Epidemiological data on HPV infection and associated disease have been reviewed in a recently published IARC Monograph 1. Cervical cancer is the second most common and lethal cancer in women worldwide, with estimated 493,000 new cases and 274,000 deaths in 2002 2. More than 80% of cervical cancers and related deaths occur in developing countries, where cervical cancer is the most common cancer, accounting for 15% of female cancers, while in developed countries it accounts for only 3.6% of new cancers 2. Cervical cancer has a relatively low incidence in Europe (less than 15 per 100,000), whereas its incidence is generally over 30 per 100,000 in developing countries 34. The incidence of invasive cervical cancer has a peak at about 20–25 years after the peak age for HPV infection prevalence, and the incidence of cervical intraepithelial neoplasia (CIN) peaks in between 1. According to the AIRTum (Associazione Italiana dei Registri Tumori) database, in Italy, during 1998–2002, cervical cancer represented 1.6% of all newly diagnosed cancers among females, with an average incidence of 9.8 cases per 100,000 females per year, and represented 0.6% of all cancer deaths among females. It has been estimated that every year 3,418 new cervical carcinomas are diagnosed in Italy; as regards mortality, in 2002 there were 370 deaths due to cervical cancer, and 1,756 deaths due to cancer of the uterus not otherwise specified 5. Survival rates in patients with cervical cancer vary in different countries, ranging from 73% at 5 years in US 6 and 63% in Europe 7 to 30.5% in sub-Saharan Africa 8. In Italy, at 5 years from diagnosis of cervical cancer, survival rates range between 78% in Ferrara and 40% in Naples, with an average national survival rate of 65% 5.

Epidemiological and experimental evidence demonstrate that persistent infection with a carcinogenic HPV genotype is a necessary cause of cervical cancer and its precancerous lesions, as suggested by zur Hausen in '70 9. Over 120 different HPV genotypes have been identified so far and at least 40 of these HPV types infect the anogenital mucosa 10. Genital HPVs are designated as high-risk or low-risk types, based on their association with cervical cancer 11121314. It should be recognized that the distinction between high-risk and low-risk HPV genotypes is continuously under revision on the basis of new knowledge in the field; therefore the definition of the oncogenicity of some HPVs could change with time, with obvious implications on the definition of the diagnostic standards for HPV infection, as outlined below.

High-risk HPV DNA is virtually detectable in all cervical cancers 15. Among high-risk or carcinogenic HPVs (currently, HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59) and probably carcinogenic HPVs (HPV types 26, 53, 66, 68, 73, 82) 45, HPV-16 is responsible for 60% of cervical cancer, HPV-18 for 10%, HPV-45 and HPV-31 for 4% each, whereas HPV-33, HPV-52, and HPV-58 altogether account for 2% of cervical cancer 1216. In high-grade squamous intraepithelial lesions (HSIL), the most common HPV types identified are similar to those in cervical cancer, except for the under-representation of HPV-18 and HPV-45 16. In low-grade squamous intraepithelial lesions (LSIL), high-risk HPV DNA can be detected in 30 to 100% of cases in different studies 17. Again, HPV-16 is the most common type, being detected in 26% of cases, followed by HPV-31 (12%), HPV-51 (11%), HPV-53 and HPV-56 (10%), and many other HPV types 17.

A recent meta-analysis investigated the worldwide prevalence and genotype distribution of HPV DNA in cervical samples from women with normal cytology 18. This meta-analysis estimated the overall HPV prevalence in women with normal cytology was 10.4% (95% confidence interval [CI], 10.2–10.7%). The prevalence was relatively high in Africa, Central America and Mexico (about 20 to 30%) and lower in northern America, Europe, and Asia (about 8 to 11%). In all world regions, HPV prevalence was highest in women younger than 35 years of age, decreasing in women of older age 18. In Africa, the Americas, and Europe, a second peak of HPV prevalence was observed in women of older age 18. Overall, the most common HPV types were HPV-16, HPV-18, HPV-31, HPV-58, and HPV-52, while the most common types in southern Europe were HPV-16, HPV-66, HPV-45, HPV-31, and HPV-42 18.

In the Italian population, the prevalence of high-risk HPV in women aged 25–70 years has been reported to be about 9%, and HPV-16 was the most common genotype, being detected in 30–60% of HPV-positive women, followed by HPV-66, HPV-45, HPV-31, and HPV-53 1920. HPV genotyping by restriction fragment length polymorphism analysis and direct sequencing showed that, in HSIL, HPV-16 and HPV-31 were the most commonly detected types, followed by HPV-67, HP-87, HPV-61, and HPV-58, while HPV-16, HPV-31, HPV-66, HPV-53, HPV-52, and HPV-56 were the most frequent types in LSIL 21. By contrast, low-risk HPV-6 and HPV-11 were responsible for about 90% of genital warts 22, whereas the other low-risk HPV types were less frequently associated with anogenital lesions. Data at recruitment from the New Technologies for Cervical Cancer Screening (NTCC) randomized trial, which is ongoing in Italy and includes women attending routine cervical screening programs, showed the rates of high-risk HPV positivity were 13.1% for women aged 25–34 and 5.8% for women aged 35–60, without a second peak of HPV prevalence 23. In younger women aged 18 to 24, the prevalence of high-risk HPV DNA detection is 17.4% and HPV-16 was the most frequent type, followed by HPV-53, HPV-84, HPV-42, HPV-62, HPV-66, and HPV-89 24. A higher prevalence has been reported in a Danish population-based study 25, which showed that about 40% and 45% of women aged 15–19 and 20–24 years, respectively, were infected with high-risk HPV types. HPV-16 was the most common HPV type, followed by HPV-31, HPV-52, and HPV-51 25. Like in the Italian population, the prevalence of infection markedly decreased with age, without significant second prevalence peak 25. The EDITH (Etude de la Distribution des Types d'HPV en France) study investigated the prevalence and distribution of HPV types in France. The results of this study showed that HPV DNA could be detected in 98% of LSIL smears, in 98% of high-grade cervical intraepithelial neoplasia (CIN2/3), and in 97% of invasive cancers 262728. In LSIL, the most prevalent genotypes were HPV-66, HPV-16, HPV-53, HPV-51 and HPV-52 26; in CIN2/3, HPV-16 was by far the most prevalent genotype (62%), followed by HPV-31, HPV-33, HPV-52, HPV-51, and HPV-58 27, while the most prevalent genotypes in invasive cervical cancer were HPV-16 (73%) and HPV-18 (19%), followed by HPV-31 (7%), HPV-33, HPV-68, HPV-45, HPV-52, and HPV-58 (4.1-2.3%) 28. While HPV-16 was the most prevalent type in both squamous cell carcinoma (74%) and adenocarcinoma (64%), HPV-18 was more prevalent in adenocarcinoma (37%) compared to squamous cell carcinoma (16%) 28.

