To our knowledge, this is the first time concurrent data on the presence of HPV in three different mucosal sites i.e. the cervical, anal and oro-pharyngeal sites, has been reported for non-immunosuppressed women. Much less is known about the natural history of HPV infection in the anus and oro-pharynx, compared to the cervix. Valid comparative data in studies of oral and anogenital HPV infection have been limited, most studies exhibiting differences in sampling techniques and HPV detection methods, whilst involving different risk groups and risk behaviours.
In this study, the aim was to investigate HPV prevalence and genotypes in a population of 100 women selected by the UK national cervical screening programme for colposcopic examination. As expected, this cohort with abnormal cytology demonstrated a high rate of HPV carriage in the cervix, with an overall prevalence of 96%. Furthermore, patients were tested for the prevalence of HPV infection at two other main mucosal sites, which are clinically relevant to the increasing rates of anal and head and neck cancers observed today. Remarkably, the rates of HPV present at these two other sites were also equally high, with an overall prevalence of 91.4% at the anus and 92.4% at the oro-pharynx. These findings have significant implications for the continuing rise in observed cancers of the anus and the oro-pharynx and strongly suggest that the oropharyngeal and anal sites may be reservoirs of HPV infection with the potential to affect the dynamics of HPV transmission within the community. It follows that post-treatment for cervical disease, a woman should not be declared "HPV negative" unless and until these other major sites have been tested.
The high oral HPV prevalence in our population contrasts to another study, in which 25.5% of women with concurrent cervical HPV infection had detectable oral HPV . In two other studies (one a prospective study in men , the other a systematic review of literature ), oral HPV prevalence of 1.3 - 4.5% has been reported. The observed differences may be due to factors such as different patient populations (men vs. women, different countries of origin), selection bias: our patients were women who were selected because of abnormal cervical smears, and sample sizes: small studies are more prone to bias. Despite this, it is clear that the majority of HPV cancers of the head and neck and anal regions are caused by a small number of HR-types and therefore, the use of the standardised, highly sensitive and specific Roche Linear Array HPV DNA detection method was valid for this study. These findings suggest that women with HPV infection causing dysplasia at the cervix require some further surveillance to detect and prevent development of cancers at these other mucosal sites. Indeed, it has been shown that there is a high risk of anal cancer development in women with a previous history of cervical, vulval or vaginal neoplasia or cancer . Further, in a prospective study, high incidences of vulval, vaginal and anal cancers were found in women who had a history of CIN3 compared to women who did not .
There are very limited data on HPV presence in the anal site in non-immunosuppressed women. It has been reported that 67% of a cohort of young women with anal cytological abnormalities demonstrated detectable HPV . A prevalence of 42% HPV positivity in the anal site in a cohort of "high-risk" HIV-women has also been reported . Studies also show a baseline prevalence of 27% in healthy HIV-women from the Hawaii HPV cohort . A follow-up study from the same group reported that 70% of women followed-up for a mean period of 1.3 years were positive for anal HPV infection at one or more visits . Another study has found that women with confirmed gynaecological neoplasia also have a high rate of biopsy proven anal intraepithelial abnormalities (38%) upon high-resolution anoscopic examination . Our study supports these data and furthermore, suggests that HPV DNA testing per se at the anal site is unlikely to be useful, as there is a high rate of positivity.
The risks of sequential incident infection from cervix to anus and vice versa, is reported to be significantly more likely than infection with a discordant HPV type or no previous HPV infection . The route of transmission from one anogenital site to an adjacent one is not clear and may be by sexual or non-sexual spread.
HPV 16 was significantly the most prevalent genotype across all three mucosal sites. Only poor to moderate concordance rates were demonstrated between the sites (data not shown), which may suggest that initial infection at each of the sites could have been a separate event. Even if infections at these sites were acquired at the same instance, differences in the amount of viral inocula, other factors such as infection persistence, systemic immunity levels and effects of the local environment could all result in high discordance in a cross-sectional study such as this. Interestingly, the strong dominance of HPV 16 at all sites is supportive of published data in which HPV 16 is the predominant type found in squamous cell carcinomas at all three mucosal sites . The finding that HPVs 56 and 66 were significantly more prevalent in LG patients across all three sites compared to HG patients is supported by a study that shows these two high-risk genotypes to be found at > 10-fold higher prevalence in LSIL lesions than in SCCs, suggesting that these are less likely to persist or progress to malignancy .
It may be that this group of non-HIV+ women possess inherent and, as yet undefined, specific inabilities to clear HPV, which accounts for both the cervical dysplasia and high rates of infection seen at all three sites. The presence and expression of high-risk HPV at non-cervical sites is likely to be insufficient for carcinogenesis on their own. Indeed, a recent study has shown that in women from a colposcopy cohort, only 8% of anal smears were positive for high-risk HPV E6 and E7 mRNA, compared to 25% at the cervix . This suggests that there is a significant difference of expression, and hence activity of high-risk E6 and E7 at the anus, compared to the cervix. This may also partially account for the fact that anal cancer rates are much lower than those of the cervix. It has also been suggested that co-factors, such as chronic oestrogen exposure , can act to promote carcinogenesis in the cervix. The presence of co-factors with weaker pro-oncogenic effects at non-cervical mucosal sites may help explain the lower frequencies of cancer observed in the anus and oro-pharynx.