The recent Nobel Prize winner in Physiology or Medicine, Harald zur Hausen, first hypothesised a role for human papillomavirus (HPV) in the development of cervical cancer. Since the 1970s HPV has been proposed as a causative factor in a variety of benign and malignant diseases. HPV is a non-enveloped double-stranded DNA virus that infects the epithelial basal layer. The majority of HPV infections occur without symptoms and are cleared by the host over 8–12 months. However, infection may persist, resulting in intraepithelial neoplasia and, over time, progression to invasive carcinoma.
One hundred and twenty different types of HPV affecting humans have now been described. HPV has been identified in various benign and malignant lesions within anogenital sites, the aerodigestive tract, skin and conjunctiva. There are extensive data in the literature regarding HPV and cervical disease. Walboomers et al. reported that the worldwide prevalence of HPV in cervical carcinomas is 99.7%. The burden of HPV in non-cervical disease has been increasingly recognised and this review will summarise the evidence available.
HPV is species- and tissue-specific (Table 1). There are approximately 40 types that affect human mucocutaneous tissue, such as those found in the anal, genital and oral tracts. In addition, they can be divided between those of malignant potential (high risk) and those associated with benign conditions (low risk). Although the focus of this article is on HPV-associated neoplastic disease, there is reference to benign conditions affecting the anogenital site, with particular emphasis on their management in pregnancy.
Table 1. Diseases commonly associated with individual HPV types
1, 2 & 4
Flat cutaneous warts
Carcinogenesis in epidermodysplasia verruciformis
Anogenital warts, respiratory papillomatosis
Anogenital neoplasia, oropharyngeal cancers
In 2010, 1172 women were diagnosed with vulval cancer in the UK. When coupled with cancer of the vagina, it accounts for 8% of gynaecological cancers in the UK. The incidence of both vulval cancer and its precursor, vulval intraepithelial neoplasia (VIN), are increasing, especially in women below the age of 50 years.[6-8] In the UK, the proportion of vulval cancer diagnosed in women under the age of 50 rose from 6% in 1975 to 14% in 2008. Vulval squamous cell carcinoma (VSCC) accounts for over 90% of vulval cancers. There appear to be at least two distinct pathways for the development of vulval cancer and VIN (Figure 1). Usual type VIN is most common and is associated with HPV, younger patients, multifocal lesions and other anogenital intraepithelial neoplasia. It generally gives rise to warty/basaloid VSCC. Differentiated VIN is usually found adjacent to invasive keratinising VSCC. It is typically HPV-negative and is seen most frequently in older women with other epithelial disorders such as lichen sclerosus or lichen simplex chronicus.
Most of the studies looking at HPV prevalence in vulval cancer and VIN have been small and results vary considerably. A meta-analysis carried out in 2009 reported HPV prevalence in vulval cancer, high-grade VIN and low-grade VIN to be 40.4%, 85.3% and 67.8%, respectively. HPV 16 is by far the most commonly found type accounting for 67.5% of VIN and 32.2% of vulval cancer. The study demonstrates the paucity of data regarding vulval neoplasia. HPV positivity in VSCC varies widely between populations, from 34.7% in some European countries to 63.2% in the USA.
The symptoms associated with VIN and vulval cancer can be quite distressing and embarrassing. However, the diagnosis of vulval cancer is often delayed. A Dutch study found that out of all the gynaecological malignancies, vulval cancer had the longest delay in diagnosis. The most common presenting symptom of VIN is pruritus, however, this is also a common symptom found in many benign conditions and therefore may not be recognised as serious. Clinical examination is essential and in order to make an accurate diagnosis a biopsy is always recommended. VIN lesions in particular can vary in size, shape, regularity, pigmentation and location on the vulva and, without a biopsy, misdiagnosis is common.
