Worldwide Prevalence of Epstein–Barr Virus in Patients with Burkitt Lymphoma: A Systematic Review and Meta-Analysis

Burkitt lymphoma (BL) is a form of B-cell malignancy that progresses aggressively and is most often seen in children. While Epstein–Barr virus (EBV) is a double-stranded DNA virus that has been linked to a variety of cancers, it can transform B lymphocytes into immortalized cells, as shown in BL. Therefore, the estimated prevalence of EBV in a population may assist in the prediction of whether this population has a high risk of increased BL cases. This systematic review and meta-analysis aimed to estimate the prevalence of Epstein–Barr virus in patients with Burkitt lymphoma. Using the appropriate keywords, four electronic databases were searched. The quality of the included studies was assessed using the Joanna Briggs Institute’s critical appraisal tool. The results were reported as percentages with a 95% confidence interval using a random-effects model (CI). PROSPERO was used to register the protocol (CRD42022372293), and 135 studies were included. The prevalence of Epstein–Barr virus in patients with Burkitt lymphoma was 57.5% (95% CI: 51.5 to 63.4, n = 4837). The sensitivity analyses demonstrated consistent results, and 65.2% of studies were of high quality. Egger’s test revealed that there was a significant publication bias. EBV was found in a significantly high proportion of BL patients (more than 50% of BL patients). This study recommends EBV testing as an alternative for predictions and the assessment of the clinical disease status of BL.


Introduction
Epstein-Barr virus (EBV) is a pathogenic double-stranded DNA human herpes virus 4 (HHV4). It was first discovered as a human-associated virus by Michael Anthony Epstein and Yvonne Barr in 1964 [1]. The virus consists of a 170-180 kb liner of double-stranded (ds) enveloped DNA with a toroid-shaped protein core, a nucleocapsid with 162 capsomers, and external virus-encoded glycoprotein spikes on the surface of the viral tegument [2]. The EBV genome encodes more than 85 genes, which are involved in the pathogenesis of infection and initiating EBV-associated human disease. There are two major types of EBV: type 1 EBV, which is found worldwide, and type 2 EBV, which is mainly detected in Africa [3]. EBV is the most frequent cause of infectious mononucleosis, with primary infections commonly occurring asymptomatically in teenagers and young adults, especially college students, while in adults, the symptoms are more severe. After primary infection, EBV establishes latent and lytic programs [4,5]. During the latent form of infection, the post-transplant lymphoproliferative disorder. "+" indicates the protein is expressed, while "−" indicates that the protein is not expressed.
Studies showed that BL expresses high levels of MYC, and more than 90% show the translocation of the MYC oncogene (8q24) onto the immunoglobulin heavy chain (IgH) (14q34). The chromosomal breakpoints of both MYC and IgH vary between sBL and eBL, giving rise to different aetiologic drivers [35]. A translocation of the MYC gene on chromosome 8, including genetic material from chromosomes 2, 14, or 22, is the classic etiology of BL. The majority of translocations (around 80%) involve the Ig heavy chain on chromosome 14, t(8;14), whereas 15% involve the kappa light chain on chromosome 2, t(2;8), and 5% involve the lambda light chain on chromosome 22 [36,37].
EBV-associated malignancies are diagnosed primarily by a biopsy of the primary tumor, with an EBER in situ hybridization test to confirm the presence of EBV [38]. However, due to the difficulty in obtaining a sample of the tumor or poor patient condition, performikng a biopsy might be challenging [39].
Many studies of EBV-associated lymphoma reveal that EBV-DNA may be found in the plasma of most patients with EBV-related malignancies [40]. DNA from EBVassociated lymphoma is derived as naked DNA fragments from apoptotic or necrotic tumor cells [35,38], whereas it is undetectable in non-EBV-associated tumors or healthy people [24].
Although plasma EBV DNA has recently become more important in the diagnosis and management of EBV-associated cancers [41], particularly Hodgkin's lymphoma (HL) [41,42] and nasopharyngeal carcinoma [43,44], there are limited data on the diagnostic and prognostic significance of plasma EBV DNA for BL. In order to identify EBV in various types of samples, methods such as the heterophile antibody test, Immunofluorescence assays, enzyme immunoassays, Western blot, and polymerase chain reaction (PCR) are used. The use of PCR to determine the EBV viral load is becoming more popular in the diagnosis of EBV-related diseases [45].
Artificial intelligence (AI) is now advancing quickly, and its application in medicine is becoming more relevant. To predict or classify based on input data, AI integrates computer science and databases. Machine learning and deep learning are two types of AI used in the medical field to evaluate medical data and acquire an understanding of the pathogenesis of diseases. Recently, an AI application used for EBV has been developed, such as a deeplearning-based EBV prediction method from H&E-stained whole-slide images (WSI) in gastric cancer [46], and deep-learning-based classifiers to detect microsatellite instability and EBV status directly from hematoxylin-and-eosin-stained histological slides [47]. In BL, artificial neural networks and various types of machine learning were used to analyze the gene expression and protein levels by immunohistochemistry of several hematological neoplasia and pan-cancer series in order to predict patients' survival and the disease subtype classification with a high accuracy [48]. There is no systematic review and metaanalysis of the prevalence of EBV in patients with BL that we are aware of. As a result, the goal of this systematic review and meta-analysis was to determine the prevalence of EBV in patients with BL, which helps in predicting whether populations are at high risk of increasing the number of BL cases corresponding to EBV infection.

