Monitoring the safety of the adjuvanted human papillomavirus vaccine HPV-16/18-AS04 in association with pregnancy and potential immune-related diseases in China: A retrospective observational cohort study Authors: Junting Yang, Adoración Navarro-Torné, Zhike Liu, Jinxin Guo, Siwei Deng, Ruogu Meng, Yu Yang, Rhed Evans, Amila Poddar, Huifeng Yun, Siyan Zhan Published: 2026 ABSTRACT ABSTRACTHuman papillomavirus (HPV) infection causes cervical cancer. The objective of the study is to monitor AS04-adjuvanted HPV-16/18 vaccine (Cervarix, GSK) safety in a real-world setting in China, using electronic health records from the Yinzhou Regional Health Information Platform (YRHIP). This retrospective observational cohort study included all permanent residents (males and females) aged 9–45 y with ≥12 months continuous YRHIP registration between 1 January 2010 and 31 December 2020. The study analyzed demographic information, diagnoses coded using International Classification of Diseases 10th Revision for potential immune-mediated diseases (pIMD) and pregnancy-related outcomes, and HPV vaccination records. Of 1,002,399 participants in the YRHIP, 891,463 met the eligibility criteria. In the age group 9–45 y, uptake of AS04-adjuvanted HPV-16/18 vaccine was 11.95/1000 persons. The incidence density for other spondylopathies, sarcoidosis, and pulmonary sarcoidosis was significantly higher in females vaccinated with any dose of AS04-adjuvanted HPV-16/18 vaccine compared with females who did not receive HPV vaccination during the study period (p < .01). No causal relationship could be established due to confounding factors and potential misclassification. The GSK safety database or EudraVigilance database were reviewed and no signals associated with these events were identified. For other pIMDs and adverse pregnancy-related outcomes, no increased incidence was found for vaccinated females. While certain pIMDs demonstrated an increased incidence in vaccinated individuals, no causal association was confirmed. Further research is needed to address residual confounding factors and validate findings. The overall safety profile of AS04-adjuvanted HPV-16/18 vaccination in Chinese girls and women aged 9–45 y could be considered as favorable.Clinical Trial Registration:Retrospectively registered on http://www.chinadrugtrials.org.cn/index.html, 22 March 2021, registration number CTR20210523.KEYWORDS: China, HPV infection, human papillomavirus, observational study, potential immune-mediated diseases, pregnancy, real-world, safety, vaccine Introduction Infection with human papillomavirus (HPV) can cause cervical cancer.1 In China, cervical cancer ranks as the sixth most common cancer in women, with an estimated 109,741 new cases annually in 2020.2 It is also the seventh leading cause of cancer-related deaths in women, resulting in approximately 59,060 deaths annually estimated in 2020.2 Among women aged 15–44 y in China, cervical cancer is the third most prevalent cancer and the third leading cause of cancer mortality.2 HPV-16 and HPV-18 are the two HPV types most commonly associated with invasive cervical cancer in China, estimated to account for a combined 75.6% of cases.3 The World Health Organization (WHO) recommends that HPV vaccination should be included in all national immunization programs, aiming to reach 90% of girls by the age of 15 y by 2030.4 The first HPV vaccine licensed in China was the AS04-adjuvanted HPV-16/18 vaccine (Cervarix, GSK).5 This was approved in July 2016 by the National Medical Products Administration (NMPA) for use in China as a two-dose vaccine schedule for women and girls aged 9–25 y. In 2018, the indication was extended for use in women aged up to 45 y.5 Clinical trials conducted in China have shown no significant differences in the incidence of adverse effects such as potential immune-mediated diseases (pIMDs) or pregnancy-related outcomes between participants vaccinated with the AS04-adjuvanted HPV-16/18 vaccine and unvaccinated participants, indicating an acceptable overall safety profile for the AS04-adjuvanted HPV-16/18 vaccine.6–8 A pooled analysis of data from 42 clinical studies also showed no increased relative risk in pIMDs and no difference in pregnancy-related outcomes between participants vaccinated with the AS04-adjuvanted HPV-16/18 vaccine and controls.9 Clinical trials typically have limited sample sizes and restricted duration of observation, which may limit their ability to detect potential rare events.5 Post-marketing safety surveillance helps address the limitations of clinical trials by providing real-world data on rare adverse events and long-term vaccine effects. A prospective post-marketing study was conducted in 2018–2020 in 3013 girls and women in China aged 9–45 y who had been vaccinated with at least one dose of HPV-16/18-AS04.10 That study reported no pIMDs considered related to vaccination and no safety concerns, although the authors commented that the sample size would preclude detection of rare adverse events.10 The Yinzhou District in China has a target to achieve 90% HPV vaccination in girls aged 9–14 y by the end of 2025.5 The Yinzhou Regional Health Information Platform (YRHIP) collects healthcare information on permanent residents and covers almost all healthcare-related activities from birth to death.11 The primary objective of this observational study was to monitor the safety of the AS04-adjuvanted HPV-16/18 vaccine in a real-world setting in China, using the YRHIP to construct a population cohort and following a stepwise approach to mitigate potential issues such as exposure uncertainty and event rates variability.5 The study had four objectives: (1) to evaluate the performance of the selected case identification method for the detection of five pIMDs of interest and two pregnancy-related outcomes of interest; (2) to estimate the pre-launch incidence of pIMDs and pregnancy-related outcomes and evaluate the uptake of AS04-adjuvanted HPV-16/18 vaccination among female participants aged 9–45 y; (3) to estimate the incidence of pIMDs and pregnancy outcomes pre- and post-launch; and (4) to compare the probability of pIMDs and pregnancy outcomes within 12 months of vaccination between vaccinated and unvaccinated participants, as well as pre-launch versus post-launch in unvaccinated participants, conditional on AS04-adjuvanted HPV-16/18 vaccine uptake and observed incidence. Participants and methods The study protocol has been described in detail in a previous publication5 and is summarized here. Study setting and design The study was conducted in Yinzhou, a district of Ningbo City, Zhejiang province, China.5 It was an observational retrospective cohort study using data from the YRHIP from the period 2010–2020. Study population We included all permanent residents (males and females) aged 9–45 y with at least 12 months of continuous registration in the YRHIP between 1 January 2010 and 31 December 2020. The observational period for pregnancy outcomes was 01 January 2012 to 31 December 2020. Male participants were included to account for potential time-confounding bias. As all eligible participants were included, the study size was determined by the database size. The pre-launch period was defined as 1 January 2010 to 31 December 2016, and the post-launch period as 1 January 2018 to the end of the study on 31 December 2020. The period 1 January to 31 December 2017 was considered to be an intermediate period. However, due to the small number of HPV vaccinations in the intermediate period, it was merged with the pre-launch period. The initial registry date was defined as the later of YRHIP registration date or 