Efficacy of the AS04-adjuvanted HPV-16/18 vaccine in young Chinese women with oncogenic HPV infection at baseline: post-hoc analysis of a randomized controlled trial

ABSTRACT Human papillomavirus (HPV) vaccines are efficacious against HPV infections and associated lesions in women HPV-naïve at vaccination. However, vaccine efficacy (VE) against oncogenic, high-risk HPV (HR-HPV) types in women infected with any other HR-HPV type at first vaccination (baseline) remains unclear. This post-hoc analysis of a phase II/III study (NCT00779766) evaluated AS04-adjuvanted HPV-16/18 (AS04-HPV-16/18) VE against HR-HPV type infection in 871 Chinese women aged 18–25 years over a 72-month follow-up period. Study participants were DNA-negative at baseline to HR-HPV type(s) considered for VE and DNA-positive to any other HR-HPV type. Initial serostatus was not considered. Baseline DNA prevalence was 14.6% for any HR-HPV type and 10.6% excluding HPV-16/18. In the total vaccinated cohort for efficacy, VE against 6-month and 12-month HPV-16/18 persistent infections (PIs) in women DNA-negative to HPV-16/18 but DNA-positive to any other HR-HPV type at baseline was 100.0% (95% Confidence Interval [CI]: 79.8–100.0) and 100.0% (95%CI: 47.2–100.0), respectively. VE against HPV-16/18 incident infections in women DNA-positive to one vaccine type but DNA-negative to the other one at baseline was 66.8% (95%CI: −18.9–92.5). VE against HPV-31/33/45 incident infections, in women DNA-positive to HPV-16/18 and DNA-negative to the considered HPV type at baseline was 71.0% (95%CI: 27.3–89.8). No HPV-16/18 PIs were observed in vaccinated women with non-vaccine HPV A7/A9 species cervical infection at baseline. These findings indicated that women with existing HR-HPV infection at vaccination might still benefit from the AS04-HPV-16/18 vaccine. However, this potential benefit needs further demonstration in the future.


Introduction
Three prophylactic human papillomavirus (HPV) vaccines have been licensed worldwide for protection against HPVrelated diseases and are currently available in over 135 countries. All three vaccines induce strong and sustained protection against anogenital infections and precancerous lesions caused by the vaccine types in subjects HPV-naïve at the time of vaccination. 1-4 The AS04-adjuvanted HPV-16/18 (AS04-HPV -16/18) vaccine provides some level of protection against oncogenic, high-risk (HR) non-vaccine types HPV-31/33/45 while a limited cross-protection against HPV-31 is observed for the quadrivalent HPV vaccine. 5,6 Most countries have approved the use of at least one HPV vaccine in women up to age 45 years. 7 A Chinese domestic HPV vaccine was also approved by the China Food and Drug Administration in December, 2019. 8 As of June 2020, HPV vaccines were included in 107 countries worldwide as part of national immunization programs (NIP), 9 targeting mainly adolescent girls. Independently, unvaccinated adult women may seek vaccination services on a user-pay basis. In China, for example, low awareness of HPV infections and related diseases and lack of availability of HPV vaccines in the NIP have led to HPV vaccines being mainly administered to adult women. 10,11 Adult women are more likely to be sexually active and may thus already have been, or currently be, HPV-infected at the time of first HPV vaccination.
National screening programs for cervical cancer were in place in >144 countries worldwide by 2018. 12 In some countries where HPV DNA testing is used for cervical cancer screening and prevention efforts, women attending screening service may be aware of an eventual HPV infection before vaccination, although screening for presence of HPV prior to vaccination is not required. 13 This raised the question about HPV vaccine benefit in this population, especially as the World Health Organization and a growing number of countries are proposing HPV testing as a primary screening tool for cervical cancer. 14 Until now, most studies focused on subjects HPV-DNAnegative to vaccine types at baseline, 7 and some of them have demonstrated vaccine efficacy (VE) in women having cleared prior HPV vaccine type infection (seropositive but DNAnegative to vaccine type). 15,16 It remains unclear whether, among women with cervical HPV infection (HPV DNApositive) at the time of first vaccination, the residual benefit of preventing infection with other HPV types to which they have not yet been exposed would be sufficient to warrant vaccination. More VE evidence in the HR-HPV infected population is needed for informed decisions by public health authorities and healthcare professionals.
Hence, we performed the current post-hoc analysis to determine VE against infections associated with HPV-16/18 and HPV-31/33/45 types (individually or in combination) in Chinese women aged 18-25 years with cervical HPV infection (DNA-positive) with any other HR-HPV type at the time of first vaccination, irrespective of their serostatus. Figure 1 summarizes the research, clinical relevance and impact of this study on the patient population.