Primary cancer of the vagina is rare. The most common causes of squamous cell vaginal cancer are HPV and irradiation. There were 281 cases of vaginal cancer in the UK in 2010. There are a limited number of studies looking at HPV prevalence in vaginal cancer and its precursor, vaginal intraepithelial neoplasia (VAIN). The combined overall HPV prevalence from a total of 15 studies has been reported as 100% in VAIN 1, 90.1% in VAIN 2/3 and 69.9% in vaginal carcinomas. HPV 16 is responsible for the majority of the HPV-related vaginal disease. VAIN may be responsible for abnormal cervical smears. Careful vaginal examination is, therefore, mandatory when a cervical cause for the abnormal smear cannot be found. Vaginal lesions may also coexist with cervical or vulval lesions and may not, therefore, be recorded as a separate entity, leading to underreporting.
The incidence of anal cancer is approximately 1100 new diagnoses each year in the UK. Although a relatively uncommon malignancy, studies have shown an increase in incidence over the last few decades.[18, 19] Both anal cancer and anal intraepithelial neoplasia have been linked to HPV infection. Overall HPV prevalence has been reported as 91.5%, 93.9% and 84.3% in anal intraepithelial neoplasia (AIN) 1, AIN 2 and AIN 3, and anal carcinomas, respectively. HPV 16 is responsible for 68–76% of anal cancers.[3, 11, 20] However, in immunocompromised individuals with anal cancers or AIN 2/3, several studies have shown proportionally higher rates of infection with other genotypes, and they are often multi-type infections.[3, 11, 20] In the absence of any effective prevention or screening, the current trends of increasing incidence appear unlikely to change until the effects of HPV prophylactic vaccination become apparent in 10–30 years.
Penile cancer is rare in the UK, with approximately 520 cases diagnosed each year. It appears that, like vulval neoplasia, the prevalence of HPV in penile cancer varies with histological type.[3, 22] The review by Backes et al. reported the overall prevalence of HPV in penile cancer to be 47.9%. Basaloid and warty subtypes are far more consistently associated with HPV infection than verrucous penile cancers (66.3% versus 22.4%). HPV 16 is again the most prevalent HPV type. These histological types have also been associated with a precancerous stage called penile intraepithelial neoplasia, which has an HPV prevalence of 75–100%.[23, 24]
HPV is associated with the development of a subset of head and neck cancers arising in the oropharynx, which includes the tonsils, base of the tongue and soft palate. The UK incidence of oropharyngeal cancer in 2009 was 1346. In the majority of head and neck cancers, the UK incidence is stable or decreasing, whereas the incidence of oropharyngeal cancer, and tonsillar cancer in particular, is increasing. This trend has also been demonstrated in many other countries including the USA, Sweden and Greece, and is known to affect younger patients with less exposure to the potentially carcinogenic effects of tobacco and alcohol, relative to HPV-negative patients.[27-30] In Scotland, incidence rates for oropharyngeal cancer are increasing faster than rates for any other cancer. The hypothesis that HPV is the fundamental reason for this is supported by a recent UK study. It reports an increase in HPV-positive cases from 15% to 57% between 1988 and 2009.
The natural history of HPV infection in the oral cavity is currently poorly defined. Oral HPV infection is believed to be sexually acquired.[27, 33] Case–control studies have shown an association between HPV-positive oropharyngeal cancer and sexual behaviours including a high lifetime number of oral sex or vaginal sex partners, early age at first intercourse and infrequent use of condoms. However, questions regarding the duration of infection, degree of oral transmission and whether productive viral infections are established in the oropharynx have yet to be answered.
There is accumulating evidence from prospective studies that HPV-positive oropharyngeal cancer responds better to treatment, including chemotherapy and radiotherapy, than HPV-negative oropharyngeal cancer, and that patients with HPV-positive disease have excellent long-term survival rates. Clinical trials to investigate de-escalation of treatment for HPV cases are now under way in the UK and USA.
There are only a few studies looking at HPV prevalence in the oesophagus, larynx and lung. The findings are inconsistent and a causative role for HPV in the development of these cancers cannot yet be reliably demonstrated.
Skin and conjunctiva
The skin of both healthy populations and immunocompromised patients harbours a very large spectrum of HPV genotypes. Genus β papillomaviruses, in particular, are involved in cutaneous lesions in humans. In 2010, Karagas et al. showed that people with squamous cell carcinoma, but not basal cell carcinoma, were far more likely to have each of the β HPV types compared with people in the control group. In addition, they found the more HPV types that were present increased the probability of having squamous cell carcinoma. However, the pathogenesis in relation to HPV still remains unclear. Further molecular and epidemiological data are required.
