REPORTS HAVE INDICATED AN increased prevalence of oral condylomas in human immunodeficiency virus (HIV)–infected individuals on highly active antiretroviral therapy (HAART).1–3 Human papillomavirus (HPV) is the etiologic agent of condylomatous lesions (warts) of the skin and mucosal tissues. Investigators examining the role of HPV in the etiology of HIV-associated oral warts have demonstrated a wide variety of HPV genotypes, including but not limited to the oral-specific HPV types 7, 13, and 32 and HPV types 6, 11, 16, 18, and 35 commonly found in the genital tract.4,5
In order to understand the pathogenesis of HPV-associated lesions in the oral mucosa of HIV-infected individuals, the incidence, prevalence, risk factors, and natural history of HPV infection in the oral cavity must be investigated. Prevalence of oral HPV was previously shown to be 11% in a Canadian cohort of HIV-infected individuals and 25% in a Baltimore cohort.6,7 The reported increased prevalence of oral condylomas since the widespread administration of HAART suggests that a drug or combination of drugs used to treat HIV may be a risk factor for oral HPV infection. The Canadian study predates the reported increase in oral warts in the HIV-positive population on HAART, and so it did not assess the impact of HAART or other antiretroviral therapy on oral HPV infections. The Baltimore study also failed to assess the impact of HIV therapy, especially HAART, on prevalent oral HPV infections. No study to date has examined the relationship between oral HPV infection and HIV therapy. In this study, we used a robust consensus primer polymerase chain reaction (PCR) and linear array genotyping assay to examine the prevalence of HPV in banked saliva of 98 HIV seropositive adults attending an outpatient clinic in New Orleans, LA. Risk factors for oral HPV infection, including HIV therapy regimens and markers of HIV disease status and immunocompetence, were evaluated.
Study participants were patients attending the HIV outpatient program at the Medical Center of Louisiana in New Orleans. Participants were enrolled in a study examining local immunity to oropharyngeal candidiasis (principal investigator Dr. Paul Fidel; protocol approved by local institutional review boards [IRB] before initiation), and informed consent was obtained before enrollment. Basic demographic characteristics were extracted from the Adult Spectrum of Disease database (sponsored by the Centers for Disease Control and Prevention and maintained locally by the HIV clinic at each participating site), which monitors HIV and health-related information for HIV-positive patients at 6-month intervals.8 Salivary DNA specimens were banked at the conclusion of the study, and 99 HIV-positive individuals with adequate specimens remaining were tested for HPV DNA. Before HPV testing, all specimens were deidentified and referred only by study number per Louisiana State University Health Sciences Center IRB protocol for exempted studies.
Collection and Processing of Whole Saliva
Participants provided 10 ml of whole, unstimulated saliva by expectorating into a sterile collection tube. The sample was clarified by centrifugation for 5 minutes at 800 × g at 4°C. Supernatants were removed and pellets were kept on ice for DNA extraction. Within 24 hours of collection, DNA was extracted from saliva cellular pellets using a Qiagen DNA extraction kit as per the manufacturer's protocol (Qiagen, Valencia, CA). DNA extracts were stored at −80°C.
HPV DNA Detection and Genotyping
Saliva DNA extracts were tested for HPV DNA by a multiplex PCR reaction containing biotinylated L1-consensus primers (PGMY09/11) and biotinylated β-globin primers (GH20/PC04) to control for sample adequacy. PCR contamination was avoided by using separate areas for DNA extraction, PCR setup, and gel electrophoresis. A master PCR mix was prepared for each run, and water blanks were inserted after every fourth sample to monitor for contamination. Reactions in which the controls did not behave as predicted were repeated. Primers and PCR reagents were kindly supplied by Roche Molecular Systems (Alameda, CA) and reactions were conducted according to Roche protocol.9 After 40 amplification cycles, PCR products were visualized by ethidium bromide–stained 2% agarose gel. Only 1 sample was β-globin negative upon gel analysis; this sample was considered indeterminate and was excluded from further analysis (number of individuals with adequate HPV results, n = 98).