Besides cervical cancer, HPV infection is associated with over 70% of cases of anal carcinoma and squamous oropharyngeal carcinoma, with 50% of vulvar, vaginal, and penile cancer, and, in immunocompromised patients, with about 90% of cutaneous squamous cell cancer 122930. In Italy, high-risk HPV DNA, mainly HPV-16, was detected in 46% of cases of penile carcinoma 31.

Several basic experimental studies have clarified the mechanisms of HPV oncogenesis 2932. In its life cycle, HPV requires infection of epidermal or mucosal epithelial cells of the basal layer, that are still able to proliferate, most probably through wounds or abrasions. In basal layer cells, viral gene expression is largely suppressed, although the limited expression of the early viral genes E1, E2, E5, and especially E6 and E7, results in enhanced proliferation of the infected cells and in their lateral expansion 33. Following entry into the suprabasal layers, expression of the late viral genes E4, L1, and L2 is initiated; the circular viral genome is then replicated and structural proteins form. In the upper layers of the epidermis or mucosa, complete viral particles are assembled and released 33. The HPV E1 and E2 proteins are involved in replication and segregation of episomal viral genome 34; the E2 protein also acts as a transcriptional activator of viral promoters through its ability to bind to specific sequences within the regulatory sequences of HPV genome, known as long control region (LCR) 35. However, in some cases, E2 may also inhibit viral gene transcription: in genital tract-associated high-risk HPV, E2 protein has been demonstrated to suppress E6 and E7 expression probably through steric interference with the binding of positively acting cellular transcription factors in the LCR 36. HPV E4 function is probably to aid the virus in its egress from the cell 37. E5 seems to be important in the early course of infection. It probably stimulates cell growth by forming a complex with the epidermal growth factor receptor or a subunit of the vacuolar ATPase, thus inhibiting the acidification of endosomes and growth factor receptor recycling 3839. However, since the E5 gene is not expressed in most HPV-positive cancers, E5 does not seem to be obligatory in late events of HPV-mediated carcinogenesis 33. A more significant role for malignant transformation can be assigned to the E6 and E7 genes and their respective proteins. They are consistently expressed in malignant tissue, and inhibiting their expression blocks the malignant phenotype of cervical cancer cells 40. They are independently able to immortalize various human cell types in tissue culture, and show synergistic activity when they are expressed together. Interestingly, in high risk HPVs, E6 and E7 protein are translated together from a polycistronic mRNA 33.

The E7 protein encoded by high-risk HPV is a small multifunctional nuclear protein that binds zinc, is phosphorylated by casein kinase II, and shares structural similarities with oncoproteins of other DNA tumor viruses 41. The most significant cellular targets of E7 are the retinoblastoma (Rb) tumor suppressor protein and its functionally related proteins, p107 and p130, especially in their hypophosphorylated active forms, which are able to inhibit cell cycle progression by binding and repressing the activity of E2F cellular transcription factors 42. Binding with high-risk E7 results in pRb proteasomal degradation, but this event is not sufficient to account for E7 transforming functions 43. Indeed, E7 has additional cellular targets, such as the cyclin-dependent kinase inhibitors p27^kip1 and p21^cip1, which are inactivated 444546, and causes genomic instability leading to chromosomal abnormalities and aneuoploidy 47.

The first transforming activity identified for the high-risk HPV E6 proteins was the ability to mediate p53 degradation through the ubiquitin proteolysis pathway 48, a property not possessed by the low-risk HPV E6 proteins. E6 first associates with the cellular ubiquitin ligase protein termed E6-associated protein (E6AP), then the E6-E6AP complex binds to p53, stimulating its ubiquitination and degradation 49. In targeting p53, the high-risk HPV E6 proteins inhibit DNA damage and oncogene-mediated cell death signals. E6 from high-risk HPV also interacts with several other cellular proteins involved in DNA replication, signal transduction (NFX1), activation of telomerase (Myc), cell proliferation (PDZ motif-containing proteins), and inhibition of apoptosis (Bak, CBP/p300) 33.

Invasive cervical cancers and precancerous lesions associated with high-risk HPV infection are characterized by integration of HPV DNA in cellular chromosomes. Viral DNA integration is associated with deletion of large segments of viral genome, but with the presence of intact E6 and E7, and with transcription of sequence downstream from the integrated LCR 50. Disruption of the viral E1 and E2 genes, as well as of downstream viral sequences, may permit higher levels of E6 and E7 transcription, whose RNA are stabilized following fusion with downstream cellular sequences 51. Persistent expression of E6 and E7 appears to be necessary to maintain the malignant phenotype and to favour the occurrence of cellular genetic and epigenetic events, which are important for cancer progression 52.

Screening programs based on cytology testing led to the reduction of the incidence of cervical cancer and cancer-related mortality of about 70–80% in many industrialized countries 1.

