Approximately 15 different human papillomavirus (HPV) types from 5 species (A5, A6, A7, A9, and A11) may cause cervical cancer. They are grouped according to phylogenetic relatedness, which also correlates with carcinogenicity [1]. The majority of HPV types associated with human cancers are from species A7 (HPV-18, -39, -45, -59, and -68) and A9 (HPV-16, -31, -33, -35, -52, and -58). Given their similarity, phylogenetically related HPV types share capsid epitopes that can elicit a cross-reactive immune response, and, although serological cross-reactivity has been demonstrated in epidemiological studies of naturally infected women with and without cervical cancer [2, 3], vaccine-elicited neutralizing antibodies were expected to be largely type specific

The recent development of and demonstration of efficacy for HPV-6, -11, -16, and -18 L1 virus-like particle vaccines against cervical and other lesions associated with vaccine types has generated enormous enthusiasm as a promising method for partial control of cervical and possibly other cancers. The possibility of cross-protection against other HPV types is an extremely important question, because it could increase the fraction of cervical cancers prevented

A recent study showed that the bivalent vaccine against HPV-16 and -18 produced with ASO4 adjuvant (Cervarix) affords partial cross-protection against 6-month persistent infection with HPV-31, -45, and -52 [4], but no data on efficacy against lesions (cervical intraepithelial neoplasia [CIN] 2–3 or adenocarcinoma in situ [AIS]) associated with nonvaccine HPV types were presented. Two articles [5, 6] in this issue of the Journal report analyses of cross-protection by the quadrivalent (HPV-6, -11, -16, and -18) vaccine (Gardasil) against lesions caused by phylogenetically related HPV types

The first analysis [5] presents efficacy results against infection and disease end points associated with 10 HPV types (31, 33, 35, 39, 45, 51, 52, 56, 58, and 59) that, when combined, are detected in ∼20% of invasive cervical cancers around the world [7, 8]. The multicentric, double-blind, randomized clinical trials included 17,622 women aged 16–26 years. The analysis presented is restricted to women who were seronegative for HPV-6, -11, -16, and -18 and DNA negative for the 14 HPV types analyzed (to approximate subjects naive to HPV infection) and who received ⩾1 vaccination (unrestricted susceptible). Separate analyses are presented for >6-month persistent HPV infection, CIN1–3/AIS, and CIN2–3/AIS, using end-of-study data after 3.6 years of follow-up

In the companion article [6], an intention-to-treat (ITT) analysis is presented for women who received ⩾1 dose of vaccine or placebo and returned for follow-up, regardless of detection of HPV infection or HPV-related disease at enrollment

Table 1 summarizes the efficacy estimates presented in both articles, with the statistically significant values indicated by boldface. There was evidence of efficacy against 6-month persistent infection with HPV-31 and against combinations including this HPV type in both the naive and the ITT populations. In addition, in the naive population there was some evidence of protection against HPV-59 in the A7 species

Table 1

Summary of efficacy estimates for individual and grouped nonvaccine human papillomavirus (HPV) types (adapted from Brown et al. [5] and Wheeler et al. [6])

Table 1

Summary of efficacy estimates for individual and grouped nonvaccine human papillomavirus (HPV) types (adapted from Brown et al. [5] and Wheeler et al. [6])

A similar pattern of significant protection against HPV-31 and its combinations was present for CIN1–3/AIS, with stronger effects with increasing severity of the lesions in the naive population but not in the ITT population. As expected, the efficacy against infection and disease was generally higher in the naive population than in the ITT population, because of the inclusion of women with prevalent infection and disease in the latter. Notably, in the ITT population there was no significant protection against CIN2–3/AIS associated with any combination of HPV types, despite a sizable number of cases (84 in the vaccine arm and 107 in the placebo arm for the combination of HPV-31 and -45, the smallest combination reported), indicating very limited potential benefit against lesions associated with nonvaccine HPV types in the context of catch-up vaccination

HPV-31 was the main type driving efficacy in the A9 species, with some contribution of HPV-58. In the A7 species, efficacy was driven by HPV-59; there was no evidence of protection against infection or lesions associated with HPV-45. Despite the fact that HPV-31 was the predominant type driving efficacy against nonvaccine types, when all nonvaccine types excluding HPV-31 were considered, there was evidence of significant protection against CIN1–3/AIS but not against CIN2–3/AIS. Thus, the only significant cross-protection against CIN2–3/AIS associated with nonvaccine HPV types was protection against HPV-31–related lesions, the type most closely related to HPV-16 (83% homology)