Samples that demonstrated the 254-base-pair β-globin band and the ∼450-base-pair HPV band were genotyped by the Roche linear array assay, which distinguishes 27 HPV genotypes commonly found in the female genital tract.9 Nylon strips lined with HPV type-specific probes and all hybridization and detection reagents were supplied by Roche, and hybridization was conducted according to Roche protocol.9 Strip results were interpreted within 24 hours of development, and HPV genotype positivity was defined as any probe on the strip having a band intensity equal to or greater than that of the β-globin probes on that strip. Specimens that demonstrated the ∼450-base-pair band on agarose gel analysis but that did not hybridize to any of the probes on the linear array strip were classified as HPV positive, unidentified genotype.
Variables were summarized using the SAS procedures FREQ for categorical variables and the MEANS for continuous variables. All hypothesis tests were 2 sided and were tested at the α = 0.05 level of significance. The variables CD4+ T lymphocyte count and peripheral HIV viral load were treated as dichotomous to reflect clinically significant cutoff values (i.e., patients with CD4+ count <200 cells/ml are considered to be immunosuppressed; patients with HIV viral loads <400 copies/ml are considered to be controlling HIV infection, and patients with viral loads >10,000 copies/ml are considered to be failing therapy). Highly active antiretroviral therapy (HAART) was defined as any antiretroviral regimen containing a minimum of 3 drugs, including 3 nucleoside reverse transcriptase inhibitors (NRTIs), 2 NRTIs and 1 protease inhibitor, or 2 NRTIs and 1 nonnucleoside reverse transcriptase inhibitor (NNRTI). These variables were extracted from the Adult Spectrum of Disease database, which tracks clinical data in 6-month intervals. Values used in this analysis were taken within 6 months before the date of enrollment in the study.
Logistic Regression Analysis
Logistic regression models for predicting the probability of having HPV were constructed first as single predictor univariate models (e.g., any HPV versus log (viral load)), and then in multiple logistic regression models involving relevant demographic and explanatory variables (covariates) potentially useful in modeling HPV status. The covariates used in the multiple logistic models were CD4+ T lymphocyte count, log of HIV viral load, race, gender, age (in years), and the binary status variables of oropharyngeal candidiasis (OPC), HAART, protease inhibitors, drug use, and regimen. Three multiple logistic regression modeling strategies were used: (1) inclusion of all potentially important variables based on biologic reasoning; (2) backward elimination variable selection; and (3) stepwise variable selection. SAS; version 9.0 (SAS Institute, Cary, NC) was used to compute the logistic regression models. Information from 95 subjects was used in the analysis. Four individuals were excluded because of incomplete survey information. Because the results were so different for Caucasian and African American individuals, separate logistic regression models were generated for these 2 categories.
Sociodemographics of Study Population
The sociodemographics of the study population are shown in Table 1. The population was predominantly male (64%), African American (65%), and 62% was on HAART at the time of enrollment. Sociodemographics of African Americans and Caucasians were similar, with the exceptions that African Americans were slightly younger and less likely to report the mode of acquisition of HIV than Caucasians.
Prevalence of HPV Genotypes in Saliva
Overall, 35% of subjects were HPV DNA positive in saliva (Table 1, Fig. 1). HPV prevalence is shown stratified by oncogenic potential of the virus types (high-oncogenic risk versus low-oncogenic risk, Fig. 1). High-risk HPV types (associated with anogenital and oral carcinomas) were more prevalent (26%) than low-risk HPV types (associated with benign anogenital and oral condylomas, 4%). Multiple HPV genotypes were identified in 7% of subjects, and 8% harbored HPV that could not be genotyped by the linear array assay (unidentified genotypes). Prevalence of any HPV, high-oncogenic-risk HPV, and multiple HPV genotypes were significantly higher in Caucasians compared to African Americans. The number of samples positive for individual HPV genotypes is shown in Figure 2. Sixteen out of the 27 HPV genotypes identified by the linear array assay were represented in the saliva specimens of the study population. The most prevalent genotype detected was HPV-16, followed by HPV-55 and -83.
