The head and neck squamous cell carcinoma (HNSCC) is the fifth most common cancer worldwide, with around 600.000 cases diagnosed in 20081. Most of these cancers arise from squamous cells in the oral cavity, oropharynx, nasopharynx, hypopharynx and larynx. Smoking and alcohol drinking are known risk factors for these cancers2, and Epstein-Barr virus has been recognized as a cause of nasopharyngeal cancer3.

The relationship of Human Papillomavirus (HPV) with a subset of HNSCC cases has been firmly established4-7. HPV positive HNSCC cases are associated with younger age, male predominance and risky sexual behaviors such as a high number of sexual partners and history of oral sex8-10. In contrast HPV-negative HNSCC is most likely related to cumulative measures of smoking and alcohol drinking and poor oral hygiene11. Cancers of the tonsils and base of the tongue are the sites more frequently associated with HPV but HPV has been also detected in oral cavity and larynx.

A systematic review that included 5,046 specimens of HNSCC of 60 studies that used PCR-based methods to detect the genotype of HPV12 showed and estimated prevalence of HPV DNA of about 25.9% of these tumors. The prevalence was significantly higher in cancer of the oropharynx (35.6%, range 11-100%) than in the oral cavity (23.5%, range 40% to 80%) or larynx (24.0%, range 0-100%) with HPV 16 being the predominant genotype found. Among the HPV-positive tumors, HPV 16 is the most frequent genotype in oropharyngeal HNSCC (86.7%) followed by larynx (69.2%) and oral (68.2%). HPV 18 is the second most frequent genotype detected: 2.8% in the oropharynx, 34.1% in the oral cavity and 17% in carcinomas of the larynx.

In 2009, the Internationl Agency for Research on Cancer evaluated the available evidence on the carcinogenicity of HPV in humans and concluded that there is sufficient evidence for the carcinogenicity of HPV 16 in the oral cavity, oropharynx, and tonsils and still limited evidence of HPV as a carcinogen in the larynx3. There are significant geographical and temporal variations in the incidence of HNSCC cancer. Although in Europe and North America, non-HPV related HNSCC has been declining, a time-trend increase in the incidence of HPV-related oropharyngeal cancer has been observed in the last decade9,13-15.

It has been suggested that regional variations in global and site-specific incidence of HNSCC seem to relate to alcohol and tobacco use and that temporal variations may be due to changes in sexual behavior and increased exposure to HPV infection16,17. In spite of a very low HPV prevalence in HNSCC cases18, Brazil has the highest incidence rate of HNSCC among South American countries1,16. Prevention of diseases caused by HPV underlies on the introduction of prophylactic vaccines which are highly effective in preventing HPV infection19-21. Currently available vaccines include HPV 16 and 18, the 2 most prevalent genotypes in HNSCC. Therefore, it is expected that these vaccines have an impact on this type of cancers22.

There are not many studies in South America evaluating the prevalence of HPV in HNSCC. The aim of this study was to determine the prevalence of HPV genotypes by age and gender in cancer of the oral cavity, oropharynx and larynx, in four laboratories of pathology from the city of Medellin in Colombia.

The causal association between HPV and a subset of HNSCC, has been demonstrated5-7 and studies of case series suggest that HPV is associated with a proportion of about 25-60% of these cases12. The confirmation of HPV association with these cancers opens the opportunity for implementing prevention strategies based on the control of HPV infection. Currently available HPV prophylactic vaccines have shown a nearly 100% efficacy in preventing infection with HPV 16 and 18, with a projected reduction in risk of cervical cancer by 50-70% and significant impact on non-genital HPV related diseases22. The ability of these vaccines to reduce the incidence of these cancers may depend partly on the proportion of cases attributed to the genotypes included in current available vaccines. Estimates of the contribution of HPV genotypes in HNSCC are scarce in Latin American countries12.

In this article we describe the demographic characteristics of 175 cases of HNSCC consecutively diagnosed over a period of 10 years (1999-2008) and the prevalence of HPV in these cases. Our results show similar prevalence of HPV found in other studies conducted in several regions worldwide12 but higher than the prevalence described in a study including cases from Brazil, Cuba and Argentina18. These findings are consistent with observations on the geographical variability of the prevalence of HPV in HNSCC worldwide. Our study also found that among the positive HPV samples, HPV 16 was present in 82% of cases, while HPV 18 was observed only in 18% of cases. These results are consistent with widely reported data which implicates HPV 16 in most cases of HNSCC12,25.

As it has been described26, we also noted that despite the high proportion of cases of HNSCC being diagnosed in people over 50 years, HPV-positive cases tend to be younger than HPV-negative (60.1 vs. 65.5, Mann-Whitney test p = 0.02), and this difference persisted even when the analysis included only cases positive for HPV 16 although this did not reach statistical significance (data not shown). The prevalence of HPV in HNSCC in this study is higher than that observed in cases from Brazil, Argentine and Cuba18. Due to the nature of our study, it was not possible to collect information on alcohol and tobacco use. It has been suggested that the proportions of anatomical sites included in the analysis can influence the observed prevalence differences.

