Cost-effectiveness analysis of HPV vaccination of men who have sex with men in Germany

Abstract

Introduction

Men who have sex with men (MSM) have a high risk of human papillomavirus (HPV) infection and HPV-related diseases. While gender neutral HPV vaccination between the ages of 9–14 years (with the option for catch-up between 15- and 17-years-of-age) has been recommended in Germany since 2018, adult MSM are currently not included and thus do not benefit from its advantages. This analysis aims to quantify the reduction in public health and health economic burden of including 18–26-year-old or 18–45-year-old MSM in the national HPV vaccination recommendation, compared to the status quo of vaccinating adolescent boys only.

Methods

We developed a dynamic transmission model of HPV, with an integrated HIV model, to analyze the potential impact of the 9-valent HPV vaccination on HPV infections and HPV-related diseases (anal, penile, and oropharyngeal cancers, and anogenital warts). By including economic outcomes, the model provides estimates of the cost-effectiveness of HPV vaccination among adult MSM in Germany.

Results

Vaccinating MSM aged 18–26 years could prevent an additional 2,583 anal, penile and oropharyngeal cancers, 709 deaths and 81,372 anogenital warts. Expanding vaccination to MSM aged 18–45 years, 4,091 cancers, 1,516 deaths and 114,117 anogenital warts could be averted. The highest reductions were found in anal cancers and anogenital warts; significant incidence reductions in cancers were seen within about 20 years. Vaccinating 18–26 and 18-45-year-old MSM resulted in Incremental Cost-Effectiveness Ratios (ICERs) of 35,300.09€/QALY and 42,088.06€/QALY, respectively, when compared to the vaccination of adolescent boys only.

Conclusions

Vaccination of MSM up to 26 and 45 years of age can profoundly accelerate beneficial public health outcomes while reducing the economic burden of HPV-related cancers and anogenital warts in a cost-effective way compared to vaccinating adolescent boys only.

Supplementary Information

The online version contains supplementary material available at 10.1007/s15010-025-02567-z.

Introduction

Human papillomaviruses (HPV) are the most common form of sexually transmitted pathogens, and nearly 100% of all women and men become infected with HPV at least once in their lifetime [1]. These infections can lead to a broad spectrum of benign, premalignant and malignant mucosal lesions in both men and women, including anogenital warts and various types of cancer [2]. Globally, HPV causes about 85% of anal, 40% of oropharyngeal, and 45% of penile cancers in men [3]. In Germany, approximately 600 new cases of HPV-associated anal carcinoma, at least 250 cases of penile carcinoma, and at least 750 cases of oropharyngeal carcinoma are diagnosed annually in men [3].

HPV is highly prevalent among men who have sex with men (MSM) regardless of Human Immunodeficiency Virus (HIV) status, and studies report substantially higher rates of HPV infections in MSM compared to heterosexual men [4]. The prevalence of HPV among MSM varies globally between 25.1 and 57% [5–7] and is substantially higher in MSM living with HIV [8, 9]. Anal HPV infections could be detected in 13% of MSM who were HIV- and 22% of MSM who were HIV+ [8]. According to studies conducted in Germany, significantly higher HPV rates, 42.4% in HIV- and 91.5% in HIV + MSM, were reported [10]. In Germany, the proportion of MSM within the total population is estimated to be around 2.5–3.4% of adult men [11].

In addition to the high personal burden for those affected by HPV, the health care system also bears significant costs associated with diagnosis, treatment, and prevention of HPV infections in men. HPV vaccination is a crucial component in preventing certain HPV infections and their related conditions [12]. However, adult MSM hardly benefit from HPV vaccination in Germany. While HPV vaccination has been recommended in Germany since 2007 for girls and universally since 2018, and thus few adult males are vaccinated. The current general recommendation by the Standing Committee on Vaccination (STIKO) includes all children and adolescents between the ages of 9 and 14 with the possibility to catch-up missed doses until the age of 17 [13]. HPV vaccination coverage rates (VCRs) in Germany, especially for boys (as they have only been included in the recommendation since 2018), are still at low levels with around 54% for 18 years old girls and around 8% for 18 year old boys in 2020 [14]. Moreover, MSM may not benefit much from herd protection from vaccination of girls. The German S3-guideline for the prevention of HPV associated neoplasia recommends HPV vaccination also for immunocompromised individuals from 18 to 26 years. Further, HPV vaccination may be considered for people living with HIV from 18 to 26 years in the light of their sexual history [15]. Nevertheless, this guideline is not binding for reimbursement purposes as only the national vaccination recommendations by STIKO form the basis for reimbursement decisions.

In other countries, such as Canada [16] and France [17], HPV vaccination recommendations already include MSM as a high-risk population. Recent data from Canada showed that the prevalence of anal infections with the 4-valent HPV vaccine types was 27% lower in young MSM that received at least one dose of HPV vaccination than in unvaccinated MSM [18].

