Introduction

Human papillomavirus (HPV) infection targets epithelial basal cells; hence, all HPV-associated diseases have an epithelial origin, are spread via skin-to-skin contact, and are not viremic. The HPV types associated with human cancers are found within the alpha genus of papillomavirus infections. The phylogenetically related HPV types are clinically focused within species 9 and species 7. Species 9 includes HPV 16, 31, 33, 35, 52, 58, and 67. Species 7 includes HPV 18, 45, 39, 70, 59, 68, and c85 [1]. Cervical cancer is the most common HPV-associated cancer, making up 5 % of all cancers globally with 100 % HPV attribution [2].

In countries without routine cervical cancer surveillance programs, the incidence of cervical cancer ranges from 50 to 90/100,000 women. Among countries with opportunistic or organized cervical cancer screening, the incidence is 5–8/100,000 women: 5/100,000 in Finland and 8/100,000 in the United States, Canada, and Western Europe [3]. Routine cytology screening has accomplished this decrease in cervical cancer incidence over the past 70 years. Regardless of conventional or liquid cytology methods, analysis of cellular morphology cannot bring the cervical cancer incidence any lower than 2–3/100,000 due to false-negative and false-positive testing [4].

Adenocarcinomas constitute about a quarter of all cervical cancers and occur at younger ages than the squamous cell cervical cancers. Adenocarcinoma is very difficult to detect at screening and work up, often with later stages of cancer present at diagnosis resulting in significantly higher mortality rates [5, 6].

Other HPV-associated cancers with lesser HPV attribution are vulvar (50 %), vaginal (90 %), penile (50 %), anal (80 %), and oropharyngeal (15–80 %) cancers, most of which occur at a younger age and with a better prognosis because of the HPV association [2, 7]. Benign HPV-associated lesions include genital warts and respiratory papillomatosis, both of which are orders of magnitude less common in incidence than abnormal cervical cytology due to HPV infection [8, 9]. While these benign and malignant diseases are germane, public health relevance centers on the fact that 88 % of all HPV-associated diseases are cervical cancer–related and constitute 92 % of all economic costs [10•].

HPV infections occur throughout a woman’s lifetime (Fig. 1) starting in early childhood for some. Infection prevalence peaks between 16 and 25 years of age, drops to a 10 % prevalence until the perimenopausal years, at which time there is an upsurge to near peak prevalences in many populations [1120]. The concordance of HPV types between male and female partners is low, ranging from 2 to 35 % [21], significantly hindering any attempt to identify the transmission source.

Fig. 1
figure 1

HPV prevalence by age. High-risk HPV prevalence exists at all ages, with varying amounts of increase in the perimenopausal age groups [1119] HPV human papilloma virus

The prevalence of seropositivity for HPV 16/18 infection is independent of any current HPV infection and continues to increase throughout the woman’s lifetime [22]. There is no one age at which an entire population is not infected with HPV and no age after which HPV is no longer an infective possibility. Hence, any protection against HPV-associated diseases will have to start very early in life and continue over the expected seven or eight decades of life.

Prevalence of HPV Type by Cervical Status

HPV infection with oncogenic types occurs without disease sequelae in 90 % of infections [23, 24]. The prevalence of oncogenic HPV infection in women with normal cytology is 5–7 %. The five most common oncogenic types are HPV 16, 18, 31, 33, and 45, with HPV 16 and 18 causing half of the infections [25]. About 2.5 % of U.S. women have a low-grade squamous intraepithelial lesion (LSIL) [8]; LSIL has a 60 % HPV attribution to HPV 16, 18, 31, 33, and 45 [26]. HPV 16 and 18 infections constitute 27 % and 9 % of women with LSIL cytology, respectively. The low-risk types of HPV, including types 6 and 11, are detected in about 15 % of LSIL cytologies [27] and never progress to a precancerous lesion.