1 January 2010 for participants who were registered in Yinzhou District before 2016, and the later of YRHIP registration date or 1 January 2017 for participants assigned to Yinzhou District after 2016. Participants were followed to their date of death, 46th birthday, last visit before loss to follow-up, or last visit before the end of the study on 31 December 2020, whichever came first. Data sources and collection Demographic data (age, sex, marital status, education level) were extracted from the health archives within the YRHIP. Diagnoses coded using International Classification of Diseases 10th Revision (ICD-10) codes were extracted from the electronic health records. Vaccine name, manufacturer, batch number, dose, and time of vaccination were extracted from the vaccination records. Date and cause of death were extracted from the death register. Maternal characteristics, health behaviors, preexisting conditions, pregnancy history, obstetric complications, interventions, and birth outcomes were extracted from the maternal records. Endpoint definitions and measurement Endpoints, including pIMDs and pregnancy-related outcomes, were identified using ICD-10 codes and Chinese terms. The onset date of pIMDs was defined as the first diagnosis date, and subjects with a pIMD code in the 12 months before the onset date were excluded. The occurrence of pIMDs was monitored for 12 months after the last AS04-adjuvanted HPV-16/18 vaccination date or to the end of follow-up. Pregnancy outcomes were monitored for 6 months after the expected end of pregnancy. The pIMDs of interest for the assessment of case identification methods were autoimmune thyroiditis, type 1 diabetes mellitus (T1DM), rheumatoid arthritis (RA), inflammatory bowel disease (IBD), and idiopathic thrombocytopenic purpura (ITP). The pregnancy outcomes of interest for the assessment of case identification methods were spontaneous abortion and live infant congenital anomaly. Potential cases of each outcome of interest in the YRHIP identified by the ICD-10 codes were confirmed by independent medical record review by two experts according to internationally recognized guidelines in China. This paper reports overall data for the 9–45 y age group for the pregnancy-related outcomes and for the following pIMD endpoints at 12 months: neuralgia and neuritis, unspecified (ICD-10 code M79.2); other neuroinflammatory disorders (neuralgia and neuritis [unspecified], radiculopathy, neuritis, or radiculitis [ICD-10 code M79.2 vs M54.1]); other spondylopathies (ICD-10 code M48); sarcoidosis (ICD-10 code D86); and pulmonary sarcoidosis (ICD-10 code D86.0). Data for other pIMD endpoints, and for separate age groups can be found in the full report published in the GSK study register (https://www.gsk-studyregister.com/en/trial-details/?id=212380#documents-section). Statistical analysis The analysis used a stepwise approach to validate the case identification method and establish baseline incidence rates before performing comparative analyses, as described elsewhere.5 For database performance, cases of the pIMDs of interest and pregnancy outcomes of interest identified using ICD-10 codes were assessed using other data types in addition to ICD-10 codes, and a positive predictive value (PPV) with exact 95% confidence intervals (CI) was calculated for each outcome. The other data types used included disease terms in Chinese, specific medications and/or examination, and/or procedures, and/or narratives. For vaccine uptake, the annual uptake of AS04-adjuvanted HPV-16/18 vaccine (number and percentage of participants vaccinated with at least one dose) in the post-launch period was calculated in female participants aged 9–45 y. For each pIMD, the baseline incidence was calculated as the total number of new pIMD episodes divided by the sum of the number of person-years, with an exact Poisson 95% CI. For each pregnancy-related outcome, baseline incidence during the pre-launch period was calculated as the percentage of pregnancies with the pregnancy outcome, together with exact 95% CI. The incidences of pIMDs were calculated for the pre- and post-launch periods in female participants and in male participants, and pregnancy-related outcomes for the pre- and post-launch periods in female participants. For each pregnancy outcome, pre- and post-launch incidence was calculated as the number of pregnancies with the given outcome divided by the total number of women at risk, together with an exact Poisson 95% CI. In the comparative analysis, incidence in vaccinated and unvaccinated female participants was compared if ≥5 cases of any pIMD were observed within 12 months after AS04-adjuvanted HPV-16/18 vaccination. Probabilities and their associated 95% CIs were calculated if the sample size was sufficient to detect a relative risk of at least 2. A Poisson regression model adjusted for age and other possible covariates was used for associations between pIMD and AS04-adjuvanted HPV-16/18 vaccination. Logistic regression adjusted for age and other possible covariates was used for associations between pregnancy outcomes and AS04-adjuvanted HPV-16/18 vaccination. The sample size calculation has been described elsewhere.5 YRHIP data collection is mandatory, and therefore, core variables (e.g., vaccination records) had low rates of missing data (<1%) and were excluded if incomplete. Covariates with missing data (e.g., education level) were retained as a separate category in the analyses. All analyses were conducted using SAS 9.2 software or a higher version. Results Participant disposition and demographics Among 1,002,399 participants identified from the YRHIP, 891,463 met the eligibility criteria and were included in the analysis. Participant disposition and numbers in each analysis set are shown in Figure 1. Table 1 summarizes the demographic characteristics of study participants in the total, pIMD, vaccination, and pregnancy sets. No missing data were observed in core variables (e.g., vaccination status) due to the mandatory nature of YRHIP reporting. Participant disposition. ICD-10, International Classification of Diseases 10th Revision; N, number of participants identified by ICD-10 code; pIMD, potential immune-mediated disease; PY, person-year, calculated as sum of person-years for all participants in the category. Total set: eligible participants who met the inclusion criteria: (a) female or male aged between 9 and 45 y; (b) participants registered in the Yinzhou Regional Health Information Platform (YRHIP) between 01 January 2010 and 31 December 2020; (c) participants with at least 12 months of continuous registration in the YRHIP database over the study period, defined by minimum two or more observations recorded in the YRHIP at 12 months or more apart. pIMDs set: eligible participants from the Total set with pIMDs diagnosed. Pregnancy set: female pregnant participants from the Total set. Repeated pregnancies for one participant were counted separately. (A total of 80,049 female participants experienced 96,201 pregnancies.) Vaccination set: female participants from the Total set with at least one dose of any human papillomavirus (HPV) vaccine (including AS04-adjuvanted-HPV-16/18 vaccine and other HPV vaccines) administered. Sub-vaccination/pIMDs and Sub-vaccination/pregnancy outcome sets relate to the intersection of the pIMDs/pregnancy sets and Cervarix vaccination set, respectively. Table 1.Demographic characteristics at inclusion.ParameterTotal setpIMD setVaccination setPregnancy setTotalFemaleMaleTotalFemaleMaleTotalFemaleMaleFemaleTotal number891,463469,611421,85283,89444,50539,38920,37020,3373380,049aAge group at