We assessed the cost-effectiveness of including boys vs girls alone in a pre-adolescent vaccination programme against human papillomavirus (HPV) types 16 and 18 in Brazil. Using demographic, epidemiological, and cancer data from Brazil, we developed a dynamic transmission model of HPV infection between males and females. Model-projected reductions in HPV incidence under different vaccination scenarios were applied to a stochastic model of cervical carcinogenesis to project lifetime costs and benefits. We assumed vaccination prevented HPV-16 and -18 infections in individuals not previously infected, and protection was lifelong. Coverage was varied from 0-90% in both genders, and cost per-vaccinated individual was varied from I$25 to 400. At 90% coverage, vaccinating girls alone reduced cancer risk by 63%; including boys at this coverage level provided only 4% further cancer reduction. At a cost per-vaccinated individual of $50, vaccinating girls alone was o$200 per year of life saved (YLS), while including boys ranged from $810 -18 650 per YLS depending on coverage. For all coverage levels, increasing coverage in girls was more effective and less costly than including boys in the vaccination programme. In a resource-constrained setting such as Brazil, our results support that the first priority in reducing cervical cancer mortality should be to vaccinate pre-adolescent girls. British Journal of Cancer (2007) Cervical cancer is the second most common cancer in women worldwide (Parkin and Bray, 2006), with the majority of cases and deaths occurring in low-resource countries where organised screening has not been feasible, for example, nearly 20 000 women in Brazil are predicted to develop cervical cancer over the next year (Ferlay et al, 2004). Vaccines designed to prevent infections with human papillomavirus (HPV)-16 and -18, responsible for roughly 70% of cases, provide an opportunity for primary prevention. Clinical trials of these vaccines have shown a high degree of efficacy at preventing types 16 and 18 associated infection and precancerous changes in women not previously infected with these types (Harper et al, 2006;Ault, 2007; Future II Study Group, 2007;Garland et al, 2007;Paavonen et al, 2007).Reductions in cervical cancer mortality by pre-adolescent HPV vaccination would not be observable for many years. Mathematical models that synthesize the best available data while ensuring consistency with epidemiological observations can project outcomes beyond those reported in clinical trials, can provide insights into cost-effectiveness, and can be modified as new information becomes available . We recently used an empirically calibrated stochastic model of cervical cancer in a cost-effectiveness analysis of pre-adolescent vaccination of Brazilian girls, with specific attention to strategies that include screening . Because HPV is sexually transmitted, vaccination of both sexes is being considered in some settings.To assess the value of including boys in a vaccination programme, a dynamic transmission model is ...