Comparison of oral human papilloma virus detection methods among Hispanic adults

Abstract Objectives Oral human papilloma virus (HPV) infection is associated with nearly three‐quarters of all oropharyngeal cancers in the United States. Research also suggests its association with periodontal disease. There are limited studies evaluating differences in HPV detection methods; however, oral rinse is considered the most sensitive detection method. We compared HPV detection by self‐collected oral rinse versus self‐collected cytobrush and assessed whether the strength of association between periodontitis and HPV is modified by the collection method. Materials and Methods Data from a cross‐sectional study of Hispanic adults in Puerto Rico (n = 346) who provided oral rinse and cytobrush samples for oral HPV detection and were clinically evaluated for periodontitis. The agreement between the oral mouthwash and cytobrush methods was assessed using the Kappa (κ) statistic. Logistic regression models were used to determine if the association between HPV infection and other risk factors varied by oral sample collection method. Results HPV prevalence was slightly higher using cytobrush than oral rinse (5.8% vs. 4.3%). The sensitivity of cytobrush to detect oral HPV was 64.7%, and the specificity was 97.4%. We observed a κ of 0.61 (95% confidence interval [CI]: 0.45–0.78), indicative of fair to good agreement between the two collection methods. The association between oral HPV infection and periodontitis severity was stronger when using the oral rinse collection method (odds ratio [OR] = 3.23, 95% CI: 1.06–9.84); the association was not statistically significant for cytobrush (OR = 1.96, 95% CI: 0.68–5.65). Conclusions These findings support the significance of choosing the most suitable collection method in oral HPV‐related studies. Selecting the most appropriate collection method is an essential criterion in oral HPV‐related studies.


| INTRODUCTION
A vast majority of oropharyngeal cancer cases in the United States are associated with oral Human Papilloma Virus (HPV) infection (Bouvard et al., 2009;Pytynia et al., 2014;Senkomago et al., 2019).
The rising rates of HPV-associated oropharyngeal cancers are a global public health concern despite the introduction of an effective HPV vaccine (Chaturvedi et al., 2011;You et al., 2019). Previous studies have compared HPV detection rates using different DNA extraction kits and amplification methods, finding nested PCR is the most appropriate method for detecting HPV DNA (Božić et al., 2020). In addition to adequate detection and extraction methods, selecting an acceptable sampling method is essential. There are various sampling methods for the detection of oral HPV, including oral rinse, nylon swab, biopsy specimen, cytobrush, and oral mucosal scraping; however, it is unclear which is the best oral HPV detection method (Chikandiwa et al., 2018;Combes et al., 2017;de Souza et al., 2018;Donà et al., 2019;Ong et al., 2014;Steinau et al., 2012). Variability in the rates of oral HPV detection through these collection methods may affect the results of HPV-related studies. Despite there is no universally agreed-upon method, oral rinse is considered the most suitable sample method since it provides the highest HPV detection rate in epidemiologic studies (Rosenthal et al., 2017). A previous study among 163 men who have sex with men found that the agreement for HPV status between oral rinse and brushing is poor (Donà et al., 2019).
Beyond its association with oropharyngeal cancers, HPV infection has also been associated with periodontal disease (Ortiz et al., 2018;Wiener et al., 2015). This chronic oral condition affects approximately 47% (~65 million) of adults in the United States (Eke et al., 2012). Nonetheless, epidemiologic studies are sparse and have yielded inconsistent results (Ortiz et al., 2018(Ortiz et al., , 2019Wiener et al., 2015). A recent review article highlights the contradictory evidence regarding the association between periodontitis and oral HPV infection. These potentially conflictive findings are attributed to the diversity of methodologies to determine periodontitis and oral HPV infection and the consideration of periodontitis severity (Ortiz et al., 2019). The biological plausibility for this association includes the possibility of periodontal tissue as a potential reservoir for oral HPV infection and facilitation of HPV persistence by chronic inflammation of the tissue (Ortiz et al., 2019;Syrjänen, 2018). A recent meta-analysis evaluating the association between HPV and periodontitis found a strong association (odds ratio [OR] = 3.65, 95% confidence interval [CI] = 1.67-8.01) between any oral HPV infection and periodontitis, although an inconclusive positive association between periodontitis and oral high-risk HPV infection (Ali et al., 2020). The aims of this study were: (1) to compare the detection of HPV by oral rinse and self-collected cytobrush and (2) to determine whether the strength of association between HPV and periodontitis is modified by the oral sample collection method in a sample of Hispanic adults in Puerto Rico.