The efficacy of conventional cytological screening for cervical cancer has never been investigated in randomised clinical trials, but evidence of its effectiveness derives from observational studies. Diagnostic performance of cervical cytology has been evaluated by several studies in the literature, whose results have been reviewed in recent meta-analyses. These studies have demonstrated that a single cytology test has a sensitivity of about 50% for identifying women with high-grade precancerous lesions or invasive cervical carcinoma (i.e., cervical intraepithelial neoplasia type 2 or higher, CIN2+) 60616263. In particular, a recent meta-analysis on 8 cervical cancer screening studies done in North American and European countries with established cytology-based screening activities, including over 60,000 women who were tested for both HPV DNA and cytology, found that the overall relative sensitivity of cytology for CIN2+ was 53.0% (95% CI, 48.6%–57.4%), by using ASC-US as cut-off for a positive result 62. Moreover, sensitivity of cytology varied considerably among studies, ranging from 18.6% to 76.7% 62. In fact, cytology testing has been shown to have poor inter-laboratory and inter-operator reproducibility 64. The meta-analysis also showed the sensitivity of cytology testing increases with patient's age and is significantly higher in women over the age of 50 than in younger women (79.3% vs. 59.6%, respectively). The overall specificity of cytology for CIN2+ is very high (96.3%; 95% CI, 96.1–96.5%), and this is an important requisite for a screening test, and is even higher in women aged ≥ 35 years than in younger women (97.1% vs. 95.9%, respectively) 62. More recent meta-analyses involving more European/North American studies yielded a pooled sensitivity of cytology at cut-off ASCUS+ for finding CIN2+ of 70% (95% CI: 60–83%). After omission of three German studies, which reported very low sensitivity data, the pooled sensitivity increased to 78% (95% CI: 69–87%) 6365.

Liquid-based cytology has been more recently introduced and is now widely used for primary screening of cervical cancer. It has the advantage of allowing automation, faster reading times, and to be used for adjunctive testing, including HPV testing. It has been suggested to be more sensitive than conventional cytology, but this feature has been rebutted by recent randomized studies, which compared the accuracy of liquid-based cytology with conventional cytology for primary screening of cervical carcinoma 6667. In these studies, liquid-based cytology showed no significant difference in sensitivity to conventional cytology for detection of CIN2+, besides a lower positive predictive value (PPV) due to a higher number of positive results. It showed however the advantage of a reduction of unsatisfactory smears 67.

Currently used molecular tests for the detection of nucleic acids of high-risk HPV in clinical specimens are based on signal-amplification and hybridization methods, such as the Hybrid Capture^® II assay (HC2, Diagene Corp., Gaithsburg, Maryland, USA), or on target amplification, like the polymerase chain reaction (PCR)-based Amplicor^® HPV test (Roche Molecular Diagnostics, Switzerland) or other PCR-based assays. The HC2 test, to date, is the only commercially available HPV DNA detection test that is approved by the FDA for cervical cancer screening in combination with cytology after the age of 30 years. The HC2 test is a nucleic acid hybridization assay with signal amplification for the qualitative detection of HPV DNA of 13 high-risk types in cervical specimens. The HC2 assay cannot identify the specific HPV type, since detection is performed with a combined probe mix, and provides a semi-quantitative estimate of HPV DNA load. PCR-based assays use consensus or general primers, such as the GP5+/6+ PCR system, the Roche Amplicor^® system, the PGMY primer set, and the SPF₁₀ primer set, to amplify a broad spectrum of HPV types. These primers target the L1 gene in a conserved region amongst HPVs. Subsequent to amplification of HPV DNA, detection of high-risk HPV types is performed by hybridization with probes specific for high-risk HPVs. HPV genotyping is performed by type-specific PCR or by amplification of HPV DNA by using consensus or general primers, followed by sequencing, restriction fragment length polymorphism analysis, or reverse hybridization of the amplicon to multiple oligonucleotide probes corresponding to high-risk and low-risk HPVs. Reverse hybridization assays, such as the linear array by Roche Diagnostics, the line probe assay by Innogenetics NV (Belgium), microchips by Biomedlab Co. (Korea) and by other companies, Multiplexed Luminex Assay (Multimetrix, Germany), etc., are able to identify co-infections by different HPV types, at variance with sequencing-based assays, which have high accuracy in the definition of HPV types but poor sensitivity in co-infections.

HPV testing has been introduced in programs for cervical cancer prevention after numerous studies had demonstrated that detection of high-risk HPVs has a higher sensitivity than cytology testing in predicting high-grade cervical precancerous lesions or invasive cervical carcinoma. A meta-analysis on the use of HPV testing to manage women with ASC-US showed that HPV testing has a greater accuracy than repeat cervical cytology 61. In fact, the overall sensitivity and specificity of HPV testing were 84.4% and 72.9%, respectively, vs. 57.6% and 81.8% of Pap-test, by using ASC-US as cut-off for a positive result; by using LSIL as cut-off, Pap test sensitivity was reduced to 45.7%, whereas specificity increased to 89.1% 61. Among the studies included in the meta-analysis, there was the large ALTS (ASCUS Low SIL Triage Study), a controlled clinical trial, which enrolled 3,488 women, randomized into three intervention groups: immediate colposcopy, repeated liquid-based cytology, or HPV testing. This trial, as confirmed also by other similar studies, demonstrated that HPV testing for ASC-US triage had a significantly higher sensitivity and similar specificity for CIN3+ than repeat cytology 6869. Based on these results, both the American Society for Colposcopy and Cervical Pathology (ASCCP) 5556 and the European Guidelines 58 currently recommend HPV testing using validated HPV assays as an acceptable method for managing women over the age of 20 years with ASC-US. According to ASCCP guidelines, women with ASC-US who are HPV DNA negative can be followed up with repeat cytological testing at 12 months, whereas those who are HPV DNA positive should be referred for colposcopic evaluation. If CIN is not identified, repeat HPV testing at 12 months is recommended. However, women with ASC-US who are HPV DNA negative have a very low probability of having CIN2+, lower than the general population with negative cytology and unknown HPV status, who is actually referred to three years. So, a more conservative management strategy than ASCCP recommendations could be appropriate.

