients than in female patients with gastric cancer (P < 0.01). However, the OR estimate for EBV-associated gastric cancer was significantly higher among females than in males (P = 0.06). According to these results, we concluded that women are more likely than men (1.5-fold) to develop EBV-associated gastric cancer. This novel finding can be explained by different genetic background, lifestyle, or hormonal conditions between the two genders.
Subgroup analyses based on the tumor anatomical location indicate an anatomic preference for EBV during gastric carcinogenesis. Indeed, EBV-associated gastric cancers were significantly more prevalent in the cardia and in the body of the stomach than in the antrum (P < 0.01) (Table 2). However, the situation was different when OR was calculated. So that the OR estimate for EBV-associated gastric cancer was remarkably higher in the antrum than in the cardia and in the body (Table 3), although the difference was not statistically significant. This feature can be justified with the fact that the various parts of the stomach have different physiological conditions.
One prominent finding of the present meta-analysis is that EBV was detected more frequently in biopsy samples than in FFPE specimens from gastric cancer patients (2.4-fold, P < 0.01). It is well documented that there are several challenges when working with FFPE samples such as the low amount of extracted nucleic acids, and fragmentation of genomes and transcripts during the processes of fixation and embedding in paraffin. Therefore, in order to prevent false-negative results, using biopsy samples is recommended.
According to Lauren's histological classification, gastric carcinoma is classified into two distinct types, namely intestinal and diffuse types. There are many differences between intestinal and diffuse types based on their epidemiology, etiology, and pathology [67]. However, the current meta-analysis showed that the prevalence of EBV was similar in intestinal and diffuse types (8.15% and 9.40%, respectively), and no significant association of EBV infection with the histological type was found (P = 0.27).
Similarly, our results did not indicate any significant difference in the prevalence of EBV-associated gastric cancer among different geographic regions, even between developed and developing countries. The same prevalence in developed and developing countries demonstrates that economic conditions are not related to EBV-associated gastric cancer risk.
To sum up, our meta-analysis suggests that the pooled prevalence of EBV among patients with gastric cancer was 8.78%. To determine the association between EBV infection and gastric cancer, a matched-pairs analysis from case-control studies was performed, and the pooled OR was calculated 18.56. This finding indicates a robust positive association between EBV infection and gastric cancer risk. We recommend using biopsy instead of FFPE samples and ISH technique instead of PCR methods to ensure validity of results. Furthermore, the pooled prevalence of EBV was obtained from data from 27 countries in the world. Therefore, conducting studies in other geographical regions is strongly recommended to obtain more reliable estimates.
EBV: Epstein–Barr virus; CI: confidence interval; OR: odds ratio; EBER: EBV-encoded small RNA; ISH: in situ hybridization; PCR: polymerase chain reaction; FFPE: formalin-fixed paraffin-embedded; ELISA: enzyme-linked immunosorbent assay; PBMC: peripheral blood mononuclear cell; IHC: Immunohistochemistry; LMP: Latent membrane protein; BARF-1: BamH1-A Reading Frame-1; NF-κB: nuclear factor kappa B; EBNA-1: Epstein–Barr nuclear antigen 1.
Ethics approval and consent to participate
Not applicable
Consent for publication
Not applicable
Availability of data and materials
All data generated or analyzed during this study are included in this article.
Competing interests
The authors have no conflict of interest.
Funding
This study was not financially supported by any individual, agency, or institution.
Authors ‘contributions
A.T and M.F designed the study. M.F performed statistical analysis. A.T wrote, reviewed and edited the manuscript. SH.M and SJ.K performed data interpretation. A.T, M.F, S.A, and F.S.M performed search strategy. All authors involved in acquisition of data, read and approved the final draft.
Acknowledgements
Not applicable.