Epidemiology of HPV 16 and cervical cancer in Finland and the potential impact of vaccination: mathematical modelling analyses

Abstract

Background: Candidate human papillomavirus (HPV) vaccines have demonstrated almost 90%-100% efficacy in preventing persistent, type-specific HPV infection over 18 mo in clinical trials. If these vaccines go on to demonstrate prevention of precancerous lesions in phase III clinical trials, they will be licensed for public use in the near future. How these vaccines will be used in countries with national cervical cancer screening programmes is an important question.

Methods and findings: We developed a transmission model of HPV 16 infection and progression to cervical cancer and calibrated it to Finnish HPV 16 seroprevalence over time. The model was used to estimate the transmission probability of the virus, to look at the effect of changes in patterns of sexual behaviour and smoking on age-specific trends in cancer incidence, and to explore the impact of HPV 16 vaccination. We estimated a high per-partnership transmission probability of HPV 16, of 0.6. The modelling analyses showed that changes in sexual behaviour and smoking accounted, in part, for the increase seen in cervical cancer incidence in 35- to 39-y-old women from 1990 to 1999. At both low (10% in opportunistic immunisation) and high (90% in a national immunisation programme) coverage of the adolescent population, vaccinating women and men had little benefit over vaccinating women alone. We estimate that vaccinating 90% of young women before sexual debut has the potential to decrease HPV type-specific (e.g., type 16) cervical cancer incidence by 91%. If older women are more likely to have persistent infections and progress to cancer, then vaccination with a duration of protection of less than 15 y could result in an older susceptible cohort and no decrease in cancer incidence. While vaccination has the potential to significantly reduce type-specific cancer incidence, its combination with screening further improves cancer prevention.

Conclusions: HPV vaccination has the potential to significantly decrease HPV type-specific cervical cancer incidence. High vaccine coverage of women alone, sustained over many decades, with a long duration of vaccine-conferred protection, would have the greatest impact on type-specific cancer incidence. This level of coverage could be achieved through national coordinated programmes, with surveillance to detect cancers caused by nonvaccine oncogenic HPV types.

Figure 1. Model Schematic of HPV 16 Natural History in Women and Men

(A) Susceptible women acquire HPV as determined by the force of infection (λ), which is the per susceptible risk of acquiring infection. Asymptomatic HPV infection can progress to LSILs, HSILs, and ICC, although most infections regress spontaneously to an immune state. Ten percent of asymptomatic HPV progresses rapidly to HSIL. Screening and treatment can prevent progression from HSIL to ICC. The model allows for benign hysterectomy at any stage and accounts for loss of detectable antibody over time. (B) Susceptible men acquire HPV as determined by the force of infection (λ), which is the per-susceptible-individual risk of acquiring infection from an infected woman. Infected men recover to an immune state.

Figure 2. Observed versus Predicted Cervical Cancer Incidence

The observed HPV 16 ICC [ 6] incidence is compared to model predictions. Including the reported changes in sexual behaviour and smoking trends among Finnish women allows the model prediction for HPV 16 ICC incidence to capture an increase in cancer incidence after 1991, but it doesn't capture the full magnitude of the change. The changes in sexual behaviour, which were reported in 1992, were implemented in the model in 1985, because they could have occurred before 1992.

Figure 3. The Impact of Varying the Target Population for HPV 16 Vaccination

(A) The effect of routinely vaccinating successive cohorts of men and women compared to vaccinating women alone at low (10%) and high (90%) coverage is shown. Vaccination of the given proportion of adolescents is assumed to occur before sexual debut at age 15 y and there is no screening. At 10% and 90% vaccine coverage vaccinating women and men has a small benefit (4% and 7%, respectively) over vaccinating women alone. Vaccinating 90% of women alone reduced ICC incidence by 91%. Voluntary vaccination among 10% of 15-y-olds and 30% of susceptible 20-y-old women would reduce HPV 16 ICC incidence by 43%. (B) The impact of vaccination at different ages on HPV 16 ICC incidence for vaccination of 90% of women alone is shown. Sexual debut for women is at 16.6 y and 17.7 y for men. Vaccination at birth and at age 15 y generated the greatest reduction in ICC incidence, to 0.6 cases per 100,000 women, with a lag seen for vaccination at birth. Vaccination at age 20 y produced a 63% decrease and at age 25 y, a 41% decrease, in cancer incidence. (C) The impact of varying duration of vaccine efficacy on the incidence of ICC for vaccination of 90% of women alone before sexual debut is illustrated. Because older women are assumed to be more likely to have persistent infections (a precursor to cancer) than younger women, a vaccine with duration of 15 y or less shifts incident infections to older women (who are more likely to progress to cancer) and there is no reduction in the incidence of ICC. Screening can ameliorate the small increase in cancer incidence seen. If women at all ages are likely to have transient infections, then ICC decreases with increasing vaccine duration and vaccine duration of 15 y reduces ICC incidence by 70%. The progression and regression rates according to age are described in Dataset S1 and Protocol S1. Screening parameters are shown in Table S3. (D) The incremental effect of adding vaccination to screening programmes at different screening intervals is shown. Ninety percent of women alone are routinely vaccinated before sexual debut at the age of 15 y, and it is assumed that vaccine efficacy is 100% with lifelong conferred protection against HPV type 16. Screening alone reduces HPV 16 cancer incidence from 7.0 to 2.8 cases per 100,000 women and vaccination added to this strategy can reduce ICC incidence further to 0.2 cases per 100,000 women. Vaccination alone reduces ICC incidence to 0.6 cases per 100,000 women. Changing the screening strategy (doubling time between screening rounds to 10 y) at the same time as vaccine introduction brings ICC incidence to 0.4 cases per 100,000 women.