Adverse events following 9-valent human papillomavirus vaccine (GARDASIL® 9) reported to the Vaccine Adverse Event Reporting System (VAERS), 2015–2024

ABSTRACT

GARDASIL 9, a 9-valent HPV vaccine approved in 2014, is widely administered for the prevention of HPV-related malignancies. Although clinical trials demonstrated a favorable safety profile, rare or delayed-onset adverse events may not be captured pre-licensure. Post-marketing surveillance using VAERS offers a complementary approach for signal detection and safety monitoring. We conducted a retrospective pharmacovigilance analysis of VAERS reports following GARDASIL 9 administration from January 1, 2015 to December 31, 2024. Adverse events were encoded using MedDRA v26.0 and categorized by System Organ Class. Disproportionality analyses using four independent algorithms (ROR, PRR, IC, EBGM) were applied to identify positive safety signals. Subgroup assessments included serious reports, death reports, and reports in pregnant individuals. Among reported events, most were non-serious and consistent with known reactogenicity patterns, including syncope, headache, and injection site reactions. However, signals were also detected for certain events not listed in product labeling, including postural orthostatic tachycardia syndrome, eye movement disorder, autoimmune thyroiditis, and posture abnormality. In 57 death reports, neurological terms showed signal elevation but lacked consistent etiologic patterns. No positive signals were detected in the 18 pregnancy-associated reports. This VAERS-based analysis supports the established safety of GARDASIL 9 while highlighting rare signals that warrant further investigation. Given the limitations of passive surveillance, integration with active monitoring systems is essential to refine safety profiles and support ongoing public health vaccination efforts.

Introduction

Human papillomavirus (HPV) is one of the most prevalent sexually transmitted infections globally, with an estimated 80% of sexually active individuals acquiring the virus during their lifetime. Persistent infection with high-risk HPV genotypes, particularly types 16 and 18, is the primary cause of cervical cancer and is also implicated in a range of other malignancies, including anal, vulvar, vaginal, penile, and oropharyngeal cancers.¹ Given the substantial disease burden and the lack of effective treatment for persistent HPV infection, prophylactic vaccination has become a cornerstone strategy in global cancer prevention efforts.^2,3

The burden of HPV-related disease is not restricted to women, as increasing evidence highlights its contribution to anogenital and oropharyngeal pathology in men, underscoring the importance of gender-inclusive vaccination strategies. Moreover, the identification of high-risk genotypes not yet covered by current vaccines continues to raise questions about the long-term epidemiological impact and genotype replacement potential.^4,5

GARDASIL 9 is a recombinant 9-valent human papillomavirus vaccine developed by Merck & Co. and approved by the U.S. Food and Drug Administration (FDA) in 2014. It was designed to extend protection beyond the original quadrivalent formulation by including five additional high-risk HPV types (31, 33, 45, 52, and 58), in addition to types 6, 11, 16, and 18.⁶ The vaccine is widely included in national immunization programs, including in the United States, where it is routinely administered to adolescents aged 9–14 years and approved for use through age 45.⁷

Prior to licensure, GARDASIL 9 underwent large-scale randomized controlled trials involving more than 15,000 participants across diverse demographics.⁸ These studies demonstrated a favorable safety profile, with the most commonly reported adverse events being local injection site reactions, including pain, swelling, and erythema, followed by systemic symptoms such as headache and fever. Comparative analyses with the quadrivalent vaccine confirmed similar tolerability, although injection site events were slightly more frequent with the 9-valent formulation. Serious vaccine-related adverse events were rare, and no unexpected safety issues emerged, which supported the FDA’s approval and subsequent expansion of its indications.

Shortly after its market introduction, GARDASIL 9 became subject to post-licensure safety surveillance via the Vaccine Adverse Event Reporting System (VAERS), a national passive reporting system co-managed by the CDC and FDA.⁹ Early spontaneous reports began to raise questions about potential associations with syncope, anaphylaxis, autoimmune reactions, and neurological conditions such as Guillain-Barré syndrome. Although causality could not be established based on VAERS data alone, the emergence of these signals emphasized the importance of real-world pharmacovigilance to detect rare but clinically relevant adverse events.