HPV testing has not been demonstrated to be a useful management option for women with a cytological result of LSIL. In fact, a meta-analysis of published studies demonstrated HPV triage of LSIL had no significantly higher sensitivity but lower specificity than repeat cytology in the detection of high-grade CIN 70. In particular, the ALTS study found that over 80% of women with LSIL were HPV-positive, whereas the prevalence of CIN2+ was considerably lower 71. Nevertheless, reflex HPV testing may be cost effective in older women with LSIL, due to considerably lower prevalence of HPV infection. In this regard, data from the randomized Italian study New Technologies for Cervical Cancer Screening (NTCC), which compared accuracy of cytology testing vs. HPV DNA testing, specificity of HPV testing for triage of LSIL was higher in women aged over 35 years than in younger women, since the proportion of HPV positive women was much lower in women aged 35–60 years than in those aged 25–34 years (42% vs. 72%) 72. So, HPV triage could be appropriate also for LSIL in women over 35 years, in order to reduce colposcopy burden and to increase the positive predictive value of the colposcopy referral.

Several studies demonstrated that HPV testing, performed at 4–6 months intervals following ablative therapy for high-grade lesions (CIN2–3), has a higher sensitivity and specificity than cytology in detecting residual disease or recurrence 737475. In a meta-analysis of 11 studies evaluating HPV DNA testing in monitoring women after treatment of CIN3, the negative predictive value of HPV testing for recurrent/residual disease was 98%, whereas the negative predictive value of cytology was 93% 76. Overall, the sensitivity of HPV testing for identifying recurrent/persistent CIN reaches 90% by 6 months after treatment and remains high for at least 24 months, at variance with cytological follow-up, which has a sensitivity of approximately 70%. These findings were confirmed in an updated meta-analysis, which showed HPV testing has a significantly higher sensitivity (ratio: 1.27; 95% CI: 1.06–1.51) and not-significantly lower specificity (ratio: 0.94; 95% CI: 0.87–1.01) than follow-up cytology in the prediction of residual/recurrent CIN 65.

HPV testing at 6–12 months for post treatment management of CIN2,3 is recommended by current guidelines of the American Society for Colposcopy and Cervical Pathology 56, whereas European Guidelines, while waiting for further data from clinical trials, recommend double testing with cytology and an HPV test at 6 months post treatment 58.

The potential role of HPV testing as a stand-alone test or in conjunction with cytology in cervical cancer screening has been investigated in several studies [reviewed in 626577]. Unfortunately, these studies were performed with different techniques and sometimes gave conflicting results.

A meta-analysis including 25 non-randomized clinical trials, which compared the performance of cytology vs. HPV test in screening programs for prediction of CIN2+ 63, showed the overall sensitivity of HPV test was 90%, when the HC2 assay was used, and 80.9%, when PCR with consensus primers was employed; overall specificity was higher in PCR-based tests than in HC2, being 94.7% and 86.5%, respectively. The accuracy of cytology testing was lower than HPV testing: in fact, with ASC-US as a cut-off, sensitivity was 72.7% and specificity 91.9%. In women aged over 30 yr, the overall sensitivity of the HC2 assay was 94.8% and specificity 86% 63.

Likewise, the meta-analysis by Cuzick et al. 62, which included only studies from European and North American countries, where screening programs are well implemented, showed the overall sensitivity and specificity of HPV testing were 96.1% and 90.7%, respectively, whereas the overall sensitivity and specificity of cytology testing were 53.0% and 96.3%, respectively. Moreover, at variance with cytology testing, sensitivity of the HPV test did not vary among the different Centres and was very high both in young women and in those over 50 years of age. Among the studies included in this meta-analysis, there was the HART (HPV in Addition to Routine Testing) trial, a multicentric screening protocol which enrolled over 10,000 women aged 30–60 years attending for routine screening. In this study, women with borderline cytology or negative cytology but positive HPV test were randomized to immediate colposcopy or follow-up with repeat HPV test, cytology testing, or colposcopy at 12 months 78. The HART study confirmed HPV testing is more sensitive than borderline or worse cytology (97.1%, 95% CI 91.2–99.1%, vs. 76.6%, 95% CI 65.1–85.1%) but less specific (93.3%, 95% CI 92.7–93.9%, vs. 95.8%, 95% CI 95.4–96.2%) for detecting CIN2+ 78. Comparison of management strategies for borderline cytology and for positive HPV test with negative cytology results at baseline showed that surveillance at 12 months was as effective as immediate colposcopy 78. Interestingly, no CIN2+ were detected in women with a positive HPV test at baseline but negative at follow-up nor in those with an initial negative HPV test and borderline or mild cytology 78. Based on these results, it was suggested that HPV test could be used as a screening test in women over 30 years of age with cytology triage of HPV-positive results. In case of normal or borderline cytology, repeat HPV testing after 12 months was indicated 78.