inclusion, yearsp-value, female vs male < .001 p-value, female vs male < .001 p-value, female vs male .932 9–1464,00530,55733,4483471172617453462345841915–1968,37134,04834,323495324752478288528796114220–24119,49265,84653,64697005052464833623356612,07325–29158,20391,60266,60115,2578620663747034694931,54830–34159,77385,10574,66819,49310,252924140594054520,69935–39158,46580,66777,79821,20111,18210,01915231521210,26740–45163,15481,78681,36898195198462137637514301Education levelp-value, female vs male < .001 p-value, female vs male < .001 p-value, female vs male .165 Illiterate and semi-illiterate38262364146230719711029290121Primary school and junior high school418,293222,971195,32236,57819,96416,614400239911123,810Senior high school/technical school149,41873,18876,23015,7697400836934473443414,342Junior college or above degree193,758108,80584,95323,20912,81910,39011,08511,0721333,547Unknown126,16862,28363,8858031412539061807180258229Marital statusp-value, female vs male < .001 p-value, female vs male < .001 p-value, female vs male < .001 Unmarried183,31983,74599,57412,5035382712169396932710,626Married608,888332,020276,86864,89035,54329,34712,19712,1821559,371Divorced/ widowed51453275187061938723272720317Unknown94,11150,57143,54058823193268911621151119735Year of inclusionbp-value, female vs male < .001 p-value, female vs male < .001 p-value, female vs male < .001 2010282,862145,801137,06139,10820,50618,602772877151322,196201198,92958,38240,54711,8336833500029132911211,6272012200,857103,70797,15013,7686986678213601358211,825201383,11441,93941,175509625522544730722810,004201422,42010,89411,526144369175231931902592201514,36368947469126561165439139101699201652182659255950327522824123838112017138,50773,85764,65096185262435652075205513,417201815,0508582646871446225257457402873201930,14316,89613,24754632721990790703005Female vs male compared using chi-square tests.aA total of 80,049 female participants experienced 96,201 pregnancies. The married status was investigated as participant registered in the Yinzhou Regional Health Information Platform (YRHIP), not until pregnancy.bParticipants required at least 12 months of continuous registration in the Yinzhou Regional Health Information Platform (YRHIP) database over the study period (01 January 2010 to 31 December 2020), thus 2019 was the last year of inclusion. The participants have been registered in the YRHIP from 01 January 2010 to 31 December 2020, but the observational period for pregnancy outcomes was 01 January 2012 to 31 December 2020 because the database started collecting data on maternal and child health as of January 2012. Inclusion criteria required at least 12 months of continuous registration in YRHIP database over the study period, thus the study participants are included only until Year 2019.pIMD, potential immune-mediated disease Demographic characteristics at inclusion. Female vs male compared using chi-square tests. aA total of 80,049 female participants experienced 96,201 pregnancies. The married status was investigated as participant registered in the Yinzhou Regional Health Information Platform (YRHIP), not until pregnancy. bParticipants required at least 12 months of continuous registration in the Yinzhou Regional Health Information Platform (YRHIP) database over the study period (01 January 2010 to 31 December 2020), thus 2019 was the last year of inclusion. The participants have been registered in the YRHIP from 01 January 2010 to 31 December 2020, but the observational period for pregnancy outcomes was 01 January 2012 to 31 December 2020 because the database started collecting data on maternal and child health as of January 2012. Inclusion criteria required at least 12 months of continuous registration in YRHIP database over the study period, thus the study participants are included only until Year 2019. pIMD, potential immune-mediated disease Uptake of AS04-adjuvanted HPV-16/18 vaccine In the pooled age group 9–45 y, the vaccination rate for the AS04-adjuvanted HPV-16/18 vaccine was 11.95 per 1000 persons (95% CI 11.57, 12.35). Algorithm performance The PPV was more than 70% for all algorithms for the five pIMDs of interest, except algorithm 2 (ICD-10 codes plus Chinese medical terms in outpatient setting) for RA. The final algorithm was ICD-10 codes plus Chinese medical terms in inpatient or outpatient settings for all the pIMDs of interest except IBD. For IBD, the final algorithm was: ICD-10 codes plus Chinese medical terms in inpatient and outpatient settings; or a set of Chinese medical terms in inpatient and outpatient settings, and ICD-10 code K51\ from an outpatient setting, and excluding a different set of Chinese medical terms from any discharge setting. The PPV was more than 70% for both pregnancy-related outcomes of interest, spontaneous abortion (90.84%) and live infant congenital anomaly (98.31%). Incidence of pregnancy-related outcomes pre- and post-launch Pre-launch, pooled across the 9–45 y age group, the pregnancy-related outcome with the highest reported incidence density was single live birth normal infant, in 2012 (Table 2). In the same age group post-launch, the pregnancy-related outcome with the highest reported incidence density was elective termination, in 2019 (Table 2). Table 2.Incidence of pregnancy-related outcomes in pregnant women aged 9–45 y, by calendar year.Outcome (ICD-10 code)YearNumber of casesNumber of pregnanciesIncidence per 1000 pregnant women (95% CI)Spontaneous abortion (O03) Pre-launch201236410,18935.72 (32.21, 39.51) 201340411,45435.27 (31.97, 38.81) 201453613,79238.86 (35.70, 42.22) 201568712,93253.12 (49.32, 57.13) 201674312,81957.96 (53.98, 62.15) 201766011,39557.92 (53.70, 62.37)Post-launch201872010,37169.42 (64.61, 74.49) 2019604744381.15 (75.04, 87.59) 2020500580686.12 (79.02, 93.63)Therapeutic abortion (O04) Pre-launch201246310,18945.44 (41.48, 49.67) 201353411,45446.62 (42.83, 50.64) 2014135513,79298.25 (93.33, 103.33) 2015183912,932142.21 (136.23, 148.34) 2016262112,819204.46 (197.51, 211.55) 2017261711,395229.66 (221.97, 237.50)Post-launch2018217610,371209.82 (202.01, 217.78) 201927247443365.98 (355.03, 377.04) 202021495806370.13 (357.70, 382.70)Ectopic pregnancy (O00) Pre-launch201212410,18912.17 (10.13, 14.49) 201319811,45417.29 (14.98, 19.84) 201436613,79226.54 (23.92, 29.36) 201550512,93239.05 (35.78, 42.53) 201662412,81948.68 (45.02, 52.55) 201758811,39551.60 (47.61, 55.82)Post-launch201848310,37146.57 (42.60, 50.81) 2019564744375.78 (69.86, 82.02) 2020389580667.00 (60.70, 73.74)Molar pregnancy (O01) Pre-launch2012810,189 20132111,4541.83 (1.14, 2.80) 20142613,7921.89 (1.23, 2.76) 20153312,9322.55 (1.76, 3.58) 20164912,8193.82 (2.83, 5.05) 20175311,3954.65 (3.49, 6.08)Post-launch20183410,3713.28 (2.27, 4.58) 20194374435.78 (4.18, 7.77) 20203758066.37 (4.49, 8.77)Elective termination (O04, O05, O06, P96.4) Pre-launch2012108710,189106.68 (100.75, 112.84) 2013123311,454107.65 (102.03, 113.47) 2014285013,792206.64 (199.91, 213.50) 2015311612,932240.95 (233.60, 248.42) 2016432712,819337.55 (329.36, 345.81) 2017400011,395351.03 (342.26, 359.88)Post-launch2018344910,371332.56 (323.49, 341.72) 201939177443526.27 (514.85, 537.66) 202027785806478.47 (465.55, 491.41)Missed abortion (O02.1) Pre-launch201217710,18917.37 (14.92, 20.10) 201321811,45419.03 (16.61, 21.70) 201445113,79232.70 (29.79, 35.80) 201550712,93239.21 (35.93, 42.69) 201665712,81951.25 (47.50, 55.21) 201761011,39553.53 (49.47, 57.82)Post-launch201856110,37154.09 (49.82, 58.62) 2019621744383.43 (77.25, 89.95) 2020399580668.72 (62.35, 75.53)Elective termination congenital anomaly (O04, O05, O06, P96.4 + Q00-Q99) Pre-launch2012010,189 2013011,454 2014013,792 2015012,932 