Dr. Chesson: Division of STD Prevention, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, Centers for Disease Control and Prevention, Mailstop US12-3, 1600 Clifton Road, Atlanta, GA 30329.

Dr. Markowitz: Division of Viral Diseases, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention, Mailstop H24-5, 1600 Clifton Road, Atlanta, GA 30329.

Author Contributions:

Conception and design: M. Brisson, H.W. Chesson, L.E. Markowitz.

Analysis and interpretation of the data: J.F. Laprise, H.W. Chesson, L.E. Markowitz, M. Drolet, D. Martin,É. Bénard, M. Brisson.

Drafting of the article: J.F. Laprise.

Critical revision for important intellectual content: J.F. Laprise, H.W. Chesson, L.E. Markowitz, M. Drolet, D. Martin,É. Bénard, M. Brisson.

Final approval of the article: J.F. Laprise, H.W. Chesson, L.E. Markowitz, M. Drolet, D. Martin,É. Bénard, M. Brisson. Obtaining of funding: M. Brisson.

Administrative, technical, or logistic support: J.F. Laprise, L.E. Markowitz, M. Drolet.

Collection and assembly of data: J.F. Laprise, H.W. Chesson, L.E. Markowitz, M. Drolet, M. Brisson.

In the United States, the routine age for human papillomavirus (HPV) vaccination is 11 to 12 years, with catch-up vaccination through age 26 years for women and 21 years for men. U.S. vaccination policy on use of the 9-valent HPV vaccine in adult women and men is being reviewed.

To evaluate the added population-level effectiveness and cost-effectiveness of extending the current U.S. HPV vaccination program to women aged 27 to 45 years and men aged 22 to 45 years.

The analysis used HPV-ADVISE (Agent-based Dynamic model for VaccInation and Screening Evaluation), an individual-based transmission dynamic model of HPV infection and associated diseases, calibrated to age-specific U.S. data.

Published data.

Women aged 27 to 45 years and men aged 22 to 45 years in the United States.

100 years.

Health care sector.

9-valent HPV vaccination.

HPV-associated outcomes prevented and cost-effectiveness ratios.

The model predicts that the current U.S. HPV vaccination program will reduce the number of diagnoses of anogenital warts and cervical intraepithelial neoplasia of grade 2 or 3 and cases of cervical cancer and noncervical HPV-associated cancer by 82%, 80%, 59%, and 39%, respectively, over 100 years and is cost saving (vs. no vaccination). In contrast, extending vaccination to women and men aged 45 years is predicted to reduce these outcomes by an additional 0.4, 0.4, 0.2, and 0.2 percentage points, respectively. Vaccinating women and men up to age 30, 40, and 45 years is predicted to cost $830 000, $1 843 000, and $1 471 000, respectively, per quality-adjusted life-year gained (vs. current vaccination).

Results were most sensitive to assumptions about natural immunity and progression rates after infection, historical vaccination coverage, and vaccine efficacy.

Uncertainty about the proportion of HPV-associated disease due to infections after age 26 years and about the level of herd effects from the current HPV vaccination program.

The current HPV vaccination program is predicted to be cost saving. Extending vaccination to older ages is predicted to produce small additional health benefits and result in substantially higher incremental cost-effectiveness ratios than the current recommendation.

Centers for Disease Control and Prevention.

In the United States, human papillomavirus (HPV) vaccine has been recommended for routine vaccination of girls and women since 2006 and boys and men since 2011. The routine age for vaccination is 11 to 12 years, with catch-up vaccination through age 26 for women and 21 years for men (

We examined 4 extended HPV vaccination scenarios (vaccination of women and men up to ages 26, 30, 40, and 45 years) compared with the current recommendation (vaccination of girls and women aged 11 to 26 and boys and men aged 11 to 21 years).

In all the simulations performed, we reproduced historical changes in the U.S. HPV vaccination program as well as vaccination coverage (

For predictions, we used the U.S. version of the HPV-ADVISE (Agent-based Dynamic model for VaccInation and Screening Evaluation) model (

To account for the substantial uncertainty around sexual behavior and natural history of HPV and associated diseases, we identified several parameter sets that simultaneously fit 776 U.S. data target points for sexual behavior, HPV epidemiology, and screening taken from the literature and population-based data sets (for data sources, see the technical appendix at

Before producing the model predictions for the extended-age vaccination scenarios, we conducted model validation. A key driver of the potential for additional benefits of mid-adult vaccination is the number of new HPV infections that occur in these age groups and whether these infections lead to HPV-associated diseases. We therefore examined the cumulative proportion of the age of acquisition of HPV infection that causes cervical cancer (in the absence of vaccination and screening) (

For population-level effectiveness, our main outcome was the number of HPV-associated outcomes averted. For the economic analysis, our main outcome was cost per QALY gained. We performed the economic analysis by using a health care sector perspective (see

Coauthors from the Centers for Disease Control and Prevention (CDC) contributed to designing the study, interpreting the findings, and editing the manuscript.

