Discussion
A retrospective population-based study of a bivalent HPV vaccine in Scotland showed reduction rates of 89% for CIN3 or worse, 88% for CIN2 or worse, and 79% for CIN1 (12). Younger age at vaccination was associated with increasing vaccine effectiveness: 86% for CIN3 or worse among girls vaccinated at age 12 to 13 years compared with 51% among girls vaccinated at age 17 years. Nationwide data demonstrated that the incidence and mortality of cervical cancer were reduced by more than half in 2014 and 2015, respectively, in Australia, where the national HPV vaccine program (quadrivalent vaccine) began in 2007 for girls aged 12 to 13 years (13). Nationwide registry data regarding cervical cancer in Sweden demonstrated that the age-adjusted incident ratio of vaccinated women (quadrivalent vaccine) compared with unvaccinated women was 0.57, whereas the ratio adjusted for other demographic variables (adjusted incident ratio) was 0.37 (14). Intriguingly, the adjusted incident ratio in women vaccinated before 17 years of age was 0.12, whereas it was 0.47 in women vaccinated between 17 and 30 years of age. This result suggests that younger age at vaccination increases efficacy, as indicated by reduction rates of 88% (< 17 years of age) and 53% (17–30 years of age). The efficacy of HPV vaccination (bivalent vaccine) was also observed in the United Kingdom, where the estimated relative reductions in cervical cancer rates according to age at vaccination were 34% for 16 to 18 years, 62% for 14 to 16 years, and 87% for 12 to 13 years. The corresponding risk reductions of CIN3 according to age at vaccination were 39% for 16 to 18 years of age, 75% for 14 to 16 years of age, and 97% for 12 to 13 years of age.
The present study is the first to show high efficacy of prophylactic HPV vaccination, demonstrating 100% protection against HPV18 and 95% against HPV16 in Japan (Table 1). Similar reductions were found in Scotland (13), and efficacies of 75% and 81% for CIN2+ and CIN3+, respectively, were found in the other Japanese study (18). Most patients in the present study received the HPV vaccine at 12 to 16 years of age (156/198, 79%), suggesting that vaccination at a younger age is important for efficacy.
In the present study, a reduction in abnormal cervical cytology among vaccinated patients was confirmed by a decrease in prevalence during the 9 to 12 years after vaccination (i.e., the J-HERS 2021 period) (Table 2). Population-based reductions of 78%, 62% and 72% were observed for LSIL, LSIL/ASCH+ and HSIL+, respectively. Age-dependent reductions of 87% were also observed for LSIL in 22- to 33-year-old patients, 74% to 88% for LSIL/ASCH+ in 16- to 33-year-old patients, and a reduction of 87% was observed for HSIL+ in 28- to 33-year-old patients (Table 2). The reduction rates of HSIL after vaccination are higher than expected because the prevalences of HPV16- and HPV18-related CIN3 in Japanese women reportedly range from 25% to 45% (5, 6).
In the present study, the reduction rates of HPV16 and HPV18 infection were 95% and 100%, respectively, among vaccinated patients. These findings clearly demonstrate that the bivalent or quadrivalent vaccine is highly protective against HPV16 or HPV18 infection in vaccinated Japanese women aged < 39 years. Population-based reductions of HPV16 and HPV18 infection were also confirmed by comparison between patients in J-HERS 2021 and J-HERS 2011 for 9 to 12 years after implementation of the national HPV vaccination program (Fig. 5). However, significant (for HPV31 and HPV58) and marginal (HPV52) decreases in prevalence were observed in J-HERS 2021, compared with J-HERS 2011. When we evaluated these HPV types according to age, we found that the prevalences of HPV16, HPV18, HPV31 and HPV58 decreased in the 22- to 27-year age group, whereas the prevalence of HPV52 decreased in the 16- to 21-year age group (Fig. 5). The decreases in HPV31 and HPV58 infections may have been related to cross-protection from the bivalent or quadrivalent vaccines because the decreases occurred in the 22- to 27-year age group, which had the highest vaccination rate. However, no explanation was found for reduction of HPV52 in the 16- to 21- year group (Fig. 5). Previous studies have revealed cross-protection from bivalent and quadrivalent vaccines for HPV16-related (HPV31, HPV33) and HPV18-related (HPV45) types in European countries (12, 13) and HPV16-related (HPV31, HPV52) and HPV18-related (HPV45) types in a Japanese study (21). Other studies demonstrated that cross-protection from these vaccines is more likely for HPV16-related types (HPV31) than for HPV18-related types (HPV45) (14–16). Our data may support a decrease in HPV16-related types (HPV31) after HPV vaccination (14–16), but not a decrease in HPV18-related types (HPV45) (12, 13, 21). HPV52 and HPV58 are the most common high-risk HPV types in Japan (5, 6). Therefore, reductions of these types are reasonable and would be welcome among the Japanese population. In the present study, the population-based reduction rate of HPV16 or HPV18 (HPV16/18) in all patients after vaccination was 44%, whereas the rate of cross-protection against HPV31 or HPV58 (HPV31/58) was 42%. Additionally, the reduction rates were 88% for HPV16/18, 62% for HPV31/58 and 67% for HPV16/18/31/58 among 22- to 27-year-old patients; these patients had the highest rate (53%) of HPV vaccination. These results suggest that protection against HPV16/18 and cross-protection against HPV31/58 may both contribute to reductions of high-grade cervical lesions in Japan.
Notably, vaccination had no effect on the prevalences of HPV6 and HPV11 in the present study. This result may have been related to the lower rate of vaccination with the quadrivalent vaccine than with the bivalent vaccine (bivalent: quadrivalent vaccine = 2:1) at the start of the national HPV vaccination program. The lower prevalences of cervical HPV6 (n = 30) and HPV11 (n = 5) infections than HPV16 (n = 82) and HPV18 (n = 33) infections may be another reason, limiting the statistical power. Increasing prevalences of HPV51 and HPV59 were observed among vaccinated patients in the present study; however, these changes were not confirmed in the comparative study between J-HERS 2011 and J-HERS 2021. Furthermore, there was no increase for HPV51 and HPV59 among 22-to 27-year age group who were highly vaccinated (Fig.3). Such inconsistent results do not support the notion of HPV type replacement according to HPV vaccination, contrary to previous studies (14–16). Analyses of larger numbers of patients are needed to obtain more definitive conclusions.