2016112,819 2017011,395 Post-launch2018010,371 201907443 202005806 Single live birth normal infant (Z37.0) Pre-launch2012798810,189783.98 (775.86, 791.94) 2013886811,454774.23 (766.46, 781.86) 2014923213,792669.37 (661.45, 677.22) 2015805212,932622.64 (614.22, 631.01) 2016653012,819509.40 (500.71, 518.09) 2017544511,395477.84 (468.63, 487.06)Post-launch2018500710,371482.79 (473.13, 492.46) 201920247443271.93 (261.85, 282.20) 202019255806331.55 (319.45, 343.83)Still birth (P95) Pre-launch20121210,1891.18 (0.61, 2.06) 20131411,4541.22 (0.67, 2.05) 2014813,792 2015612,932 2016712,819 2017211,395 Post-launch20181210,3711.16 (0.60, 2.02) 20191174431.48 (0.74, 2.64) 202035806 Live infant congenital anomaly (Q00-Q99 [above] +P00) Pre-launch201219510,18919.14 (16.57, 21.99) 201326111,45422.79 (20.13, 25.69) 201422313,79216.17 (14.13, 18.42) 20153912,9323.02 (2.15, 4.12) 20162912,8192.26 (1.52, 3.25) 201717511,39515.36 (13.18, 17.79)Post-launch201822610,37121.79 (19.07, 24.79) 20196874439.14 (7.10, 11.57) 202066580611.37 (8.80, 14.44)Still birth congenital anomaly (P95 + Q00-Q99) Pre-launch2012010,189 2013011,454 2014013,792 2015012,932 2016012,819 2017011,395 Post-launch2018010,371 201907443 202005806 Premature birth (healthy) (P07) Pre-launch201230310,18929.74 (26.53, 33.22) 201336611,45431.95 (28.81, 35.34) 201436713,79226.61 (23.99, 29.43) 201535112,93227.14 (24.41, 30.09) 201632812,81925.59 (22.92, 28.47) 201737811,39533.17 (29.96, 36.63)Post-launch201829010,37127.96 (24.87, 31.32) 2019117744315.72 (13.02, 18.81) 2020138580623.77 (20.00, 28.02)Only 205 participants received human papillomavirus (HPV) vaccination in the intermediate period (01 January 2017 to 31 December 2017), so this period was merged with the pre-launch period.Incidence and CI were reported only for groups and categories for which the number of participants or the number of cases were ≥10.95% CI = Exact Poisson confidence interval.Incidence of pregnancy-related outcomes per 1000 people, calculated as n/N * 1000, among which n referred to the total number of gestation outcomes cases, and N referred to the number of pregnancies among the studied female population.CI, confidence interval; ICD-10, International Classification of Diseases 10th Revision. Incidence of pregnancy-related outcomes in pregnant women aged 9–45 y, by calendar year. Only 205 participants received human papillomavirus (HPV) vaccination in the intermediate period (01 January 2017 to 31 December 2017), so this period was merged with the pre-launch period. Incidence and CI were reported only for groups and categories for which the number of participants or the number of cases were ≥10. 95% CI = Exact Poisson confidence interval. Incidence of pregnancy-related outcomes per 1000 people, calculated as n/N * 1000, among which n referred to the total number of gestation outcomes cases, and N referred to the number of pregnancies among the studied female population. CI, confidence interval; ICD-10, International Classification of Diseases 10th Revision. Incidence of pIMDs pre- and post-launch Pre-launch, pooled across the 9–45 y age group, the pIMD with the highest reported incidence density was neuralgia and neuritis [unspecified], radiculopathy, neuritis, or radiculitis, in 2015 (Table 3). In the same age group post-launch, the pIMD with the highest reported incidence density was other spondylopathies, in 2019 (Table 3). Table 3.Incidence density of pIMDs in participants aged 9–45 y, by calendar year.pIMD (ICD-10 code)YearMale Female Cases (n)Person-yearsIncidence density95% CICases (n)Person-yearsIncidence density95% CINeuralgia and neuritis, unspecified (M79.2) Pre-launch2011178156,075.22114.0597.92, 132.08229168,556.27135.86118.84, 154.63 2012175210,102.9183.2971.41, 96.58247233,275.23105.8893.10, 119.94 2013244271,192.4789.9779.04, 101.99290302,308.4195.9385.21, 107.62 2014930294,255.00316.05296.09, 337.001280326,650.44391.86370.72, 413.88 20151307291,835.88447.85423.95, 472.751716322,961.53531.33506.55, 557.01 20161031234,165.52440.29413.87, 467.941269257,467.45492.88466.19, 520.69 2017880249,773.42352.32329.46, 376.341083276,907.59391.11368.20, 415.06Post-launch20181147246,617.53465.09438.62, 492.741376275,171.63500.05474.04, 527.12 20191065244,210.83436.10410.35, 463.031350275,640.63489.77464.05, 516.54 2020885240,440.98368.07344.26, 393.101053274,437.31383.69360.91, 407.54Neuralgia and neuritis (unspecified), Radiculopathy, Neuritis or radiculitis: brachial NOS, lumbar NOS, lumbosacral NOS, thoracic NOS, Radiculitis NOS (M79.2 vs M54.1) Pre-launch2011268156,075.22171.71151.78, 193.53353168,556.27209.43188.17, 232.43 2012265210,102.91126.13111.41, 142.25394233,275.23168.90152.64, 186.41 2013300271,192.47110.6298.46, 123.87392302,308.41129.67117.16, 143.15 2014949294,255.00322.51302.34, 343.661299326,650.44397.67376.38, 419.85 20151310291,835.88448.88424.95, 473.811719322,961.53532.26507.46, 557.96 20161029234,165.52439.43413.04, 467.061268257,467.45492.49465.81, 520.29 2017880249,773.42352.32329.46, 376.341092276,907.59394.36371.35, 418.41Post-launch20181153246,617.53467.53440.98, 495.241378275,171.63500.78474.74, 527.86 20191070244,210.83438.15412.34, 465.141349275,640.63489.41463.69, 516.17 2020897240,440.98373.06349.09, 398.251061274,437.31386.61363.73, 410.54Other spondylopathies (M48) Pre-launch2011220156,075.22140.96122.96, 160.85413168,556.27245.02221.98, 269.80 2012239210,102.91113.7599.79, 129.12423233,275.23181.33164.47, 199.44 2013227271,192.4783.7073.17, 95.33472302,308.41156.13142.37, 170.86 2014662294,255.00224.97208.18, 242.761008326,650.44308.59289.86, 328.21 2015674291,835.88230.95213.86, 249.041099322,961.53340.29320.50, 360.98 2016590234,165.52251.96232.06, 273.11938257,467.45364.32341.41, 388.35 20171034249,773.42413.98389.17, 439.941468276,907.59530.14503.43, 557.89Post-launch20181429246,617.53579.44549.86, 610.192021275,171.63734.45702.89, 767.06 20191556244,210.83637.15605.98, 669.512120275,640.63769.12736.84, 802.44 20201063240,440.98442.10415.98, 469.441505274,437.31548.39521.11, 576.74Sarcoidosis (D86) Pre-launch201114156,075.228.974.90, 15.0525168,556.2714.839.60, 21.89 201277210,102.9136.6528.92, 45.8071233,275.2330.4423.77, 38.39 201379271,192.4729.1323.06, 36.3083302,308.4127.4621.87, 34.03 2014117294,255.0039.7632.88, 47.65123326,650.4437.6531.30, 44.93 2015217291,835.8874.3664.80, 84.93248322,961.5376.7967.53, 86.96 2016277234,165.52118.29104.78, 133.07303257,467.45117.68104.81, 131.70 2017323249,773.42129.32115.61, 144.21341276,907.59123.15110.43, 136.93Post-launch2018606246,617.53245.72226.57, 266.06647275,171.63235.13217.36, 253.94 2019892244,210.83365.26341.72, 389.991098275,640.63398.34375.17, 422.57 2020912240,440.98379.30355.12, 404.691271274,437.31463.13438.07, 489.24Pulmonary sarcoidosis (D86.0) Pre-launch201113156,075.228.334.44, 14.2421168,556.2712.467.71, 19.04 201267210,102.9131.8924.71, 40.5059233,275.2325.2919.25, 32.62 201372271,192.4726.5520.77, 33.4375302,308.4124.8119.51, 31.10 2014109294,255.0037.0430.42, 44.68100326,650.4430.6124.91, 37.23 2015207291,835.8870.9361.60, 81.28213322,961.5365.9557.39, 75.43 2016270234,165.52115.30101.97, 129.90289257,467.45112.2599.68, 125.95 2017304249,773.42121.71108.42, 136.18319276,907.59115.20102.91, 128.55Post-launch2018584246,617.53236.80218.01, 256.79602275,171.63218.77201.66, 236.95 2019881244,210.83360.75337.36, 385.341072275,640.63388.91366.02, 412.86 2020894240,440.98371.82347.88, 396.961233274,437.31449.28424.60, 475.02Only 205 participants received human papillomavirus (HPV) vaccination in the intermediate period (01 January 2017 to 31 December 2017), so this period was merged with the pre-launch period.Inclusion