Since VCRs for boys are still at a low level in Germany, it is crucial to additionally protect MSM who are currently, neither directly nor indirectly, protected from HPV infection and related diseases. A randomized, placebo-controlled study of the quadrivalent HPV vaccine in Japanese men aged 16–26 years demonstrated its efficacy against persistent infections and diseases associated with vaccine types [19]. Furthermore, a recent study evaluated HPV vaccine effectiveness (VE) against anal HPV among men who have sex with men (MSM) aged 18–45 in the United States from 2018 to 2023. The results indicated that VE against anal 4vHPV-type prevalence was notably higher among MSM aged 18–26 who were vaccinated before turning 18. In contrast, lower VE was observed in MSM aged 27–45 who received the vaccine between ages 18 and 26. The study concluded that early vaccination is most effective but remains beneficial when administered in older ages [20]. While the HPV vaccination is most effective in children and adolescents, adults can also benefit from vaccination [13]. This is especially true for high-risk groups, such as MSM. Prophylactic HPV vaccine efficacy has been demonstrated in an MSM clinical trial population and the absolute risk of HPV-related diseases in MSM is high [21]. However, the extent of the benefit of adding adult MSM to the national recommendation is still unclear, highlighting the need for modeling studies [22].

This study investigates the public health and health economic impact of including 18–26-year-old and 18–45-year-old MSM were in the national HPV vaccination recommendation, compared to the current recommendation of vaccinating adolescent boys only. We developed a dynamic transmission model to analyze the potential impact of the 9-valent HPV vaccination on HPV infection and associated disease burden of anal, penile, and oropharyngeal cancers as well as anogenital warts. We also provide forecasts of disease incidence and estimates of cost-effectiveness. With no established threshold for cost-effectiveness in Germany, we will use the 1x Gross Domestic Product (GDP), 2xGDP, and 3xGDP thresholds, based on a 2022 GDP per capita of €46,149.

Methods

General model description and calibration

Our model is a dynamic transmission model that captures the heterogeneity of transmission of HPV infections related to sexual behavior by dividing the population into age- and sexual activity-specific demographic groups. Such demographic models are extensively used throughout the literature and are based on the seminal work of Hethcote [23]. The population of the model consists of MSM, and not the general male population in Germany. However, we will assume that MSM are a representative demographic subgroup of the male population in the following sense: The age distribution, birth rate, and all-cause mortality among MSM is the same as the general male population in Germany (see Supplementary Material A (Table A1)). The MSM population is divided into 59 age groups: 1-year age groups through age 49, 5-year age groups through age 90, and one final age group for 90+. This population is also stratified into four sexual behavior classes: those with 1 or less partners in the last 6 months, those with 2–5 partners in the last 6 months, those with 6–10 partners in the last 6 months, and those with 11 or more partners in the last 6 months.

Once divided into demographic groups, the population is further subdivided into relevant epidemiological states, following the standard Susceptible– Infected– Recovered– Susceptible model format. The force of infection for the susceptible population of a certain age is computed by calculating the expected number of infected partners those individuals would have across the various age groups; what proportion of those partners are infected; and finally, the probability that transmission occurs. In the absence of age-specific sexual partnership data (data that gives the age of both partners in the partnership), the age distribution of partners is inferred by assuming a specific level of age-related mixing in the population [24]. The structure of the infection model is the same across the multiple tumor sites considered in the model (Fig. 1). HIV infection occurs as populations transition from one age group into the next. Thus, a proportion of the model population, as they age, acquire HIV. These acquisition probabilities are estimated directly from HIV incidence data (See Table A2 of the Supplementary Material A) in Germany. Hence, HIV, in this model, is treated as static, and we do not attempt to capture the natural history of HIV and AIDS, nor do we try to reproduce the dynamic history of HIV infection in Germany. Further information on the parameterization of sexual behavior and HIV acquisition are available in Supplemental Material A (Tables A2-A11), and a full listing of the equations is available in Supplemental Material C.

Fig. 1.

Flow diagram of the epidemiological portion of the model

HPV infections are assumed to progress to disease at some rate, which can vary by tumor site. For anal, oropharyngeal, and penile disease, we assume that prior to invasive cancer, there is a pre-cancerous state in which individuals are still infectious. This pre-cancerous state is not necessarily to be equated with a pre-cancerous lesion. In this state (which is what some HPV infections progress to, rather than directly to invasive cancer), the population can either progress to cancer or regress to infection.

We have independently simulated the various types of HPV without accounting for any interactions or co-infection between them. This methodology is consistent with previous modeling efforts in the field, which have similarly treated HPV types in isolation. For the HPV types included in the 9vHPV vaccine—specifically types 31, 33, 45, 52, and 58—we have modeled them as an “average” type, which we denote as HPV31+. This involves averaging the cancer attribution data across these types to establish the model targets. This approach eases some of the computational burden by simplifying the representation of HPV-types that make smaller contributions to the total burden of HPV-related cancers.

