ABSTRACT
Extensive evidence confirms the high effectiveness of human papillomavirus (HPV) vaccines in young girls; however, real-world data on vaccine effectiveness (VE) among women vaccinated at ages 18–45 remain limited, particularly in China, where HPV status at vaccination is often unknown. We conducted a population-based matched test-negative design study using linked cervical cancer screening and immunization records in Shenzhen, China. Cases positive for vaccine-type high-risk HPV (hrHPV) DNA were matched 1:1 with negative controls by age, screening date, household registration, and screening type. VE for at least one dose was estimated using conditional logistic regression and stratified by time since vaccination, vaccine valency, and age at vaccination. Among 7,060 matched pairs, overall VE increased significantly with time since vaccination: 6.4% (95% CI: −3.9% to 15.7%) at <1y, 17.9% (5.5% to 28.7%) at 1–<2y, and 40.2% (29.6% to 49.2%) at ≥2y. Among women vaccinated ≥2y earlier, VE against HPV16/18 was 48.4% (33.7% to 59.8%) for 2-/4-valent vaccines and 54.4% (26.5% to 71.7%) for the 9-valent vaccine. Significant protection at≥2y was observed across all age-at-vaccination groups: 45.7% (25.0% to 60.7%) for ages 18–26, 39.6% (23.5% to 52.4%) for ages 27–35, and 35.5% (12.6% to 52.4%) for ages 36–45. In conclusion, HPV vaccination confers moderate protection against vaccine-type hrHPV DNA positivity among Chinese women vaccinated at ages 18–45. However, as this outcome reflects a single-time-point surrogate virologic endpoint rather than persistent infection or clinical disease, adult women should continue regular cervical cancer screening.
KEYWORDS: Human papillomavirus, vaccine effectiveness, test-negative design, real-world study, adult women, catch-up vaccination
Background
Cervical cancer remains a leading gynecological malignancy worldwide, primarily resulting from persistent infection with high-risk human papillomavirus (hrHPV).1 Prophylactic HPV vaccines have proven safe and effective in preventing hrHPV infection, cervical intraepithelial lesions, and subsequent cancer.2 Vaccine effectiveness (VE) is highly dependent on the timing of administration.3 Pre-licensure trials have established that vaccination before sexual debut confers optimal protection,4 a finding corroborated by a meta-analysis of 65 real-world studies across 14 high-income countries, which showed the greatest benefit when vaccination was initiated between ages 9 and 14.2 Consequently, most national immunization programs prioritize routine vaccination for adolescents (typically aged 9–14) accompanied by catch-up vaccination for young adults (generally under 26 y of age).5
China ranks second globally in cervical cancer incidence, accounting for approximately 18.2% of new cases and 17.3% of cervical cancer deaths.6,7 This burden is compounded by persistently low HPV vaccination coverage.8 China’s HPV vaccination strategy has followed a distinctive trajectory: initial authorization in 2016 was restricted to women aged 16 y and older, and eligibility was not expanded to females aged 9–45 until 2020–2022.9,10 This landscape has been substantially reshaped by the recent successive launches of domestically developed, clinically effective, and more affordable 2-valent (2 v) and 9-valent (9 v) HPV vaccines.11,12 A pivotal policy shift subsequently occurred in November 2025, when HPV vaccination was formally incorporated into China’s national immunization program.13 However, the historical pattern of low coverage and vaccination at older ages may delay the attainment of population-level herd immunity. Consequently, robust real-world evidence on VE among adult women is urgently needed to inform this expanding public health effort.
Although randomized controlled trials (RCTs) in China have demonstrated the efficacy of HPV vaccination among women aged 27–45 y,9,11 these trials were typically conducted in well-characterized, HPV-naïve populations with high protocol adherence. Therefore, their findings may not be generalizable to real-world settings, highlighting the need to evaluate the real-world VE in this catch-up age group to optimize cervical cancer prevention strategies.
Designing a robust real-world study requires careful consideration of two critical methodological priorities: the selection of evaluation endpoints and the choice of study design. Although high-grade cervical lesions and cervical cancer represent the ultimate clinical endpoints, their long latency periods preclude their use for short- to mid-term assessments. In contrast, hrHPV infection is a necessary precursor in the carcinogenic pathway to cervical cancer and reflects early population-level immunity, which is fundamental to cancer elimination initiatives. Accordingly, the World Health Organization recommends monitoring hrHPV prevalence to assess VE in the near-term (5–10 y) after vaccination.14 Real-world studies often employ repeated cross-sectional surveys in regions with high vaccination coverage.14 However, the relatively low vaccination coverage in China limits the sensitivity of such trend analyses for detecting true population-level shifts.15 For example, a study of 1.07 million adult Chinese women (conducted in 2020–2021) reported that the prevalence of vaccine-targeted HPV types decreased (by 10% and 13% for 2 v- and 4 v-HPV vaccines, respectively), but the overall HPV prevalence increased by 23% to 26%.16 Nonetheless, owing to the lack of individual-level data, it was not possible to causally attribute this specific decline to vaccination rather than to secular trends. To overcome these limitations, we adopted a matched test-negative design (TND). This approach estimates individual-level protection against vaccine-type hrHPV DNA positivity while controlling for biases related to healthcare-seeking behavior and has been extensively validated in VE studies.17–19 For example, Kamolratanakul and colleagues employed a TND case-control study to assess the effectiveness of the HPV vaccine against vaccine-type hrHPV positivity among girls aged 9–18.20
Shenzhen, a major metropolis in southern China, is characterized by a large migrant population and a youthful demographic profile—factors associated with high sexual activity and an elevated risk of HPV exposure.21 Leveraging the city’s robust health information infrastructure, we conducted a real-world study using linked cervical cancer screening and immunization records to assess the effectiveness of HPV vaccination against hrHPV DNA positivity among adult women. This study aimed to bridge the evidence gap regarding vaccine performance in adult women and to provide critical data needed to inform integrated cervical cancer prevention strategies.
Methods
Study design and data source
We conducted a population-based, retrospective, matched TND case-control study in Longhua District, Shenzhen, China. Longhua is one of the largest districts in Shenzhen, with a population exceeding 3.3 million. The study included women aged 18 to 51 y; the upper limit of 51 y accounts for the maximum age of individuals who received their first dose at age 45—the upper age limit for HPV vaccination in China.
Data were derived from linked cervical cancer screening and HPV vaccination records. Screening data were extracted from the Shenzhen Maternal and Child Health Information System, which systematically captures records from all regional healthcare institutions. As this system does not record HPV vaccinations, immunization details were retrieved from the Shenzhen Vaccination Information System. This registry is linked to the national immunization database, enabling verification of doses administered outside Shenzhen. During the study period, HPV vaccination coverage among adult women in Shenzhen was 23.4% for those aged 18–26, 22.0% for ages 27–35, 20.2% for ages 36–44, and 3.2% for ages 45–50.22
This retrospective observational study utilized de-identified data from administrative health databases, involving no new sample collection, primary data acquisition, or patient intervention. All research data were kept strictly confidential. The study was approved by the Ethics Committee of Shenzhen Longhua District Maternity and Child Healthcare Hospital (Approval No: SRE-PCFR/2024007), which granted a waiver of informed consent. This study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for case-control studies.