Despite the clear advantage of a vaccine affording some degree of cross-protection, several notes of caution need to be considered in interpreting these results. Multiple infections were common, and, in this context, women in the placebo arm, who did not receive the benefit of protection against HPV-16– and HPV-18–related lesions, were referred more frequently for colposcopy because of such lesions, making detection of lesions associated with other types more likely in the placebo than in the vaccine arm. This could have introduced bias that would make efficacy estimates against nonvaccine types appear larger than they really are. This bias could partly explain the stronger protection observed with increasing severity of the lesions, but it should not operate in the analyses restricted to persistent HPV infection detected in swab samples. In fact, the authors present data showing similar efficacy against infection when only swab samples are considered, which is reassuring but does not completely remove possible bias in the efficacy estimates for CIN2–3/AIS. Additional analyses restricted to women without multiple infections could help clarify the magnitude of the cross-protective efficacy

On the other hand, neither of the 2 studies used an according-to-protocol (ATP) analysis, and therefore women could have acquired infections before completing their vaccination schedules. A higher efficacy could be expected in an ATP analysis or in the context of vaccination programs targeting sexually inexperienced women

Given the complexity of evaluating histologically confirmed lesions in the context of multiple infections and the difficulties related to attributing a specific lesion to an HPV type, persistent HPV infection outcomes are more appropriate for evaluating efficacy against nonvaccine types, because they establish unequivocally that infection has been present for some time, in contrast to the incidental presence of an HPV type in a lesion in which multiple types are detected [9]. The importance of virological persistence outcomes is further emphasized by the fact that the protective efficacy against nonvaccine HPV types is of a lower magnitude than that for vaccine types, and the incidence of lesions associated with such types is much lower than that of HPV-16 or -18, indicating that, to use CIN2–3/AIS outcomes, trials even larger than those reported in this issue would be required

When the cross-protection afforded by the quadrivalent vaccine is compared with that reported for the bivalent vaccine [4], the patterns are roughly similar, despite some methodological differences. The exception is HPV-45 (table 2), for which Cervarix but not Gardasil appears to have a clear effect, raising the possibility that the cross-protection of the bivalent vaccine is more pronounced for HPV types of the A7 species, such as HPV-45

Table 2

Comparison of the efficacy of the quadrivalent human papillomavirus (HPV) vaccine (Gardasil) and the bivalent HPV vaccine (Cervarix) against ⩾6-month persistent infection with nonvaccine HPV types

Table 2

Comparison of the efficacy of the quadrivalent human papillomavirus (HPV) vaccine (Gardasil) and the bivalent HPV vaccine (Cervarix) against ⩾6-month persistent infection with nonvaccine HPV types

In the context of strictly type-specific efficacy, a vaccine against HPV-16 and -18 would be expected to prevent only up to 70% of cervical cancers and possibly fewer, considering multiple infections. Furthermore, the impact could be even lower in developing countries, where the predominance of HPV-16 is less marked [8]. In areas where screening has not been implemented, an affordable vaccine could be the only viable intervention for the prevention of cervical cancer, given the relative simplicity of adding a new vaccine to already-functioning vaccination programs, compared with the complexity of implementing screening programs. In this context, cross-protection, even if limited, may help prevent more cancers

In developed countries, where effective screening is available, the contribution of modest cross-protection would be insufficient to consider discontinuation of screening programs in vaccinated cohorts. However, modifications of screening, including the change of the primary screening method to HPV testing, could help reduce the cost of screening, providing at the same time a mechanism for surveillance of the impact of the vaccine [10, 11]

An important aspect of this discussion is that, in vaccinated cohorts, in the absence of broad cross-protection, the role played by nonvaccine HPV types in the development of lesions is still uncertain, with the possibility that nonvaccine HPV types will replace vaccine types as causal agents of cervical precancer and cancer. Although no direct evidence from vaccine trials has been presented to support the existence of this phenomenon, long-term follow-up of participants in clinical trials is essential to rule this out. The article by Wheeler et al. [6] includes an interesting discussion of this possibility

A study conducted among participants in the Gardasil trials demonstrated that the quantity of antibodies required to neutralize a nonvaccine HPV type (HPV-45) is 1 or 2 orders of magnitude greater than that required for the corresponding phylogenetically related vaccine type (HPV-18) [12]. In the context of waning levels of antibodies, this may indicate that cross-protective efficacy might not only be lower but also be of shorter duration than efficacy against vaccine types. Brown et al. [5] and Wheeler et al. [6] did not report levels of neutralizing antibody against nonvaccine HPV types elicited by the quadrivalent vaccine, but close follow-up will be required for detection of nonvaccine type–related breakthrough infection and disease

