Prevalence of human Papillomavirus associated oropharyngeal and oral squamous cell carcinoma in Asian countries: A systematic review and large-scale meta-analysis

Human Papillomavirus (HPV) associated with oropharyngeal and oral squamous cell carcinoma (OPSCC and OSCC) is escalating over the years. Hence, the present review aims to determine the prevalence of HPV-OSCC and HPV-OPSCC in Asian countries over the last decades. An electronic search was conducted using nine online databases to identify English-language articles on the prevalence of HPV-OPSCC and HPV-OSCC in Asian countries from January 2011 to June 2022. The risk of bias was assessed using the JBI critical appraisal checklist and the level of evidence was determined based on the OCEBM guideline. Single-arm meta-analysis was used to estimate the weighted mean prevalence of HPV-OPSCC and HPV-OSCC among patients in Asia. Subgroup analysis meta-regression and Egger’s tests were also conducted. 59 eligible studies were included with a higher prevalence of HPV-OPSCC (32.6%-37.4%) as compared to HPV-OSCC (10.9%-23.5%). Subgroup analysis revealed that the weighted mean prevalence of HPV-OPSCC was significantly higher (P<0.001) among East Asians, while the weighted mean prevalence of HPV-OSCC was significantly higher (P<0.001) among South Asians. All studies showed a low to moderate risk of bias with the level of evidence ranked between 2 and 3. The diagnostic tools utilised and geographical locations significantly affect the findings.


Introduction
Squamous cell carcinoma of the head and neck is the sixth most prevalent cancer in the world [1], with oropharyngeal squamous cell carcinoma (OPSCC) and oral squamous cell carcinoma (OSCC) are both malignant neoplasms that occurred commonly in the head and neck region.In contrast to other head and neck cancers, both the incidence and prevalence of OPSCC and OSCC rose considerably over the years [2,3].Cancer of the oropharynx and oral cavity has long been linked to tobacco chewing or smoking, as well as the consumption of alcoholic beverages [3].They also mainly afflict older age groups between the fifth and sixth decades.However, the incidence of OPSCC and OSCC in young population has increased in several countries over the last few decades [4], and one of the risk factors is attributed to the rise in human papillomavirus (HPV)-related oropharyngeal squamous cell carcinoma (HPV-OPSCC) and HPV-related oral squamous cell carcinoma (HPV-OSCC), particularly high-risk HPV types such as HPV-16 and HPV-18 [5,6].HPV is one of the most prevalent infectious agents transmitted sexually across the world, and the main risk factors are sexually acquired behaviours [7].HPV infection can cause clinical illnesses, such as anogenital warts, cervical neoplasia, cervical cancer, and other anogenital malignancies, even though most infections are asymptomatic and resolve within two years [8].
It is worth noting that, in terms of aetiological variables and gender, the demographic pattern of this disease among young patients differs [9].Men who are non-smokers, non-drinkers, and have a decent socioeconomic position are more likely to develop oropharyngeal cancers associated with HPV [10].The bulk of the literature has shown that the incidence and prevalence of HPV-related OP-SCC and OSCC were high and continue to escalate [11,12].A rapid increase in HPV-OPSCC and HPV-OSCC prevalence would have crucial ramifications for patients, healthcare professionals, and commissioners, since patients would have to deal with substantial treatments and demand greater assistance from healthcare systems [12].Nonetheless, current evidence on the prevalence of HPV-OPSCC and HPV-OSCC in Asia is relatively scarce in the literature as compared to the United States of America and other Western Europe countries [13,14].Previous systematic studies on the prevalence of HPV-related OPSCC and OSCC were either too geographically restricted, focusing solely on South-Central Asia [15], or too general, pooling data from all cases of head and neck cancer without fo-cusing on the oral cavity and oropharyngeal regions [16].Furthermore, the reporting quality of the primary studies included in the previous review was not thoroughly evaluated, which might result in low-quality primary studies being included, and contribute to errors and bias [16].
To the best of the authors' knowledge, there is still a paucity of a well-conducted systematic review and meta-analysis on the prevalence of HPV-related OPSCC and OSCC in Asian countries.Although HPV prevalence rates in OSCC have been reported to be lower than in OP-SCC [17,18], it is unknown if this is only a lag period, or whether HPV prevalence in OSCC is now emerging, imitating the surge of HPV in OPSCC.Therefore, a comprehensive systematic review on this issue was justified since new research has been published and no systematic review evaluating the prevalence of HPV associated OPSCC and OSCC among Asian populations has been reported recently.The objectives of the present study were to determine the prevalence of HPV-related OSCC and OPSCC in Asian countries over the last decades, as well as critically assess the quality of currently available evidence.