Study design
The primary study (www.ClinicalTrials.gov; NCT00779766) design, methods, and results of event-driven analyses have been previously presented. 17,18 This post-hoc analysis reports VE against HPV-16/18 and HPV-31/33/45 infections (individually or in combination) in women with cervical infection (DNA-positive) with any HR-HPV type other than the one(s) considered for VE at the time of first vaccination (baseline). Subjects were followed up to 72 months with incident infection and persistent infection (PI) as endpoints.

Participants
Healthy women aged 18-25 years were enrolled at four sites (Binhai, Jintan, Lianshui and Xuzhou Centers for Disease Control and Prevention) in Jiangsu Province, China. Women were included in the trial regardless of their HPV DNA status, HPV-16/18 serostatus, or cytology results at baseline. Virgins were not enrolled in the study due to cultural and ethical considerations. Written informed consent was obtained from each participant prior to the performance of any study-specific procedures.

Procedures
Briefly, study participants were randomized in a 1:1 ratio to receive either the AS04-HPV-16/18 vaccine (Cervarix, GSK) or aluminum hydroxide [Al(OH) 3 ] at 0, 1, and 6 month intervals in a double-blind manner. AS04 is a proprietary GSK Adjuvant System containing monophosphoryl lipid A (50 μg MPL; produced by GSK) adsorbed on aluminum salt (500 μg Al 3+ ). Cervical cytology samples for HPV DNA testing were collected every six months at each study visit. Cervical cytology was tested using the ThinPrep PapTest (Cytyc Corporation, Boxborough, MA, USA) and reported according to the Bethesda 2001 classification system. Histopathological analysis was performed by a panel of gynecological pathologists at the Cancer Hospital, Chinese Academy of Medical Sciences (CICAMS), Beijing, China. Final case ascertainment of women assumed to meet the criteria for efficacy endpoints was reviewed in a blinded fashion by an independent review committee. A broad-spectrum PCR assay, SPF10-LiPA25 (version 1 based on licensed Innogenetics SPF10 technology; Labo Biomedical Products, Rijswijk, Netherlands) and type-specific PCR for HPV-16 and HPV-18 DNA were used to test cervical samples and biopsy material for HPV DNA from 14 HR-HPV types (16,18,31,33,35,39,45, 51, 52, 56, 58, 59, 66, and 68) and 11 non-HR-HPV types.

Statistical analysis
Previously published reports of this research disclose the efficacy, immunogenicity, and safety analyses in women DNAnegative at baseline (or both at baseline and Month 6), and seronegative at baseline (or irrespective of serostatus for the considered HPV type) in the according to protocol cohort for efficacy (ATP-E) and the total vaccinated cohort for efficacy (TVC-E). 17,18 The ATP-E included women with available efficacy data, who received three doses of vaccine or control and had a normal or low-grade cytology at baseline, while TVC-E For VE against individual HPV type, subjects were DNA-negative to the corresponding HPV type at baseline. For VE against combined types, subjects were DNA-negative to at least one HPV type at baseline (subjects were in the analysis of at least one single type). Besides meeting the above criteria for baseline status, subjects included in the relevant analyses must have sample available after vaccine dose 1 for the evaluation of the virological endpoints, including incident, 6-month and 12-month HPV persistent infections. 1 High-grade cytology included squamous cells cannot exclude high-grade squamous intraepithelial lesions, high-grade squamous intraepithelial lesions, atypical glandular cells, or malignancy 2 VE against incident, 6-month and 12-month persistent infections associated with HPV-16, HPV-18 and HPV-16 and/or HPV-18 was assessed. 3 VE against incident, 6-month and 12-month persistent infections associated with HPV-31, HPV-33, HPV-45, and HPV-31 and/or HPV-33 and/or HPV-45 was assessed. 4 VE against incident, 6-month and 12-month persistent infections only associated with HPV-16 were assessed. 5 VE against incident, 6-month and 12-month persistent infections only associated with HPV-18 were assessed. HR-HPV, high-risk human papillomavirus; N, number of subjects; VE: vaccine efficacy.
was similar to ATP-E but included women who received at least one dose of vaccine or control. 17,18 Here, we specifically report data from women with HR-HPV infection (i.e. HR-HPV DNApositive, irrespective of HPV serostatus) at baseline in the TVC-E. Subjects were evaluated for subsequent development of HPV infection with the types they were DNA-negative to at baseline. The evaluated endpoints were incident infection, 6-month PI and 12-month PI associated with specific HPV type(s).
Incident infection was defined as the first detection by PCR of an episode of infection by HPV type(s) in a subject previously negative for the considered HPV type(s) and may have been transient or become persistent. Six-month PI was defined as at least two positive HPV DNA PCR assays for the same HPV type(s) with no negative DNA sample between the two positive DNA samples, over an interval of approximately six months. Twelve-month PI was defined as the detection by PCR of the same HPV type(s) at all available timepoints over an interval of approximately 12 months. Due to the limited sample size and the insufficient number of cervical intraepithelial neoplasia of grade 2 or worse (CIN2+) cases, VE against histological endpoints is not presented.
Endpoint cases were calculated as number of subjects reporting at least one event in each group, and VE was calculated using a conditional exact method, as previously described. 17,18 Case counting started on the day after the first dose and ended at the time of an endpoint event or until the end of the 72-month follow-up. Statistical analyses were carried out with SAS 9.4 on the SDD platform. A similar method was used for the analysis in the ATP-E cohort.