criteria required at least 12 months of continuous registration in the Yinzhou Regional Health Information Platform (YRHIP) database over the study period (01 January 2010 to 31 December 2020), so the earliest that a pIMD case could be identified to meet the criteria would be in 2011.Cases: Denotes the number of cases (n) in the respective category in the pIMDs set.Incidence density per 100,000 person-year for the given category calculated as n/person-years *100,000, in which person-years was the sum of person-years for all participants in the category. For example, if the number of cases (n) is 8, and person-years = 80,000, incidence density would be shown as 10.95% CI = Exact Poisson confidence interval for incidence density (per 100,000 person-years).CI, confidence interval; ICD-10, International Classification of Diseases 10th Revision; NOS, not otherwise specified; pIMD, potential immune-mediated disease. Incidence density of pIMDs in participants aged 9–45 y, by calendar year. Only 205 participants received human papillomavirus (HPV) vaccination in the intermediate period (01 January 2017 to 31 December 2017), so this period was merged with the pre-launch period. Inclusion criteria required at least 12 months of continuous registration in the Yinzhou Regional Health Information Platform (YRHIP) database over the study period (01 January 2010 to 31 December 2020), so the earliest that a pIMD case could be identified to meet the criteria would be in 2011. Cases: Denotes the number of cases (n) in the respective category in the pIMDs set. Incidence density per 100,000 person-year for the given category calculated as n/person-years *100,000, in which person-years was the sum of person-years for all participants in the category. For example, if the number of cases (n) is 8, and person-years = 80,000, incidence density would be shown as 10. 95% CI = Exact Poisson confidence interval for incidence density (per 100,000 person-years). CI, confidence interval; ICD-10, International Classification of Diseases 10th Revision; NOS, not otherwise specified; pIMD, potential immune-mediated disease. Incidence of pIMDs and pregnancy-related outcomes by vaccination status In female participants vaccinated with AS04-adjuvanted HPV-16/18, the pIMD with the highest reported incidence density was other spondylopathies, followed by sarcoidosis and pulmonary sarcoidosis (Table 4). Table 4.Incidence of pIMDs in participants aged 9–45 y by vaccination status.pIMD (ICD-10 code)Vaccination statusNumberPerson-yearsCases (n)Incidence density per 100,000 person-years95% CIa lower limit95% CIa upper limitNeuralgia and neuritis, unspecified (M79.2)Received AS04-adjuvanted HPV-16/18 within 12 months3541306915488.79273.82804.91 Received AS04-adjuvanted-HPV 16/18 more than 12 months ago146552661140.93419.822466.69 No HPV vaccine449,2742,683,3459202342.93335.97350Neuralgia and neuritis (unspecified), Radiculopathy, Neuritis or radiculitis: brachial NOS, lumbar NOS, lumbosacral NOS, thoracic NOS, Radiculitis NOS (M79.2 vs M54.1)Received AS04-adjuvanted HPV-16/18 within 12 months3541306915488.79273.82804.91 Received AS04-adjuvanted HPV-16/18 more than 12 months ago146552661140.93419.822466.69 No HPV vaccine449,2742,683,3459605357.95350.84365.17Other spondylopathies (M48)Received AS04-adjuvanted HPV-16/18 within 12 months35413069351140.52795.661582.65 Received AS04-adjuvanted HPV-16/18 more than 12 months ago1465526163042.481748.844893.87 No HPV vaccine449,2742,683,34510,547393.05385.6400.61Sarcoidosis (D86)Received AS04-adjuvanted HPV-16/18 within 12 months35413069311010.18687.371430.82 Received AS04-adjuvanted HPV-16/18 more than 12 months ago1465526112091.711048.693711.77 No HPV vaccine449,2742,683,3453725138.82134.4143.35Pulmonary sarcoidosis (D86.0)Received AS04-adjuvanted HPV-16/18 within 12 months35413069311010.18687.371430.82 Received AS04-adjuvanted HPV-16/18 more than 12 months ago1465526101901.55915.543469.15 No HPV vaccine449,2742,683,3453516131.03126.74135.43CI, confidence interval; HPV, human papillomavirus; ICD-10, International Classification of Diseases 10th Revision; NOS, not otherwise specified; pIMD, potential immune-mediated disease.aExact Poisson confidence interval. Incidence of pIMDs in participants aged 9–45 y by vaccination status. CI, confidence interval; HPV, human papillomavirus; ICD-10, International Classification of Diseases 10th Revision; NOS, not otherwise specified; pIMD, potential immune-mediated disease. aExact Poisson confidence interval. Live birth was the pregnancy-related outcome with the highest incidence density in women vaccinated with AS04-adjuvanted HPV-16/18 (Table 5). Table 5.Incidence of pregnancy-related outcomes in pregnant women aged 9–45 y by vaccination status.Outcome (ICD-10 code)Vaccination statusNCases (n)Incidence95% CIper 1000Lower limitUpper limitSpontaneous abortion (O03) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy4433272.2351.54105.82 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,758518654.1652.7455.78Therapeutic abortion (O04) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy443134302.48261.49392.24 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,75816,344170.68168.31174.07Ectopic pregnancy (O00) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy4432351.9234.7480.43 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,758381839.8738.6541.24Molar pregnancy (O01) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy443613.546.1030.54 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,7582983.112.783.49Elective termination (O04, O05, O06, P96.4) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy443178401.81357.12522.15 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,75826,579277.56274.74282.62Missed abortion (O02.1) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy4433681.2659.18117.01 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,758416543.5042.2244.93Elective termination congenital anomaly (O04, O05, O06, P96.4 + Q00-Q99) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy44300.000.0016.09 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,75810.010.000.07Live birth (Z37.0) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy443180406.32361.51528.21 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,75854,891573.23570.09584.88Still birth (P95) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy44300.000.0016.09 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before95,758750.780.620.98Live infant congenital anomaly (Q00-Q99 (above) +P00) Received AS04-adjuvanted-HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy443715.807.5533.62 Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy443715.807.5533.62 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,758127513.3112.6114.08Still birth congenital anomaly (P95 + Q00-Q99) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy44300.000.0016.09 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,75800.000.000.07Premature birth (healthy) (P07) Received AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy4431431.6018.8054.53 Did not receive AS04-adjuvanted HPV-16/18 vaccine during pregnancy or within 60 d before pregnancy95,758262427.4026.3928.5295% CI = Exact Confidence Interval.N, number of pregnancies among female participants; n, number of cases in the respective category.CI, confidence interval; HPV, human papillomavirus; ICD-10, International Classification of Diseases 10th Revision. Incidence of pregnancy-related outcomes in pregnant women aged 9–45 y by vaccination status. 