The proportion of cancers of the oral cavity, larynx and oropharynx in that study was 31%, 49% and 19% respectively, which is similar to the proportion of cases included in our study. HNSCC is a malignant epithelial tumor with squamous differentiation characterized by the formation of keratin or the presence of intercellular bridges or both. Histopathologic characteristics that could be evaluated in the study biopsies and resections, demonstrated well to moderate differentiated SCC, mostly keratinizing tumors, moderate to severe peritumoral inflammatory infiltrate and expansive or infiltrative pattern of invasion regardless of anatomic sites. HPV has been associated as a risk factor for non-keratinizing HNSCC occurring in the tonsillar (oropharyngeal and oral) and sinunasal areas.

In our study, the proportion of oral cavity HPV-positive cases was higher than the proportion of larynx and oropharynx cases (sinunasal area not included). But these differences were not statistically significant and remained insignificant after adjustment by age. The lymphoid tissue of the tonsillar ring in the head and neck consists of the palatine, the pharyngeal and the lingual tonsils, which means it is distributed mainly along the oropharynx but the oral cavity too. It is worth noting that among the evaluated biopsies many were incisional biopsies that represent mainly the superficial area of the tumor and did not include representation of the non-neoplasic tissues, where tonsillar tissue may be recognized more easily.

Even though tonsillar tissue was not observed in these oral cavity biopsies, we cannot exclude from the histological slides that in some cases the tumor could involve or originated from tonsils and that this could explain the higher prevalence of HPV DNA in cases of oral cavity. In the same direction, maybe our oropharynx cases did not include as many tumors from the tonsils and thus the HPV prevalence may have been underestimated. Even further, it has been reported that recognition of non-keratinizing morphology has low reproducibility27. Another reason may be the methods used to determine exposure to HPV. For the assessment of prevalence estimates in HNSCC, serological tests that detect antibodies against L1 protein, E6 and E7 have been used.

Although these essays are useful biomarkers to define the role of HPV in HNSCC they may have some limitations to estimate the fraction of these cancers attributable to HPV. For example, antibodies against L1 are markers of current or previous exposure to HPV infection at any anatomical site, including anogenital infections and its correlation with the presence of DNA at the site of the lesion is very poor28. Antibodies against the E6 and E7 proteins correlate with invasive cancer at least in cervical cancer, but only 60%-70% of HNSCC cases are positive for these antibodies29. In cervical cancer, the most accurate method in the assessment of the exposure to HPV has been PCR of HPV DNA extracted from tumor microdissected lesions30, but the use of paraffin blocks requires the implementation of proven methods with high sensitivity in paraffin-embedded material.

Although fresh frozen tumor biopsies produce good quality DNA, they have the limitation that identification and microdisection of the tumor tissue cannot be conducted. Indeed, strength of our study is the use of paraffin blocks, which were re-cut for central pathology review and extraction of DNA from the dissected material of tumors identified by the study pathologist. We also evaluated the quality of DNA by amplifying a segment of 209 bp of the β-globin gene and use 2 type specific PCR tests that amplify small size fragments (96 pb for HPV 16 and 115 pb for HPV 18) of the E6 gene which have shown to be highly sensitive in this type of sample24.

Although we cannot completely exclude the possibility of false negatives due to technical artifacts, the use of highly sensitive type-specific PCR for HPV 16 and 18 may help to overcome this limitation, especially since these two are the predominant genotypes in HNSCC. Potential limitations of our study include possible selection bias of samples. The cases were collected from only four local laboratories, but we cannot exclude the likelihood that these cases do not represent the entire population of cases of HNSCC diagnosed in the geographic area. Compared with Brazil, Cuba and Argentina, the incidence rate of HNSCC in Colombia is lower31. Likewise, cigarette smoking and annual cigarette consumption per person is lower in Colombia32,33. It would be useful to evaluate HPV prevalence in HNSCC cases in the context of these epidemiological profiles to ascertain the fraction attributed to each risk factor in HNSCC.

The HPV prevalence in these series of cases of four laboratories of the city of Medellin, Colombia was 18.9%. HPV was found in 23.9%, 17.5% and 13.3% of oral cavity, larynx and oropharynx cases respectively. Twenty-seven and six samples were HPV 16 (15.4%) and HPV 18 (3.42%) respectively. Among HPV positive cases, 82% were HPV 16 and 18% HPV 18. No other HPV genotypes were identified. According to these data, the mass introduction of prophylactic HPV vaccines may have the potential to reduce the number of HNSCC cases in Colombia, but the magnitude of this effect is difficult to predict at this time because clinical trials which are specifically designed to evaluate the efficacy of these vaccines in the prevention of HPV oral infection and on cancers of the head and neck have not conducted.