All high-grade squamous intraepithelial lesions (HSIL) are caused by oncogenic HPV infections. HPV 16, 18, 31, 33, and 45 occur in frequency distributions more closely related to squamous cell cancer than in LSIL cytologies: 45 %, 7 %, 9 %, 7 %, and 2 %, respectively [28]. The frequency distribution for HPV types associated with squamous cell carcinoma of the cervix around the world is HPV 16 (62 %), HPV 18 (8 %), HPV 45 (6 %), HPV 31 (5 %), and HVP 33 (4 %) [19]. HPV 58 and 52 replace HPV 31 and 45 in Asian populations at similar distribution frequencies. Adenocarcinoma of the cervix, on the other hand, while also always associated with HPV infection, is attributed to a different distributional frequency of oncogenic types. HPV 16, 18, and 45 constitute nearly 90 % of all adenocarcinomas in all parts of the world [19].

Natural History of HPV Infections

HPV is an 8000–base pair circular double-stranded DNA virus that is encapsulated by a protein coat made up of L1 and L2 proteins. All HPV infection must occur at the level of the basement membrane; hence, destructive trauma removing parts of the entire stratified epithelium must occur before HPV infection can take place. In response to the wound, basal epithelial cells rapidly spread laterally to repair the wound. Infection into these basal cells is processed by two sets of receptors. The first receptor, a heparin sulfate proteoglycan, is on the basement membrane and binds to the outer L1 protein coat of the HPV virion (Fig. 2) [29]. This binding causes a transformational change in the outer protein coat allowing the L2 proteins to be exposed through the L1 layer. A protease convertase enzyme (eg, furin) cleaves part of the L2 protein exposing the neutralizing L2 epitope, but more importantly exposing a previously unexposed region of the L1 protein. This newly exposed L1 region binds to the cell receptor on the basal epithelial cell causing episomal internalization of the double-stranded DNA into the human epithelial cell nucleus.

Fig. 2
figure 2

HPV infection of epithelial basal cells and antibody function. HPV infection occurs via two separate receptor mediations. High antibody titers prevent interaction with the basement membrane receptor. Low antibody titers can only prevent interaction if there is high avidity [29]. HPV human papilloma virus; HSPGs heparan sulfate proteoglycans (From Day et al. [29], with permission.)

Most HPV infections are not recognized by the immune system for many reasons. During HPV replication, there are no double-stranded RNA intermediates to invoke innate immune responses. The proteins that are synthesized during HPV replication are dependent on the level of epithelial stratification; hence, low concentrations of early nonstructural proteins (eg, E4, E6, and E7) in the nuclei of the parabasal and intermediate cells are not detected by the Langerhans’ or dendritic cells. The entire process of HPV replication occurs within epithelial stratification and does not invoke cytolysis. There are no cytokines or antigens released to invoke innate immunity responses; there is no viremic phase of infection; there is little systemic antigen presentation and, hence, little opportunity for immune memory to be reactivated at a future infection [30, 31].

Natural history studies of HPV infection show that 70 % of HPV infections are no longer detectable after 1 year and 90 % are no longer detectable at 2 years. Only 5 % of infections progress to a cervical intraepithelial neoplasia (CIN) grade 2 or 3 precancerous stage within 3 years. Of those CIN 3 lesions that do not regress, only 20 % progress to invasive carcinoma within 5 years, and only 40 % progress to invasive carcinoma within 30 years (Fig. 3) [23]. At this time, there is no technology to predict which of the CIN 3 lesions will progress and which will just persist.

Fig. 3
figure 3

HPV persistence and progression. Within 3 years, 5 % of HPV infections progress to CIN 2/3; 20 % of CIN 3 progresses to cancer within 5 years; and 40 % of CIN 3 progresses to cancer within 30 years [23, 24] HPV human papilloma virus; CIN cervical intraepithelial neoplasia (From Schiffman and Rodríguez [23] with permission.)