Participants
Case and control populations
Cases were women who tested positive for vaccine-type hrHPV (types 16, 18, 31, 33, 45, 52, and 58) and had concurrent cytology results of normal, inflammatory, or low-grade squamous intraepithelial lesion (LSIL). This definition aimed to capture women with a higher likelihood of recent infection. As most adult women in China are unaware of their HPV status at the time of vaccination, and progression to high-grade squamous intraepithelial lesion (HSIL) or worse typically requires prolonged infection,23 individuals presenting with HSIL or worse lesions were excluded to minimize bias from pre-existing, persistent infections. Controls were women who tested negative for all vaccine-type hrHPV genotypes. This group comprised both those who were completely HPV-negative and those who tested positive only for non-vaccine-type HPV. For all participants, the date of their first valid HPV test was recorded as the index date.
Participant selection and exact matching
The participant selection process is illustrated in Figure 1. A total of 266,863 cervical cancer screening records from 2021 to 2024 were extracted. We excluded 104,985 ineligible records based on the following criteria: (1) age at screening ≥52 y; (2) incomplete data on key variables, duplicate entries, or logical errors; (3) pooled hrHPV-positive results lacking specific genotyping; (4) positivity for low-risk vaccine types (HPV 6/11) without concurrent vaccine-type hrHPV infection; (5) histological evidence of HSIL or worse; and (6) repeat screening records for the same individual (retaining only the first positive record for cases). Of the remaining 161,818 individual records, 9,428 (5.83%) tested positive for vaccine-type hrHPV. To minimize confounding related to healthcare access and demographic characteristics, we performed 1:1 exact matching between cases and controls. Matching criteria included age at screening (±6 months), screening date (year and month), household registration (Shenzhen vs. non-Shenzhen), and screening type (organized vs. opportunistic). This procedure yielded 9,174 matched case-control pairs.
Figure 1.
Study design flowchart.
Measurement
Cervical cancer screening and HPV DNA genotyping
Screening was conducted according to the Chinese expert consensus on HPV nucleic acid testing.24 Two types of HPV DNA assays were used: a semi-genotyping hybrid capture-based assay (providing individual genotyping for HPV 16/18 but pooled results for 11 other high-risk types) and a 21-type HPV genotyping assay based on qualitative polymerase chain reaction (PCR) and membrane hybridization. Individuals who tested positive using the semi-genotyping hybrid capture-based assay were recommended to undergo further testing with the 21-type genotyping assay. All participants testing positive for HPV 16 and/or 18 underwent direct histological assessment. A ThinPrep Cytology Test was performed for those testing positive for non-16/18 high-risk types (i.e., HPV 31, 33, 45, 52, and 58). Participants with cytology results showing atypical squamous cells of undetermined significance (ASC-US) or worse were subsequently referred for histological assessment.24 The 9 v‑HPV vaccine specifically targets five of the seven vaccine‑type hrHPV genotypes (31, 33, 45, 52, and 58), whereas the remaining two (16 and 18) are targeted by all three vaccine types: 2 v-, 4-valent (4 v), and 9 v-HPV vaccines.
Screening types were categorized into two groups based on health-seeking behavior. Self‑paid testing recommended by physicians during clinical visits for gynecological symptoms or high-risk sexual behaviors was classified as opportunistic screening. Government-funded periodic screening (provided every five years for women aged ≥30 y) and voluntary testing performed during routine health checkups were classified as organized screening.
Vaccine exposure assessment
Vaccination status was defined based on the timing of dose administration. Women were considered vaccinated if they had received at least one dose ≥1 month prior to the screening date. Women with no vaccination record or those who received their first dose within one month before screening were classified as unvaccinated. To ensure a rigorous assessment of cross-protection, cases vaccinated with 2 v- or 4 v-HPV vaccines who tested positive for non-cross-reactive types (HPV 52 or 58) were excluded, whereas those testing positive for potentially cross-reactive types (HPV 31, 33, or 45) were retained.25
Data assurance
To ensure data accuracy, a blinded protocol was implemented. Staff who extracted screening records and performed matching were unaware of vaccination status, whereas personnel who retrieved vaccination records were blinded to screening outcomes. A random quality audit of 10% of the dataset was conducted to verify consistency.
Variables and bias control
Variables potentially confounding the association between HPV vaccination and HPV infection risk were categorized into three domains: socioeconomic status (SES), healthcare-seeking behaviors, and sexual behavior characteristics.
The SES variables available in the administrative health records included residential sub-district, housing type (commercial housing, urban village, or dormitory), Shenzhen household registration, education level, and ethnicity. These SES factors influence both sexual behaviors and healthcare-seeking patterns, thereby affecting access to vaccination and medical resources as well as health awareness.26 Concerns about HPV infection may also drive proactive healthcare-seeking behaviors,27 including HPV vaccination and participation in cervical cancer screening. Accordingly, the screening type (organized vs. opportunistic) was considered a behavioral indicator.
Sexual behavior characteristics included age at screening, marital status, number of pregnancies, and specific behaviors (e.g., age at sexual debut, number of sexual partners, and condom use).28 We developed a directed acyclic graph (DAG) using the DAGitty platform (https://www.dagitty.net/dags.html) to visualize putative causal relationships and identify a minimal sufficient adjustment set, while avoiding adjustment for mediators (Supplementary Figure S1).29,30 The identified minimal sufficient adjustment set included age at screening, screening type, marital status, number of pregnancies, educational level, ethnicity, household registration, residential sub-district, housing type, and unmeasured sexual behaviors.
Within these administrative health records, direct measures of sexual behaviors (e.g., age at sexual debut, number of sexual partners, or condom use) were unavailable due to privacy constraints. However, as indicated in the DAG, sexual behaviors are linked to healthcare-seeking behavior, suggesting that screening type (organized vs. opportunistic) may serve as a proxy measure for latent sexual risk.
Study size
Power Analysis and Sample Size (PASS) software was used to estimate the minimum sample size required for 1:1 matching. Randomized trials have reported a VE of 61.6% against incident HPV infections in HPV-naïve populations.31 However, as real-world populations may include individuals with pre-existing infections, we conservatively prespecified a VE of 30% to account for this attenuation effect. Assuming an effective vaccination rate of 10% (considering only doses administered prior to screening), a two-sided α of 0.05, and a power (1 − β) of 0.80, the minimum required sample size was 2,744.