Protocol and registration
The current systematic review was registered at the International Prospective Register of Systematic Reviews (PROS-PERO) database, University of York with a registration number (CRD42021275003).The study was performed based on Preferred Reporting System for Systematic Reviews (PRISMA) guidelines, which specify a systematic selection of articles to be included [19].The PIO framework was, Patient or population (P) -patients diagnosed with OSCC or OPSCC, Intervention or indicator (I) -HPV testing, Outcome (O) -Prevalence of HPV-OSCC and HPV-OPSCC in Asian countries.Hence, the PIO question was 'What is the prevalence of patients diagnosed with HPV-OSCC and HPV-OPSCC in Asian countries?In this framework, Asian countries are divided into six regions: North Asia, Central Asia, East Asia, South Asia, Southeast Asia, and West Asia, with a total of 48 countries.

Study selection
After the removal of duplicated articles using EndNote software version x9, the articles were screened independently based on the title and the abstract by two investigators (KWSH, YJT).Three investigators (JLSW, WLLK, WNAN) subsequently conducted a full-text evaluation to select eligible studies based on the inclusion and exclusion criteria.The inclusion criteria were: (1).randomized or non-randomized controlled trials, prospective or retrospective cohorts, case-control, or cross-sectional studies.(2).presented in English language only.(3).prevalence of HPV-OSCC or HPV-OPSCC.(4).Asian countries.(5).HPV detected using polymerase chain reaction (PCR)based methods, in situ hybridization (ISH) or other HPV detection methods.On the other hand, the exclusion criteria of this study were: (1).expert opinions, short communications, reviews, case reports, case series or animal studies.(2).prevalence of HPV associated with tumours other than OSCC or OPSCC.(3).sample obtained using serology or mucosal brushings.
Calibrations between investigators were performed to assess interrater reliability.To compare the investigators' decisions on inclusion and exclusion, the average concordance was calculated using the Kappa value [20].With the assistance of the sixth investigator (GSSL), any conflicts that developed throughout the search were addressed and resolved.

Data extraction
Four investigators (KWSH, YJT, JLSW, WNAN) used a modified excel spreadsheet extraction form to extract and document the parameters of each article.The following information was extracted: author, country, year, type of studies, total samples, age of patients, clinical signs and symptoms, radiographical signs and symptoms, TNM stage, histological results, specific site, gender predilection, type of HPV, prevalence, and prognosis/ survival.If any discrepancies were identified, the fifth investigator (GSSL) double-checked the accuracy of the filled data, and a further discussion with all investigators was convened.

Risk of bias assessment
The risk of bias for each included study was independently evaluated by five investigators (KWSH, YJT, JLSW, WLLK, WNAN) using four quality assessment tools.Cross-sectional studies were assessed using the Joanna Briggs Institute (JBI) critical appraisal checklist for analytical cross-sectional studies [21].Either a 'yes', 'no', 'un-clear' or 'not applicable' was assigned for each domain and the studies were categorized as 'include', 'exclude' or 'seek further info'.The cohort and case-control studies were assessed by using the JBI checklist for cohort and casecontrol studies, respectively [22,23].On the other hand, randomised clinical study was assessed by using the JBI checklist for randomized controlled trials [24].The Oxford Centre for Evidence-Based Medicine (OCEBM) guideline was used to establish the level of evidence in each study [25].Any discrepancies in study selection, data extraction, or quality assessment were resolved via discussion among all investigators until a consensus was achieved.