Data sharing
Anonymized individual participant data and study documents can be requested for further research from www.clinicalstudy datarequest.com.     (Table 1).

VE against HPV-16/18 infections in women with HR-HPV infections at baseline
Overall, the vaccine had high efficacy against incident infection, 6-month and 12-month PIs associated with HPV- 16   Individual and combined vaccine efficacy against infections with HPV-31, HPV-33, and HPV-45 in women DNA-negative to the considered HPV type at baseline, using conditional exact method. For individual type, subjects were DNA-negative to the corresponding HPV type at Month 0. For combined types, subjects were DNAnegative to at least one HPV type at Month 0 (subjects were in the analysis of at least one single type). Follow-up starts at day after dose 1. AS04-HPV-16/18 v, AS04-adjuvanted HPV-16/18 vaccine; CI, confidence interval; HPV, human papillomavirus; HR-HPV, high-risk human papillomavirus; N, number of subjects included in each group; n, number of subjects reporting at least one event in each group; TVC-E, total vaccinated cohort for efficacy; VE, vaccine efficacy women DNA-negative to HPV-16/18 but DNA-positive to any other 12 HR-HPV types ( Table 2). We also analyzed the efficacy against infection related to HPV-16/18 in women with HPV-16 or HPV-18 infection at baseline. In 223 women DNA-negative at baseline to one vaccine type but DNA-positive to the other one, VE against incident infection endpoints associated with HPV-16/18 was 66.8% (95% CI: −18.9 to 92.5). There were not enough cases to evaluate 6-or 12-month PI-related endpoints in this subgroup (Table 2).
Similar results were obtained for the ATP-E cohort (Supplementary table 1).
Similar results were obtained for the ATP-E cohort (Supplementary table 2).

HPV-16/18 associated CIN2± in women with HR-HPV infections at baseline
In all 871 subjects DNA-positive to any 14 HR-HPV types at baseline, irrespective of serostatus, only five CIN2+ cases associated with HPV-16 and/or −18 (all five cases were associated with HPV-16) were reported in subjects DNA-negative to the corresponding HPV type at baseline.

