The results of the present analysis indicate that, assuming overall vaccination coverage of 70%, the vaccination of both girls and boys using the quadrivalent HPV vaccine was associated with notable incremental clinical benefits versus a strategy of girls-only vaccination. This is in contrast to other modeling studies, which concluded that the incremental impact of vaccinating boys was limited [18, 19, 21]. According to the present analysis, vaccination of boys and girls led to a 40% reduction in the incidence of HPV-related cancers and 58% reduction in the incidence of genital warts in females versus girls-only vaccination. Similarly, in males the incremental benefits associated with vaccination of boys and girls were a 65% reduction in the incidence of HPV-related carcinoma, including a 66% reduction in the incidence of HPV-related head and neck cancer and a 71% reduction in genital warts. Moreover, the incremental benefit of vaccinating both boys and girls was greatest in instances where the vaccination coverage rates in girls are sub-optimal, a finding that concurs with the findings from other modeling studies [18, 19, 21].
Differences in model structure and assumptions related to the natural history of HPV transmission and development of the disease, as well as differences in clinical outcomes, (used to assess the population-level clinical benefits) make direct comparisons between the outcomes of different models challenging. For example, Brisson et al. present their results in terms of HPV infection. They report the relative reduction in HPV-16/18 prevalence at equilibrium compared with no vaccination and the relative reduction in the incidence of vaccine-type infections over the first 70 years after the start of the vaccination, whereas our analysis reported the relative reduction of HPV-related diseases incidence rather than infections at equilibrium (100 years) . Equilibrium was also assumed to be achieved at different time points across studies (from 50 years in Smith et al. to 100 years in the present study in line with previous work) [17, 18]. Given the different approaches used in the current analysis and that of previously published analyses a detailed structural analysis and comparison of each model would identify the key differences in terms of underlying epidemiology, assumptions used and drivers of results. However, a detailed comparison of different available HPV models is beyond the scope of the present analysis.
Whilst previous modeling studies have focused on the reduction of the incidence of HPV infections and cervical cancer, data relating to the impact on vulvar, vaginal, penile, anal and head and neck cancer have until now been lacking. While girls-only vaccination would substantially reduce the incidence of HPV-related cancer in females and in some extent in males (due to herd immunity), vaccination of boys in addition to girls is associated with a substantial incremental benefit for both males (direct benefits) and females (indirect benefits). Indeed, in our base case, the estimated proportion of the maximum possible vaccine-conferred benefit to males (in terms of male HPV-related carcinomas) from gender-neutral vaccination, which would be achieved by a girls-only vaccination program, was as high as 71%. Nevertheless, this proportion may be reduced to 64% in scenarios assuming a lower vaccine coverage among girls (50%), and even further if a lower coverage rate occurred in conjunction with a waning effect (32 year long duration of protection). Such proportions refer to “proportional benefit achieved” as described by Smith et al., which primarily applied these calculations in terms of HPV-16 infection incidence data. The results suggest that in Europe, vaccination of boys and girls could prevent over 5,500 cases of HPV-related cancer annually (versus girls-only vaccination). Here, we present the number of HPV cancer cases avoided that are specifically due to HPV types 16 and 18. A previous analysis in the UK setting comparing quadrivalent and bivalent HPV vaccines assumed both vaccines provided some cross-protection against carcinomas caused by non-vaccine HPV subtypes. If vaccination does provide some cross protection for non-vaccine HPV types then potential clinical benefits estimated here may be conservative. However, research on cross-protection is currently ongoing and as such it was not included in this analysis .
With regard to the impact on genital warts, vaccination of boys and girls would reduce the incidence of genital warts by approximately 90%, which would likely lead to considerable economic benefits in terms of costs and resource use, indeed the economic burden of genital warts has been reported as being comparable to that of HPV related cancer and it has also been estimated that currently up 10% of visits to sexual health clinics are due to genital warts . Additionally, the current analysis does not capture benefits in terms of quality of life or costs savings, which are also likely to be substantial.