Clinical Value of Prophylactic HPV Vaccines

While cervical cancer prevention is a noble goal, there is insufficient knowledge about duration of protection of the current HPV vaccines to state that there will be either personal or population benefit in direct cancer prevention due to their use. The most effective method for preventing cervical cancer continues to be cervical cytology or HPV testing used in a routine organized system. The only clinical value prophylactic HPV vaccines have at this time is the prevention of type-specific HPV infections, the reduction in numbers of abnormal screening cytologies, the reduction in numbers of colposcopies and biopsies needed for diagnostic follow-up of abnormal cytology screens, and a reduction in the number of excisional treatments necessary to remove the CIN 2/3 lesion detected. Table 1 shows the phase 3 data for both Cervarix (GlaxoSmithKline, London, U.K.) and Gardasil (Merck & Co., Whitehouse Station, N.J.) randomized controlled trials measured at 3.6 years. While there are differences in the level of prevention afforded by each vaccine, both vaccines show a modest reduction in overall abnormal cytology resulting in a reduction of excisional therapies. This benefit is real and will produce quality-of-life gains for women who chose to be vaccinated at a time that is within 3.5 years of their maximal sexual activity. Phase 2 trial data extend the duration of vaccine efficacy to 5 years for Gardasil and 9.4 years for Cervarix [32, 33•].

Table 1 Clinical value of HPV vaccines: protection unrelated to HPV type [51, 52]

Prevention of HPV Type-Specific Persistent Infection and Intraepithelial Neoplasias

Both HPV prophylactic vaccines prevent vaccine-relevant type-specific persistent infections. Both prophylactic vaccines were studied for almost 4 years in the large phase 3 trials used for regulatory approval. Through phase 2 trials, smaller populations were studied for a longer time period. Gardasil prevents HPV 6/11/16/18 persistent infections for 5 years at a 96 % efficacy (95 % CI of 83, 100) [32]. Cervarix prevents HPV 16/18 persistent infections for 9.4 years at a 100 % efficacy (95 % CI of 84, 100) [33•]. Cervarix exhibits intraspecies cross protection showing efficacy against seven of the oncogenic HPV types within species 9 and species 7 of papillomavirus types: HPV 16/18/31/33/45/51/52 (Table 2) [34••].

Table 2 Cervarix efficacy against persistent infection for HPV types in addition to HPV 16/18 [34]

While cervical intraepithelial neoplasia grade 2 (CIN 2) is a recognized histologic misclassification [35], nonetheless, both sets of clinical trials for Cervarix and Gardasil were based on a squamous outcome end point of CIN 2 or CIN 3 related to HPV 16 or 18, as a cancer surrogate. Adenocarcinoma in situ (AIS) was grouped with the precancerous squamous composite end point and denoted as CIN 2+. Table 3 shows that both vaccines offer nearly 100 % efficacies against CIN 2+ caused by HPV 16 or 18, with a substantial efficacy of 93 % for Cervarix against CIN 3+ attributed to any HPV type.

Table 3 HPV Vaccine efficacy for 3.6 years from phase 3 trials [51, 52]

Immunogenicity of the HPV Vaccines

High antibody titers to L1 and low antibody titers with high avidity for L1 can block HPV infection in vitro [29]. Low antibody titers with low avidity may not block HPV infection and antibody avidity for new HPV infections may decrease over time [36].

In vivo, antibody titers were measured by different immunoassays throughout the three phases of the HPV vaccine trials. The pseudovirion neutralization assay (PBNA), developed by the U.S. National Institutes of Health (NIH), is considered the most accurate reflection of the neutralizing ability of the induced antibodies. Proprietary reports of immune responses have used competitive Luminex immunoassay (cLIA [Invitrogen, Grand Island, NY]) in milliMerck units/mL and enzyme-linked immunosorbent assay (ELISA) in ELISA units/mL.

Data from 181 women in the Brazilian cohort of the phase 2 study of Cervarix efficacy show high and sustained antibody titers, measured by PBNA, four times above titers induced by natural infection for both HPV 16 and HPV 18 lasting at least 9.4 years (Figs. 4 and 5) [37]. Cervarix-induced antibody titers for HPV 16 and HPV 18 were detectable in all recipients at the end of the trial with no loss.

Fig. 4
figure 4

Cervarix anti–HPV-16 titers last for at least 9.4 years (n = 181). PBNA neutralizing antibody titers for the Brazilian women continuing in the 9.4-year study stay at least four times above the titer induced by natural infection [37] HPV human papilloma virus; PBNA pseudovirion-based neutralization assay (From Roteli-Martins et al. [37], with permission.)

Fig. 5
figure 5

Cervarix anti–HPV-18 titers last for at least 9.4 years (n = 181). PBNA neutralizing antibody titers for the Brazilian women continuing in the 9.4-year study stay at least four times above the titer induced by natural infection [37] HPV human papilloma virus; PBNA pseudovirion-based neutralization assay (From Roteli-Martins et al. [37], with permission.)