Statistical analysis
A single-arm meta-analysis based on the DerSimonian-Laird random-effects model was used to estimate the weighted mean prevalence of HPV-OPSCC and HPV-OSCC among patients in Asian countries from each study.With a significance level of 0.05 and 95 % confidence intervals, the analysis was carried out using the OpenMeta [Analyst] software (CEBM, Oxford, UK) (CI).The upper limit was defined as 1.0 if the predicted upper limit of the 95 % confidence interval was greater than 1.0.The Higgins' I 2 statistic was also used to evaluate the degree of data heterogeneity across all included studies, with I 2 : <30% = acceptable heterogeneity, I 2 : 30-60% = moderate heterogeneity, I 2 : >60% = substantial heterogeneity [26].Subgroup analysis and meta-regression were conducted to assess the effect of Asian regions and sample size on the overall prevalence rates of HPV-OPSCC and HPV-OSCC.Furthermore, Egger's test was used to identify publication bias.

Study selection
The initial literature search yielded 2194 items with a search period covering January 2011 to June 2022 (Figure 1).1489 articles were discarded after duplication was eliminated, followed by 561 articles that were dismissed based on titles and abstracts.The remaining articles were selected for an in-depth full-text assessment according to the inclusion and exclusion criteria.Finally, 59 articles were included in the present review.During the study selection process, the average inter-investigators Kappa score for preliminary article screening (titles and abstracts) and the second screening (full-text assessment) were 0.83 and 0.79, respectively, indicating 'perfect' agreement [27].The reasons for article exclusion are depicted in Figure 1.The characteristics of the included studies are summarized in Table 1.In general, a total of 64773 samples were included in the current review with the majority of the primary studies originating from India, followed by China and Japan.Most of the primary articles were published in the years 2014 and 2016 with 42 cross-sectional studies, 13 case-control studies, 3 cohort studies, and 1 randomized controlled trial.The patients included in the current re-view ranged from 18 to 93 years old, and the majority are males.Most of the primary data obtained identified HPV-16 and HPV-18 subtypes.

Risk of bias assessment
All cross-sectional studies were given a 'Yes' for domains 1, 4, 7, and 8, while four studies in domain 2, one study in domain 3, seven studies in domain 5, and fifteen studies in domain 6 were given a 'No'.Besides, two studies were given 'Unsure'.On the other hand, all case-control studies were rated 'Yes' for domains 1,2, 3 and 8, with three studies in domain 6, four studies in domain 7, and one study each for domains 4, 5, and 10 were rated 'No'.Two studies each in domains 4 and 5, and one study each in domains 7 and 9 were rated 'Unclear'.All cohort studies were rated 'Yes' in each domain except for two studies were given 'unclear' in domain 10.In addition, the only randomised controlled study in the present review was rated 'Yes' for all domains except for domains 5 and 6.In general, all included studies demonstrated a low to moderate risk of bias.Most included studies were ranked as Level 3 with three studies ranked Level 2 based on the evidence of OCEBM (Table 2).