Randomized longitudinal studies 79, which have been conducted more recently, were not included in the meta-analyses. One of these studies is the NTCC study, an ongoing clinical trial in Italy nested in the routine activity of organized screening programmes, which enrolled approximately 95000 women, who were randomly assigned to conventional cytology or to an experimental arm that followed two phases depending on the period of recruitment 23. During the phase 1 of recruitment, women were randomly assigned to the conventional arm for screening by conventional cytology or to the experimental arm, which used both HPV DNA testing with HC2 and liquid-based cytology. HPV-positive women were managed differently according to age. Among women aged 35–60 years, those who had a cytological abnormality or were HPV-positive were referred to colposcopy. Among women aged 25–34 years, all women with abnormal cytology were referred for colposcopy but those who were HPV-positive with normal cytology were retested after 1 year and referred for colposcopy only if HPV positivity persisted or if cytology became ASC-US+. During the phase 2 of recruitment, women were randomly assigned to conventional cytology with referral for colposcopy if cytology indicated ASC-US+ or to HPV DNA testing alone with referral for colposcopy if the test was positive 23. Results at recruitment showed that the screening method in the experimental arm was more sensitive than conventional cytology in women aged 35–60 years 80. In particular, referral based on HPV testing with HC2 at a high cut-off value of 2 pg/mL showed a higher sensitivity than conventional cytology for detection of CIN2+ (relative sensitivity, 1.41%; 95% CI, 0.98–2.01), but a lower positive predictive value (PPV; relative PPV, 0.75; 95% CI, 0.45–1.27). The use of the 2 pg/mL cut-off led to a higher PPV than the 1 pg/mL cut-off 80. Adding liquid-based cytology improved sensitivity only marginally but increased false positives 80. In women younger than 35 years, HPV testing alone as primary test, with triage of HPV-positive women by cytology, appeared to be the best approach, since the use of liquid-based cytology did not increased the sensitivity for CIN2+ detection and had a significantly lower PPV as compared with conventional cytology. Moreover, addition of liquid-based cytology as a primary screening test to HPV testing had a negligible effect on sensitivity but strongly reduced PPV compared with HPV testing only 81. The use of more restricted referral criteria (i.e., cut-off of 2 pg/mL and positive results at repeat HPV testing) was predicted to increase PPV to that obtained with conventional cytology 81.

Results from the second recruitment phase of the NTCC study showed that, in women aged 35–60 years, HPV testing alone with a cut-off of 2 pg/mL achieved a substantial gain in sensitivity over conventional cytology with only a small reduction in PPV (relative sensitivity, 1.81; 95% CI, 1.20–2.72; relative PPV, 0.99; 95% CI, 0.67–1.46) 23, confirming the results of the first phase of the study 80. In women aged 25–34 years, direct referral to colposcopy after a positive HPV test led to detection of a significantly greater number of CIN2+ and CIN3+ lesions than in the first phase of the study 23. The increased sensitivity of HPV testing for CIN2+ in the second phase of the study as compared with the first was suggested to be related to a high rate of high-grade lesions in young women, which spontaneously regress 23. So, in order to avoid overtreatment, it was suggested that HPV-positive women aged 25–34 years should be referred to colposcopy only if cytology is also abnormal or if infection persists after 1 year 23. In older women, high-grade lesions are less likely to regress 7880, therefore a more aggressive management strategy for HPV-positive results appears to be justified. On the other hand, referral to colposcopy of women with abnormal cytological findings (ASC-US+) but a negative HPV test may led to a high-rate of false positive CIN2+ results and, as a consequence, to overtreatment 82.

Persistence of the additional lesions detected at baseline by HPV testing in older women was also suggested by the recently published results from two longitudinal randomized controlled trials, that included only women at least 30 83 or 32 84 years of age. One of the trials is the Population Based Screening Study Amsterdam (POBASCAM), which aims to assess whether primary HPV DNA testing is more effective than cytological testing in the setting of a regular screening programme. This study enrolled women aged 30–56 years and attending the regular cervical cancer screening programme in the Netherlands. Women were randomly assigned to the intervention group (screening with HPV DNA testing combined with Pap test) or control group (Pap test only, HPV DNA test results blinded). After 5 years, combined cytological and HPV DNA testing were done in both groups and follow-up data for ≥ 6.5 years were available for 17,155 women. Management strategies were more conservative than in the NTCC study. In fact, women with HSIL or worse were immediately referred to colposcopy, irrespective of the HPV result, whereas repeat testing after 6 and 18 months was advised to women with normal cytology but positive HPV test and to those with ASC-US, ASC-H, or LSIL. Women with repeat positive cytology and HPV test results were referred to colposcopy 83. In this study, detection of HPV DNA was done by GP5+/6+ PCR followed by enzyme immunoassay detection of 14 high-risk HPV types with a cocktail of oligonucleotide probes. The other trial was conducted in Sweden and enrolled 12,527 women aged 32 to 38 years, who, like in the POBASCAM study, were randomized to have an HPV test plus Pap test or a Pap test alone 84. In this study, HPV test was based on GP5+/6+ PCR followed by reverse dot blot hybridization to identify 14 high-risk HPVs. Women with a positive HPV test and a normal Pap test were offered a second HPV test at least 1 year later, and those who were found to be infected with the same high-risk HPV type were offered colposcopy with cervical biopsy. To avoid ascertainment bias, a similar number of Pap smears and colposcopy with biopsies were performed in randomly selected women in the control group. All women were followed for a mean of 4.1 years with annual Pap smears and HPV tests, with colposcopy in cases of persistent high-risk HPV infection 84. Both studies showed that a significantly higher number of CIN2+ or CIN3+ lesions were detected at baseline in the intervention group than in the control group, whereas the number of lesions detected in the subsequent follow-up visit, when both HPV test and cytology were done, was significantly lower in the intervention group than in the control group 8384, thus indicating that implementation of HPV DNA testing in cervical cancer screening leads to earlier detection of high-grade lesions, thus allowing longer screening intervals. In the Dutch study, the 5-year cumulative risk of CIN3+ was 0.1% (95% CI, 0.1–0.2%) after a combined negative HPV DNA and cytological result at baseline and 0.8% (95% CI, 0.6–1.0%) after a negative cytology at baseline, but without HPV testing 83. After a negative HPV test at baseline, the 5-year cumulative risk of CIN3+ was estimated as 0.2% (95% CI, 0.1–0.3%) 83. Likewise, in a longitudinal cohort study which enrolled over 20,000 women in Portland who underwent simultaneous screening with a Pap test and HPV testing with HC2, the 5-year cumulative risk of CIN3, was 4.4% for women who were HPV-positive at baseline, but only 0.24% among women with a negative HPV test and 0.16% when both the HPV test and Pap smear were negative 85.