Epidemiological parameters in the model with no known, precise value were calibrated using Maximum Likelihood Estimation and cancer incidence data among German males as a target [25]. This cancer data was not stratified by sexual orientation, HIV status, or HPV association and thus required some adjustment. To account for sexual orientation and HIV status, we used Standardized Incidence Ratios (SIR) from the literature, and HPV-attributable fractions for the various diseases. Given the large and computationally intensive nature of this model, calibration approaches that require tens of thousands of model runs, such as Markov Chain Monte Carlo (MCMC) methods, were not feasible. Instead, we employed a maximum likelihood approach to identify a single best fitting set of parameters. This function gives the likelihood of a given set of model parameters based on how closely the modeled outcomes align with the actual data, assuming that the data follows a normal distribution with a mean derived from the modeled outcomes and a variance based on the observed data. To ensure the best-fitting parameters were achieved, the likelihood was maximized using the Nelder-Mead method across a high-dimensional parameter space multiple times. While this approach has its limitations, it remains the most computationally feasible option for a model of this size.

The model outcomes that were matched to the data were: the age specific incidence of anal, penile, and oropharyngeal cancers in the HIV- MSM population; the population level incidence of anal, penile, and oropharyngeal cancers in HIV + MSM; the age specific incidence of genital warts in HIV- MSM, and the overall incidence of genital warts in the HIV + MSM population. To produce the model estimates of these targets for a given set of parameters, the endemic equilibrium of HPV infection and disease was estimated by running the model forward for a long time period until such an equilibrium is reached. Further information on calibration results, including model fits, are available in Supplemental Material A (Figures A1-A14).

Vaccination strategies and analyses

This analysis focuses on the cost-effectiveness of HPV vaccination for adult MSM. We assume the use of a 9-valent HPV vaccine (9vHPV) that has a prophylactic effect against incident HPV infection, and persistent HPV infection associated with nine genotypes (Types 6, 11, 16, 18, 31, 33, 45, 52, and 58) and thus the associated HPV-related diseases. Long-term follow-up studies provide data on 14 and 8 years of protection for the 4-valent and 9vHPV vaccine, respectively [26, 27]. For this model, protection against HPV infection with one of the nine genotypes covered in the 9vHPV is assumed to be lifelong, and waning of vaccine protection was explored in sensitivity analysis. The properties that we assume for the 9vHPV vaccine are given in Table 1.

Table 1.

Vaccine properties used in the base case analysis

Parameter/HPV Type	Site	Value	Assumed duration in this model	Reference
Efficacy against infection
HPV16	Anus	0.762	Lifelong	[21]
HPV18	Anus	1	Lifelong	[21]
HPV31+	Anus	0.762	Lifelong	Assumed, consistent with other published modeling [28].
HPV16	Penis	0.411	Lifelong	[29]
HPV18	Penis	0.621	Lifelong	[29]
HPV31+	Penis	0.621	Lifelong	[29]
HPV16	Oropharynx (Extended)	0.411	Lifelong	Assumed to be comparable with penile efficacies [28, 30].
HPV18	Oropharynx (Extended)	0.621	Lifelong	Assumed to be comparable with penile efficacies [28, 30].
HPV31+	Oropharynx (Extended)	0.621	Lifelong	Assumed to be comparable with penile efficacies [28, 30].
HPV6/11	Anogenital warts	0.49	Lifelong	[29]
Efficacy against persistent infection 1
HPV16	Anus	0.938	Lifelong	[21]
HPV18	Anus	1	Lifelong	[21]
HPV31+	Anus	0.938	Lifelong	Assumption [28]
HPV16	Penis	0.787	Lifelong	[29]
HPV18	Penis	0.96	Lifelong	[29]
HPV31+	Penis	0.96	Lifelong	Assumption [28]
HPV16	Oropharynx (Extended)	0.787	Lifelong	Assumed to be equal to penile efficacies [28]
HPV18	Oropharynx (Extended)	0.96	Lifelong	Assumed to be equal to penile efficacies [28]
HPV31+	Oropharynx (Extended)	0.96	Lifelong	Assumed to be equal to penile efficacies [28]
Efficacy against anogenital warts
HPV6/11	Genital Warts	0.843	Lifelong	[29]

¹Persistent infection in the clinical trials is defined as two positive tests for HPV of the same type at least 6 months apart. Efficacy against persistent infection amounts to a shorter duration of breakthrough infections

In the base case, we examine three vaccination scenarios.

• Vaccinating adolescent boys only: First, as a baseline, we assume that only adolescent boys are vaccinated against HPV. This scenario (as well as the other scenarios) refer to the MSM subgroup only, not the total male population. Since we assume it is a representative subgroup, we assume that adolescent MSM are vaccinated at the same rate as the general adolescent male population. The uptake rates for adolescent vaccination are based on a database (IQVIA™ Vaccine Analyzer) containing a nationally representative panel of 460 office-based physicians from 353 practices [31], and the most recent year of data is assumed to persist throughout the time horizon (see Supplementary Material A (Table A12)). Adult MSM do not receive any vaccination in this strategy.