Statistical analysis
Table 3 shows the prevalence of HPV-OPSCC and HPV-OSCC using different diagnostic tools.The prevalence rates of both HPV-OPSCC and HPV-OSCC were calculated only if there are three or more studies reporting the number of OPSCC and OSCC patients or tissue samples diagnosed with HPV.For OPSCC, 14 studies used p16 Immunohistochemistry (IHC) to determine the prevalence of HPV-OPSCC, 8 studies used in-situ hybridization (ISH) test, and 18 studies used polymerase chain reaction (PCR) test.On the other hand, 11 studies used p16 IHC to determine the prevalence of HPV-OSCC, while 29 studies used PCR test to determine the prevalence.4 studies were excluded from the analysis of the prevalence of HPV-OSCC using PCR test due to zero value.Moreover, the analysis of HPV-OPSCC and HPV-OSCC identified using EasyChip HPV blot, as well as the analysis of HPV-OSCC detected with ISH test were not performed owing to a paucity of data.
Sensitivity analyses were performed for the prevalence of HPV-OPSCC and HPV-OSCC.The highest and lowest weighted mean prevalence of HPV-OPSCC diagnosed using p16 IHC were 33.neity for HPV-OPSCC diagnosed with IHC, ISH and PCR were 90.43%, 97.67% and 98.06%, respectively.On the other hand, the weighted mean prevalence of HPV-OSCC diagnosed using p16 IHC was significantly higher (P<0.001)among South Asians [12.1%, CI: (6.Meta-regression was performed to evaluate the effect of the sample size of each study on the prevalence of HPV-OPSCC and HPV-OSCC (Appendix).No significant differences were found for both HPV-OPSCC [P-values: IHC (0.840), ISH (0.490), and PCR (0.960)] and HPV-OSCC [P-values: IHC (0.214) and PCR (0.794)], signifying that the sample size of each study does not have any direct effect on the degree of data heterogeneity.Egger's test revealed no evidence of significant publication bias in the prevalence of HPV-OPSCC [P-values: IHC (0.121), ISH (0.083), and PCR (0.072)] and HPV-OSCC [P-values: IHC (0.087) and PCR (0.091)], respectively.