A direct comparison between Pap test and HPV test (by using the HC2 assay) as stand alone tests for cervical cancer screening was evaluated in a recent randomized trial in Canada, involving 10,154 women aged 30–69 years 86. Both tests were performed on all women in a randomly assigned order but, to the aim of the study, only one first test was considered for statistical evaluation. Women with abnormal Pap test results or a positive HPV test at 1 pg/mL cut-off underwent colposcopy and biopsy, as a random sample of women with negative tests. Results at baseline testing showed sensitivity of HPV testing for CIN2 or CIN3 was 94.6% (95% CI, 84.2–100%), significantly higher than sensitivity of Pap test, which was 55.4% (95% CI, 33.6–77.2%). HPV test specificity was 94.1% (95% CI, 93.4–94.8%) and 96.8% for Pap testing (95% CI, 96.3–97.3%) 86.

Recommendations on the use of HPV testing in cervical cancer screening programs (grading of recommendations is reported in Table 2)

• HPV testing is recommended for ASC-US triage since HPV test is more sensitive than repeat cytology for detection of CIN2+ (IA).

• HPV testing is, in general, not a management option in case of LSIL (ID), even if it may be cost effective in older women with LSIL (IB). Local assess is needed to identify a good triage test for women with LSIL (research need).

• HPV testing, alone or in combination with cytology, is recommended for follow-up after treatment of high-grade cervical lesions, since it is more accurate than repeat cytology in diagnosing residual disease or relapse (IA).

• HPV testing, both based on hybridization (HC2) and PCR amplification with detection of high-risk HPVs, has been widely used in women aged >30 years in addition to Pap cytology because of the high NPV of a combined negative test (I). Women with normal Pap test result and negative HPV test have a very low risk to develop cervical cancer and should not perform the subsequent follow-up visit earlier than 5 years (IA). However, combination of HPV testing and Pap cytology is not a cost-effective strategy (I), so, strategies based on HPV testing as a stand alone test should be preferred in women aged >30 years (IA).

• Follow-up with HPV testing or repeat cytology at 12 months is recommended for cytology-negative, HPV-positive women aged ≥ 30 years (IA). Women who are persistently HPV-positive on repeat testing should undergo colposcopy, whereas those who are negative on both tests may be rescreened in 5 years (IA).

• Using the HPV test in screening programs leads to earlier detection of CIN2+ and CIN3+ and reduction of the number of high-grade lesions detected at subsequent follow-up in women aged ≥ 30 years (I), thus allowing longer screening intervals. The extent of long-term protection against high-grade lesions after a negative HPV DNA test has been estimated to be longer than 5 years, but further research is need to establish the most appropriate screening interval (research need).

• HPV testing as a stand alone test is more effective than cytology alone in identification of CIN2+ lesions in women aged ≥ 35 years (I) and might be used as a primary screening test in this age category. First follow-up results from ongoing controlled randomized trials, which compare longitudinal performance of Pap test vs. HPV testing in screening programs, confirm the higher efficacy of HPV screening that Pap screening (I). Local assessment is needed before screening policies based on primary HPV testing can be recommended (research need).

• In women aged < 35 years, HPV testing with immediate referral to colposcopy leads to over diagnosis of CIN2+ as compared with more conservative management strategies (I). Thus, in women younger than 35 years of age with a positive HPV test, cytology triage or repeat HPV testing at 12 months could be appropriate management strategies to avoid over diagnosis and over treatment of CIN2+, but this hypothesis needs confirmation in longitudinal studies (research need).

In women younger than 25 years of age, 20% of high-risk HPV infections persist, whereas the risk of persistent infection is over 50% in women older than 55 years 18. Only a few women with HPV infection will develop cancer. Thus, a single positive HPV test in the absence of clinically significant lesions does not justify treatment 87. Several longitudinal studies demonstrated that persistence of high-risk HPV is necessary for cancer initiation and progression and that the majority of high-risk HPV infections and associated low-grade lesions spontaneously regress within 6–18 months 1888990919293. Screening strategies based on repeat HPV testing at 12 months following detection of a high-risk HPV could be useful to identify women at risk of cancer, who should undergo colposcopy, and to avoid overtreatment of lesions which will regress 9495.

From a practical point of view, persistence can be defined as the detection of the same HPV type (or, with a higher degree of certainty, the same intratypic variant) two or more times over a certain period. There is no consensus as to the length of time that implies persistence, but at least 6 months to 1 year is the time frame that is usually chosen 1. A prospective study of type-specific HPV natural history, with a median follow-up of 5.1 years, showed particularly pronounced persistence of HPV-16 and markedly increased risk of CIN3+ occurrence when HPV-16 persisted, compared with any other HPV type 96. Evaluation of type-specific HPV persistence was used as a management strategy for women with a positive HPV test but normal Pap test in a controlled randomized trial in Sweden, as above reported 84.