Under current recommendations and base-case assumptions, HPV vaccination is predicted to prevent 32 million diagnoses of anogenital warts, 13 million diagnosed cases of CIN2/3, 653 000 cervical cancer cases, and 769 000 cases of noncervical HPV-associated cancer over 100 years in the United States (

Our base-case results predict that the current recommended HPV vaccination strategy in the United States is cost saving and would produce substantial QALY gains (

Cost-effectiveness results are most sensitive to natural immunity after infection, rate of progression to cervical lesions, and assumptions about historical vaccination coverage and vaccine efficacy (

Our results suggest that the current vaccination strategy in the United States will substantially reduce HPV-associated diseases and is cost saving, whereas vaccinating mid-adult women and men through age 45 years is predicted to produce small additional reductions in HPV-associated diseases and ICERs above $463 000 per QALY gained in 90% of base-case simulations (median, $1.5 million per QALY gained). The ICERs for mid-adult vaccination are highly sensitive to assumptions about the natural history of HPV, historical vaccination coverage, and vaccine efficacy.

To our knowledge, this is the first published study to examine the cost-effectiveness of vaccinating mid-adult women and men against HPV, taking into account herd effects. However, Kim and colleagues (

Our analysis has 4 main strengths. First, HPV-ADVISE was calibrated to highly stratified U.S. data to represent country-specific sexual behavior, HPV epidemiology, health care resource use, and cervical cancer screening. Second, the model was validated with pre- and postvaccination data not used during the calibration process. Our short-term predictions are consistent with postvaccination data from the United States (

Some limitations and several uncertainties also are related to the analysis. First, the long-term herd effects on mid-adult women and men from vaccinating younger cohorts remain uncertain. If our model overestimates the herd effects of the current program, vaccinating mid-adult women and men might produce greater benefits and lower ICERs than predicted. However, our model reproduces short-term postvaccination herd effects (

In conclusion, our results suggest that the current vaccination program in the United States will substantially reduce HPV-associated diseases and is cost saving, whereas vaccinating mid-adult women and men through age 30, 40, or 45 years is predicted to produce small additional reductions in HPV-associated diseases and to result in substantially higher ICERs than the current program. Future research priorities should include estimating the herd effects produced by the current U.S. HPV vaccination program (that is, the reduction of HPV-associated outcomes in unvaccinated mid-adult women and men) and the percentage of HPV-associated disease burden due to acquisition of infection after age 26 years, and identifying subgroups of women and men in the United States who would benefit most from mid-adult vaccination.

By a contract from the CDC (00HCVGEB-2018-25176), a

The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Current author addresses and author contributions are available at

Percent age of change in incidence under the current U.S. HPV vaccine recommendation and extended vaccination of women and men up to age 45 years for different outcomes.

Base-case assumptions. The term

Cost-effectiveness acceptability curves: percentage of the parameter sets in which the estimated incremental cost per QALY gained by mid-adult vaccination (vs. current vaccination) is equal to or less than a given cost per QALY–gained threshold.

We separated the 50 parameter sets into those with lower probability of natural immunity after clearance in women (≤40% vs. >40%) and faster progression to CIN1/2/3 (e.g., average median time from infection to CIN1, 10 vs. 15 months; average median time from infection to CIN3, 32 vs. 36 months) (22 of 50 parameter sets) and those with higher natural immunity and slower progression (28 of 50 parameter sets). CIN1/2/3 = cervical intraepithelial neoplasia of grade 1, 2, or 3; QALY = quality-adjusted life-year.

ICERs for Mid-adult Vaccination

Vaccination Strategy | Additional Cost, | QALYs Gained, | ICER vs. Current Recommendation, | Cost-Effectiveness Frontier ICER vs. Nondominated Strategy, | ||
---|---|---|---|---|---|---|

Median | 10th-90th Percentile | Median | 10th-90th Percentile | |||

Current recommendation | 0 | 0 | ||||

Women and men aged ≤26 y | 1400 | - | - | 44 000 to - | Dominated | |

Women and men aged ≤30 y | 3700 | 5 | 830 000 | 124 000 to - | 830 000 | 124 000 to - |

Women and men aged ≤40 y | 10 800 | 6 | 1 843 000 | 382 000 to - | Extended dominated | |

Women and men aged ≤45 y | 15 000 | 10 | 1 471 000 | 463 000 to - | 1 746 000 | 415 000 to - |

ICER = incremental cost-effectiveness ratio; QALY = quality-adjusted life-year.