Discussion
The current study comprehensively summarised, reviewed, and presented concrete evidence on the prevalence of HPVrelated OPSCC and OSCC in Asian countries which included 59 primary studies using various diagnostic techniques, particularly p16 IHC, HPV ISH, and HPV PCR.Comparisons of prevalence patterns across different geographical regions can reveal a great deal about the global burden of HPV-related oropharyngeal and oral cavity cancers.Based on the present findings, HPV-OPSCC was shown to have a higher prevalence than HPV-OSCC which corroborates with other previous similar studies [28,29].Despite the fact that the previous studies reported a worldwide HPV-OPSCC prevalence rate of approximately 45.8% to 52.9% [29,30], with the Asian region was found to exhibit a high prevalence of 51.1 % [29], these findings contradicted the results of the current study.Conversely, the prevalence of HPV-OSCC was observed to range from 10.9% to 23.5% in the present study, which is consistent with previous systematic reviews indicating prevalence rates of 23.5% [28], and 24.2% [29], respectively, but contradicts the finding of another meta-analysis which revealed a higher prevalence of 34.5% [31].Nonetheless, one explanation for the vast differences in prevalence rates across various studies might be attributed to reporting standards as some review studies only included primary research that employed HPV PCRbased diagnostic tools.Hence, a direct comparison of the present findings may not be conceivable.
To explore the potential rationale underlying the present findings, a deeper understanding of the epidemiology of both HPV-associated OPSCC and OSCC is needed.It has been reported that oral benign papillomatous lesions appear to be strongly related to low-risk HPV, primarily HPV-6 and 11, while High-risk HPV, such as HPV-16 and 18 subtypes, is linked to OPSCC and OSCC [32].Most HPV-associated OPSCC and OSCC differ from HPV-negative oral malignancies in terms of demographic features, clinical response, and overall survival rates [33].There is a distinct trend in the literature that HPV-negative OPSCC and OSCC patients are often older and have a history of alcohol consumption and smoking, whereas those HPV-positive patients on the other hand are often younger, male, and sexually active patients [34,35].Furthermore, a direct relationship has also been discovered between high-risk HPV 16-positive individuals and frequent sexual activity [36].Although oral sexual behaviour may explain some of the epidemiological differences in HPV-OPSCC and HPV-OSCC among various ethnic groups and genders [34], such a conjecture may not be appropriate to extrapolate to Asian countries since collecting comprehensive sexual behaviour histories can be uncomfortable for both patients and researchers, limiting data availability.As a result, more in-depth research into the association between sexual behaviour and epidemiological differences in HPV-associated OPSCC and OSCC in Asian countries is needed to elicit the relationship between these two variables [37].
The present study revealed that the prevalence of HPV-OPSCC was highest among East Asians, followed by West Asians, South Asians and Southeast Asians which is consistent with the findings of Shaikh MH et al. [38].On the other hand, South Asians had the greatest prevalence of HPV-OSCC, followed by Southeast Asians, East Asians, and lastly West Asians, contrasting a previous systematic review that indicated HPV-OSCC to be most common among Southeast Asians, followed by East Asians, and finally South Asians [38].This implies that HPV-OSCC cases have been on the rise in South Asia over the past ten years, outpacing those in Southeast, East, and West Asia.Though minimal investigation on the prevalence of HPV-OSCC in the Asian population based on different geographical regions has been conducted, the current finding must be considered in light of genetic variations and ethnic subpopulations throughout the Asian nations.Additionally, more research is required to examine the potential confounding effects of risk factors, such as cigarette and tobacco use, which is prevalent in South Asia rendering it difficult for the current study to draw evidence on the specific role of HPV infection in OSCC [39].As primary studies derived from West Asia were scarce and led to highly uneven regional contributions, the current finding mostly reflects the situation in South, Southeast, and Eastern Asia, and disregards the evident disparity in the prevalence of HPV-related OPSCC and OSCC.Hence, the au-thors speculate that prevalence differs widely across Asian countries, and that a generalisation of the Asian figure is likely to be an underestimation or overestimation.
One interesting finding is that while the reported East Asia demonstrated a greater incidence of HPV-OPSCC fits with the overall observation of a rise in developed countries globally [40], the causes for this specific increase remain unknown.Evidence suggested that the occurrence of HPV infection is likely followed by the emergence of OPSCC after a few years or even decades, and this relationship between increased OPSCC incidence rates and higher HPV prevalence has been established [41].Therefore, the plausible explanation for this observation could be that while tobacco use significantly reduced in developed East Asian countries such as Japan, South Korea, China, Taiwan, and Hong Kong [42], sexual behaviour has changed over time [6], which has coincided with the rise in HPV-positive OP-SCC.The number of oral sex partners has been found as the most predictive factor of HPV-OPSCC, which might explain the high prevalence of HPV-OPSCC in East Asia, since premarital sexual behaviour and multiple sex partners have increased in recent years [43,44].Besides, it has also been suggested that prophylactic vaccinations targeting specific HPV16 and HPV18 may have the potential to avert a significant portion of head and neck squamous cell carcinomas globally [29,45], particularly OPSCC and OSCC.Thus, with the increased adoption of HPV vaccines, a decline in HPV-OPSCC and HPV-OSCC could be predicted.Nevertheless, the disparity in prevalence between HPV-OSCC and HPV-OSCC among Asian population is still unknown.Considering how the various Asian nations differ in terms of social norms, religions, and cultural backgrounds, it is not surprising that such a difference, notably in developed countries, contributed to greater societal acceptance of more sexually active behaviours, particularly those involving oral contact.Consequently, this argument may provide an explanation for the trend of a higher prevalence of HPV-OPSCC in East Asians.
The great range of currently available HPV detection tools necessitates a review of their benefits and drawbacks in testing protocols.According to the current meta-analyses, the prevalence of HPV-OPSCC and HPV-OSCC in Asian regions varies depending on the diagnostic tool used, with those detected with the HPV PCR test exhibiting a greater prevalence which is in-line with a previous metaanalysis [31].This can be explained by the diagnostic tool's reliability, since PCR has been shown to have higher sensitivity than ISH in identifying HPV DNA [31].Moreover, it has been reported that HPV PCR has a sensitivity of 97-98% and a specificity of 84-87%, whilst HPV ISH has a sensitivity of 85-88% and a specificity of 83-88% [46,47].Prior studies have also underlined the significance of p16 IHC expression as a surrogate marker for HPV infection in head and neck squamous cell carcinomas (HNSCC), which also predicts the survival outcome of patients with HNSCC [48,49].It is worth noting that PCR cannot tell if the HPV came from tumour cells or non-tumour tissues, and both PCR and ISH-based methods can only show the presence of HPV without confirming the viral activity, but p16 IHC can identify cancer cells with transcriptionally active HPV [50].
Although testing for viral E6/E7 mRNA expression in OPSCC and OSCC is a valid alternative to HPV association testing, it is less widely utilised and lacks standardisation [51].Nevertheless, the ideal standard for HPV identification in HNSCC remains the combination of p16 IHC followed by HPV PCR or HPV ISH [46], because p16+ HNSCC with HPV-independent carcinogenesis can occur in a limited number of cases [44].Despite the majority of the studies included in the present review employed HPV PCR as a diagnostic tool, followed by p16 IHC, and HPV ISH, it is still reasonable to infer that the reported prevalence of HPV-OPSCC and HPV-OSCC can be used as a guide for future public health predictions as the percentage does not differ much among various diagnostic tools.
The current study opted to limit data synthesis and focus on HPV prevalence over the last decade since HPV-OPSCC and HPV-OSCC have escalated over the previous years and are thus influenced by the time of diagnosis.One limitation of the present review is the lack of extensive quantitative measurements of the associating factors such as patients' age, gender, educational background, and social behaviour due to the pooling of the primary data.Multivariate regression using these covariates would have been fascinating to be explored if the available data are not limited.Nevertheless, the present review is still unable to separate the data for oropharyngeal and oral cavity sublocations, which might offer relevant insight into explaining the prevalence of HPV in different subsites [52].Some Asian regions were also under-represented in the present study.
In addition, despite the great reliability and specificity of HPV diagnostic techniques, the lack of false-positive results may also render the present data to a certain level of biasness.Due to the high degree of data heterogeneity, varied study designs, and the extensive pooling of primary data across different countries in Asia with various patients' inclusion and exclusion criteria, it is challenging to assimilate solid conclusions drawn from the current meta-analysis.Nonetheless, data heterogeneity is still present in the current review despite stratifying data based on diagnostic tools and sub-regions in Asia.Other methods employed to mitigate heterogeneity include performing subgroup analysis, meta-regression, and study elimination.