Discussion
As part of the debate on the added-value of prophylactic HPV vaccination in already sexually active women infected with HPV, this post-hoc analysis supplied data on Chinese 18-25year-old women who, at the time of first vaccination, had cervical HR-HPV infection (regardless of their serostatus). It evidences that the AS04-HPV-16/18 vaccine may provide preventive benefit to this population.
The FUTURE II (NCT00092534), VIVIANE (NCT00294047) and PATRICIA (NCT00122681) studies have already demonstrated significant VE against vaccine type HPV infection or related neoplasia in subjects with previous or existing HPV infection but DNA-negative for the targeted HPV type at baseline. 16,19,20 The FUTURE II study, using the quadrivalent HPV vaccine, provided results from women seropositive or DNA-positive to one to three HPV vaccine types before vaccination. 19 The VIVIANE study, using the AS04-HPV-16/ 18 vaccine, had a subset analysis consisting of women with a history of previous HPV infection or disease. 20 However, seropositivity is considered to be associated with prior HPV exposure even though the infection has been cleared, while DNA status informs more about ongoing HR-HPV infection, which significantly affects VE. Moreover, HPV antibody titers are not routinely tested. Thus, our study merely included subjects who were DNA-positive at the time of first vaccination (regardless of their serostatus). The findings provide real-world practical and instructive implications for both clinicians and infected women.
In women DNA-positive to one of the HPV-16/-18 types at baseline but DNA-negative to the other type, our results displayed a statistically non-significant VE against HPV-16/18 incident infections. The PATRICIA study group also analyzed VE against CIN2+ associated with HPV-16/18 in women who were HPV DNA-positive to one vaccine type at baseline but DNA-negative and seronegative to the other vaccine type. The VE was 90.0% (95% CI: 31.8 to 99.8) for HPV-16/18. 16 Considering DNA status might have more impact on efficacy compared with serostatus, the negative lower bound of the CI for the VE observed in our analysis could be explained by the limited sample size available. It is also possible that efficacy is higher against high-grade lesions than against infections.
As demonstrated in our study, among women DNApositive to the vaccine types (HPV-16/18), including those simultaneously infected with both types, cross-protection against HPV-31/33/45 incident infections is evident. We also observed decreasing VEs against HPV-31/33/35 infections that became persistence; VE for incident infection > VE for 6-month PI > VE for 12-month PI. These changes in VE estimates may be linked to the underestimation of nonvaccine HPV types in the control group due to the broadspectrum HPV PCR methodology used. This method is known to potentially underestimate the prevalence of genotypes present at low relative concentrations in multiple infections, a scenario that is more likely to arise in the control group when considering vaccine and cross-protected HPV types. 21 This bias is stronger for 12-month PI than for 6-month PI and incident infection, and may explain the decreasing VEs observed in this study. 21 The L1 protein of non-vaccine A9 species (HPV-31/33/35/ 52/58) and non-vaccine A7 species (HPV-39/45/59/68) display amino acid homology with L1 proteins of HPV-16 and HPV-18 types, respectively. 22 Based on the phylogenetic relatedness, it is interesting to explore the impact of the preexistence of non-vaccine A7/A9 species infections at first vaccination on VE against endpoints related to HPV-16/18. However, most published studies have focused on cross-protective efficacy and a possible cross-reactive immune response related to A7/A9 species. 22,23 In contrast, our data indicate high VE against HPV-16 or HPV-18 infections in women DNA-positive to any other non-vaccine HR-HPV A9 or A7 species at vaccination, respectively.
HPV prevalence studies among Chinese women have shown two peaks in HR-HPV infections: a first peak in those aged 15-24 years, soon after the onset of sexual debut, and a second in those aged >40 years. 24 The second peak is thought to be associated with new HPV exposure and/or the reactivation of a latent infection, and coincides with the peak of cervical cancer incidence observed in women up to 50-55 years old in China. 25 As evidenced by our study, vaccinating women with preexisting HR-HPV infection could reduce the risk of infection with and transmission of HPV type to which they have not yet been exposed. Moreover, even though HPV vaccines have no therapeutic impact on preexisting infection, 19 vaccination could prevent eventual reinfection later in life. Vaccination of infected women on an individual basis would confer protection to women from the second attack peak, protect their partners from possible infection and may slow down the progression of the HPV infection 26 at the scale of the Chinese population. Further evidence is required on this point.
The HPV-FASTER protocol proposed offering HPV vaccination to women in a broad age range (9-45 years) to contribute to faster and greater population impact, based on the results of high VE from phase III clinical trials among adult women. However, vaccination of HR-HPV infected women has been controversial. [27][28][29] The results of our analysis lend support to this concept, suggesting vaccinating all women, irrespective of their HPV DNA status. Within the context of global investment and commitments toward the elimination of cervical cancer, this would allow to achieve maximum health benefit and equity to the entire population.
The main strength of our analysis was the specific focus on HR-HPV DNA-positive women at baseline, ignoring confounding factor of serostatus that is moreover not examined routinely, apart from clinical trial testing. An unavoidable limitation is that the findings were generated from a post-hoc analysis of a randomized controlled trial, which was not designed to analyze efficacy in women HR-HPV positive at vaccination, and randomization was not controlled for HPV-infection at baseline. Besides, the small number of women with ongoing HR-HPV infection available for inclusion in the present analysis (i.e., 15% of the total cohort) limits its statistical power. Also, due to insufficient endpoint cases, we evaluated efficacy mainly on a virologic perspective in the form of incident infections, with some subsets based on 6-month and 12-month PIs. However, as demonstrated in previous clinical trials, efficacy against PI is an acceptable surrogate endpoint for efficacy in the prevention of cervical cancer. 30 In conclusion, our analysis indicated that women with existing HR-HPV infection at the time of first vaccination might still benefit from the AS04-HPV-16/18 vaccine. However, this potential benefit needs further demonstration in the future.