The evidence for a role of HPV in ocular disease is controversial. HPV has been implicated in lid warts, conjunctival and lacrimal sac papillomas, conjunctival and corneal dysplasia, as well as squamous cell carcinoma of the conjunctiva. However, the major studies examining the role of HPV infection in ocular disease have mixed findings. In a 2008 review, Hughes et al. found that 80% of conjunctival papillomas were HPV-positive. Thirty-two per cent of dysplasias, including conjunctival intraepithelial neoplasia and carcinoma in situ, were HPV-positive. HPV 16 and 18 were most prevalent, although HPV 6 and 11 were also identified. Thirty-one per cent of conjunctival squamous carcinomas were HPV-positive and this was almost exclusively HPV 16 and 18.
HPV disease in pregnancy
The innate immune defences normally activated in pregnancy can contribute to increased clinical severity of some maternal infections. Often women will see deterioration in symptoms during pregnancy. The warts may become more prolific and more friable, and much harder to treat. The treatment of genital warts in pregnancy is also a challenge because the treatment options are more limited (Box 1). Transmission of HPV to the neonate is a rare but associated risk. A small number of neonates with HPV infection at delivery will go on to have recurrent respiratory papillomatosis (RRP), which can lead to airway obstruction, repeated surgery and, rarely, malignancy. Quantifying the risk of RRP is very difficult because the condition is so uncommon. A Danish study found over a 20-year period the overall rate of RRP was 3.5 per 1 000 000 person-years. It also reported that seven of every 1000 births with a maternal history of genital warts resulted in RRP. Caesarean delivery has so far not been shown to reduce the risk and should only be performed for genital warts if the birth canal is obstructed or if vaginal delivery would result in excessive bleeding.
Box 1. Management of genital warts in pregnancy
Warts often grow in size and number or may appear for the first time during pregnancy.
The aims of treatment are symptom relief and minimising the number of lesions present at delivery.
Treatment options include laser, diathermy, surgical excision and trichloroacetic acid.
Most topical treatments should be avoided because of possible teratogenic effects.
Only in severe cases should caesarean delivery be considered.
There is an extremely small risk of their child developing recurrent respiratory papillomatosis.
Anogenital intraepithelial neoplasia
The NHS Cervical Screening Programme has guidance (Box 2) for managing cervical intraepithelial neoplasia in pregnancy. The primary aim of management in pregnancy is to identify invasive disease. Cervical carcinomas can present in pregnancy and enquiring about the patient's smear history and a speculum examination of the cervix should be performed routinely following an episode of antepartum haemorrhage. There are currently no guidelines for the management of vulval, vaginal or anal intraepithelial neoplasia in pregnancy.
Box 2. NHS Cervical Screening Programme Guidelines in Pregnancy
If a woman has been called for routine screening during pregnancy it should be deferred.
If a previous test was abnormal and in the interim the woman becomes pregnant, then the test should not be delayed but should be taken in the mid-trimester unless there is a clinical contraindication.
If a pregnant woman requires colposcopy or cytology after treatment (or follow-up of untreated cervical intraepithelial neoplasia [CIN] 1), her assessment may be delayed until after delivery. Unless there is an obstetric contraindication, however, assessment should not be delayed if a first follow-up cytology or colposcopy is required following treatment for cervical glandular intraepithelial neoplasia, or treatment for CIN 2/3 with involved or uncertain margin status.
A woman who meets the criteria for colposcopy should be examined in the colposcopy clinic even if she is pregnant.
If there is a suspicion of invasive disease then a biopsy is indicated. Punch biopsies suggesting only CIN cannot reliably exclude invasion. Therefore, cone, wedge and diathermy loop biopsies are preferred. Beware of the increased risk of haemorrhage.