Conclusions
The association between HPV and cancers of the oropharyngeal and oral cavity has gained broad consensus.Although the prevalence of HPV-OPSCC and HPV-OSCC in the Asian region has been consistent over the last decade, the present findings are strongly influenced by the diagnostic tools utilised and geographical locations.Other con-founding variables must be considered, such as age, gender, race, and specific anatomical sites of the oropharyngeal and oral cavities.Nevertheless, future studies into the standardisation of diagnostic tools and the implementation of cancer prevention and treatment programmes emphasising the importance of HPV vaccination are warranted.

Table 1 .
Characteristics of the included studies.HPV of all types in the OSCC group was higher than in the control group (55/200 vs 2/68, OR=11.5, 95% CI=2.6-50.2).HPV-16 and HPV-18 were the main types detected, with the HPV-6 was the only low-risk type identified. of OSCC examined, none of them was found to be positive for HPV18 in this study.The findings of this study revealed that there is a low viral detection of HPV18 in Malaysian OSCC by using saliva samples, suggesting that prevalence of HPV18 may not be important in this group of Malaysian OSCC.oral cancer patients revealed the presence of HPV type 16 and 18 infection.The results suggested that HPV-16 and -18 do not play an important role in oral carcinogenesis in the population from Gujarat, West India.

Table 2 .
Risk of bias assessment and level of evidence of each included study using JBI and OCEBM checklists, respectivelyThe Risk of Bias Using the Joanna Briggs Institute Checklist for Analytical Cross-Sectional Studies