Despite strong evidence from pre- and post-licensure studies, HPV vaccine hesitancy persists in various populations, often driven by public concerns about long-term safety and media amplification of anecdotal adverse outcomes. Such sociocultural factors can significantly influence vaccination uptake, particularly among adolescents and their caregivers, thereby affecting herd immunity goals. In this context, reinforcing safety signals through robust post-marketing analyses becomes crucial not only for scientific understanding but also for public reassurance.

As a cornerstone of U.S. post-marketing surveillance, VAERS offers broad coverage and timely detection of potential vaccine safety concerns through spontaneous reporting by healthcare providers, manufacturers, and the public. However, its limitations include underreporting, variability in report quality, lack of a clearly defined denominator population, and absence of an internal unvaccinated control group, all of which preclude direct estimation of incidence rates or causal inference. These characteristics necessitate cautious interpretation and underscore the need for complementary analytical methods and external data validation to strengthen signal assessment.

Given the broad demographic uptake of GARDASIL 9 and the limitations inherent in pre-licensure studies – particularly their inability to capture delayed or low-frequency outcomes – complementary post-marketing analyses using large-scale observational data have become increasingly essential. VAERS provides a practical platform for this purpose, aggregating reports from healthcare providers, manufacturers, and the general public. Its utility lies in early signal detection, especially for vaccines newly introduced or applied across heterogeneous populations. However, due to underreporting and non-standardized reporting practices, supplemental analytical strategies are required to interpret these signals appropriately.

This study aimed to conduct a comprehensive pharmacovigilance assessment of adverse events reported to VAERS following GARDASIL 9 vaccination over the past decade. We applied disproportionality analyses to detect potential safety signals, characterized demographic and clinical reporting patterns, and compared our findings to those described in regulatory documents and existing literature. This investigation seeks to refine the understanding of GARDASIL 9’s post-marketing safety and identify signals warranting further clinical or mechanistic investigation.

Materials and methods

Data source and extraction

This study utilized data derived from the VAERS. The dataset collects spontaneous reports of adverse events occurring after vaccination from healthcare professionals, manufacturers, and members of the public.¹⁰ The database is fully open-access and can be downloaded freely from the official VAERS website (https://vaers.hhs.gov/data.html), ensuring transparency and reproducibility. The dataset contains structured information including patient demographics, vaccine product identifiers, administration dates, concurrent vaccines, detailed adverse event descriptions, and health outcomes. For this analysis, all VAERS reports from January 1, 2015, to December 31, 2024, were downloaded directly from the official CDC VAERS data portal. Reports involving GARDASIL 9, as identified through the vaccine name field, were retained for further processing.

To ensure clinical relevance and dataset consistency, inclusion was restricted to reports that explicitly involved GARDASIL 9 as the administered product and contained adequate core data fields, such as age, sex, date of vaccination, and at least one described adverse event. Cases were excluded if they lacked essential identifiers, contained duplicate entries, or described conditions clearly unrelated to vaccine exposure. Specifically, exclusion criteria encompassed events linked to product quality complaints without clinical manifestations, diagnostic or screening procedures without adverse outcomes, congenital anomalies or hereditary conditions not plausibly related to immunization, and any events attributable to external causes such as trauma, poisoning, or surgical interventions. Where applicable, medical judgment was applied to determine lack of temporal or pathophysiological association with the vaccination event.

MedDRA encoding and SOC selection

All reported adverse events were standardized using the Medical Dictionary for Regulatory Activities (MedDRA), version 26.0.¹¹ Individual Preferred Terms (PTs) were extracted and subsequently mapped to their corresponding System Organ Classes (SOCs) to facilitate signal clustering and classification. In line with previous pharmacovigilance studies, SOCs not directly reflecting adverse drug reactions were systematically excluded from downstream analyses. These included categories such as product issues, social circumstances, procedures and investigations, and congenital disorders, which do not represent pathologic responses to vaccine administration. This filtering step allowed for a focused evaluation of clinically meaningful adverse events and enabled subsequent statistical analyses to concentrate on biologically plausible vaccine-related outcomes.

Descriptive statistics

Descriptive analyses were performed to summarize key demographic and clinical features of the VAERS reports following GARDASIL 9 vaccination. Age and sex distributions were examined, revealing a predominance of adolescent and young adult females, consistent with the vaccine’s target population. The median time to symptom onset was calculated from the reported date of vaccination to the date of event onset when available, allowing assessment of temporal proximity to vaccine administration.