• MSM 18–26 (5%) scenario: The second strategy adds vaccination of MSM up to the age of 26, assuming an initial annual uptake of 5% at age 18 (that is to say, 5% of unvaccinated adult MSM get vaccinated annually) that decreases with age at a rate commensurate with an 80% drop in uptake by age 45 (though vaccination in this strategy ceases at age 26). This strategy also includes vaccination of adolescent boys as described above.

• MSM 18–45 (5%) scenario: The third strategy that we consider is the same as the MSM 18–26 (5%) strategy, but vaccination of adult MSM is extended through age 45. This strategy also includes vaccination of adolescent boys as described in the first scenario.

In all strategies that involve vaccination of adult MSM, we also assume that MSM living with HIV have a 10% higher uptake, due to their increased interactions with the healthcare system and their possibly amplified awareness of their higher risk of acquiring HPV infection. Details on the vaccination coverage rate (VCR) achieved in the model can be found in Supplementary Material B (Figures B1-B3).

When these strategies are implemented and run in the model, we collect incident cases over time for each disease and relevant subgroup, as well as health-economic results. These results include discounted treatment costs (at 3%), vaccination costs, and quality-adjusted life years (QALYs) lost due to disease. The parameters used for these calculations are depicted in Table 2. We also express the number of averted cases of a particular cancer and anogenital warts as a percentage of the total number of disease cases seen in the baseline, vaccinating adolescent boys only scenario. Details on sources and calculations for these costs and QALYs can be found in Supplementary Material A (Tables A13-A15).

Table 2.

Health-economic parameters used in this analysis. Costs for palliative care only incurred by deaths due to cancer

Disease	Parameter	Duration	Value	Source
Cost Parameters 1
Anal Cancer	Treatment Costs	-	€ 9,994.53	[32]
	Cost of Inpatient Palliative Care	-	€ 6,482.87	[32]
Penile Cancer	Treatment Costs	-	€ 7,292.64	[32]
	Cost of Inpatient Palliative Care	-	€ 4,956.05	[32]
Oropharyngeal Cancer	Treatment Costs	-	€ 12,863.33	[32]
	Cost of Inpatient Palliative Care	-	€ 8,344.44	[32]
Anogenital Warts	Treatment Costs	-	€ 430.82	[32]
Utility Parameters
Anal Cancer	Treatment Utility2	1 year	0.735	[28, 30]
	Survivor Utility	4.5 years	0.9	[30, 32]
Penile Cancer	Treatment Utility2	1 year	0.845	[28, 30]
	Survivor Utility	4.5 years	0.9	[30, 32]
Oropharyngeal Cancer	Treatment Utility2	1 year	0.74	[28, 30]
	Survivor Utility	4.5 years	0.9	[30, 32]
Anogenital warts	Treatment Utility	1 year	0.9	[28, 30]

¹All costs were inflated to 2020 Euro. ²Disease disutilities were computed from the sources as a weighted average of disutility by disease severity. These values represent the proportion reduction in the quality of life for individuals in these disease states

In the main analysis, each adult MSM vaccination strategy is compared to the baseline strategy of vaccinating adolescent boys only. We compare both the public health outcomes (averted cases and deaths) as well as the health economic outcomes (costs averted, QALYs gained), producing incremental cost-effectiveness ratios (ICERs). We also consider the number needed to vaccinate (NNV) for the various disease endpoints. Further details on the NNV calculations are shown in Supplementary Material B (Table B1).

Sensitivity analyses

As with any model, it is important to understand how uncertainty in parameter values can impact the health economic conclusions. To that end, we have selected the following parameters for a one-way deterministic sensitivity analysis. The size and computational complexity of the model made a probabilistic sensitivity analysis infeasible. We assessed the impact of the following on the health economic outcomes:

1. Utility associated with diagnosed disease: ±25%, up to 1.

2. Discount rate: 0 − 5%.

3. Cost of treatment of disease ± 20%.

4. Duration of vaccine protection: 20 years.

We also considered the impact of lowering the VCR among adolescent boys; in this case, we investigated the impact of a 25% and 50% reduction in this VCR.

In the base case, the ICERs are computed by comparing the various adult MSM vaccination scenarios to vaccinating adolescent boys only; however, we will also include ICERs comparing the age groups included in vaccination of adult MSM, i.e. health economic analyses will be repeated to show the comparison of MSM 18–45 scenario to the MSM 18–26 scenario. Since there is uncertainty in what uptake of the vaccine will be achieved in adult MSM, we also considered the impact of vaccination scenarios in adult MSM with an initial annual uptake of 10% at age 18 (MSM 18–26 (10%) and MSM 18–45 (10%) scenarios). The calculations of these scenarios are displayed in Supplementary Material B. In the development of the manuscript and supplemental materials, 2022 CHEERS criteria on standards in reporting health economic modeling results were considered [33]. The 2022 CHEERS criteria can be found in Supplementary Material A (Table A16).