Statistical analysis
Categorical variables were reported as frequencies and percentages, with differences between cases and controls evaluated using the chi-square test. The association between HPV vaccination and hrHPV DNA positivity was estimated using conditional logistic regression, adjusting for educational level, marital status, number of pregnancies, residential sub-district, housing type, and ethnicity. Missing covariate values were coded as “unknown.” To reduce bias from preexisting hrHPV infections, a “buffer-period” analysis was performed based on the interval between the first vaccine dose and the screening date. We categorized this interval into three lag time categories (<1 y, 1–<2 y, and ≥2 y) to identify the most reliable VE estimate.32–34 Subgroup analyses were stratified by HPV genotype and age at vaccination. VE was calculated as (1 − adjusted odds ratio) × 100%. A two-sided p-value < .05 was considered statistically significant. All analyses were conducted using SPSS Statistics, version 27.0.1 (IBM Corp., Armonk, NY, USA).
Sensitivity analysis
Two sensitivity analyses were performed to evaluate potential residual confounding. First, a negative-control analysis was conducted using non-vaccine-type HPV (nvt-HPV) as the outcome.35,36 Since vaccination is not expected to protect against nvt‑HPV types, an adjusted odds ratio (aOR) close to 1 (i.e., VEnvt ≈ 0) would indicate minimal bias. Second, to examine whether including nvt-HPV-positive individuals in the control group affected the results, the analysis was repeated using only pan-HPV-negative controls.
Results
Participant characteristics
Between 2021 and 2024, the overall HPV DNA positivity rate was 13.6%, increasing from 9.4% in 2021 to 17.4% in 2024 (Figure 2). The positivity rate was higher among women undergoing opportunistic screening than among those undergoing organized screening, suggesting potential differences in baseline risk profiles or sexual behavior patterns between the two populations.
Figure 2.
Detection of HPV DNA in cervical cancer screening: data from Longhua District, Shenzhen, China (2021–2024).
Following the selection and matching process (Figure 1), a total of 14,120 individuals were included in the analysis, comprising 7,060 vaccine-type hrHPV-positive cases and 7,060 matched controls. As shown in Table 1, the matched groups were well-balanced in terms of the number of pregnancies, residential sub-district, and housing type, but differed significantly in ethnicity, screening type, marital status, educational attainment, and time since the first vaccination.
Table 1.
Characteristics of study participants.
| Matched Cases (n = 7060) |
Matched Controls (n = 7060) |
χ2 | p value | |
|---|---|---|---|---|
| Age at screening (years) | ||||
| 18–26 | 1881 (26.6) | 1881 (26.6) | – | Matching |
| 27–35 | 2482 (35.2) | 2482 (35.2) | ||
| 36–51 | 2697 (38.2) | 2697 (38.2) | ||
| Year at cervical cancer screening | ||||
| 2021 | 594 (8.4) | 594 (8.4) | – | Matching |
| 2022 | 2171 (30.6) | 2171 (30.6) | ||
| 2023 | 2212 (31.3) | 2212 (31.3) | ||
| 2024 | 2083 (29.5) | 2083 (29.5) | ||
| Household registration | ||||
| Shenzhen residence | 1390 (19.7) | 1390 (19.7) | – | Matching |
| Non-Shenzhen residence | 5670 (80.3) | 5670 (80.3) | ||
| Types of screening | ||||
| Organized screening | 2106 (29.8) | 2106 (29.8) | – | Matching |
| Opportunistic screening | 4954 (70.2) | 4954 (70.2) | ||
| Ethnicity | ||||
| Han ethnicity | 6902 (97.7) | 6930 (98.1) | 13.2 | .010 |
| Ethnic minorities | 112 (1.6) | 68 (1.0) | ||
| Unknown | 46 (0.7) | 62 (0.9) | ||
| Types of screening institutions | ||||
| Hospital | 5113 (72.4) | 4575 (64.8) | 95.2 | <.001 |
| Community health service center | 1947 (27.6) | 2485 (35.2) | ||
| Marital status | ||||
| Married | 5219 (73.9) | 5464 (77.4) | 24.7 | <.001 |
| Single | 1620 (22.9) | 1421 (20.1) | ||
| Divorced | 148 (2.1) | 110 (1.6) | ||
| Unknown | 73 (1.1) | 65 (0.9) | ||
| Educational level | ||||
| Senior high or below | 4131 (58.5) | 3842 (54.4) | 31.5 | <.001 |
| College or university | 2656 (37.6) | 2887 (40.9) | ||
| Graduate or above | 183 (2.6) | 190 (2.7) | ||
| Unknown | 90 (1.3) | 141 (2.0) | ||
| Number of Pregnancies | ||||
| 0 | 6827 (96.7) | 6822 (96.6) | 0.116 | .944 |
| 1 | 217 (3.1) | 233 (3.3) | ||
| 2 | 16 (0.2) | 15 (0.2) | ||
| Residential sub-district | ||||
| Longhua street | 1516 (21.5) | 1495 (21.2) | 2.9 | .823 |
| Minzhi street | 1976 (28.0) | 1968 (27.9) | ||
| Dalang street | 1027 (14.5) | 1046 (14.8) | ||
| Guanhu street | 1009 (14.3) | 980 (13.9) | ||
| Guanlan street | 517 (7.3) | 558 (7.9) | ||
| Fucheng street | 704 (10.0) | 718 (10.2) | ||
| Others | 311 (4.4) | 295 (4.2) | ||
| Housing types | ||||
| Commercial housing | 2548 (36.1) | 2511 (35.6) | 0.5 | .764 |
| Urban village | 4460 (63.2) | 4500 (63.7) | ||
| Factory dormitory | 52 (0.7) | 49 (0.7) | ||
| §HPV vaccination | ||||
| No | 6339 (89.8) | 6038 (85.5) | 96.0 | <.001 |
| Yes | 721 (10.2) | 1022 (14.5) | ||
| Time since vaccination of first dose | ||||
| §Not vaccinated | 6339 (89.8) | 6038 (85.5) | 59.3 | <.001 |
| <1 y | 372 (5.3) | 405 (5.7) | ||
| 1–<2 y | 201 (2.8) | 278 (3.9) | ||
| ≥2 y | 148 (2.1) | 339 (4.8) | ||
Data are presented as number (%).
¶Number of Pregnancies was defined as the number of pregnancies between 2016 and the date of cervical cancer screening.
§Non-vaccination records or buffer period between 0–1 month defined as not receiving any of HPV vaccination.
Vaccine effectiveness against vaccine-type hrHPV DNA positivity
Overall, 721 (10.2%) individuals in the case group and 1,022 (14.5%) individuals in the control group had received at least one dose of HPV vaccine (Table 1). The unadjusted VE against vaccine-type hrHPV DNA positivity (types 16/18/31/33/45/52/58) varied by time since vaccination: 6.5% (95% CI: −3.8% to 15.8%) at <1 y, 18.1% (95% CI: 5.7% to 28.8%) at 1–<2 y, and 40.7% (95% CI: 30.2% to 49.6%) at ≥2 y after vaccination. This trend remained after adjusting for confounders (educational level, marital status, number of pregnancies, ethnicity, residential sub-district, and housing type), yielding an overall VE of 40.2% (95% CI: 29.6% to 49.2%) at ≥2 y after vaccination (Table 2).