95% CI = Exact Confidence Interval. N, number of pregnancies among female participants; n, number of cases in the respective category. CI, confidence interval; HPV, human papillomavirus; ICD-10, International Classification of Diseases 10th Revision. Probability of occurrence of pIMD within 12 months after AS04-adjuvanted HPV-16/18 vaccination, by vaccination status There was no significant difference in the incidence density for neuralgia and neuritis, unspecified and other neuroinflammatory disorders (neuralgia and neuritis [unspecified], radiculopathy, neuritis, or radiculitis) within 12 months of vaccination in female participants vaccinated with any dose of AS04-adjuvanted HPV-16/18 vaccine, compared with female participants who did not receive HPV vaccination during the study period (Table 6). The incidence density for other spondylopathies, sarcoidosis, and pulmonary sarcoidosis was significantly higher in female participants vaccinated with any dose of AS04-adjuvanted HPV-16/18 vaccine compared with female participants who did not receive HPV vaccination during the study period (p < .01) (Table 6). The incidence density of sarcoidosis and pulmonary sarcoidosis significantly increased over time in the post-launch period (2018–2020). Table 6.Probability of pIMDs in participants aged 9–45 y who received HPV-16/18-AS04 vaccination, compared with those not receiving vaccination.pIMD (ICD-10 code)AS04-adjuvanted HPV-16/18 vaccination No HPV vaccination (reference group) Univariate regression Multivariable regression numberPerson-yearsnumberPerson-yearsCoefficientp-valueRR (95% CI)Coefficientp-valueaRR (95% CI)Neuralgia and neuritis, unspecified (M79.2)1530763413765,9270.09.731.09 (0.66, 1.82)0.22.401.24 (0.75, 2.07)Neuralgia and neuritis (unspecified), Radiculopathy, Neuritis or radiculitis: brachial NOS, lumbar NOS, lumbosacral NOS, thoracic NOS, Radiculitis NOS (M79.2 vs M54.1)1530763421765,9270.09.731.09 (0.66, 1.81)0.22.411.24 (0.75, 2.06)Other spondylopathies (M48)3530765033765,9270.55<.011.73 (1.24, 2.41)0.65<.011.92 (1.38, 2.68)Sarcoidosis (D86)3130762607765,9271.09<.012.96 (2.08, 4.22)1.03<.012.81 (1.97, 4.01)Pulmonary sarcoidosis (D86.0)5530762507765,9271.7<.015.46 (4.18, 7.14)1.64<.015.15 (3.94, 6.74)aRR, adjusted rate ratio from multivariable regression; CI, confidence interval; HPV, human papillomavirus; ICD-10, International Classification of Diseases 10th Revision; NOS, not otherwise specified; pIMD, potential immune-mediated disease; RR, rate ratio from univariate regression. Probability of pIMDs in participants aged 9–45 y who received HPV-16/18-AS04 vaccination, compared with those not receiving vaccination. aRR, adjusted rate ratio from multivariable regression; CI, confidence interval; HPV, human papillomavirus; ICD-10, International Classification of Diseases 10th Revision; NOS, not otherwise specified; pIMD, potential immune-mediated disease; RR, rate ratio from univariate regression. Due to the low disease incidence, no pIMD related to vasculitis, skin disorders, blood disorders, liver disorders, gastrointestinal disorders, endocrine disorders, and kidney disorders met the prespecified sample size threshold for the conditional analysis comparing probability of occurrence of these pIMDs among vaccinated and unvaccinated participants. Probability of occurrence of pregnancy-related outcomes, by vaccination status Women vaccinated with any dose of AS04-adjuvanted HPV-16/18 vaccine during or within 60 d before pregnancy had a significantly higher likelihood of live birth and a significantly lower likelihood of therapeutic abortion and elective termination compared with women who did not receive any HPV vaccine during or within 60 d before pregnancy (Table 7). There was no significant difference in the incidence rate for other pregnancy outcomes by AS04-adjuvanted HPV-16/18 vaccination status (Table 7). Table 7.Probability of pregnancy-related outcomes in pregnant women aged 9–45 y who received HPV-16/18-AS04 vaccination, compared with those not receiving vaccination.Outcome (ICD-10 code)AS04-adjuvanted HPV-16/18 vaccination No HPV vaccination (reference group) Univariate regression Multivariable regression CasesNumberCasesNumberCoefficientp-valueRR (95% CI)Coefficientp-valueaRR (95% CI)Spontaneous abortion (O03)32443179223,177−0.07.7050.93 (0.65, 1.34)−0.15.4300.86 (0.60, 1.24)Therapeutic abortion (O04)134443691523,1770.02.8611.02 (0.83, 1.25)−0.23.0310.80 (0.65, 0.98)Ectopic pregnancy (O00)23443141323,177−0.17.440.84 (0.55, 1.29)−0.27.2240.77 (0.50, 1.18)Elective termination (O04, O05, O06, P96.4)178443996623,177−0.12.2350.89 (0.74, 1.08)−0.31<.010.74 (0.60, 0.90)Missed abortion (O02.1)36443154523,1770.21.2241.24 (0.88, 1.75)0.16.3901.17 (0.82, 1.66)Live birth (Z37.0)180443877623,1770.12.2371.12 (0.93, 1.36)0.33<.011.39 (1.13, 1.70)Live infant congenital anomaly (Q00-Q99 +P00)744335323,1770.04.9281.04 (0.49, 2.21)0.27.5021.31 (0.61, 2.82)Premature birth (healthy) (P07)1444353123,1770.33.2321.39 (0.81, 2.39)0.37.1921.44 (0.83, 2.50)aRR, adjusted rate ratio from multivariable regression; CI, confidence interval; HPV, human papillomavirus; ICD-10, International Classification of Diseases 10th Revision; RR, rate ratio from univariate regression.Exact logistic confidence interval.The logistic model was applied for each pregnancy outcome separately, in which multivariable analysis included AS04-adjuvanted HPV-16/18 vaccination as exposure variable, each pregnancy outcome as outcome variable, and age at onset date and calendar year as confounding factors. Satisfying the conditions of RR > 1 and p < .05 simultaneously would indicate significantly increased risk of pregnancy outcome compared to reference group. Probability of pregnancy-related outcomes in pregnant women aged 9–45 y who received HPV-16/18-AS04 vaccination, compared with those not receiving vaccination. aRR, adjusted rate ratio from multivariable regression; CI, confidence interval; HPV, human papillomavirus; ICD-10, International Classification of Diseases 10th Revision; RR, rate ratio from univariate regression. Exact logistic confidence interval. The logistic model was applied for each pregnancy outcome separately, in which multivariable analysis included AS04-adjuvanted HPV-16/18 vaccination as exposure variable, each pregnancy outcome as outcome variable, and age at onset date and calendar year as confounding factors. Satisfying the conditions of RR > 1 and p < .05 simultaneously would indicate significantly increased risk of pregnancy outcome compared to reference group. Discussion This observational cohort study used data from the YRHIP database to assess the incidence of pIMDs and pregnancy-related outcomes in Chinese women and girls aged 9–45 y pre- and post-launch of the AS04-adjuvanted HPV-16/18 vaccine. The results showed a statistically significant increase in incidence density for other spondylopathies, sarcoidosis, and pulmonary sarcoidosis in women vaccinated with AS04-adjuvanted HPV-16/18 compared with women who did not receive HPV vaccination during the study period. There were no significant differences in the incidence of neuralgia and neuritis (unspecified) and other neuroinflammatory disorders. No causal relationship could be established, as outcomes were not validated (except for the five pIMDs assessed for the case-identification algorithm), and potential misclassification of diagnoses cannot be excluded. Furthermore, the association between the event and exposure to vaccination was not assessed by a physician. There is a theoretical possibility that the immunostimulatory adjuvant system AS04 could have the potential to activate previously unrecognized autoimmune disease in susceptible individuals.9,12,13 However, no such safety signal has been observed