Risk Factors for Detection of Oral HPV in HIV-Infected Individuals
Sociodemographic and HIV disease-associated variables were assessed as possible risk factors for detection of oral HPV DNA. The results of univariate analysis for the entire population are shown in Table 2. The only factor significantly associated with detection of HPV DNA was race, with Caucasians having an increased risk of almost fourfold over African Americans [OR (CI) = 3.90 (1.60–9.54), P < 0.01]. Lower CD4+ T lymphocyte counts (below 200 cells/ml) and higher HIV plasma viral loads (above 10,000 copies/ml) did not increase risk for oral HPV infection. Oral HPV was detected equally in those on HAART and those not on HAART. When patients on HAART were further classified by HAART efficacy, HPV infection was detected more frequently in those on effective HAART (57%) or marginally effective therapy (54%) compared to those on clinically ineffective HAART (29%) [OR (CI) = 2.00 (0.92–4.31)]. This observation was approaching statistical significance (P = 0.079).
Risk Factors for Oral HPV DNA Detection in HIV-Infected African Americans Versus Caucasians
Because the African American and Caucasian populations examined in this study demonstrated differences in oral HPV infection status, they were examined separately for risk factors for oral HPV DNA. Table 3 shows the risk factor analysis for oral HPV DNA detection in African Americans. Among African Americans, being on HAART was marginally protective against oral HPV infection [OR (CI) = 0.34 (0.11–1.09), P = 0.07]. Interestingly, risk of oral HPV increased with increasingly effective therapy. HPV prevalence was only 7% in those on ineffective HAART, 38% in those on marginally effective therapy, and 50% in those on effective therapy [OR (CI) = 3.92 (1.21–12.56)]. In multivariate analysis adjusting for HIV viral load, CD4+ T lymphocyte count and age, individuals who reported acquisition of HIV by sexual contact [OR (CI) = 4.04 (1.49–10.98), P = 0.006] and those on effective HAART [OR (CI) = 4.10 (1.03–1.59) P = 0.040] were at increased risk for oral HPV detection.
Table 4 shows the risk factor analysis for oral HPV DNA detection in Caucasians. Although not statistically significant, males were more likely than females to be orally infected with HPV (P = 0.095), as were those with concurrent oropharyngeal candidiasis (P = 0.081). In contrast to the African American population, being on HAART therapy significantly increased the likelihood of having an oral HPV infection [OR (CI) = 6.75 (1.43–31.90), P = 0.016]. This increased risk was regardless of the effectiveness of HAART. There was a marginally significant positive association between oral HPV infection and therapy regimens containing a protease inhibitor (P = 0.079). Plasma HIV viral load did not predict oral HPV infection, in contrast to the African American population. In a multivariate model controlling for HIV viral load, CD4+ T lymphocyte count and age, being on HAART was still significantly associated with the presence of oral HPV DNA [OR (CI) = 7.27 (1.24 − 46.3), P = 0.029].
The reported increase in prevalence of oral warts in HIV-infected individuals on HAART1–3 prompted the examination of the oral HPV genotype prevalence and HIV-associated risk factors for oral HPV infection. We used a convenience sample of banked saliva DNA specimens for determining oral HPV infections to take a preliminary look at the impact of HAART on oral HPV. Previous reports have indicated that oral rinses provide a more complete analysis of total oral HPV infection than do swabs or brushes,10,11 and saliva compares favorably to oral rinses (unpublished observation), therefore providing a reliable sample for the comprehensive detection of HPV infection of the oral cavity.
The oral HPV prevalence of 37% found in this study is marginally higher than that reported for 2 other HIV cohorts.6,11 The 11% prevalence in the Canadian cohort most likely reflects the chosen sampling method (swabs of 3 oral sites); the 25% prevalence reported in the Baltimore cohort might reflect temporal or population differences compared to this New Orleans cohort.
Volter et al.5 have reported the detection of HPV-16 and -55, 2 of the most prevalent HPV genotypes detected in this study, in oral lesions of HIV-infected patients. HPV-16 was also the most prevalent genotype detected in the Canadian and Baltimore HIV cohorts, despite the fact that the spectrum of genotypes examined varied among these 3 studies.6,11 The Baltimore study reported high prevalence of HPV-89, which was not identified by the Roche linear array assay used in this study. It is likely that a proportion of the indeterminate genotypes detected in this study were HPV-89, as well as other mucosal genotypes (e.g., HPV-61, -62, and -72) not identified by the technology used in this study. It is also likely that a proportion of these HPV infections is due to oral-specific genotypes (HPV-7, -13, and -32). Although the PGMY primer set used in this study is capable of amplifying these genotypes (unpublished data), the prevalence of oral-specific HPV genotypes in HIV-infected individuals remains unknown. The development of an oral HPV specific linear array assay would enable a more thorough analysis of the HPV genotypes present in the oral cavity.