Table 2.
Vaccine effectiveness of at least one dose of HPV vaccination against vaccine-type high-risk HPV DNA positivity.
| Cases n (%) | Matched Controls n (%) | Unadjusted VE%(95%CI) | P value | Adjusted¶ VE%(95%CI) | P value | |
|---|---|---|---|---|---|---|
| HPV vaccination (2 v/4 v/9 v) against vaccine-type high-risk HPV DNA positivity* | ||||||
| Time since vaccination of first dose | ||||||
| Not vaccinated$ | 6339 (89.8) | 6038 (85.5) | 1 (reference) | 1 (reference) | ||
| <1 y | 372 (5.0) | 405 (5.7) | 6.5 (−3.8 to 15.8) | .206 | 6.4 (−3.9 to 15.7) | .216 |
| 1–<2 y | 201 (3.2) | 278 (3.9) | 18.1 (5.7 to 28.8) | .005 | 17.9 (5.5 to 28.7) | .006 |
| ≥2 y | 148 (2.1) | 339 (4.8) | 40.7 (30.2 to 49.6) | <.001 | 40.2 (29.6 to 49.2) | <.001 |
| 2 v/4 v-HPV vaccine against HPV-16/18 | ||||||
| Time since vaccination of first dose | ||||||
| Not vaccinated$ | 2843 (91.7) | 2676 (86.4) | 1 (reference) | 1 (reference) | ||
| <1 y | 125 (4.0) | 145 (4.7) | 12.3 (−4.9 to 26.7) | .151 | 10.0 (−7.7 to 24.8) | .251 |
| 1–<2 y | 68 (2.2) | 97 (3.1) | 22.0 (0.80 to 38.7) | .043 | 20.2 (−1.6 to 37.2) | .067 |
| ≥2 y | 63 (2.0) | 181 (5.8) | 49.4 (35.0 to 60.6) | <.001 | 48.4 (33.7 to 59.8) | <.001 |
| 2 v/4 v-HPV vaccine against HPV-31/33/45 (cross-protection) | ||||||
| Time since vaccination of first dose | ||||||
| Not vaccinated$ | 564 (9.3) | 558 (9.2) | 1 (reference) | 1 (reference) | ||
| <1 y | 8 (1.3) | 7 (1.1) | −6.1 (−113.2 to 47.2) | .868 | −6.6 (−114.2 to 47) | .858 |
| 1–<2 y | 14 (2.3) | 15 (2.5) | 4.0 (−63.2 to 43.5) | .881 | 0.80 (−68.9 to 41.7) | .977 |
| ≥2 y | 23 (3.8) | 29 (4.8) | 12.0 (−33.5 to 42.0) | .548 | 10.0 (−36.7 to 40.8) | .620 |
| 9 v-HPV vaccine against HPV-16/18/31/33/45/52/58 | ||||||
| Time since vaccination of first dose | ||||||
| Not vaccinated$ | 5877 (9.4) | 5744 (9.1) | 1 (reference) | 1 (reference) | ||
| <1 y | 233 (3.7) | 251 (4.0) | 4.8 (−8.5 to 16.5) | .461 | 5.1 (−8.3 to 16.7) | .438 |
| 1–<2 y | 114 (1.8) | 160 (2.5) | 17.7 (1.0 to 31.6) | .039 | 18.0 (1.2 to 31.9) | .036 |
| ≥2 y | 55 (0.9) | 124 (2.0) | 39.2 (20.8 to 53.4) | <.001 | 39.4 (20.9 to 53.5) | <.001 |
| 9 v-HPV vaccine against HPV-16/18 | ||||||
| Time since vaccination of first dose | ||||||
| Not vaccinated$ | 2635 (95.8) | 2534 (92.1) | 1 (reference) | 1 (reference) | ||
| <1 y | 73 (2.7) | 99 (3.6) | 16.7 (−5.1 to 34.0) | .123 | 17.8 (−3.8 to 34.9) | .100 |
| 1–<2 y | 26 (0.9) | 63 (2.5) | 42.7 (15.7 to 61.1) | .005 | 43.9 (17.3 to 61.9) | .003 |
| ≥2 y | 17 (0.6) | 55 (2.0) | 53.7 (25.4 to 71.3) | .002 | 54.4 (26.5 to 71.7) | .001 |
| 9 v-HPV vaccine against HPV-31/33/45/52/58 | ||||||
| Time since vaccination of first dose | ||||||
| Not vaccinated$ | 5713 (94.9) | 5661 (94.1) | 1 (reference) | 1 (reference) | ||
| <1 y | 171 (2.8) | 166 (2.8) | −3.0 (−20 to 11.5) | .699 | −3.0 (−20 to 11.6) | .703 |
| 1–<2 y | 92 (1.5) | 107 (1.8) | 5.8 (−15.8 to 23.3) | .572 | 4.7 (−17.2 to 22.5) | .649 |
| ≥2 y | 42 (0.7) | 84 (1.4) | 33.8 (10.4 to 51.2) | .008 | 34.0 (10.6 to 51.3) | .007 |
VE: vaccine effectiveness; 2 v: 2-valent; 4 v: 4-valent; 9 v: 9-valent.
*vaccine-type high-risk HPV types included type 16/18/31/33/45/52/58.
$Not vaccinated was defined as no vaccination records or vaccinated within one month at the time of cervical cancer screening.
¶Controlling for marital status, number of pregnancies since vaccine introduction, educational level, ethnicity, residential sub-district, and housing type.
Vaccine effectiveness by vaccine type
Table 2 presents the VE estimates stratified by vaccine type. For the 2 v- and 4 v-HPV vaccines, the adjusted VE against the primary targeted types (HPV 16/18) was 48.4% (95% CI: 33.7% to 59.8%). However, cross-protection against non-targeted types (HPV 31, 33, and 45) was not statistically significant.
For the 9 v-HPV vaccine, the overall VE against all seven targeted high-risk types (HPV 16/18/31/33/45/52/58) was 39.4% (95% CI: 20.9% to 53.5%). Notably, the 9 v-HPV vaccine conferred stronger protection against HPV 16/18 (VE: 54.4%; 95% CI: 26.5% to 71.7%) than against the other five types (HPV 31, 33, 45, 52, and 58), for which the VE was 34.0% (95% CI: 10.6% to 51.3%).