Results

Base case

A dynamic transmission model to analyze the potential impact of the 9vHPV vaccination on HPV infection and associated disease burden showed a reduction of a total of 2,583 averted anal, penile, and oropharyngeal cancers, 709 deaths, as well as 81,372 averted anogenital warts in a MSM 18–26 (5%) vaccination scenario in addition to vaccinating adolescent boys only. In a MSM 18–45 (5%) vaccination scenario, the numbers rose to 4,091 averted cancer cases, 1516 averted deaths, and 114,117 averted anogenital warts. Almost a third of the total averted cancer cases came from MSM living with HIV. All results are summarized in Table 3. Moreover, the results for the scenario with 10% uptake are similar, and are shown in Supplementary Material B (Table B2).

Table 3.

The main public health and economic results

Outcome/Disease	MSM Population	Vaccination Strategy
		Vaccinating adolescent boys only (Baseline)	MSM 18–26 (5% Uptake)	MSM 18–45 (5% Uptake)
N Averted Cases (% averted in addition to vaccinating adolescent boys only scenario) 1
Anal Cancer	HIV-	-	1,230 (11.86%)	2,004 (19.33%)
Anal Cancer	HIV+	-	670 (12.51%)	1,114 (20.81%)
Oropharyngeal Cancer	HIV-	-	540 (18.48%)	760 (26.00%)
Oropharyngeal Cancer	HIV+	-	120 (22.86%)	182 (34.67%)
Anogenital Warts	HIV-	-	77,166 (34.21%)	108,283 (48.01%)
Anogenital Warts	HIV+	-	4,206 (38.24%)	5,834 (53.04%)
Penile Cancer	HIV-	-	20 (15.26%)	27 (20.61%)
Penile Cancer	HIV+	-	3 (13.04%)	4 (17.39%)
N Averted Deaths (% averted in addition to vaccinating adolescent boys only scenario) 1
Anal Cancer	HIV-	-	305 (11.05%)	506 (18.32%)
Anal Cancer	HIV+	-	173 (11.77%)	291 (19.81%)
Oropharyngeal Cancer	HIV-	-	185 (15.70%)	271 (22.98%)
Oropharyngeal Cancer	HIV+	-	41 (20.33%)	64 (31.91%)
Penile Cancer	HIV-	-	4 (14.28%)	5 (19.88%)
Penile Cancer	HIV+	-	< 1 (12.49%)	< 1 (18.20%)
NNV per vaccination strategy
Anal Cancer	HIV-	9,562	9,100	8,244
Anal Cancer	HIV+	1,353	1,250	1,007
Oropharyngeal Cancer	HIV-	38,456	35,776	31,388
Oropharyngeal Cancer	HIV+	17,171	13,707	10,030
Anogenital Warts	HIV-	182	167	161
Anogenital Warts	HIV+	8	8	8
Penile Cancer	HIV-	946,514	868,165	745,151
Penile Cancer	HIV+	393,912	356,931	288,870
Cost per Capita 2
Vaccination	Total	€ 62.972340	€ 122.522638	€ 174.561377
Anal Cancer	Total	€ 32.165318	€ 30.502436	€ 29.092110
Oropharyngeal Cancer	Total	€ 9.080931	€ 8.431183	€ 8.029869
Anogenital Warts	Total	€ 36.896731	€ 30.449492	€ 27.484450
Penile Cancer	Total	€ 0.191225	€ 0.179977	€ 0.174467
QALYs lost per Capita 2
Anal Cancer	Total	0.010647	0.010145	0.009732
Oropharyngeal Cancer	Total	0.002854	0.002687	0.002579
Anogenital Warts	Total	8.8E-05	8.37E-05	8.17E-05
Penile Cancer	Total	0.003874	0.003109	0.002742
Cost-Effectiveness3
Incremental Costs	Total	--	€ 50.7791818	€ 98.0357278
Incremental QALYs	Total	--	0.0014385	0.0023293
ICER	Total	--	35,300.09 €/QALY	42,088.06 €/QALY
Cost-Effectiveness (Vaccinating adolescent boys only vs. MSM 18–26 vs. MSM 18–45) 3
Incremental Costs	Total	--	€ 50.7791818	€ 47.256546
Incremental QALYs	Total	--	0.0014385	0.0008908
ICER	Total	--	35,300.09 €/QALY	53,049.56 €/QALY

¹Relative frequency as compared to vaccinating adolescent boys only strategy. ²These values are divided among the MSM population, not the general population of Germany. ³The incremental values here are based on comparison with the vaccinating adolescent boys only strategy

In terms of averted cancer cases, the highest number of averted cases was seen in anal cancers. When vaccinating MSM 18–26 (5%) compared to vaccinating adolescent boys only, up to 1,230 cases and 305 deaths could be averted, and up to 2,004 cases and 506 deaths in the MSM 18–45 (5%) scenario. In addition to vaccinating adolescent boys only, vaccinating adult MSM would correspond to an additional 12% and 19% of averted anal cancer cases by vaccinating MSM 18–26 and MSM 18–45, respectively.

