Alessandro Curto, Katelijne van de Vooren, Maurizio Bonati, Livio Garattini
Department of Public Health. 'Mario Negri' Institute for Pharmacological Research, Italy

We refer to a recent article [1] discussing the economic burden of HPV-related diseases in Italy, limiting the analysis to four specific virus genotypes only. This choice seems to be closely related to the quadrivalent vaccine, the only one erroneously mentioned in the article as used in the Italian HPV immunization program. In fact, both the bivalent (Cervarix@) and the quadrivalent (Gardasil@) vaccines are used in Italy in the campaign against cervical cancer [2]. Most of the 20 Italian regions, which fund and implement vaccination programs on their territory, ran single tenders to exploit competition between the two vaccine manufacturers.
More in detail, we are really concerned about many methodological flaws and “inputs” regarding the Italian setting that could weaken the study results. Here below we have listed the main ones.
• The 12 studies included in the review were not retrievable from the text, and it was only possible to find them in an Appendix. Surprisingly, one third of them concerned the American setting [4,5,6,7] and were related to respiratory recurrent papillomatosis (RRP), around 20% of the non-cervical costs estimated in the study, i.e. those that should make a difference between the two vaccines. The inclusion of non-Italian studies is not in line with the search criteria reported in the text.
• The prevalence rates of the nine HPV-related diseases considered are essential to adequately estimate their economic burden. The authors mainly used old data from a WHO/ICO report that mainly refers to 1998-2002 for seven out of the nine diseases [3]. In addition, the costs of head and neck cancer treatment (the highest excluding cervical cancer) seem largely overestimated referring to the same report [3], which held there were 6,223 incident cases, estimating at 15-20% the range of fraction associated to all the HPV genotypes. However, the authors report almost double the number of incident cases (12,106, see Table 2) and an even larger fraction associated to the four genotypes analysed (26,4%, see Table 3). Accordingly the lifetime cost per incident case would drop steeply from 59.1 million euros to 17.2-23.0 million euros.
• The authors claim in the methods that tariffs "represent the actual direct costs" to the NHS, in order to justify this feature as a criterion for inclusion. But this arguable statement is contradicted in their Conclusions, when they claim that "DRG tariffs do not capture the actual direct costs". They also raise as a limit of the study the exclusion of indirect and direct non-medical costs, which is a nonsense according to the NHS perspective adopted.
• The authors mix up inflation and discount rates in their calculations, although they are two completely different concepts in economic theory. They also adjust national tariffs for inflation even though the tariffs have not been updated since 2006 in Italy, and they have remained the same.
• The three columns reporting the fractions and the total attributed to the four genotypes in Table 4 give inconsistent numbers for most diseases (e.g. rows h, q, r, s, u. v, z, y), particularly for non-cervical conditions. This is important information since the article is focused on the quadrivalent vaccine.
• The formula reported in the Methods seems meaningless, providing no information in practice. The references in Table 1 include studies and tariff lists, so we wonder why the authors, in view of the paucity of secondary data, did not use only tariffs to estimate the costs of all the nine HPV-related diseases. They also report in the Results that “an independent oncologist” validated all the inputs used in the cost analysis of non-cervical malignancies, a critical task which was not reported in the Methods and could have further biased the study.
To conclude, we wonder whether the results of this model focused on four specific virus genotypes do actually provide “plausible estimates on the Italian welfare system”, as claimed by authors.

References
1 Baio G, Capone A, Marcellusi A, Mennini FS, Favato G. (2012) Economic burden of human papillomavirus-related diseases in Italy. PLoS One. 7:e49699. Epub 2012 Nov 21.
2 Garattini L, van de Vooren K, Curto A. (2012) Pricing HPV vaccines, lessons from Italy. Pharmacoeconomics 30:213-217.
3 WHO/ICO Information Centre on HPV and Cervical Cancer (HPV Information Centre). Human Papillomavirus and Related Cancers in Italy. Summary Report 2010. Full text available for download from the WHO’s website: http://www.who.int/hpvcen.... Last accessed on 19/12/2012.
4 Armstrong LR, Derkay CS, Reeves WC (1999) Initial results from the national registry for juvenile-onset recurrent respiratory papillomatosis. RRP Task Force. Arch Otolaryngol Head Neck Surg 125(7): 743–748.
5 Armstrong L, Preston E, Reichert M, Phillips DL, Nisenbaum R, et al. (2000) Incidence and prevalence of recurrent respiratory papillomatosis among children in Atlanta and Seattle. Clinical Infectious Diseases 31: 107–109.
6 B. Bishai D, Kashima H, Shah K (2000) The cost of juvenile-onset recurrent respiratory papillomatosis. Archives of Otolaryngology – Head and Neck Surgery 126: 935–939.
7 Hu D, Goldie S (2008) The economic burden of noncervical human papillomavirus disease in the United States. Am J Obstet Gynecol 198: 500.e1–500.e7e7.