Vaccine effectiveness according to age at first dose
Stratified analysis by age at vaccination revealed age-dependent protection patterns (Table 3). The overall VE was highest in the 18–26 y age group, showing significant protection both at 1–<2 y (VE: 23.3%; 95% CI: 3.7% to 38.8%) and at ≥2 y after vaccination (VE: 45.7%; 95% CI: 25.0% to 60.7%). In contrast, among women vaccinated at ages 27–35 or 36–45 y, significant protective effects were only observable after ≥2 y, yielding VEs of 39.6% (95% CI: 23.5% to 52.4%) and 35.5% (95% CI: 12.6% to 52.4%), respectively.
Table 3.
Vaccine effectiveness of at least one dose of HPV vaccination against vaccine-type high-risk HPV DNA positivity stratified by age at vaccination of first dose.
| Cases n(%) | Matched Controls n(%) | Unadjusted VE%(95%CI) | P value | Adjusted¶ VE%(95%CI) |
P value | |
|---|---|---|---|---|---|---|
| Age 18–26 y | ||||||
| Time since vaccination of first dose | ||||||
| Not vaccinated | 1631 (87.9) | 1504 (80.0) | 1 (ref) | 1(ref) | ||
| <1 y | 137 (6.0) | 169 (9.0) | 12.6 (−3.5 to 26.2) | .119 | 14.3 (−1.5 to 27.7) | .075 |
| 1–2 y | 76 (4.3) | 114 (6.1) | 21.9 (2.1 to 37.7) | .032 | 23.3 (3.7 to 38.8) | .022 |
| >2 y | 37 (2.0) | 94 (5.0) | 44.9 (23.8 to 60.1) | <.001 | 45.7 (25.0 to 60.7) | <.001 |
| Age 27–35 y | ||||||
| Time since vaccination of first dose | ||||||
| Not vaccinated | 2254 (90.7) | 2099 (84.5) | 1 (ref) | 1(ref) | ||
| <1 y | 123 (4.6) | 136 (5.5) | 7.3 (−10.8 to 22.4) | .407 | 5.8 (−12.6 to 21.2) | .51 |
| 1–2 y | 76 (2.9) | 93 (3.7) | 20.4 (−1.8 to 37.8) | .069 | 19.6 (−2.9 to 37.2) | .083 |
| >2 y | 32 (2.7) | 157 (6.3) | 40.4 (24.4 to 53.0) | <.001 | 39.6 (23.5 to 52.4) | <.001 |
| Age 36–45 y | ||||||
| Time since vaccination of first dose | ||||||
| Not vaccinated | 2479 (92.0) | 2435 (90.4) | 1 (ref) | 1(ref) | ||
| <1 y | 112 (3.8) | 100 (3.7) | −3.2 (−24.3 to 14.4) | .745 | −4.6 (−26.2 to 13.2) | .635 |
| 1–2 y | 61 (2.6) | 71 (2.6) | 9.8 (−16.1 to 29.9) | .424 | 8.1 (−18.3 to 28.6) | .512 |
| >2 y | 42 (1.6) | 88 (3.3) | 36.9 (14.6 to 53.4) | .003 | 35.5 (12.6 to 52.4) | .005 |
VE: vaccine effectiveness.
$Not vaccinated was defined as no vaccination records or vaccinated within one month at the time of cervical cancer screening.
¶Controlling for marital status, number of pregnancies since vaccine introduction, educational level, ethnicity, residential sub-district, and housing type.
Sensitivity analyses
To assess potential residual confounding, we performed a negative control outcome analysis using non-vaccine-type HPV as the outcome. As shown in Table S1, the VE estimates against non-vaccine types (VEnvt) were not statistically significant in any of the post-vaccination intervals or age groups considered. Specifically, the adjusted VEnvt was −19.2% (95% CI: −67.7% to 15.2%) for <1 y, −0.2% (95% CI: −50.5% to 33.3%) for 1–<2 y, and −5.3% (95% CI: −52.9% to 27.5%) for ≥2 y after vaccination. While these non-significant findings suggest the absence of strong systematic bias, the wide confidence intervals indicate that some residual confounding cannot be entirely ruled out. Additionally, a second sensitivity analysis that restricted the control group exclusively to pan-HPV-negative individuals yielded VE estimates consistent with those of the primary analysis, confirming the robustness of the findings (Tables S2 and S3).
Discussion
This study provides real-world evidence from a low-coverage, high-prevalence setting, demonstrating that HPV vaccination is associated with a lower odds of vaccine-type hrHPV DNA positivity among adult Chinese women. The moderate yet statistically significant overall VE of 40.2% at ≥2 y post-vaccination supports a measurable virologic benefit. Despite attenuation with increasing age at vaccination, the protective effect remained significant even among women vaccinated at ages 36–45 y. These results provide real-world data relevant to catch-up vaccination strategies for adult women up to age 45.37 However, scaling up coverage to achieve immediate reductions in vaccine-type HPV prevalence faces substantial challenges, including a large target population and persistent barriers such as vaccine hesitancy, cost, and accessibility issues.
The results of this study provide several critical insights. First, the overall VE of 40.2% at ≥2 y observed in this study is lower than that reported in RCTs among Chinese women, in which the VE against incident HPV 16/18 infection in HPV-naïve populations was 61.6%.31 This disparity is likely attributable to the inclusion of women with pre-existing HPV infection in this real-world study, as well as differences in exposure risks compared with controlled settings. Nevertheless, these estimates align closely with those from international real-world studies that adjusted for time since vaccination. For example, a U.S. study using National Health and Nutrition Examination Survey (NHANES) data reported an overall VE of 58% among women aged 18–35 y at HPV testing, which decreased to 42% when the analysis was restricted to women tested within four years after vaccination, similar to our findings.38 The higher overall VE observed in the U.S. largely reflects a younger vaccination age (routine adolescent immunization) and a longer follow-up period allowing pre-existing infections to clear over time. Thus, the lower VE in our study cohort primarily reflects the impact of older age at vaccination and the lag time necessary for pre-existing infections to clear.
Second, although VE declined with increasing age at vaccination in this study, vaccination provided significant protection at ≥2 y post-vaccination even among women aged 27–35 y (VE: 39.6%) and 36–45 y (VE: 35.5%) at the time of vaccination. These estimates are notably higher than those reported for U.S. adult women vaccinated at an even younger age, among whom VE estimate was non-significant for those aged 28–35 y at HPV testing (vaccinated at age 23–26 y).38 This divergence may be partially attributed to cross-country differences in sexual behavior patterns. In the NHANES cohort, women initiated sexual activity at a relatively young age (median 17 y, IQR: 15–19 y), and 45% had at least three lifetime sexual partners, both of which increase the risk of HPV exposure prior to vaccination. In contrast, Chinese women typically have a later age at sexual debut and more conservative sexual behaviors,39 which may result in a lower baseline prevalence of cumulative HPV infection.3 Consequently, a larger proportion of adult Chinese women may remain HPV-naïve at the time of vaccination, thereby preserving higher vaccine effectiveness. Although these findings indicate that extending vaccination to women aged 18–45 y may be beneficial, the moderate VE (approximately 40%) underscores that protection is not absolute, particularly among women with higher exposure risk. Therefore, integrating vaccination with regular cervical cancer screening remains essential for comprehensive prevention in adult women.