Vaccination of MSM significantly decreases the NNV for the various diseases in the HIV- population, especially in oropharyngeal cancer, reducing the NNV by over 18% from 38,356 to 31,388 when MSM up to age 45 are vaccinated. The reduction in NNV for penile cancer was similar. Approximately 10% reductions were seen in the other diseases in the HIV- population. In the HIV + population, the decrease in NNV for oropharyngeal cancer was nearly 40% (17,171 to 10,030) when MSM up to age 45 were vaccinated. In the same group, the other diseases reached an approximate 25% drop in NNV.

There was a significant benefit realized in the burden of anogenital warts. Compared to vaccinating adolescent boys only, the MSM 18–26 (5%) scenario resulted in 77,000 less anogenital warts, corresponding to an additional 34.2% of averted anogenital warts. For the MSM 18–45 (5%) scenario, 108,000 anogenital warts were averted, corresponding to an additional 48% reduction compared to vaccinating adolescent boys only.

In all scenarios, the QALYs lost due to HPV-related disease were the highest for anal cancer. Correspondingly, the greatest benefit in quality of life was due to the averted anal cancers and deaths. Combining the cost and QALY data, we computed the ICERs comparing the various strategies to the vaccinating adolescent boys only strategy. Vaccinating adult MSM 18–26 (5%) resulted in an ICER of 35,300.09 €/QALY for all investigated cancers and anogenital warts combined. Increasing the MSM vaccination age up to age 45 years also increased the cost per QALY gained over the vaccinating adolescent boys only strategy, with an ICER of 42,088.06 €/QALY. Corresponding results with a 10% uptake are shown in Supplementary Material B (Table B3).

Scenario and sensitivity analyses

We considered several scenarios and sensitivities to understand how these cost-effectiveness results depend on various assumptions. The tornado diagram for these different sensitivity analyses is presented in Fig. 2 below.

Fig. 2.

Tornado diagram for the various sensitivities that were run for the ICER of vaccinating adolescent boys only vs. adult MSM up to age 26 or 45

Regarding health-economic parameters, the model is most sensitive to changes in the disutility associated with HPV-related disease. The discounting rate is the next most influential parameter, but the range of variation is substantially less than what is seen with disease disutility. Assuming waning vaccine protection leads to a decrease in the ICER across the various vaccination strategies. This is explained by the fact that waning protection increases the benefit of herd immunity among the adult MSM resulting from vaccination in those strategies. Uncertainty in disease cost does not have a significant impact on the results.

Additional scenarios that were considered looked at the variation of the assumed uptake achieved among the adult MSM population. Assuming that the base uptake is doubled to 10%, the ICERs increase to 36,841.90 €/QALY for the MSM 18–26 (10%) scenario compared to vaccinating adolescent boys only. Considering the MSM 18–45 (10%) scenario, the ICER increases to 51,554.48 €/QALY.

In the main analysis, both scenarios were compared to vaccinating adolescent boys only. To show a comprehensive overview of cost-effectiveness estimates, ICERs were also computed to compare MSM 18–45 vaccination scenario to MSM 18–26 scenario. This comparison showed comparable results. ICERs were estimated at 53,049.56 €/QALY and 88,852.04 €/QALY for the 5% and the 10% uptake scenario, respectively (Supplementary Material B (Table B4)).

To better understand how the coverage of HPV vaccination in adolescent boys affects the value of vaccination of adult MSM, we considered scenarios where the uptake in adolescent boys decreases from our base-case estimates of uptake. When this coverage is reduced by 25%, the ICER comparing the vaccination of adolescent boys only and the adult MSM 18–26 (5%) scenario decreases to 29,612.89€/QALY; when the coverage is reduced by 50%, this ICER is reduced to 24,398.55€/QALY. Similar drops were seen in the ICER comparing the vaccination of adolescent boys only and adult MSM 18–45 (5%) scenario, where 25% and 50% drop in adolescent VCR resulted in ICERs of 36,858.11€/QALY and 31,723.81€/QALY, respectively. These results are summarized in Fig. 3.

Fig. 3.

ICERs for the various strategies when the vaccine uptake among adolescent boys is varied

Discussion

The dynamic transmission model shows that 9vHPV vaccination of MSM 18–45 would remarkably reduce HPV-related disease burden among MSM in Germany and would likely be cost-effective. Across different scenarios, the results showed that 9vHPV vaccination of MSM up to the age of 45 years would measurably decrease HPV-related anal, penile, and oropharyngeal cancer cases and deaths, and the incidence of anogenital warts among MSM. In the total MSM population, 2,583 to 4091 cases of anal, penile, and oropharyngeal cancers combined were averted along with 709 to 1,516 averted deaths in addition to vaccinating adolescent boys only as the base case, depending on the considered age group. If there were no HPV vaccination among adolescent boys or MSM, there would be 28,034 cases of anal cancer, 5728 cases of oropharyngeal cancer, and 224 cases of penile cancer among the MSM population over the 100-year time horizon (see Supplemental Material B (Table B5)). While the vaccinating adolescent boys only scenario averted an estimated 44% of total anal cancer cases, the MSM 18–26 (5%) and 18–45 (5%) scenarios added an additional reduction by about 12% and 20%, respectively. Of the total 224 penile cancer cases, about 31% were estimated to be reduced by vaccinating adolescent boys only scenario, and additionally 15% (n = 23) and 20% (n = 31) in the MSM 18–26 (5%) scenario and MSM 18–45 (5%) scenario, respectively.