Cervical cancer is characterized by overexpression of E6/E7 mRNA of high-risk HPVs. A multiplex nucleic acid sequence based amplification system (NASBA) assay was developed for the identification of E6/E7 mRNA from HPV types 16, 18, 31, 33, and 45 (PreTect™ HPV Proofer kit, NorChip AS, Norway; now also distributed as NucliSENS EasyQ^® HPV by bioMérieux SA, Marcy l'Etoile, France). Analysis of E6/E7 mRNA expression with the PreTect™ HPV Proofer kit in 204 histologically confirmed invasive cervical squamous cell carcinomas from Norwegian women demonstrated the presence of oncogenic transcripts of HPV-16, HPV-18, HPV-31, HPV-33, and HPV-45 in 97% of HPV-positive cases, thus confirming the expression of oncogenic E6/E7 in most invasive cancers and the involvement of the above HPV types in most cases 97. A cross-sectional study was performed in 4,136 women >30 years of age to compare the performance of HPV E6/E7 mRNA detection with detection of HPV-DNA by Gp5+/6+ consensus PCR 98. No significant difference in sensitivity and specificity was observed between the two tests for detection of HPV in high-grade squamous intraepithelial lesion (HSIL). On the contrary, only a small proportion of the HPV DNA-positive women with a normal, ASC-US, or LSIL diagnosis had a detectable E6/E7 mRNA expression 98. A comparative study between HPV testing with Gp5+/6+ consensus PCR and high-risk HPV E6/E7 mRNA testing with the PreTect™ HPV Proofer kit showed the latest test had equal sensitivity and significantly higher specificity for CIN2+ than PCR 99. Moreover, E6/E7 mRNA-positive women had a significantly higher risk to have a CIN2+ within 2 years than negative women 99. Interestingly, quantitative analysis of E6/E7 mRNA in cervical cancers, but not HPV DNA load, correlated with patient survival 100. The APTIMA HPV^® Assay, developed by Gen-Probe Incorporation (San Diego, CA), employs transcription-mediated amplification and target capture to detect HPV E6/E7 mRNA from 14 high-risk HPV types, but it does not differentiate among the 14 high-risk types. Preliminary results obtained with this assay suggest it has higher sensitivity and specificity than HPV testing for identification of CIN2+ lesions 101.

As above reported, higher cut-off values improve the specificity of the HC2 test 808182 and correlate with an increased risk of HPV persistence and development of high-grade cervical lesions 78. A more accurate quantification of viral load can be achieved by using real-time PCR to detect specific HPV types. This method allowed demonstrating that high HPV-16 DNA load is associated with CIN3 and invasive cervical carcinoma or with an increased risk to develop CIN2–3 during follow-up 92102103104105106. Thus, measurement of HPV-16 load by real-time quantitative PCR has been proposed as a marker of risk for CIN progression. Contrasting results have been obtained with other high-risk HPV types, although, generally, a positive relationship between viral load and severity of cervical lesions has been reported 107108109.

Women with HPV-16 and HPV-18 infection have a higher probability to develop cancer than women with infection by other high-risk HPV types 110111112113. Some studies demonstrated that detection of HPV-16 and HPV-18 types represents a risk factor for CIN2–3 111112, but also identification of other high-risk HPV types could be useful to select at risk patients 114115. Investigation of intra-typic variants of HPV, which are defined by a difference in the L1 gene sequence less than 2% of the reference HPV prototype, have been proven to be useful in epidemiological studies. In multiethnic populations, Asiatic-American and other non-European variants of HPV-16 and HPV-18 are associated with a higher risk of persistent infection and development of cervical precancerous lesions and invasive cancer 116117118.

Investigation of high-risk HPV E6 e E7 variants also gave interesting results. The T350G sequence variant of the HPV16 E6 gene is associated with increased viral persistence and oncogenicity 119120. The aminoacid change caused by this mutation could affect E6 degradation activity on p53, besides changing its immunogenicity. Moreover, the L83V mutation of HPV16 E6 increases activation of the MAPK pathway and cooperates with Notch 1 in tumorigenesis 121. HPV-16 sequence variants may also be associated with different oncogenic potential, even in the absence of virus integration, due to derepression of the E6 and E7 promoter 122123. Thus, typing, subtyping, and sequencing of HPV oncogenes and the upstream regulatory region could be useful for epidemiological purposes but also as prognostic markers.

Demonstration that some new HPV types, defined as intermediate-risk, such as HPV-26, 53, 66, 73, and 82, may be associated with the risk of cervical cancer has suggested to extend the number of HPV types included in HPV DNA tests. To assess the clinical impact of such an extended test, the results from two randomized trials which used HPV test for cervical cancer screening and ASC-US triage, respectively, were analyzed 124. In the cancer screening protocol, addition of other high-risk HPV types, besides the 12 most common, did not ameliorate test sensitivity. Likewise, when HPV testing was used for ASC-US triage, test sensitivity was the greatest with 17 genotypes, but specificity was very low 124. Thus, at the moment, increasing the number of high-risk HPV types, besides those already present in the FDA-approved HC2 test, does not seem to be advantageous neither for screening nor for triage of an abnormal Pap smear. On the other hand, reduction of the number of high-risk types included in the HPV test to those most common in invasive cervical cancer (e.g., HPV-16, HPV-31, HPV-33, etc.) could improve HPV test specificity 125.

Integration of HPV-16 and other high-risk HPVs into the genome of infected cells is an important step in tumorigenesis, since it promotes expression of E6/E7 oncogenes. Clinical studies demonstrated HPV DNA integration is associated with a higher risk of treatment failure and with a shorter disease-free interval than cases without integration 126127. Thus, assessment of HPV genome integration, by Southern-blot or quantitative real-time PCR analysis of E2/E6 ratio, could represent a good prognostic marker.

In high-risk HPV infection, E7 binds to and degrades RB which results in substantial up regulation of p16-INK4A synthesis. In spite of the high level of p16-INK4A synthesis, this protein remains functionally inactive, as E7 induces cyclin A and cyclin E expression, thereby functionally bypassing its interference with the cell cycle. Antibodies that are directed against p16-INK4A allow selective staining of high-risk HPV-infected histological sections or of cytological smears, but not other cervical epithelia, suggesting that detection of this marker could provide diagnostic support to distinguish true CIN/dysplasia from immature metaplasia or other non-neoplastic changes of the cervix 128129130131 and to improve interpretation of histology 132. Sensitivity and specificity for CIN2+ of p16-INK4A overexpression testing by immunostaining in cervical cell samples from HPV-DNA-positive women were estimated in a nested sub-study of the NTCC trial 133. This study demonstrated that HPV testing with p16-INK4A triage improves test specificity for CIN2+. In fact, p16-INK4A triage allows maintaining all the gain in sensitivity obtained by HPV testing alone with respect to conventional cytology, but with referral to colposcopy similar to that of conventional cytology 133.