in the data collected to date from the clinical trial program or post-marketing surveillance.12,14,15 The incidence of autoimmune disease may be influenced by demographic factors, such as age and sex, and may change over time. Many autoimmune diseases are more common in women than men.16 Potential underlying mechanisms may include the presence of two X chromosomes, stimulation of the immune system by estrogen, or female-biased expression of immune regulators such as vestigial-like family member 3.17,18 The incidence of autoimmune diseases such as IBD and RA increased from 1990 to 2019 in China, with incidence increasing with age.19 This may reflect a variety of factors, such as increased industrialization, urbanization, and exposure to environmental pollution.20 Ankylosing spondylitis, an inflammatory rheumatic disorder that causes back pain, has increased in prevalence in China in recent decades.21,22 Sarcoidosis is characterized by the formation of immune granulomas in many organs, including the lungs (pulmonary sarcoidosis). The cause is unknown, although genetic susceptibility, environmental factors, and autoimmune activation have been suggested as potential factors,23 and a study in China reported an association with autoimmune disease.24 Pulmonary sarcoidosis and interstitial lung disease increased in incidence worldwide between 1990 and 2019, with an estimated 24.2 million new cases in 2019.25,26 This increase may reflect a number of factors: including aging of the population, because incidence increases with age; environmental factors such as air pollution, occupational exposure and smoking; and changes in diagnostic guidelines and improvements in diagnostic techniques.26 The observed increase in disease burden may reflect demographic trends.26 A safety assessment was conducted using the GSK safety database to evaluate whether the results of the present study could represent a potential safety signal. Two reports of sarcoidosis, two of spondyloarthropathy, and three cases of collapsed vertebra were retrieved from the GSK safety database. These events were not identified as having either disproportionate reporting or unexpected time-to-onset distribution, and the individuals concerned had underlying risk factors which suggested that a causal association with the AS04-adjuvanted HPV-16/18 vaccine was unlikely. No cases of pulmonary sarcoidosis, spinal stenosis, diffuse idiopathic skeletal hyperostosis, kissing spine, spondylopathy traumatic, and stress fracture were reported in the GSK safety database. Furthermore, reports from the EudraVigilance database identified no cases of pulmonary sarcoidosis, spinal stenosis, diffuse idiopathic skeletal hyperostosis, kissing spine, spondylopathy traumatic, stress fracture (fatigue fracture), and spinal compression fracture (collapsed vertebra), and disproportionate reporting was not identified for sarcoidosis and spondyloarthropathy. Although this study reported a statistically significant association with pulmonary sarcoidosis in women vaccinated with AS04-adjuvanted HPV-16/18, this cannot be considered as a safety signal based on the review of data currently available to GSK, including clinical study data and spontaneous report data in the GSK safety database. No change in event frequency, distribution (age and country), duration, severity, or outcome was noted, and therefore, it can only be considered as a statistical finding. No causal relationship could be established due to confounding factors and potential misclassification. The company concluded that the available evidence did not suggest a new potentially causal association, or a new aspect of a known association, of the AS04-adjuvanted HPV-16/18 vaccine with the events of sarcoidosis, pulmonary sarcoidosis, and other spondyloarthropathies, and therefore, these events do not warrant further investigation. This study had several strengths. Vaccination data were accessible through linkage in the same YRHIP platform, allowing ascertainment of exposure. It used electronic health records covering 95% of permanent residents of Yinzhou District to monitor vaccine safety in a large sample in a real-world setting in China and assessed ICD-10 codes against other source data as a method of case identification for selected pIMDs and pregnancy-related outcomes of interest. It used a stepwise approach, as described elsewhere,5 to mitigate some of the inherent limitations of a retrospective database study. As with any retrospective database study, the present study also had some limitations. The uptake rate of the AS04-adjuvanted HPV-16/18 vaccine in China is low, and thus the sample size was insufficient to perform comparative analyses on more than a few outcomes. This low AS04-adjuvanted HPV-16/18 vaccine uptake rate resulted in a relatively small sample of vaccinated female participants. This may have limited the ability to detect rare outcomes, as the incidence rate of autoimmune diseases is 1–50 per 100,000 person-years.13 Yinzhou District is an affluent and developed region of China, and the results may not be generalizable to the entire Chinese population. Differences in healthcare-seeking behavior (selection bias) may confound results, e.g., if vaccinated participants are generally more health-conscious and more likely to seek healthcare (healthy vaccinee bias), or if people with existing health concerns are disproportionately more likely to seek vaccination (confounding by indication).27 Lifestyle and socioeconomic factors are recognized as potential confounders in observational research, such as improvements in healthcare services, physician availability and diagnostic capabilities, lifestyle changes, improvements in awareness and training among healthcare practitioners, and improvements in reporting, and the study could not adjust for these. Demographic changes, such as increases in disease incidence over time associated with aging of the population and/or environmental change, could also confound the results if they affect the reported incidence of outcomes over time. Apart from the specific outcomes of interest used for assessment of case identification, the ICD-10 codes were not validated, leading to the possibility of misclassification. Finally, there was no physician assessment of association between event and vaccine exposure, so causal relationships between vaccination and outcomes could not be established. This retrospective observational study evaluated the incidence of pIMDs and pregnancy-related outcomes pre- and post-launch of the AS04-adjuvanted HPV-16/18 vaccine in Yinzhou District, Ningbo, China. Compared with women who did not receive any HPV vaccine, women vaccinated with AS04-adjuvanted HPV-16/18 vaccine had a higher likelihood of live birth (adjusted rate ratio from multivariable regression 1.39; 95% CI 1.13, 1.70) and also had a lower likelihood of therapeutic abortion and elective termination. There was no significant difference in the likelihood for other pregnancy-related outcomes by AS04-adjuvanted HPV-16/18 vaccination status. The incidence of sarcoidosis, pulmonary sarcoidosis, and other spondylopathies was significantly higher in female participants vaccinated with AS04-adjuvanted HPV-16/18 vaccine compared with female participants who had not received HPV vaccination. However, as the study did not validate diagnoses other than the small number of endpoints used for algorithm evaluation, no causal relationship could be established due to confounding factors and potential misclassification. In light of these results, the GSK safety database and EudraVigilance database were reviewed and no signals associated with