Reporting Guidelines and Protocol Registration
This systematic review and meta-analysis were carried out according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [49] and Metaanalysis of Observational Studies in Epidemiology (MOOSE) [50] guidelines. This study protocol (PROSPERO: CRD42022372293) was submitted to the International Prospective Registry of Systematic Reviews database at the University of York, York, UK.

Eligibility Criteria
The study looked for published studies on the prevalence of Epstein-Barr virus among Burkitt lymphoma patients. The screening was carried out to find possible studies that looked at the presence of EBV in Burkitt lymphoma patients without any restrictions.

Literature Search
In total, 3981 studies were retrieved from four electronic databases: PubMed, Scopus, Web of Science, and Google Scholar. The most recent search was in January 2021, for studies on the prevalence of Epstein-Barr viruses among Burkitt lymphoma patients. Burkitt, Burkitt's, African Lymphoma, Epstein-Barr, EBV, Human Herpesvirus 4, HHV4, HHV-4, and EB virus were used in the search utilizing a combination of Boolean logical operators ('AND' & 'OR') and the 'Advanced' and 'Expert' search options (Table S1). To ensure a thorough method, the references of the included papers were also examined. To organize and filter out duplicate studies, EndNote X9 software was used.

Study Selection
Two authors (M.J.A.-K. and N.H.I.) independently screened the research title and abstract, followed by the entire text, of all studies retrieved from the literature search to determine the matched studies to be included. Excluded studies include review articles, case studies, non-human studies, views, and viewpoints. Data from news accounts and press releases and information acquired from blogs and databases were not considered.

Data Extraction
The data from the included studies were accessed independently by two authors (M.J.A.-K. and N.H.I.). Before the data extraction procedure, all non-English language studies were translated into English using Google Translate. The data extracted from each of the eligible studies was imported into a predetermined Excel spreadsheet. The following are the extracted data from the selected studies: author name, study type, country, number of BL patients, participants' age, number of EBV positives in BL, sample type, and EBV detection method. Any discrepancies, or confusing or unfounded data were discussed among the authors in order to reach an agreement. If the problem remains, the corresponding or first author of each study was emailed for clarification.

Quality Assessment and Publication Bias
The quality of the included studies was assessed using Joanna Briggs Institute's critical appraisal tools. The studies were defined as poor-quality (high risk of bias), moderatequality (moderate risk of bias), or high-quality (low risk of bias) if the overall score was ≤49%, 50-69%, or ≥70%, respectively [51,52]. Egger's test was used to verify the funnel plot's asymmetry. To evaluate publication bias, a funnel plot was constructed to compare the prevalence estimate against the standard error.

Data Analyses
To address the inconsistency among the included studies, a tau-squared test was used to assess heterogeneity (I 2 ), where p < 0.05 was regarded as statistically significant. A greater homogeneity was regarded as an I 2 value close to zero, where I 2 values between 25-50% indicated low heterogeneity, 51-75% indicated moderate heterogeneity, and >75% indicated significant heterogeneity. Based on the critical assessment tools, two authors M.J.A.-K. and N.H.I.) evaluated the quality of each of the included studies by using the critical assessment tools.
Sensitivity analyses and Galbraith plots were also used to assess the quality of the results and identify potential causes of heterogeneity, respectively. The following strategies were used to conduct sensitivity analyses: excluding small studies (n < 100); excluding low-quality studies (high risk of bias); excluding studies that did not disclose the prevalence of EBV in patients with BL; only considering cross-sectional studies; and excluding outlier studies. All analyses and plots were generated by using RevMan software (version 5.3.5), RStudio (version 1.1.463), and the metafor package (version 2.0-0) of R software (version 3.5.1) [53].