Clinical outcomes were categorized based on standard VAERS fields, including hospitalization, emergency department visits, life-threatening conditions, disability, and death. Each outcome was summarized by frequency and percentage, and distributions were cross-tabulated by age group and vaccine usage pattern. To facilitate visualization and comparative interpretation, the proportions of each outcome were displayed in bar charts, and serious event frequencies were further evaluated across subgroups.¹²

All descriptive statistics were processed using validated scripts in R software (version 4.2.3), and missing or ambiguous data were handled using listwise deletion. This step laid the foundation for subsequent signal detection analyses and allowed for clear profiling of the reported event burden across strata of interest.

Disproportionality analysis

Signal detection was performed using a disproportionality analysis framework that included four independent statistical algorithms: Reporting Odds Ratio (ROR), Proportional Reporting Ratio (PRR) with chi-square testing, Bayesian Confidence Propagation Neural Network (BCPNN) for Information Component (IC) estimation, and Empirical Bayes Geometric Mean (EBGM). These methods were selected to ensure robustness of signal identification across frequentist and Bayesian paradigms.

ROR was calculated as the odds of reporting a specific event following GARDASIL 9 relative to all other vaccines. PRR was defined as the proportion of a given event among GARDASIL 9 reports compared to the proportion of the same event among reports for all other vaccines, with statistical significance assessed using chi-square ≥4. The BCPNN model provided the IC and IC025 values to quantify signal strength with 95% credibility. EBGM, derived from the Multi-Item Gamma Poisson Shrinker algorithm, was also applied, with the lower bound of the 90% credibility interval (EB05) used as the primary threshold.^13–17

For each method, a signal was considered positive when the lower bound of the confidence or credibility interval exceeded the respective predefined threshold (e.g., ROR >2 with lower bound >1, PRR >2 with chi-square ≥4, IC025 >0, or EB05 >1), as recommended by established pharmacovigilance guidance from the European Medicines Agency and prior literature.^15,18 Among these, ROR was selected as the primary analytic metric in this study, given its extensive validation, interpretability, and continued endorsement as the preferred approach for vaccine signal detection by global regulatory bodies including the WHO-Uppsala Monitoring Centre and EMA.¹⁹ PRR, IC, and EBGM were retained for secondary comparison and consistency checking, ensuring robustness through triangulation of signal strength across methods.

Results

Descriptive and signal detection analysis

From January 1, 2015, to December 31, 2024, a total of 23,499 adverse event (AE) reports related to GARDASIL 9 were retrieved from the VAERS database, spanning all 50 U.S. states, the District of Columbia, and multiple U.S. territories. The most frequent reporting states included California (8.3%), Texas (6.0%), New York (3.8%), and Florida (3.2%), while only 12.4% of entries lacked state-level information, supporting the dataset’s broad national representativeness (Supplementary Figure S1). Annual reporting volumes peaked in 2016 (n = 3,378), followed by a gradual stabilization with a mean annual report count exceeding 2,000 since 2017 (Figure 1). Among the 17,025 cases with documented gender, females accounted for 64.49% of the reports, in line with HPV vaccine uptake patterns. Age information was available for 15,386 individuals, revealing a predominance of AE reports in recipients younger than 18 years (73.52%), with markedly fewer reports in older age groups (Figure 2).

Figure 1.

Yearly trends in reported adverse events linked to GARDASIL 9 in the VAERS database from 2015 to 2024.

Figure 2.

Age stratification of individuals submitting adverse event reports associated with GARDASIL 9 to VAERS.

Most AEs were categorized as non-serious (92.50%), while serious events – including death, hospitalization, life-threatening conditions, disability, and prolonged hospitalization – constituted 7.50% (Table 1). Specific clinical outcomes were available in 9,230 reports, among which hospitalization (n = 1,266) was most frequently observed, followed by life-threatening events (n = 211) and persistent disability (n = 540). Recovery was documented in 77.33% of evaluable cases, and only 57 deaths were reported. Additionally, GARDASIL 9 was administered alone in 70.61% of the AE reports. Regarding temporal distribution, the majority of AEs occurred within 30 days following vaccination, with 92.56% reported during this early post-vaccination period.