The majority of HPV genotypes detected in saliva were classified as high-oncogenic risk HPV genotypes, which are associated with anogenital malignancies and have also been detected in oropharyngeal malignancies.12–21 Although high-risk HPV infection appears to be common in HIV-infected persons, it is yet unclear whether this will lead to increases in oral malignancies in this population as HAART extends the lives of HIV patients. Benign wart lesions are usually associated with infection with low-risk genotypes of HPV (e.g., HPV-6 and -11); however, the high-risk HPV types 16 and 18 have been detected in HIV-associated oral warts.4,5 HPV 6 and 11 were only rarely seen in this study, which may implicate oral-specific low-risk HPV genotypes in the etiology of oral warts in HIV-positive individuals.
The risk of oral HPV infection in the context of HAART was different between Caucasians and African Americans. For Caucasians, simply taking a HAART regimen was a significant predictor of oral HPV infection. A dichotomy was seen in African Americans with those on HAART trending toward a decreased risk of oral HPV infection, while those on effective HAART had an increased risk of oral HPV infection. Compliance to the prescribed HIV therapy regimen may be responsible for the differences observed between Caucasians and African Americans, as well as the dichotomy observed within the African American group. The data could be explained if effective HAART in the African American group is a marker of compliance, while the Caucasian group is more uniformly compliant, and oral HPV infection is associated with a drug or drugs making up the HAART regimen. Unfortunately, we were not able to formally measure compliance in this population. We were able to obtain measures of plasma HIV viral load, but a high viral load can indicate either noncompliance or treatment failure in a compliant individual. This is the first report that attempts to examine the impact of HIV therapy, particularly HAART, on oral HPV infections.
Previous studies have shown that sexual risk factors (oral sex practices, homosexuality, previous history of sexually transmitted diseases) can impact oral HPV infection rates.6,11 In this study, homosexuality was not a significant risk factor for oral HPV infection. Detailed data on sexual behavior were not available for this population. However, because the population was exclusively HIV seropositive, it is likely that many individuals in the group did exhibit high-risk sexual behavior. This fact may have masked any potential impact of sexual practices, lifetime number of sex partners, age of first coitus, and other known risk factors for HPV infection. It is distinctly possible, however, that a specific sexual behavior or other confounder that differs among Caucasians and African Americans could account for the differences in oral HPV infection seen in this study. Future studies must include a detailed analysis of the impact of specific sexual practices, as well as other oral health and hygiene measures, on oral HPV infection.
The cause of the high prevalence of oral HPV infection in HIV-infected individuals is unclear. It is possible that HIV-infected individuals exhibiting high-risk sexual practices frequently acquire oral HPV infection due to multiple exposures. Alternatively, the high HPV detection rates could be due to increased HPV replication and/or persistence rather than increased HPV acquisition. If persistence of oral HPV leads to HPV-related disease, as in the genital tract, then increased persistence of HPV could explain the increased prevalence of oral warts in HAART-treated HIV-positive individuals. This persistence could be a result of increased immunosuppression from advancing HIV disease, though the general lack of association with markers of immunosuppression (low CD4+ T lymphocyte counts and high HIV viral loads) contradicts this theory. This does not rule out local immunosuppression that is not being restored by HAART; however, the decrease in other opportunistic infections of the oral cavity (oral candidiasis, oral hairy leukoplakia) since HAART administration would indicate the unlikely scenario that the local immune deficiency is HPV-specific. It is also possible that the individuals who are placed on HAART have already been immunosuppressed long enough to surpass some critical threshold for developing HPV-related disease, which cannot be reversed with therapy.
We have shown that saliva can be used to study the epidemiology of HPV infection of the oral cavity in HIV-infected people. Longitudinal studies are needed to better delineate the role of specific HPV genotypes in oral lesions and the mechanisms by which these lesions occur, particularly in the increased-risk setting of HAART. Further studies are also warranted to determine the precise role of HIV therapy regimens and the covariables associated with racial differences in oral HPV infections.
1. Leigh J. Oral warts rise dramatically with use of new agents in HIV. HIV Clin 2000; 12:7.
2. Greenspan D, Canchola AJ, MacPhail LA, Cheikh B, Greenspan JS. Effect of highly active antiretroviral therapy on frequency of oral warts. Lancet 2001; 357:1411–1412.