In the model, MSM living with HIV comprised only about 5% of the total MSM population. Nevertheless, the HPV burden was much higher in this population, so HPV vaccination was more beneficial. Around one third (30.70%) of the cancer cases prevented came from this subpopulation, highlighting the significance of HPV vaccination in this subgroup. Further research is needed in this area in order to determine the precise level of disproportionality of disease burden among HIV + MSM. This is because the model is limited in terms of capturing HIV dynamics and natural history. We avoided a full co-infection model in an effort to constrain the size of the model, but this did require major simplifications to HIV dynamics. This is a substantial limitation, but the long time-horizon of the model likely mitigates the impact of our static assumptions, since over this period of 100-years (2022–2122), the dynamics of HIV in a full HIV/HPV coinfection model may be at or near an equilibrium.

Since vaccination of adolescent boys generally captures males before they engage in sexual activity, it is not surprising that a large proportion of the potential benefit among MSM is realized in the baseline scenario. After 100 years, compared to vaccinating adolescent boys only, the total incidence of penile cancer was 4–5% and 5–7% lower, and for oropharyngeal cancer 3–4% and 4–5% lower for MSM 18–26 and MSM 18–45, respectively. This is also significant when weighed against the fact that 25-31% of oropharyngeal and penile cancers are due to HPV [34, 35].

Moreover, a cohort study analyzing data from 2007 to 2017 in Giessen, Germany, showed a significant increase in the overall oropharyngeal squamous cell carcinoma incidence and also for HPV-associated cancers of the tonsils and oropharynx [36], which indicates an increased population burden also for the following years. The model does not capture this increasing trend since it assumes a stable baseline incidence. As such, our results may represent a conservative estimate of the benefit of HPV vaccination for oropharyngeal cancers, although it is unclear if this increasing trend is being seen in MSM, or just non-MSM males.

The differences in incidence after 100 years were smaller in anal cancer. Importantly, vaccination of MSM significantly accelerates these incidence drops. After a relatively short period of about 20 years, a drop in incidence was seen for all three cancer types (compared to the vaccinating adolescent boys only scenario), in all scenarios for both considered age groups and independent of HIV status (results are shown in Supplementary Material B (Figures B4-B11). This acceleration in the reduction of HPV-related cancers in MSM highlights the need to facilitate access for MSM as soon as possible. From a public health perspective, the vaccination scenarios of adult MSM achieve the cancer incidence outcomes of adolescent boys vaccination 16–50 years earlier, depending on the disease and vaccination strategy.

Compared to vaccinating adolescent boys only, the incidence of anogenital warts decreased in HIV- MSM by 42% considering the vaccination scenario of MSM 18–26 (5%) and by up to 78% considering the vaccination scenario of MSM 18–45 (10%). The additional vaccination of adult MSM averted 81,372 to 162,388 cases of anogenital warts. This drop was similar among MSM living with HIV and will begin to be seen after a few years.

Forecasting disease burden into the future with a model comes with many limitations. The most relevant of these limitations is where we have extrapolated certain facets of the model into the future. We highlight here: sexual behavior, HIV risk, management of HPV-related disease, and efficacy of the vaccine in HIV + populations. Sexual behavior (in terms of volume and age-based mixing) is an unpredictable social activity that could change significantly over the next century. If these changes impact HPV transmission, our results may change. Research into the management and treatment of HIV/AIDS is ongoing, and subsequent improvements could result in a normalization of risk for HPV-related disease among MSM living with HIV, rendering our results an overestimate. No independent costing was performed to establish treatment costs, but to our knowledge, the prices for the relevant services have not changed significantly, excepting inflation. Also, higher vaccination coverage rates in adolescent boys will increase the estimated ICERs, making vaccination of adult MSM less cost-effective. Finally, clinical trials for screening for anal disease are ongoing [37], which, if widely implemented, could decrease the burden of anal disease and affect our estimates of the benefit for vaccination, particularly in the short term.