Data from the phase 2 study of 114 vaccinated women document Gardasil-induced antibody titers for anti-HPV 6/11/16/18 for 5 years [32]. Figures 6, 7, 8 and 9 show the Gardasil-induced antibody kinetics. The geometric mean titers for HPV 6/11/18 become equivalent to naturally induced titers within the 5-year time frame. More worrisome for Gardasil long-term protection is the inability to detect any anti–HPV 16 and 18 antibodies a short time after the three-dose vaccination series. Figure 10 shows that 14 % of women lose all detectable anti–HPV 16 antibodies within 8.5 years after three doses of the monovalent HPV 16 prototype Gardasil vaccine [38], while 35 % of women lose all detectable anti–HPV 18 antibodies within 5 years after three doses of the quadrivalent Gardasil vaccine [39].

Fig. 6
figure 6

Gardasil anti–HPV-16 titers over 5 years (n = 78). Gardasil-induced anti–HPV-16 titers remain tenfold higher than titers induced by natural infection for 5 years HPV human papilloma virus; cLIA chemiluminescent immunoassay; GMT geometric mean titers; PPI per-protocol immunogenicity; PCR polymerase chain reaction (From Olsson et al. [39], with permission.)

Fig. 7
figure 7

Gardasil anti–HPV-18 titers over 5 years (n = 82). Gardasil-induced anti–HPV-18 titers over 5 years. Arrow indicates time point at which induced antibody titers are no different from antibody titers induced by natural infection [32] HPV human papilloma virus; cLIA chemiluminescent immunoassay; GMT geometric mean titers; PPI per-protocol immunogenicity; PCR polymerase chain reaction (From Olsson et al. [39], with permission.)

Fig. 8
figure 8

Gardasil induced anti–HPV-6 titers over 5 years. Arrow indicates time point at which induced antibody titers are no different from antibody titers induced by natural infection [32] HPV human papilloma virus; cLIA chemiluminescent immunoassay; GMT geometric mean titers; PPI per-protocol immunogenicity; PCR polymerase chain reaction (From Olsson et al. [39], with permission.)

Fig. 9
figure 9

Gardasil anti–HPV-11 titers over 5 years (n = 83). Gardasil-induced anti-HPV 11 titers over 5 years. Arrow indicates time point at which induced antibody titers are no different from antibody titers induced by natural infection [32] HPV human papilloma virus; cLIA chemiluminescent immunoassay; GMT geometric mean titers; PPI per-protocol immunogenicity; PCR polymerase chain reaction (From Olsson et al. [39], with permission.)

Fig. 10
figure 10

Percentage of females (15–24 years) who lose Gardasil-induced antibody titers over time. Substantial proportions of women no longer have detectable antibody titers to vaccine relevant type-specific oncogenic HPV types after 5 years [38, 39] HPV human papilloma virus

There appears to be a time-dependent relationship with the loss of anti–HPV 18 antibodies titers that could extrapolate to complete loss of prevention against HPV 18 long before there is an opportunity to prevent any HPV 18 cervical cancers. Low antibody titers to HPV 6 and 11, as well as a 10 % loss of seroconversion within 5 years, may lead to a time-limited genital wart protection. This may be seen in Australia, which has reported a genital wart prevalence reduction initially after Gardasil vaccination [40], if the antibody response is insufficient for long-term protection.

Supporting the lack of duration of Gardasil-induced immune response is the lack of detection of any HPV 18 antibodies in 38 % of men within 2 years of vaccination (Fig. 11). All men had initially seroconverted after the three dose vaccination series, but went on to lose detectable antibodies for HPV 6/11/18 in a very short time interval [41]. Specifically, the men lost all detectable antibodies to HPV 6 and HPV 11 one to 3 years earlier than did a similar proportion of women.