Subgroup and Sensitivity Analyses
For subgroup analysis, the prevalence of EBV in patients with BL was analyzed through four-time interval trends (1969-1982, 1983-1995, 1996-2008, and 2009-2021); methods of EBV detection (nucleic acid hybridization, polymerase chain reaction (PCR), Immunofluorescence, in situ hybridization (ISH), ISH+PCR, and southern blot); and geographical locations (Sub-Saharan Africa, Northern Africa, Southern America, Southern Asia, Northern America, Europe, Eastern Asia, and South-eastern Asia). The studies were categorized based on the sociodemographic index (SDI). To measure social and economic development, the SDI, which ranges from zero to one, employs data on the world's economies, educational systems, and fertility rates. The SDI is divided into five categories: high SDI (lower bound to upper bound: 0.805129 to 1), high-middle SDI (lower bound to upper bound: 0.689504 to 0.805129), middle SDI (lower bound to upper bound: 0.607679 to 0.689504), low-middle SDI (lower bound to upper bound: 0.454743 to 0.607679), and low SDI (lower bound to upper bound: 0 to 0.454743) [54]. To identify the source of heterogeneity and check the robustness of the results, sensitivity analyses were performed using the following strategies: (1) excluding small studies (<100); (2) excluding low-quality studies (high risk of bias); (3) considering only cross-sectional studies; (4) considering only case-control studies; (5) considering only cohort studies; (6) considering only studies where the age was less than 18 years old; and (7) excluding the outlier studies.

Selection and Inclusion of Studies
From the database search, 3981 studies qualified for initial screening, and then 2130 studies were excluded due to being duplicate studies (n = 1778), review articles (n = 259), case reports (n = 86), and non-human studies (n = 7). Therefore, 1851 studies were further assessed for eligibility by a detailed screening of the titles, abstracts, and full text. Finally, after excluding 1716 studies because they did not comply with the objective of this study, 135 studies were eligible to be included in this systematic review and meta-analysis, as illustrated in the PRISMA flow diagram (Figure 1).

Study Characteristics
Our literature search yielded 135 studies  published between 1969 and 2021, which examined the prevalence of EBV in patients with BL. Detailed characteristics and references of the included studies are presented in Table 2. Overall, this meta-analysis reports data from 4837 patients with BL lymphoma (34.7% female). The ages of these patients ranged from 2.1 ± 2.5 to 47.7 ± 31.8 years (mean ± SD; range, 0.7-98.0). The studies came from eight different regions, and these region groupings were based on the geographic regions defined under the Standard Country or Area Codes for Statistical Use (known as M49) of the United Nations Statistics Division [55]: region unspecified (n = 414),

Quality Assessment
In Tables S2-S4, the quality assessment of the included studies was presented in detail. Generally, of the included studies, 65.2%, 29.6%, and 5.2% were high-, moderate-, and low-quality studies, respectively. The funnel plot and Egger's test results revealed evidence of a publication bias for the prevalence of EBV in BL (p = 0.0034) (Figure 4).

Quality Assessment
In Tables S2-S4, the quality assessment of the included studies was presented in detail. Generally, of the included studies, 65.2%, 29.6%, and 5.2% were high-, moderate-, and low-quality studies, respectively. The funnel plot and Egger's test results revealed evidence of a publication bias for the prevalence of EBV in BL (p = 0.0034) (Figure 4).

Discussion
EBV was found to be associated with human cancer when it was discovered in BL. This was a result of BL cell isolation. EBV has been extensively characterized due to purported links to a variety of human diseases, including BL, HL, post-transplant and AIDS-related lymphomas, and nasopharyngeal carcinoma [7,191,192]. Our results revealed a high prevalence of EBV in patients with BL, at 59.4% in all BL patients worldwide. However, as shown in our study, the prevalence of EBV in patients with BL varies by region; we found the highest prevalence in Sub-Saharan Africa (76.5%) and Northern Africa (69.3%), while the prevalence in Southern America (58.4%), Southern Asia (54.7%), Northern America (54.3%), Europe (49.7%), Eastern Asia (29.5%), and Southeastern Asia (29.1%) were lower than the pooled prevalence. We can explain the variations in EBV prevalence among patients with BL worldwide, as more than 95% of people in the world acquire the Epstein-Barr virus, a herpes virus belonging to the gamma subfamily, within the first ten years of life. Primary exposure to infections occurs in childhood in Africa and other developing countries, probably as a result of different cultural norms compared to developed countries [115,193].
The Epstein-Barr virus infection persists asymptomatically for the entirety of the host's life, maintaining the immune system and this deceptive virus constantly in balance. In our study, the incidence of BL was higher in children (≤18) at 64.9% compared to adults; this corresponds to many studies that report that BL is more common in children