| Study design and study subjects
A subset of 346 consecutive Hispanic participants recruited within a cross-sectional study performed in Puerto Rico between 2014 and 2016 who provided oral rinse and cytobrush samples of the gums comprised the study population of interest (Ortiz et al., 2018). The parent cross-sectional study was designed to determine the prevalence of oral HPV infection in the San Juan Overweight Adults Longitudinal Study (SOALS) (Ortiz et al., 2018). SOALS is a prospective cohort study designed to evaluate the bidirectional association between the progression of periodontitis and glucose abnormalities among overweight and obese adults aged 40-65 years (Andriankaja et al., 2015). The Institutional Review Board of the University of Puerto Rico Medical Sciences Campus approved the study, and all adults provided written consent before participation in the study.

| Data collection procedures
Participants were asked to complete a self-administered questionnaire consisting of 10 questions regarding their current oral hygiene practices: time since last tooth brushing, type of toothbrush and toothpaste used, and mouthwash use. In addition, participants completed a computer-assisted self-interview to gather information on oral HPV infection risk factors, sexual behavior, and drug use.

| Oral samples for HPV determination
Self-collected oral mouthwash samples were collected using the NHANES methodology (Gillison et al., 2012). A 50-ml sterile collection cup filled with 10 ml of Scope (original mint flavor) was provided to each participant. They were asked to rinse/gargle with the mouthwash for 30 seconds, spit the mouthwash into the collection cup while trying not to spill the liquid, and close the cup tightly. Afterward, participants used a cytobrush (QIAGEN) to self-collect an oral sample (Read et al., 2012).
They gently brushed the gums' inner and outer areas for at least 10 strokes. Samples were obtained using the digeneHc2 DNA collection device from QIAGEN. Both samples were performed during the visit and supervised by the study coordinators.

| HPV typing
HPV typing was performed at the University of California at San Francisco HPV Virology Core Laboratory using polymerase chain reaction (PCR) by dot-blot hybridization with modified L1 consensus primers (MY09/MY11) that identify 38 different type-specific probes, including oncogenic and non-oncogenic HPV types (Bouvard et al., 2009;Ortiz et al., 2018). After thawing, the samples were processed for DNA using the Gentra Pure gene Buccal Cell Kit (QIAGEN) following the manufacturer's instructions. The quality of the DNA was evaluated through β-globin gene amplification. A sample was considered unsatisfactory if it was negative for β-globin.
From this analysis, only five samples, using the oral rinse method, were unsatisfactory. A sample was considered HPV positive if it was positive for the consensus probes or any specific HPV type probe.
Amplification of a solution containing all the above components, except for sample DNA, or DNA from cell lines with and without HPV, was used as control.

| Periodontal assessment
Periodontitis was evaluated by clinical measurements. These measurements include clinical attachment loss (CAL) and probing depth (PD) at six sites for all teeth, excluding the third molars (Gillison et al., 2012). Periodontitis was defined using the Centers for Disease Control/American Academy of Periodontology definition (Borrell & Talih, 2012;Palefsky et al., 2001). Severe periodontitis was defined as ≥2 interproximal sites with CAL ≥6 mm (not on the same tooth) and ≥1 interproximal site with PD ≥5 mm. Full mouth clinical exams were conducted by a dental examiner previously calibrated by the NHANES reference examiner (The Centers for Disease Control and Prevention [CDC, 2018]).

| Statistical analysis
Sensitivity and specificity were estimated using 95% binomial confidence intervals (CIs). Receiver operating characteristic (ROC) analysis was used to determine the area under the curve. The Kappa (κ) statistic quantified the level of agreement between the oral mouthwash and cytobrush methods. Logistic regression models were used to assess if the association between HPV infection and severe periodontitis varied by oral sample collection method. The models were adjusted by age, gender, time since last tooth brushing, and number of sexual partners as they were significantly associated (p < .05) in the bivariate analysis to oral HPV infection and periodontitis. The addition of other potential confounders did not change the effect estimates.