Base-case assumptions: mid-adult vaccine uptake rates, 2.6% and 1.9% for women and men, respectively; 9-valent cost per dose, $205 up to age 18 years and $225 in persons aged 19 to 45 y; vaccine efficacy, 95%; discount rate, 3%; time horizon, 100 y; U.S. population. Model predictions are represented as the median of model predictions generated by the 50 best-fitting parameter sets.

Versus current recommendation (the current recommendation is cost saving vs. no vaccination).

ICERs are estimated as the median over the 50 best-fitting parameters sets, not median costs divided by median QALYs gained.

The 10th and 90th percentiles were generated by using the 50 best-fitting parameter sets.

A dash (–) means that no median gains in QALYs could be measured because of small incremental gains and the population size of the model.

“Women and men aged ≤26 y” is more costly and does not result in measurable median gains in QALYs compared with “Current recommendation.”

“Women and men aged ≤40 y” is a less efficient use of resources than “Women and men aged ≤45 y” (extended dominance), because the ICER for the former exceeds that for the latter. Thus in the frontier analysis, “Women and men aged ≤45 y” is compared with “Women and men aged ≤30 y.”

Sensitivity Analysis: ICERs for Vaccination Through Age 30 Years Versus Current Recommendation

Sensitivity Analysis | ICER, | ||
---|---|---|---|

All Parameter Sets ( | Lower Immunity and Faster Progression ( | Higher Immunity and Slower Progression ( | |

Base case | 830 000 | 404 000 | 2 308 000 |

Vaccination of mid-adult women only | - | 601 000 | - |

High mid-adult coverage | 747 000 | 507 000 | 1 487 000 |

Low historical vaccination coverage | 336 000 | 318 000 | 410 000 |

Stopping mid-adult catch-up after 40 y | 616 000 | 296 000 | 1 697 000 |

Switching to a 2-dose regimen for all ages | 546 000 | 261 000 | 1 492 000 |

Low vaccine efficacy | 481 000 | 366 000 | 835 000 |

Vaccination cost per dose | |||

$176 | 644 000 | 310 000 | 1 775 000 |

$235 | 867 000 | 423 000 | 2 417 000 |

Maximum health care costs | 821 000 | 395 000 | 2 253 000 |

Maximum disease burden | 753 000 | 317 000 | 1 211 000 |

Disutility in cervical cancersurvivors | 499 000 | 276 000 | 658 000 |

Discount rate | |||

1.5% | 479 000 | 307 000 | 699 000 |

0% | 327 000 | 275 000 | 389 000 |

50-ytime horizon | 932 000 | 449 000 | 2 054 000 |

CDC = Centers for Disease Control and Prevention; CIN1/2/3 = cervical intraepithelial neoplasia grades 1, 2, and 3; HPV = human papillomavirus; HPV-ADVISE = Agent-based Dynamic model for VaccInation and Screening Evaluation; ICER = incremental cost-effectiveness ratio; QALY = quality-adjusted life-year.

Discount rate, 3%; time horizon, 100 y. Model predictions are represented as the median estimate generated by the 50 best-fitting parameter sets.

We separated the 50 parameter sets into the parameter sets with lower probability of natural immunity after clearance in women (≤40% vs. >40%, on the basis of Beachler and colleagues [25]) and faster progression to CIN1/2/3 (e.g., average median time from infection to CIN1, 10 vs. 15 mo; average median time from infection to CIN3, 32 vs. 36 mo) (22 of 50 parameter sets) and those with higher natural immunity and slower progression (28 of 50 parameter sets).

Base case: mid-adult vaccine uptake rates, 2.6% and 1.9% for women and men, respectively; 9-valent cost per dose, $225 in persons aged 19 to 45 y; vaccine efficacy, 95%.

A dash (–) means that no median gains in QALYs could be measured because of small incremental gains and the population size of the model.

75% of base-case vaccination uptake rates for historical coverage.

Vaccine efficacy is assumed to be 85% (vs. base-case vaccine efficacy of 95%) against persistent infections for all HPV types included in the vaccine.

Vaccination costs are based on the CDC vaccine price list as of 1 August 2018 (

Maximum estimates from the U.S. literature (see

We assumed an average 0.24 annual disutility (vs. none in the base case) for life after cervical cancer survival, on the basis of Elbasha and colleagues (