Since there is no established official cost per QALY threshold in Germany [38], the threshold of 1-3xGDP level is used. Based on the 2022 GDP of 46,149 €/capita, 9vHPV vaccination of MSM was found to be cost-effective. Especially the reduction of anal cancer cases had a significant effect on the estimated costs of HPV in Germany. In comparison to vaccinating adolescent boys only, both vaccination strategies (MSM 18–26 and 18–45) were cost-effective at the 1xGDP level. ICERs of 35,300.09€/QALY and 36,841.59€/QALY were found with VCRs of 62% and 75% for 5% and 10% vaccination scenarios of MSM 18–26, respectively. For the vaccination scenario of MSM 18–45 (5%), ICERs were estimated at 42,088.06 and 51,554.82€/QALY with VCRs of 73% and 87% in MSM 18–45 (10%), respectively, when compared to vaccinating adolescent boys only. Comparing vaccination scenario 18–45 to vaccination scenario MSM 18–26, ICERs were cost-effective at a 2xGDP level with ICER of an estimated 53,049.56€/QALY and 88,852.04€/QALY, depending on the uptake scenario.

Other studies that modeled HPV vaccination specifically in MSM populations from England, Australia, and the United States reported that a targeted HPV vaccination of MSM would be an effective way of reducing the burden of HPV-related diseases and would likely be cost-effective [30, 39–41]. The approaches are not directly comparable to our model as the approaches, input parameters, MSM vaccination age (up to 26, 27 or older or up to 40), analyzed HPV-related diseases, and HPV vaccine (2vHPV and 4vHPVv) differed. Nevertheless, all four studies conclude that HPV vaccination in MSM would reduce the HPV-associated disease burden and would likely be cost-effective. Zhang et al. (2017) highlighted that in Australia, vaccination of MSM in addition to the national boys vaccination program would lead to a faster reduction of HPV-related diseases but that the additional benefits of vaccinating adult MSM would decrease over time as it was assumed that the number of unvaccinated MSM would be reduced by 75% in the next decade by the boys vaccination program [41]. However, in Germany, the uptake among boys is still low at around 8% [14]. Even with higher uptake scenarios, the final cumulative VCR achieved after 100 years among adult MSM was 75% and 87% by vaccination of 18–26-year-old MSM and 18–45-year-old MSM, respectively. Considering the progress of girls vaccination, where the VCR is about 54% after 13 years [14], VCRs among MSM of 62% (18–26-year-old MSM) and 73% (18–45-year-old MSM) seem more probable and would be likely cost-effective with an ICER 35,300.09€/QALY and 42,088.06€/QALY, respectively.

The calibration of the model is limited by two main factors: The lack of prevalence data for targets, and the lack of multiple parameter sets. Taken together, we may be missing regions of the parameter space that have feasible fits to the data that may produce different predictions. The limitations are largely technical, in that more advanced methods of calibration require increased computational capacity on top of an already immense model, and future work will need to address these challenges. However, despite these limitations, we believe that this work is still an important first step in investigating the need for HPV vaccination in this potentially vulnerable subgroup of the population of Germany. This means that ongoing validation of the calibrated values will be necessary as more of the relevant data becomes available. One validation that has become recently available to this work is the ANCHOR study [42], which estimated the rate of progression from HSIL to anal cancer to be 402 per 100,000 person-years (262 to 616), which is remarkably consistent with the calibrated values (200–300 per 100,000-years) in the model. The model values may be lower since the precancerous states are aggregated across the two grades of precancers.

The strength of this study is that it integrates detailed, Germany-specific data into well-established modeling methods, and yields results that are consistent with other modeling that has been done. The high level of detail and granularity in age and sexual activity structure allows the model to capture a detailed picture of the situation in Germany as it relates to HPV and HPV-related diseases in MSM.

Overall, the study´s results were robust concerning several scenarios and sensitivity analyses. Assumptions around disease utility and discount rates had the largest effect on the results, which is consistent with other modeling analyses for HPV-related disease.

Conclusions

The dynamic transmission model shows that the vaccination of adult MSM 18–45 could result in a profound acceleration of beneficial public health outcomes for HPV-related cancers and anogenital warts when compared to the vaccination of adolescent boys only. Moreover, the realization of these additional benefits is likely to be cost-effective, especially if vaccination of adolescent boys is low. However, future increasing coverage of HPV vaccination among adolescent boys could change these results.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

All authors attest they meet the ICMJE criteria for authorship. The study concept and design were developed by C.P.,C.W., R.W., M.R. and A.L. C.W. and C.P. were involved in data acquisition. C.P. performed the analyses. All authors were involved in data interpretation. C.P., C.W., and A.L. wrote the manuscript. All authors provided critical review and revision of the manuscript and approved the final version of the manuscript.

Funding

This study was funded by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

Cornelia Wähner, Miriam Reuschenbach and Agnes Luzak are full-time employees of MSD Sharp & Dohme GmbH. Regine Wölle is a retired employee of MSD Sharp & Dohme GmbH. Cody Palmer is full-time employee of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. Jens Peter Klussmann received a research grant and honoraria as member of medical advisory boards from MSD Sharp & Dohme GmbH. Alexander Kreuter received honoraria as speaker and honoraria as member of medical advisory boards from MSD Sharp & Dohme GmbH. Julian Witte reports grants from MSD Sharp & Dohme GmbH during the conduct of the study; grants from Sanofi, GSK, Viatris, Seqirus, Pfizer, Janssen, and DAK-Gesundheit outside the submitted work.