| Characteristics of the study population
The mean age of participants was 53.6 ± 6.7 years; 73% were women, and 49% were married or living with someone. Nearly 15% of participants were current smokers and binge drinkers. Approximately 1 in 5 participants had at least 10 lifetime sexual partners, while the vast majority reported less than 2 lifetime oral sex partners. Nearly 20% of participants had severe periodontitis. On the interview day, less than two-thirds of participants brushed their teeth within 5 or more hours before the sample collection, and nearly half used mouthwash before study participation (Table 1).

| Prevalence of oral HPV and method of assessment
The overall HPV prevalence of oral HPV infection in the study population was 4.3% (n = 15) using the oral rinse detection method T A B L E 1 Demographic, lifestyles, and clinical characteristics of the study population (n = 346) and 5.8% (n = 20) using cytobrush. The prevalence of high-and lowrisk types was also lower for oral rinse (0.6% and 2.6%, respectively) than for cytobrush (0.9% and 3.5%, respectively [data not shown]).
According to the area under the ROC curve, the diagnostic performance for cytobrush was 0.83. The κ statistic between collection methods was 0.61 (95% CI = 0.71-0.94).
Some differences were observed in the agreement between collection methods among persons positive to any consensus probes.
Only 9 of these 25 individuals (36%) had the same results with oral rinse and cytobrush (

| Association of oral HPV infection and periodontitis
When using the oral rinse, participants with severe periodontitis had significantly higher adjusted odds of oral HPV infection (OR = 3.23, 95% CI = 1.06-9.84) than those without periodontitis after controlling for age, gender, time since last tooth brushing, and number of sexual partners. Based on cytobrush, individuals with severe periodontitis also had a higher, but statistically nonsignificant odds of oral HPV infection (OR = 1.96, 95% CI = 0.68-5.65, Table 3).

| DISCUSSION
The prevalence of HPV infection was slightly lower with the oral rinse method (4.3%) than the cytobrush (5.8%); the cytobrush detected a higher proportion of low-risk and high-risk infections among HPV positive individuals. Whether the prevalence of HPV types in the gums is somewhat higher than in the oral cavity remains unknown.
Our study also showed fair to good agreement between cytobrush and oral rinse as collection methods for detecting oral HPV (κ: 0.60, 95% CI = 0.71-0.94). Contrary to our findings, a study among men who have sex with men participating in a longitudinal study for oral HPV infection in Italy found a significantly lower rate of positivity for any HPV infection in oral brushing than oral rinse. Moreover, the Italian study found a lower agreement between rinse and oropharyngeal brushing for HPV detection (κ = 0.14, 95% CI: 0.07-0.21). The differences between these studies may be due to variations in cytobrush collection methodology; our samples were self-collected, while samples in the Italian study were collected by otolaryngologists.
The participants in our study gently brushed the gums' inner and outer areas for at least 10 strokes. In contrast, participants in the Italian study used only two cytobrushes-one to collect a sample from the oral cavity (i.e., hard palate, gums, front two-thirds of the tongue, and floor of the mouth below the tongue) and the second from the oropharynx (i.e., soft palate, base of the tongue, and tonsils or tonsillar region). However, it is unclear how many strokes were done with each cytobrush. Finally, the Italian study was performed on a smaller sample of men who have sex with men, whereas we T A B L E 2 HPV types detected among oral HPV-positive cases by collection method (n = 25)   (Ortiz et al., 2018(Ortiz et al., , 2019. For example, we showed a positive association between severe periodontitis and oral HPV infection in Puerto Rico (OR = 2.9, 95% CI: 1.0-8.4) based on the entire sample of the parent study (Ortiz et al., 2018). Although a study based on the 2009-2010 and the 2011-2012 NHANES data did not find an association between periodontitis and oral HPV infection (OR = 1.04, 95% CI: 0.63-1.73), periodontitis severity was not assessed (Wiener et al., 2015). Both of these studies collected oral HPV specimens using oral rinse samples following the NHANES U54GM133807.

CONFLICT OF INTERESTS
The authors declare that there are no conflict of interests.

AUTHOR CONTRIBUTIONS
Conceptualization, supervision, funding acquisition, and project administration was performed by Ana P. Ortiz and Cynthia M. Pérez.
Study design and data acquisition was performed by José T A B L E 3 Logistic regression models of the association between periodontitis and oral HPV infection by collection method (n = 346) All the authors read and approved the final manuscript.

DATA AVAILABILITY STATEMENT
The data are not publicly available due to privacy or ethical restrictions.