Both antibody-mediated and cell-mediated immune responses are essential for clearance of HPV and HPV-related cervical lesions. The immune response against HPV (and HPV virus-like particles, VLPs) is mainly directed against type-specific immunodominant conformational epitopes, although cross-reactivity has been documented, such as between HPV-6 and HPV-11, HPV-31 and HPV-33, HPV-18 and HPV-45. Serum levels of anti-HPV antibodies are stable, even after virus clearance, so HPV serological assays simply demonstrate previous exposure to the virus. Sensitivity of HPV serology is low (50–60%) while specificity is high (about 90%). The low sensitivity of the test is in part explained by lack of seroconversion in some individuals with documented HPV infection 134. On the other hand, high anti-HPV antibody titre is more common in women with persistent infection than in those who clear infection 134135136. Moreover, seropositivity for different high-risk HPV types is associated with high-grade CIN 137. Unfortunately, no standardized and validated HPV serological assays are available 138139.

Cell-mediated immune response is important for HPV clearance, as suggested by the high incidence of persistent HPV infection in AIDS or transplanted patients. Persistent infection and cancer progression is also associated with abnormal response of CD4+ T cells 140 and CD8+ T cells 141.

Evaluation of vaccine efficacy in phase II-III clinical trials was based on the demonstration of reduction of the incidence of persistent HPV infection (two positive detections at 4–6 months intervals) and on reduction of the incidence of precancerous lesions (CIN2 and CIN3) caused by HPV vaccine types. International standards for treatment of CIN2/3 are ablative therapy. Moreover, in prospective studies, it is not ethical to allow a woman to develop invasive disease in order to demonstrate efficacy of prevention strategies. In this regard, WHO and other international agencies indicated CIN2/3 as clinical endpoint surrogate to demonstrate efficacy of cancer prevention.

Randomised clinical trials demonstrated that three doses vaccination is effective for the prevention of persistent infection and for precancerous lesions caused by HPV genotypes included in the vaccine. Results on efficacy reported in published studies are mainly two types: (i) per protocol analysis, which demonstrates the pure efficacy of the vaccine, obtained by analysis, in the experimentation context, of clinical data of women who did not violate the protocol (woman who received three doses of vaccine or placebo as expected), who resulted negative for all HPV genotypes included in the vaccine both at enrolment or during the administration of the vaccine. (ii) modified intention to treat analysis, where evaluation of efficacy is more similar to the clinical setting in the general population, since it includes all enrolled women, as long as they receive at least the first dose of vaccine or placebo and who resulted negative for all HPV genotypes of the vaccine merely at the time of administration of first vaccine dose.

Data on safety of the quadrivalent HPV VLP vaccine reported a higher rate of side effects in the vaccination group than in the placebo group. Side effects included local reaction at the site of injection (erythema, pain, edema) (altogether, 86.8% in the vaccine group vs. 77.4% in the placebo group, respectively) and systemic side effects, such as fever (13.5% vs. 10.2%, respectively) 149. Also data on safety of the bivalent vaccine reported a higher rate of injection site symptoms and systemic side effects (fatigue, headache, myalgia) in the vaccine group than in the control group 146. No increased incidence of spontaneous abortion, late fetal deaths, or congenital anomalies was observed in pregnancies occurring in women receiving either type of vaccines as compared with controls 146148.

• in ASCUS triage (IA).

• in the follow-up after treatment of CIN2/3 (IA).

• as a screening test for cervical cancer, as an alternative to cytology, in women older than 30 years, with cytology or repeat HPV testing as triage tests (IA).

Efficacy in cervical cancer screening has been proven for both PCR-based and signal amplification-based HPV tests (I).

• Evaluation of HPV persistence may be useful to identify patients at risk for cervical cancer, who should be referred for colposcopy (A).

• HPV typing and subtyping can be used for epidemiological evaluation of HPV distribution and diffusion (B), for identification of persistent infection (B), for identification of HPV-16 and HPV-18, which are associated with a higher risk of precancerous lesions (B) or invasive cervical cancer (B), and for monitoring the efficacy of vaccination (A). But, there is insufficient evidence to define the best protocol to manage women with positive tests (I).

• Detection of high-risk HPV E6/E7 mRNA may be useful to identify women at risk for cervical cancer and with worse prognosis (IIB). Further research is needed to define the best protocol to manage women with a positive test and to evaluate test efficacy in primary screening programs.

• Measurement of HPV-16 DNA load in women with normal cytology may be useful to identify women at risk for cervical cancer (IIB). But, there is insufficient evidence to define the best protocol to manage women with positive tests (I).

• In epidemiological surveys, HPV serological testing and measurement of anti-HPV cellular immune response could be useful to monitor the duration of vaccine protection (B).

• Quadrivalent HPV VLP vaccine is safe and effective in the prevention of cervical and vulvar precancerous lesions, in situ adenocarcinoma, invasive cervical cancer and genital warts associated with vaccine HPV types in women aged 15–26 yrs (I).

• HPV vaccination with the AS04 adjuvanted bivalent vaccine is safe and efficacious for the prevention of incident and persistent HPV-16 e HPV-18 infection and CIN1+ correlated with HPV vaccine types in women aged 15–26 yrs (I)

• HPV vaccine should be administered before genital HPV exposure, since the benefit of vaccination declines in women already positive for HPV DNA (IA).

• In epidemiological surveys, HPV serological testing and HPV genotyping could be useful to monitor vaccine efficacy (research need).

• Cervical cancer screening is recommended both in vaccinated and non-vaccinated women, in agreement with current guidelines for cervical cancer prevention (III).

• Further research is needed to assess the duration of vaccine protection, variations in the prevalence of HPV types in vaccinated women and in the general population, and surveillance on side-effects of vaccination (research need).