Third, the VE in our study was negligible or limited within the first two years after vaccination but became more pronounced thereafter, a trend particularly evident in the older age groups. It is crucial to note that this pattern should not be interpreted as an increase in intrinsic VE over time. Rather, the delayed apparent effectiveness likely results from reduced misclassification of pre-existing infections. Since the vaccine prevents new infections but does not clear existing ones, and because natural clearance of HPV typically takes 6 to 24 months, pre-existing infections can mask the vaccine’s protective effect during the initial post-vaccination period. These findings suggest that in real-world evaluations of adult women with unknown baseline HPV infection status, a buffer period of at least two years may be methodologically important to accurately assess VE against incident-like infections.
Fourth, regarding vaccine valency, the VE estimate against HPV 16/18 was numerically higher for the 9 v-HPV vaccine (54.4%) than for the 2 v- and 4 v-HPV vaccines (48.4%). However, this difference should not be interpreted purely as a valency effect. Vaccine type in this real-world setting is heavily confounded by age at vaccination (9 v-HPV recipients were primarily aged 18–26 y), availability period, socioeconomic status, and potentially differences in dose completion rates.22 Furthermore, the effectiveness of the 9 v-HPV vaccine against the additional five HPV types (31, 33, 45, 52, and 58) was lower (34.0%) than that against HPV 16/18. This discrepancy may be driven by the epidemiology of HPV types 52 and 58, which are more prevalent in China and have longer clearance times than those of HPV 16/18.40 Consequently, a higher proportion of women may have harbored persistent infections of these types prior to vaccination, leading to an underestimation of VE within the two-year period immediately after vaccination. Extending the buffer period to three years increased the overall VE of the 9 v-HPV vaccine from 39.4% (95% CI: 20.9% to 53.5%) at ≥2 y post-vaccination to 52.9% (95% CI: 17.0% to 73.2%) at ≥3 y post-vaccination (data not shown), highlighting the need for longer observation periods when evaluating protection against these specific genotypes. Furthermore, this study did not show any significant cross-protection of the 2 v/4 v-HPV vaccines against HPV types 31, 33, and 45, possibly due to limited statistical power.
Limitations
This study has several limitations. First, the primary endpoint was based on a single-time-point HPV test, which reflects a surrogate virologic endpoint of recent hrHPV prevalence rather than persistent infection, HSIL, or cervical cancer. Further studies with longitudinal follow-up are needed to evaluate long-term clinical endpoints. Second, the lack of information on baseline HPV infection status precluded definitively distinguishing between pre-existing and incident infections, although we mitigated this limitation by applying a buffer period. Third, potential misclassification in the control group is possible. As this study relied on the results of a single HPV test, some HPV-negative controls may have had a recent HPV infection that had cleared naturally. However, given the low positivity rate of vaccine-type hrHPV and the use of a TND design, the impact of misclassification related to transient or recently cleared infections on VE estimation is likely to be minimal.41 Fourth, and importantly, direct measures of sexual behavior (e.g., age at sexual debut, number of sexual partners) were unavailable due to privacy constraints. Although we used screening type as a proxy, and our negative-control analysis showed no significant association with vaccination, the wide confidence intervals for the negative-control estimates indicate that we cannot definitively rule out meaningful residual confounding. Finally, the findings may not be generalizable to all Chinese women.
Conclusions
In summary, our real-world findings support the feasibility of catch-up vaccination for women aged 18–45 y to reduce the future risk of vaccine-type hrHPV DNA positivity. However, this outcome is a surrogate virologic endpoint and does not directly establish effectiveness against persistent infection, cervical precancer, or long-term clinical outcomes. Given these limitations and the continuing challenges of achieving universal vaccine coverage, adult women should continue regular cervical cancer screening after receiving the HPV vaccine.
Supplementary Material
Acknowledgments
We deeply appreciated the medical professionals and participants who have contributed to this database.
Conceptualization: BL, ZL, WY, JC, XL, WY, GL, YZ; Methodology: ZL, JC, XL, WY, GL, YZ; Data curation: BL, ZL, JC, WY, SL, JL, FH, XP, WS, LZ; Project administration: BL, ZL, JC, WY, WS, LZ, GL, YZ; Assisted with the analysis and writing-original draft preparation: BL, ZL, JC, WY, GL, YZ; Writing-review and editing: BL, ZL and YZ. All authors read and approved the final manuscript.
Biographies
Gang Liu is the Director of the Immunization Planning Division at the Shenzhen Center for Disease Control and Prevention, a Master’s Supervisor, and a Chief Physician. His primary work focuses on immunization program management and related policy research. He serves as a member of the National Immunization Program Technical Working Group of China.
Yuying Zhang is a Senior Researcher of the Department of Child Healthcare at the Longhua Maternal and Child Health Hospital in Shenzhen. She holds a postdoctoral degree, serves as a Master’s Supervisor, and holds the title of Associate Professor. Her primary research direction is epidemiology. Her main work focuses on child population health and vaccination.
Funding Statement
This study was supported by the Shenzhen Longhua District Medical Research Grant [No. 2024033], the Natural Science Foundation of Guangdong Province [No. 2023A1515010425], Shenzhen Science and Technology Innovation Program [No. JCYJ20220531091200001, JCYJ20210324123005014].
Disclosure statement
No potential conflict of interest was reported by the author(s).
During the preparation of this manuscript, the authors used Gemini 3.1 Pro and GPT-5.5 to polish the language, improve readability, and correct grammatical errors. After using this tool, the authors carefully reviewed and edited the content as needed and take full responsibility for the final content of the publication.
Data availability
The datasets generated and analyzed during the current study are not publicly available but are available from the corresponding author upon reasonable request.
Ethics approval and consent to participate
The study was approved by the Ethics Committee of Longhua District Maternity and Child Healthcare Hospital (approval number:SRE-PCFR/2024007) in accordance with the ethical standards of the 1964 Declaration of Helsinki and its subsequent amendments. All research data were kept strictly confidential. Written informed consent was waived owing to the retrospective study design.