Previous studies have shown that vaccine coverage in girls is a key driver of outcomes in both males and females [18, 19]. Vaccine efficacy as well as compliance were also shown to be of particular interest when assessing the results. In Europe, vaccine coverage varies widely depending on setting due to differences in vaccination policy and modes of implementation (e.g. school based, invitation-based or available on request, and whether a catch-up program is in place for adolescent girls and young women). Countries with school-based vaccination programmes such as the UK have coverage rates of 80–90%, but school based programs only cover 13% of young women living in the EU. In settings where vaccination is administered on demand, such as France and Germany coverage rates are around 50%. In addition to substantial variations in vaccine coverage there are also marked differences across Europe with regard to uptake of cervical screening, which again is influenced by policies implemented on a national level. The proportion of women screened has been found to vary notably according to both age group and setting. For example, in Norway, Sweden and The Netherlands screening rates are high due to organized population-based programs, whereas in many other EU countries cervical screening remains opportunistic (e.g. France, Germany) with unequal access to screening and lower coverage or variation from one region to another (e.g. Spain, Italy) .
The results of the current analysis, together with the results of previous modeling analyses suggest that vaccination of boys and girls would be associated with the greatest benefit in settings where vaccine uptake among girls is low such as those countries that do not have a nationally coordinated vaccination program for females [18, 26]. However, whether it is more feasible/more efficient to implement a strategy of vaccinating both boys and girls or increase vaccine uptake among girls only is an important policy decision that needs to be addressed on a national level given the variety of different vaccine implementation strategies (and hence coverage rates) in place across Europe. Indirect protection (herd immunity) in males is strongly dependent on vaccine coverage in females so the vaccination strategy used and coverage rate achieved is a key factor in determining the incremental benefit of the vaccination of boys at a national level. Additionally, ethical considerations are warranted regarding the type of vaccination program implemented (e.g. a consumer based approach versus a partially or fully subsided voluntary program versus compulsory vaccination).
The analysis presented here is associated with both strengths and limitations. Limitations of the current analysis include the fact that it does not consider the incidence of precancerous states such as cervical, vulvar, vaginal, anal or penile intraepithelial neoplasia, or capture temporal trends in HPV-related disease, such as the increasing incidence of head and neck cancer and anal cancer. Moreover, there is substantial uncertainty in the proportion of head and neck carcinoma attributable to HPV, which may be a contributing factor in the differences in the magnitude of clinical benefit reported across different studies. The present analysis was based on a proportion of 19% of head and neck cancers being attributable to HPV-16/18. This figure might be overestimated even if it is in line with estimates assumed in another recent modeling study by Smith et al. 2011 . Given the magnitude of the burden of the disease of this subset of HPV-related conditions among males in particular, this is an area that potentially warrants further investigation.
Additionally, this analysis does not consider the quality of life benefit associated with the reduction in the incidence of HPV-related disease, which is also likely to be substantial. A further limitation of this analysis is that it is an exploratory analysis that presents mean findings relating to Europe as a whole and also that the model used here and applied to the European setting was based on a US-based dynamic transmission model with input data derived from the US setting, which may potentially limit its applicability to the European setting. In particular, the US-base case scenario (screening alone) is supposed to be consistent with what would be a European base case (screening alone). This is a strong underlying assumption given the specificities of screening implementation in the US and Europe in particular, in addition it is assumed that sexual behavior patterns and the age-structure of the population is similar between settings. Another limitation concerns the structure of the model in that it consists of a number of independent submodels (according to disease type), and incorporates the assumption that only subjects who are at risk of developing the disease can become persistently infected. As such, this means the transmission dynamics for female-only conditions (cervical, vaginal and vulvar cancer) are different from those where both males and females may be affected (head and neck and anal cancer) and from the male only penile cancer submodel. In addition, within Europe there are wide variations in vaccine uptake rates, screening coverage, HPV prevalence and transmission rates, and as shown here, vaccination uptake rates are a key driver of outcomes.
One of the key strengths of the analysis is that it assesses the benefits of male vaccination in all carcinomas that have an established causal link with HPV 6, 11, 16 and 18 whereas many previous analyses have focused primarily on the impact of vaccination in terms of cervical cancer incidence. The current analysis incorporates an extended number of HPV-related disease endpoints including subtypes of head and neck cancer and as such reflect the potential maximum clinical benefits that could be gained from different HPV vaccination scenarios (in comparison with a number of previous analyses that have focused largely on cervical cancer and genital warts only). It is also likely that the potential maximum clinical benefit reported in the current analysis would have been even greater if pre-cancerous states had been included in the analysis. Another strength is that this is the first analysis to present the potential public health impact at EU level of vaccinating boys and girls.
Further country-specific analyses that fully deal with uncertainty are required in order to guide policy decisions relating to the incremental benefits of vaccination of boys and girls. Until such data are available on a country by country basis, a pooled European-wide analysis may provide useful estimates, as well as serving as a valuable comparator for such analyses.