The cross-protection reported in this issue of the Journal is definitely good news. However, 30% maximum protection against precursor lesions could reduce the burden of cervical cancer an additional 6%, compared with that afforded by the HPV-16/18 vaccine. This could represent ∼30,000 cases worldwide but would still not afford protection against at least 25% of cancers, making the development of safe, effective, and low-cost polyvalent vaccines a major priority for both academia and industry, to ensure success in the fight against cervical cancer

Acknowledgment

I thank Dr. Allan Hildesheim for his useful comments on the manuscript

References

1
Schiffman
M
Herrero
R
De Salle
R
, et al.  . 
The carcinogenicity of human papillomavirus reflects viral evolution
Virology
 , vol. 
337
 (pg. 
76
-
84
)
2
Wang
S
Schiffman
M
Shields
TS
, et al.  . 
Seroprevalence of human papillomavirus-16, -18, -31, and -45 in a population-based cohort of 10000 women in Costa Rica
Br J Cancer
 , vol. 
89
 (pg. 
1248
-
54
)
3
Combita
AL
Bravo
MM
Touze
A
Serologic response to human oncogenic papillomavirus types 16, 18, 31, 33, 39, 38 and 59 virus-like particles in Colombian women with invasive cervial cancer
Int J Cancer
 , vol. 
97
 (pg. 
796
-
803
)
4
Paavonen
J
Jenkins
D
Bosch
FX
, et al.  . 
Efficacy of a prophylactic adjuvanted bivalent L1 virus-like-particle vaccine against infection with human papillomavirus types 16 and 18 in young women: an interim analysis of a phase III double-blind, randomised controlled trial. HPV PATRICIA Study Group
Lancet
 , vol. 
369
 (pg. 
2161
-
70
)
5
Brown
DR
Kjaer
SK
Sigurdson
K
, et al.  . 
The impact of quadrivalent human papillomavirus (HPV; types 6, 11, 16, and 18) L1 virus-like particle vaccine on infection and disease due to oncogenic nonvaccine HPV types in generally HPV-naive women aged 16–26 years
J Infect Dis
 , vol. 
199
 (pg. 
926
-
35
(in this issue)
6
Wheeler
CM
Kjaer
SK
Sigurdson
K
, et al.  . 
The impact of quadrivalent human papillomavirus (HPV; types 6, 11, 16, and 18) L1 virus-like particle vaccine on infection and disease due to oncogenic nonvaccine HPV Types in sexually active women aged 16–26 years
J Infect Dis
 , vol. 
199
 (pg. 
936
-
44
(in this issue)
7
Muñoz
N
Bosch
FX
de San Jose
S
, et al.  . 
Epidemiologic classification of HPV types associated with cervical cancer: implications for prevention. IARC Multicenter Cervical Cancer Study Group
N Engl J Med
 , vol. 
348
 (pg. 
518
-
27
)
8
Smith
JS
Lindsay
L
Hoots
B
, et al.  . 
Human papillomavirus type distribution in invasive cervical cancer and high-grade cervical lesions: a meta-analysis update
Int J Cancer
 , vol. 
121
 (pg. 
621
-
32
)
9
Jenkins
D
A review of cross-protection against oncogenic HPV by an HPV-16/18 AS04-adjuvanted cervical cancer vaccine: importance of virological and clinical endpoints and implications for mass vaccination in cervical cancer prevention
Gynecol Oncol
 , vol. 
110(3
  
Suppl 1):S18–25
10
Schiffman
M
Integration of HPV vaccination, cytology and HPV testing
Cancer
 , vol. 
111
 (pg. 
145
-
53
)
11
Franco
EL
Cuzick
J
Cervical cancer screening following prophylactic HPV vaccination
Vaccine
 , vol. 
26(Suppl
  
1):A16–23
12
Smith
JF
Brownlow
M
Brown
M
, et al.  . 
Antibodies from women immunized with Gardasil cross-neutralize HPV 45 pseudovirions
Hum Vaccin
 , vol. 
3
 (pg. 
109
-
15
)
(See the articles by Brown et al and Wheeler et al, on pages 926–35 and 936–44, respectively.)
Potential conflicts of interest: R.H. is the co–principal investigator of the Costa Rica HPV vaccine trial, which is sponsored and funded by the US National Cancer Institute (NCI; contract N01-CP-11005) with funding support from the National Institutes of Health Office for Research on Women’s Health and is conducted with support from the Ministry of Health of Costa Rica. Vaccine was provided for the trial by GlaxoSmithKline (GSK) Biologicals, under a clinical trials agreement with the NCI. GSK also provided support for aspects of the trial associated with regulatory submission needs of the company under FDA BB-IND 7920. The NCI and Costa Rica investigators are responsible for the design and conduct of the study and the collection, management, analysis, and interpretation of the data and are not funded by vaccine manufacturing companies
Financial support: none reported