HPV disease in males/implications for sexual partners
With the introduction of the HPV vaccine, knowledge of the virus and its link with cervical cancer is improving in the general population. The vast majority of people, however, would not be aware of the risk of non-cervical HPV-related disease. It is very important, therefore, for healthcare workers to inform patients about the risks to both the patients and their sexual partners. The uptake of cervical screening is currently only 79% so every opportunity should be taken to check a woman's cervical smear history. Female sexual partners should be strongly encouraged to attend routine cervical screening and both male and female partners should be told to report any unusual symptoms or lesions in the anogenital area. All should also be aware of and report persistent or unusual oropharyngeal symptoms. The ubiquitous nature of the infection should be emphasised, as should the fact that sexually active adults will most likely have been exposed to HPV already.
HPV vaccine discussion
The full impact of the HPV vaccine on non-cervical disease is not yet known, but the uptake of the vaccine and the choice of vaccine are likely to have the greatest significance. There are a few studies that have looked at the incidence of anogenital intraepithelial neoplasia post-vaccination.[44-46] Kjaer and colleagues have shown that after 42 months the vaccine was 100% effective against HPV 6/11/16/18-related high-grade vulval and vaginal lesions. Similar efficacy has been shown for low-grade cervical, vulval and vaginal intraepithelial neoplasia. Smith et al. predicted that the first-generation prophylactic HPV vaccines have the potential to prevent around 45% of all HPV-associated cancers, including approximately 70% of invasive cervical cancers and 25% of non-cervical cancers. This was based on systematic reviews and worldwide cancer statistics produced by the International Agency for Research on Cancer in 2010. One of the problems in predicting the effect of the vaccine is a possible overestimation in disease attributed to HPV, as the presence of HPV DNA in tumour tissue does not necessarily indicate a causal relationship (Box 3). Where polymerase chain reaction detection of HPV DNA is used alone, it is reasonable to assume that there would be false positives and therefore overestimation of the aetiological significance. Smeets et al. demonstrated this by comparing HPV DNA polymerase chain reaction detection alone with more detailed analysis using detection of oncogene expression. More evidence is required to establish the full burden of HPV disease (Table 2).
Table 2. Summary of level of evidence for role of human papillomavirus (HPV) in carcinogenesis
Level of evidence
Adapted from International Association for Research on Cancer Monograph.
bBasaloid and warty tumours; sSquamous carcinoma; vVerrucous carcinoma.
Many reviews on this topic have shown variations in HPV prevalence in different populations. When combining these data, care must be taken in interpreting them, especially if they will influence health policy in a particular country. Recent UK-based studies have shown the importance of population-specific data and have indicated that the potential influence of the HPV vaccine may actually be underestimated in the UK. There is continuing debate about whether boys should be vaccinated as well as girls. The arguments for and against have been summarised by Hibbitts and Cuschieri (Box 4).
Box 3. Establishing causal links between HPV and cancer
Certain criteria must be met to confirm a causative link between HPV and a cancer. The basis for this are Koch postulates published in 1890 (for bacterial disease). These postulates can be adapted for HPV and cancer:
Is HPV infection found in affected patients?
Are viral genes found in cancer cells?
Can the cancer be linked to the presence of an active viral gene product, for example, an oncogene?
Does prevention of the infection stop the cancer, for example, by vaccination?
Box 4. Should we vaccinate boys?
Herd immunity obtained by vaccinating only women is likely to be insufficient to eradicate the targeted HPV types.
Increased protection from non-cervical HPV disease, particularly in men who have sex with men risk groups.
Reduced disease burden on the NHS, particularly if a quadrivalent vaccine is used.
Increased costs of extending the vaccination programme to prevent relatively few non-cervical malignancies.
Targeted screening instead of widespread vaccination may be more appropriate for high-risk groups.
HPV causes disease in many non-cervical sites, the full extent of which has not been fully established. Our traditional anatomical approach to the management of cancers may be replaced by a system related to disease aetiology that will allow improved management strategies.
The management of HPV-associated disease in pregnancy necessitates a modified approach. The treatment as well as the counselling given should be individualised to the patient.
HPV vaccination has been a huge step forward in cancer prevention, but there are many social, economic and ethical questions still to be answered.