these events were identified. For other pIMDs and adverse pregnancy related outcomes, no increased incidence was found for vaccinated females in the study. While certain pIMDs demonstrated an increased incidence in vaccinated individuals, no causal association was confirmed. Further research is needed to address residual confounding factors and validate findings. The overall safety profile of AS04-adjuvanted HPV-16/18 vaccination in Chinese girls and women aged 9–45 y remains favorable. Conclusions In this retrospective observational study evaluating the safety of the AS04-adjuvanted HPV-16/18 vaccine using electronic health records from the YRHIP, the incidence of sarcoidosis, pulmonary sarcoidosis, and other spondylopathies was found to be significantly higher in female participants vaccinated with AS04-adjuvanted HPV-16/18 vaccine compared with female participants who had not received HPV vaccination. However, no causal relationship could be established due to the inability to assess confounding factors and potential misclassification of outcomes. In addition, no signals associated with these events were identified in the GSK safety database and EudraVigilance database. The overall safety profile of AS04-adjuvanted HPV-16/18 vaccination in Chinese girls and women aged 9–45 y remains favorable. Acknowledgments The authors thank Xin Chen and Yixin Sun from Peking University, and Tianfei Yi from Yinzhou District Center for Disease Control and Prevention, Ningbo, for their valuable contributions in data collection during the case verification process. The authors also thank Nadia Meyer, Raja Amrinder, and Sandeep Gunapu from GSK for their contribution to the study. The authors also thank Enovalife Medical Communication Service Center for editorial assistance, manuscript coordination and writing support (Carole Nadin [Fleetwith Ltd]), on behalf of GSK. Biographies Huifeng Yun, GSK, Rockville, Maryland, USA, huifeng.x.yun@gsk.com: Doctor Huifeng Yun currently serves as senior director in Vaccine and infectious disease at GSK. As the head of the non-respiratory viral vaccine epidemiology with over twenty years’ experience in pharmacoepidemiology and pharmacovigilance, she directed the team that designed and executed epidemiologic studies, supporting core functions across multiple investigational assets and marketed products. Before joining GSK, she was an associated professor in the Department of Epidemiology, University of Alabama at Birmingham (UAB). Dr. Yun earned her medical degree in China, MS in biostatistics and PhD in Epidemiology from UAB. To date, Dr. Yun has contributed over 120 peer-reviewed publications. Siyan Zhan, Department of Epidemiology and Biostatistics, School of Public Health, Peking University, Beijing, China, siyan-zhan@bjmu.edu.cn: Professor Siyan Zhan is currently the Dean of the School of Public Health at Peking University, where she also serves as Chair of the Department of Epidemiology and Biostatistics. She holds concurrent appointments as Director of the Clinical Epidemiology Research Center at Peking University Third Hospital (dual-affiliated), Director of the Center for Intelligent Public Health at Peking University’s Institute for Artificial Intelligence, Associate Director of the Peking University Evidence-Based Medicine Center, and Director of the Post-Marketing Drug Safety Research Center at Peking University Health Science Center. She assumed the deanship of the School of Public Health in January 2022. Professor Zhan earned her Ph.D. from The Chinese University of Hong Kong, following a six-year undergraduate program and subsequent Master’s degree at Beijing Medical University (now Peking University Health Science Center). After completing her Master’s in 1989, she joined the faculty, was promoted to full professor in 2001, and became a doctoral advisor in 2004. Her research focuses on pharmacoepidemiology, clinical epidemiology, and evidence-based medicine. In her current roles, Professor Zhan has dedicated herself to advancing public health education and research, driving the development and practical application of epidemiology and evidence-based medicine disciplines Funding Statement This work was sponsored by GlaxoSmithKline [GSK study identifier: 212380], which was involved in all stages of study design, data interpretation, and writing of the study report. GSK also funded all costs associated with the development and publishing of the manuscript. Disclosure statement Adoración Navarro-Torné is employed by GSK and holds financial equities in GSK. Rhed Evans is employed by GSK and holds financial equities in GSK. Amila Poddar is employed by GSK. Huifeng Yun is employed by GSK and holds financial equities in GSK. Siyan Zhan, the primary investigator from Peking University, has received funding from GSK, which is the manufacturer of Cervarix. These authors declare no other financial and non-financial relationships and activities. Junting Yang has nothing to disclose. Zhike Liu has nothing to disclose. Jinxin Guo has nothing to disclose. Siwei Deng has nothing to disclose. Ruogu Meng has nothing to disclose. Yu Yang has nothing to disclose. These authors declare no financial and non-financial relationships and activities and no conflicts of interest. Data availability statement GSK did not have access to any data in the Health and Family Planning Bureau of Yinzhou District Platform. The data generated in this study are not publicly available. The raw data cannot be shared in the public domain by GSK because GSK did not have access to raw data. For requests for access to anonymized subject level data, please contact the corresponding author from the Peking University. Consent to participate Not applicable. The study used only aggregated statistical data with no identifiable individual information included in any publications. Consent for publication Not applicable. The research utilized fully anonymized electronic health records from the YRHIP, with all personal identifiers removed prior to analysis. Ethical statement The study was conducted in accordance with all applicable regulatory requirements, including all applicable participant privacy requirements and the guiding principles of the Declaration of Helsinki. The local health authority (Health and Family Planning Bureau of Yinzhou District) has approved the use of the databases for post-marketing vaccine surveillance (Yinzhou Ethics Committee for Disease Control and Prevention 2022–03). The study protocol was approved by the ethics committee of the Peking University Health Science Center (IRB00001052– 20098). The need for informed consent was waived. Trademark statement Cervarix is a trademark owned by or licensed to the GSK group of companies. Abbreviations adjusted rate ratio; confidence interval; human papillomavirus; inflammatory bowel disease; International Classification of Diseases 10th Revision; idiopathic thrombocytopenic purpura; National Medical Products Administration; not otherwise specified; potential immune-mediated disease; positive predictive value; person-year; rheumatoid arthritis; rate ratio; type 1 diabetes mellitus; World Health Organization; Yinzhou Regional Health Information Platform References Associated Data Data Availability Statement GSK did not have access to any data in the Health and Family Planning Bureau of Yinzhou District Platform. The data generated in this study are not publicly available. The raw data cannot be shared in the public domain by GSK because GSK did not have access to raw data. For requests for access to anonymized subject level data, please contact the corresponding author from the Peking University.