Discussion
EBV was found to be associated with human cancer when it was discovered in BL. This was a result of BL cell isolation. EBV has been extensively characterized due to purported links to a variety of human diseases, including BL, HL, post-transplant and AIDS-related lymphomas, and nasopharyngeal carcinoma [7,191,192]. Our results revealed a high prevalence of EBV in patients with BL, at 59.4% in all BL patients worldwide. However, as shown in our study, the prevalence of EBV in patients with BL varies by region; we found the highest prevalence in Sub-Saharan Africa (76.5%) and Northern Africa (69.3%), while the prevalence in Southern America (58.4%), Southern Asia (54.7%), Northern America (54.3%), Europe (49.7%), Eastern Asia (29.5%), and South-eastern Asia (29.1%) were lower than the pooled prevalence. We can explain the variations in EBV prevalence among patients with BL worldwide, as more than 95% of people in the world acquire the Epstein-Barr virus, a herpes virus belonging to the gamma subfamily, within the first ten years of life. Primary exposure to infections occurs in childhood in Africa and other developing countries, probably as a result of different cultural norms compared to developed countries [115,193]. The Epstein-Barr virus infection persists asymptomatically for the entirety of the host's life, maintaining the immune system and this deceptive virus constantly in balance. In our study, the incidence of BL was higher in children (≤18) at 64.9% compared to adults; this corresponds to many studies that report that BL is more common in children [194,195]. Our results revealed that the incidence of BL among males is much higher than in females (34.7%), which is commensurate with several studies that report that BL is more prevalent in males compared to females [104,120,123,195]. This result is in agreement with Yakimchuk et al., which reported that estrogen has an anti-proliferative effect on BL cells through estrogen receptor β (ERβ) signalling [196]. Our study revealed a significant publication bias for EBV prevalence in patients with BL, and that is in agreement with some studies exploring the prevalence of EBV in different diseases, such as multiple sclerosis (p < 0.05) [197] and breast cancer (p = 0.006) [198], while that is in disagreement with some studies such as for gastric carcinoma (p = 0.912) [199], Hodgkin's lymphoma (p = 0.162) [200], and EBV-associated epithelial tumors (p = 0.23617) [201].
Interestingly, our study shows a significant decline in EBV prevalence over four time periods (13 years), with the prevalence decreasing from 64.2% in the period from 1969 to 1982, to 54% in the period from 2009 to 2021. This decrease in incidence could be attributed to the development and widespread use of EBV vaccines, as well as improved sanitation, living habits, and personal hygiene [202,203]. There are many methods used to detect EBV, but these methods are different depending on whether they are faster, are more sensitive, or provide more informative than previous assays [204]. Our study revealed that the most used method in EBV detection was the microscopic examination (in situ hybridization (ISH) in 59 studies and Immunofluorescence in 18 studies) method followed by molecular methods (PCR in 17 studies, nucleic acid hybridization in nine studies, ISH+PCR in nine studies, and Southern blot in seven studies), and, finally, immunoassay methods in seven studies. This result confirms that ISH is the methodology of choice for the detection of EBV in tissue sections [205][206][207]. Our results revealed a higher prevalence of EBV in patients with BL in both low and middle SDI countries, at 82.7% and 60.1%, respectively. A study showed that the highest incidence and mortality burden occurred in EBV-attributed BL in low and low-middle SDI areas [208]. The reasons for the increases in the burden of malignancies related to EBV infection appear to be growing populations, an increase in life expectancy, and changing age structure [209].

Conclusions
In conclusion, based on the comprehensive systematic and meta-analysis of the available data on the prevalence of EBV in patients with BL until January 2021, the prevalence was 59.4% in all patients with BL. Due to factors such as cultural habits, personality hygiene, limited use of developed EBV vaccines, and malaria endemic areas, Sub-Saharan Africa (76.5%) and Northern Africa (69.3%) revealed the highest prevalence (hot spots) in comparison to the rest of the world. Countries with middle and low SDI have a higher prevalence of EBV in patients with BL. Despite the fact that the EBV prevalence in patients with BL has decreased significantly from 64.2% in 1969 to 1982 to 54% from 2009 to 2021, as well as there being a higher incidence in younger (≤18) patients than adults, EBV detection should be used as a routine test in hot spots as well as in all young people because it will help in predicting whether populations are at a high risk of increasing the number of BL cases corresponding to EBV infection.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/diagnostics13122068/s1, Figure S1: Subgroup analyses estimating the prevalence of EBV at (A-D) different times, using (E-K) different EBV detection methods, in (L-S) different regions, and based on the (T-X) socio-demographic index (SDI); Figure S2: Sensitivity analyses (A) excluding small studies, (B) excluding low-and moderate-quality studies, (C) considering only cross-sectional studies, (D) considering only case-control studies, (E) considering only cohort, (F) considering only studies where the age was less than 18 years old, and (G) excluding outlier studies estimating the prevalence of EBV in patients with BL; Table S1: Search strategies; Table S2: Quality assessment of the included cross-sectional studies; Table S3: Quality assessment of the included case-control studies; Table S4: Quality assessment of the included cohort studies.

Conflicts of Interest:
The authors declare no conflict of interest.