Fig. 11
figure 11

Percentage loss of Gardasil-induced antibody titers over time in men compared with women of the same age ranges. a HPV 18: Nearly 40 % of men lose all detectable antibodies to HPV 18 within 2 years of the complete vaccination series. b HPV 11: Nearly 5 % of men lose all detectable HPV 11 antibodies within 2 years of the complete vaccination series. c HPV 6: Nearly 10 % of males lose all detectable antibody titers to HPV 6 within 2 years of the complete vaccination series. HPV human papilloma virus

Vaccine Implementation

Both HPV vaccines exhibit efficacy when used in a three-dose regimen. Regulatory approval by the U.S. Food and Drug Administration (FDA), European Medicines Agency (EMEA), and other international bodies was based on this three-dose data. Clinical practice often does not reflect the rigor of randomized controlled trials, though. Experience in the United States shows that while $1.4 billion was spent in 2008 on Gardasil vaccinations by the Centers for Disease Control and Prevention as well as private insurances, over 60 % of those who received Gardasil received less than three doses or three doses at ineffective timing intervals [42]. This significant wastage could be avoided in the future if prophylactic HPV vaccines requiring less than three doses are implemented.

A Cervarix randomized controlled trial of women 18–26 years old resulted in 384 women receiving only one dose. Vaccine efficacy of one dose after at least 4 years was 100 % against 12-month persistent HPV 16/18 infection [43••]. The study population was HPV 16/18 DNA negative at baseline regardless of serostatus and regardless of entry cytology. A one-dose implementation product would reduce the wastage that occurs when women do not complete the three-dose series in a timely manner. This option has to be a high priority for regulatory approval because the World Health Organization (WHO) and the Pan American Health Organization (PAHO) cannot implement a one-dose program without appropriate regulatory approvals.

Duration of Vaccine Efficacy

Prevention of persistent HPV infection is the mechanism of action of these vaccines. Should enough persistent HPV infection be prevented, then subsequent type-specific disease development may be prevented as well. While 70 % of HPV infections are cleared at a year, and 90 % are cleared by 2 years via natural immune mechanisms, the efficacy of the vaccines must be sufficiently broad to prevent the 5 % of persistent infections from developing into CIN 3, the recognized cancer precursor. Modeling exercises show that no reduction in cervical cancer will occur if the vaccines do not offer at least 15 years of at least 90 % efficacy against at least HPV 16 [44]. If less protection is achieved, women will continue to be infected with HPV, only at a later point in time, going on to develop cervical cancer should she not have access to screening programs.

Conclusions

The prevention of cervical cancer is dependent on many factors: time the screening program has been in place, compliance, willingness, economics, politics, access, implementation, efficacy, and duration of efficacy to name the most prominent. The program of repeated routine cervical screening in industrialized countries has taken 70 years to reach levels of reduced cervical cancer incidence of 5–8/100,000 [45, 46], still with significant pockets of populations not being routinely screened. If HPV vaccination alone is used as a cervical cancer prevention program, the lowest incidence Gardasil could provide would be 14/100,000 and Cervarix, due to its cross protection, would be 9.5/100,000 [47]. These results would be expected after 60 years of 100 % efficacy, 100 % population coverage, and no need for boosters due to lifetime duration of efficacy [48]. Even with the newer Gardasil formulation with five more oncogenic types added, the lowest incidence of cervical cancer it could achieve would be 9.3/100,000 assuming the same 100 % efficacy, 100 % population coverage and lifetime duration after three appropriately timed doses.

At this time, prophylactic HPV vaccination cannot eradicate cervical cancer. Neither can organized screening programs. The combination of early in life vaccination with the current prophylactic HPV vaccines followed by organized screening in the third and subsequent decades of life may, over 60 years, provide an incremental decrease in the incidence of cervical cancer by 1–2/100,000 women. This will only be true if the duration of vaccine efficacy is at least 15 years.

Rather than spend billions in prophylactic vaccination of all women, a wiser strategy may be to spend resources on therapeutic vaccine development that is independent of HPV type for the relatively small number of women who develop CIN 3 lesions. Prophylactic vaccination of all women cannot be a wise investment until we have a one dose regimen that has lifetime efficacy or a minimum number of one dose boosters. Routine screening must continue. Without vaccine efficacy being lifetime in duration, countries without cervical cancer screening programs will benefit most from the development of routine screening programs that range from one to seven times in a lifetime [49], not mandated population vaccination.