Supplementary material
Supplemental data for this article can be accessed online at https://doi.org/10.1080/21645515.2026.2693440
References
- 1.de Martel C, Georges D, Bray F, Ferlay J, Clifford GM.. Global burden of cancer attributable to infections in 2018: a worldwide incidence analysis. Lancet Global Health. 2020;8(2):e180–14. doi: 10.1016/s2214-109x(19)30488-7. [DOI] [PubMed] [Google Scholar]
- 2.Drolet M, É B, Pérez N, Brisson M, Ali H, Boily M-C, Baldo V, Brassard P, Brotherton JML, Callander D, et al. Population-level impact and herd effects following the introduction of human papillomavirus vaccination programmes: updated systematic review and meta-analysis. Lancet (London, England) . 2019;394(10197):497–509. doi: 10.1016/s0140-6736(19)30298-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ellingson MK, Sheikha H, Nyhan K, Oliveira CR, Niccolai LM. Human papillomavirus vaccine effectiveness by age at vaccination: a systematic review. Hum Vaccin Immunother. 2023;19(2):2239085. doi: 10.1080/21645515.2023.2239085. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Zeglin RJ, Fetteroll JL. Human papillomavirus vaccine timing associated with eventual human papillomavirus diagnosis in women. Int J STD AIDS. 2020;31(10):976–981. doi: 10.1177/0956462420937168. [DOI] [PubMed] [Google Scholar]
- 5.Guillaume D, Waheed DE, Schleiff M, Muralidharan KK, Vorsters A, Limaye RJ. Global perspectives of determinants influencing HPV vaccine introduction and scale-up in low- and middle-income countries. PLOS ONE. 2024;19(1):e0291990. doi: 10.1371/journal.pone.0291990. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gu X, Sun G, Zheng R, Zhang S, Zeng H, Sun K, Wang S, Chen R, Wei W, He J. Incidence and mortality of cervical cancer in China in 2015. J Natl Cancer Cent. 2022;2(2):70–77. doi: 10.1016/j.jncc.2022.01.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Singh D, Vignat J, Lorenzoni V, Eslahi M, Ginsburg O, Lauby-Secretan B, Arbyn M, Basu P, Bray F, Vaccarella S. Global estimates of incidence and mortality of cervical cancer in 2020: a baseline analysis of the WHO Global cervical cancer elimination initiative. Lancet Global Health. 2023;11(2):e197–e206. doi: 10.1016/s2214-109x(22)00501-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Zhou L, Li Y, Wang H, Qin R, Han Z, Li R. Global cervical cancer elimination: quantifying the status, progress, and gaps. BMC Med. 2025;23(1):67. doi: 10.1186/s12916-025-03897-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zhao C, Zhao Y, Li J, Li M, Su Y, Mi X, La Tu SY, Shen D, Ren L, Li Y, et al. The eight-year long-term follow-up on the effectiveness of the quadrivalent human papillomavirus vaccine in Chinese women 20–45 years of age. Hum Vaccin Immunother. 2022;18(5):2052700. doi: 10.1080/21645515.2022.2052700. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lv H, Wang S, Liang Z, Yu W, Yan C, Chen Y, Hu X, Fu R, Zheng M, Group T, et al. Immunogenicity and safety of the 9-valent human papillomavirus vaccine in Chinese females 9–45 years of age: a phase 3 open-label study. Vaccine. 2022;40(23):3263–3271. doi: 10.1016/j.vaccine.2022.02.061. [DOI] [PubMed] [Google Scholar]
- 11.Qiao YL, Wu T, Li RC, Hu Y-M, Wei L-H, Li C-G, Chen W, Huang S-J, Zhao F-H, Li M-Q, et al. Efficacy, safety, and immunogenicity of an Escherichia coli-produced bivalent human papillomavirus vaccine: an interim analysis of a randomized clinical trial. J Natl Cancer Inst. 2020;112(2):145–153. doi: 10.1093/jnci/djz074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zhu FC, Zhong GH, Huang WJ, Chu K, Zhang L, Bi Z-F, Zhu K-X, Chen Q, Zheng T-Q, Zhang M-L, et al. Head-to-head immunogenicity comparison of an Escherichia coli-produced 9-valent human papillomavirus vaccine and Gardasil 9 in women aged 18–26 years in China: a randomised blinded clinical trial. Lancet Infect Dis. 2023;23(11):1313–1322. doi: 10.1016/s1473-3099(23)00275-x. [DOI] [PubMed] [Google Scholar]
- 13.Wang J, Song L. China’s integration of HPV vaccine into national immunization program: current challenges and future path. J Cancer Educ. 2025. doi: 10.1007/s13187-025-02758-0. [DOI] [PubMed]
- 14.Franceschi S, Clifford GM, Baussano I. Options for design of real-world impact studies of single-dose vaccine schedules. Vaccine. 2018;36(32):4816–4822. doi: 10.1016/j.vaccine.2018.02.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chen J, Zhang Z, Pan W, Song Y, Zheng L, Li L, Ye J, Cao L, Yu W. Estimated human papillomavirus vaccine coverage among females 9–45 years of age — China, 2017–2022. China CDC Wkly. 2024;6(19):413–417. doi: 10.46234/ccdcw2024.080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Chen Q, Wang S, Lin Q, Zhang G. Temporal changes in prevalence and geographical distribution of human papillomavirus infection in mainland China: a systematic review and analysis of 1.07 million females from 2010 to 2021. BMC Public Health. 2025;25(1):3083. doi: 10.1186/s12889-025-24323-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Oliveira CR, Ortiz AM, Sheth SS, Shapiro ED, Niccolai LM. Effectiveness of HPV vaccine by age at vaccination and number of doses: protocol for a population-based matched case–control study. BMJ Open. 2021;11(4):e043093. doi: 10.1136/bmjopen-2020-043093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Yang Y, Zhang L, Hartwig S, Jiang P, Zhao H, Meng R, Liu Z, Liu Z, Ding K, You X, et al. Post-marketing study design to evaluate the effectiveness of the 9-valent and 4-valent HPV vaccines on serious HPV-related cervical disease in China. Hum Vaccin Immunother. 2024;20(1):2418168. doi: 10.1080/21645515.2024.2418168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wang W, Kothari S, Baay M, Garland SM, Giuliano AR, Nygård M, Velicer C, Tota J, Sinha A, Skufca J, et al. Real-world impact and effectiveness assessment of the quadrivalent HPV vaccine: a systematic review of study designs and data sources. Expert Rev Vaccines. 2022;21(2):227–240. doi: 10.1080/14760584.2022.2008243. [DOI] [PubMed] [Google Scholar]
- 20.Kamolratanakul S, Niyom SL, Isarankura N, Ruengprasertkit C, Horthongkham N, Sripradit K, Dhitavat J, Muangnoicharoen S, Sunthornchart S, Pitisuttithum P. Long-term effectiveness of at least one dose of human papillomavirus vaccine in adolescents: a test-negative case–control study. Int J Infect Dis. 2026;165:108489. doi: 10.1016/j.ijid.2026.108489. [DOI] [PubMed] [Google Scholar]