3. King MD, Reznik DA, O'Daniels CM, Larsen NM, Osterholt D, Blumberg HM. Human papillomavirus-associated oral warts among human immunodeficiency virus-seropositive patients in the era of highly active antiretroviral therapy: an emerging infection. Clin Infect Dis 2002; 34:641–648.
4. Greenspan D, de Villiers EM, Greenspan JS, de Souza YG, zur HH. Unusual HPV types in oral warts in association with HIV infection. J Oral Pathol 1988; 17:482–488.
5. Volter C, He Y, Delius H, et al. Novel HPV types present in oral papillomatous lesions from patients with HIV infection. Int J Cancer 1996; 66:453–456.
6. Coutlee F, Trottier AM, Ghattas G, et al. Risk factors for oral human papillomavirus in adults infected and not infected with human immunodeficiency virus. Sex Transm Dis 1997; 24:23–31.
7. Kreimer AR, Alberg AJ, Daniel R, et al. Oral human papillomavirus infection in adults is associated with sexual behavior and HIV serostatus. J Infect Dis 2004; 189:686–698.
8. Farizo KM, Buehler JW, Chamberland ME, et al. Spectrum of disease in persons with human immunodeficiency virus infection in the United States. JAMA 1992; 267:1798–1805.
9. Gravitt PE, Peyton CL, Apple RJ, Wheeler CM. Genotyping of 27 human papillomavirus types by using L1 consensus PCR products by a single-hybridization, reverse line blot detection method. J Clin Microbiol 1998; 36:3020–3027.
10. Lawton G, Thomas S, Schonrock J, Monsour F, Frazer I. Human papillomaviruses in normal oral mucosa: a comparison of methods for sample collection. J Oral Pathol Med 1992; 21:265–269.
11. Gillison ML, Lowy DR. A causal role for human papillomavirus in head and neck cancer. Lancet 2004; 363:1488–1489.
12. Brandsma JL, Abramson AL. Association of papillomavirus with cancers of the head and neck. Arch Otolaryngol Head Neck Surg 1989; 115:621–625.
13. Palefsky JM, Silverman S Jr, Abdel-Salaam M, Daniels TE, Greenspan JS. Association between proliferative verrucous leukoplakia and infection with human papillomavirus type 16. J Oral Pathol Med 1995; 24:193–197.
14. Mineta H, Ogino T, Amano HM, et al. Human papilloma virus (HPV) type 16 and 18 detected in head and neck squamous cell carcinoma. Anticancer Res 1998; 18:4765–4768.
15. Schwartz SM, Daling JR, Doody DR, et al. Oral cancer risk in relation to sexual history and evidence of human papillomavirus infection. J Natl Cancer Inst 1998; 90:1626–1636.
16. Smith EM, Hoffman HT, Summersgill KS, Kirchner HL, Turek LP, Haugen TH. Human papillomavirus and risk of oral cancer. Laryngoscope 1998; 108:1098–1103.
17. Aggelopoulou EP, Skarlos D, Papadimitriou C, Kittas C, Troungos C. Human papilloma virus DNA detection in oral lesions in the Greek population. Anticancer Res 1999; 19:1391–1395.
18. Gillison ML, Koch WM, Capone RB, et al. Evidence for a causal association between human papillomavirus and a subset of head and neck cancers. J Natl Cancer Inst 2000; 92:709–720.
19. Badaracco G, Venuti A, Morello R, Muller A, Marcante ML. Human papillomavirus in head and neck carcinomas: prevalence, physical status and relationship with clinical/pathological parameters. Anticancer Res 2000; 20:1301–1305.
20. Lindel K, Beer KT, Laissue J, Greiner RH, Aebersold DM. Human papillomavirus positive squamous cell carcinoma of the oropharynx: a radiosensitive subgroup of head and neck carcinoma. Cancer 2001; 92:805–813.
© Copyright 2005 American Sexually Transmitted Diseases Association
21. Shin KH, Park KH, Hong HJ, et al. Prevalence of microsatellite instability, inactivation of mismatch repair genes, p53 mutation, and human papillomavirus infection in Korean oral cancer patients. Int J Oncol 2002; 21:297–302.