- 21.Weng R, Zhang C, Wen L, Luo Y, Ye J, Wang H, Li J, Ning N, Huang J, Chen X, et al. Chlamydia-related knowledge, opinion to opportunistic testing, and practices of providers among different sexually transmitted infections related departments in hospitals in Shenzhen City, China. BMC Health Serv Res. 2022;22(1):601. doi: 10.1186/s12913-022-08012-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lin Z, Liang X, Su L, Peng W, Chen H, Fang Y, Chen S, Yang W, Chen W, Zhang L, et al. Coverage with the first dose of human papillomavirus vaccination among females aged 9–50 years in Shenzhen, China: a surveillance based on administrative health records in 2023. Vaccines . 2024;12(1). doi: 10.3390/vaccines12010075. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Gravitt PE. The known unknowns of HPV natural history. J Clin Invest. 2011;121(12):4593–4599. doi: 10.1172/jci57149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Branch of Cancer Prevention and Control, Chinese Preventive Medicine Association; Chinese Obstetrics and Gynecology Association Colposcopy and Cervical Neoplasia Committee; Chinese Society of Colposcopy and Cervical Pathology of China Health Birth Science Association; Beijing Medical Doctor (Technician) Society of Laboratory Medicine . Chinese expert consensus on the use of human papillomavirus nucleic acid testing for cervical cancer screening (2022). Zhonghua Yi Xue Za Zhi. 2023;103(16):1184–1195. doi: 10.3760/cma.j.cn112137-20230117-00096. [DOI] [PubMed] [Google Scholar]
- 25.Bogaards JA, van der Weele P, Woestenberg PJ, van Benthem BHB, King AJ, van Benthem BHB. Bivalent human papillomavirus (HPV) vaccine effectiveness correlates with phylogenetic distance from HPV vaccine types 16 and 18. J Infect Dis. 2019;220(7):1141–1146. doi: 10.1093/infdis/jiz280. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lamot M, Kirbiš A. Understanding vaccine hesitancy: a comparison of sociodemographic and socioeconomic predictors with health literacy dimensions. Vaccines . 2024;12(10):1141. doi: 10.3390/vaccines12101141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Barnard-Mayers R, Kouser H, Cohen JA, Tassiopoulos K, Caniglia EC, Moscicki A-B, Campos NG, Caunca MR, Seage GRS, Murray EJ. A case study and proposal for publishing directed acyclic graphs: the effectiveness of the quadrivalent human papillomavirus vaccine in perinatally HIV infected girls. J Clin Epidemiol. 2022;144:127–135. doi: 10.1016/j.jclinepi.2021.12.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Sauvageau C, Gilca V, Ouakki M, Kiely M, Coutlée F, Mathieu-Chartier S, Defay F, Lambert G. Sexual behavior, clinical outcomes and attendance of cervical cancer screening by HPV vaccinated and unvaccinated sexually active women. Hum Vaccin Immunother. 2021;17(11):4393–4396. doi: 10.1080/21645515.2021.1961470. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Tennant PWG, Murray EJ, Arnold KF, Berrie L, Fox MP, Gadd SC, Harrison WJ, Keeble C, Ranker LR, Textor J, et al. Use of directed acyclic graphs (DAGs) to identify confounders in applied health research: review and recommendations. Int J Epidemiol. 2021;50(2):620–632. doi: 10.1093/ije/dyaa213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Suzuki E, Shinozaki T, Yamamoto E. Causal diagrams: pitfalls and tips. J Epidemiol. 2020;30(4):153–162. doi: 10.2188/jea.JE20190192. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Zhu FC, Chen W, Hu YM, Hong Y, Li J, Zhang X, Zhang Y-J, Pan Q-J, Zhao F-H, Yu J-X, et al. Efficacy, immunogenicity and safety of the HPV-16/18 AS04-adjuvanted vaccine in healthy Chinese women aged 18–25 years: results from a randomized controlled trial. Intl J Cancer. 2014;135(11):2612–2622. doi: 10.1002/ijc.28897. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Hariri S, Schuler MS, Naleway AL, Daley MF, Weinmann S, Crane B, Newcomer SR, Tolsma D, Markowitz LE. Human papillomavirus VChinese experaccine effectiveness against incident genital warts among female health-plan enrollees, United States. Am J Epidemiol. 2018;187(2):298–305. doi: 10.1093/aje/kwx253. [DOI] [PubMed] [Google Scholar]
- 33.Kjaer SK, Dehlendorff C, Belmonte F, Baandrup L. Real-world effectiveness of human papillomavirus vaccination against cervical cancer. J Natl Cancer Inst. 2021;113(10):1329–1335. doi: 10.1093/jnci/djab080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Bao W, He X, Huang Y, Liu R, Li Z. The clinical effectiveness of single-dose human papillomavirus vaccination. Vaccines . 2024;12(9):956. doi: 10.3390/vaccines12090956. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Dema E, Shing JZ, Checchi M, Beddows S, Liu D, Sierra MS, Haas CB, Soldan K, Field N, Kreimer AR, et al. Developing, validating, and testing non-vaccine–preventable human papillomavirus to control for differences in sexual behavior when evaluating HPV vaccination. Cancer Epidemiol Biomarkers Prev. 2025;34(7):1093–1102. doi: 10.1158/1055-9965.Epi-24-1775. [DOI] [PubMed] [Google Scholar]
- 36.Etievant L, Sampson JN, Gail MH. Increasing efficiency and reducing bias when assessing HPV vaccination efficacy by using nontargeted HPV strains. Biometrics. 2023;79(2):1534–1545. doi: 10.1111/biom.13663. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Chinese Preventive Medicine Association Women’s Health Branch . Guidelines for comprehensive prevention and control of cervical cancer. Beijing: People’s Medical Publishing House; 2023. [Google Scholar]
- 38.Li C, Hall TG, Hall JJ, He WQ. Effectiveness of quadrivalent HPV vaccination in reducing vaccine-type and nonvaccine-type high risk HPV infection. Epidemiol Infect. 2023;151:e37. doi: 10.1017/s0950268823000213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Zhao FH, Tiggelaar SM, Hu SY, Xu L-N, Hong Y, Niyazi M, Gao X-H, Ju L-R, Zhang L-Q, Feng X-X, et al. A multi-center survey of age of sexual debut and sexual behavior in Chinese women: suggestions for optimal age of human papillomavirus vaccination in China. Cancer Epidemiol. 2012;36(4):384–390. doi: 10.1016/j.canep.2012.01.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Yang X, Li Y, Tang Y, Li Z, Wang S, Luo X, He T, Yin A, Luo M. Cervical HPV infection in Guangzhou, China: an epidemiological study of 198,111 women from 2015 to 2021. Emerg Microbes Infect. 2023;12(1):e2176009. doi: 10.1080/22221751.2023.2176009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Baum U, Kulathinal S, Auranen K. Exposure misclassification bias in the estimation of vaccine effectiveness. PLOS ONE. 2021;16(5):e0251622. doi: 10.1371/journal.pone.0251622. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated and analyzed during the current study are not publicly available but are available from the corresponding author upon reasonable request.