Table 1.

Summary of demographic features and clinical outcomes in VAERS reports concerning GARDASIL 9 administration.

Characteristics	Vaccine-induced AE reports (n = 23499)
Number of events	Available number*, n	Case number, n	Case proportion, %
Gender, n (%)	17025
F		10980	64.49%
M		6045	35.51%
Age (years), n (%)	15386
<18		11339	73.70%
18≤ and <28		2705	17.58%
28≤ and <38		691	4.49%
38≤ and <48		464	3.02%
48≤ and <58		107	0.70%
58≤ and <68		51	0.33%
≥68		29	0.19%
Seriousc/non-serious status, n (%)	23499
Nonserious		21728	92.50%
Serious**		1771	7.50%
Outcomes, n (%)	9230
death		57	0.62%
hospitalization		1266	13.72%
life-threatening illness		211	2.29%
prolongation of existing hospitalization		18	0.20%
permanent disability		540	5.85%
recover		7138	77.33%
GARDASIL® 9 given alone, n (%)	23499
Alone		16592	70.61%
Vaccine co-administration		6907	29.39%
Onset time,n(%)	16883
0–30 days		15676	92.85%
31–60 days		268	1.59%
61–90 days		174	1.03%
91–120 days		74	0.44%
121–150 days		77	0.46%
151–180 days		53	0.31%
181–360 days		215	1.27%
>360 days		346	2.05%

*Available number indicates that missing data has been removed.

**Serious reports are those describing death, life-threatening illness, hospitalization or prolongation of existing hospitalization, or permanent disability.

The most frequently reported MedDRA PTs included syncope, dizziness, headache, and a range of injection site reactions such as pain, erythema, and swelling (Figure 3). These PTs consistently appeared in both serious and non-serious reports.

Figure 3.

Ranking of Preferred Terms (PTs) by their proportional representation among total GARDASIL 9-related VAERS reports.

Disproportionality signal detection using the ROR method identified 112 positive signals. The strongest associations were observed for pallor (ROR: 9.00, 95% CI: 8.47–9.58), loss of consciousness (ROR: 5.67, 95% CI: 5.37–5.99), and syncope (ROR: 5.46, 95% CI: 5.23–5.71), suggesting that vasovagal and neurocardiogenic responses are prominent in the AE profile (Figure 4). The volcano plot further highlighted that several PTs such as fatigue, pyrexia, and injection site swelling demonstrated both statistical significance and high effect sizes (Figure 5).

Figure 4.

Most frequently reported adverse events with statistically elevated RORs in the post-marketing data for GARDASIL 9.

Figure 5.

The volcano plot highlighted the several PTs.

To improve robustness and minimize false-positive signals, multiple detection algorithms were employed, including PRR, BCPNN (Bayesian Confidence Propagation Neural Network), and EBGM (Empirical Bayes Geometric Mean). Cross-validation across these four methods revealed that a core set of 164 PTs (23.3%) was simultaneously identified by all algorithms (Figure 6), suggesting high reliability. In contrast, BCPNN alone detected a significant number of exclusive signals (n = 202), while ROR and PRR did not contribute any unique signals in this analysis.

Figure 6.

Intersection of adverse event signals concurrently detected by four disproportionality algorithms following GARDASIL 9 vaccination.

The tabulated results from the SOC-classified PTs (Table 2) demonstrate that signal enrichment was not limited to general and local symptoms. Notably, several organ-specific SOCs exhibited strong signal strength across multiple algorithms. For example, eye movement disorder (SOC: Eye Disorders) had an ROR of 8.68 and an IC025 of 2.82, while postural orthostatic tachycardia syndrome (POTS) within the Cardiac Disorders SOC exhibited an ROR of 7.89 and was consistently flagged by all signal detection methods. Within the Reproductive System SOC, premature menopause (ROR: 12.36) and infertility, female (ROR: 14.27) were of particular note. Moreover, immune-mediated conditions such as autoimmune thyroiditis, immune thrombocytopenic purpura, and autoimmune disorder also surpassed signal detection thresholds. These findings imply potential immunologic or neuroendocrine axes of reactogenicity following GARDASIL 9 exposure.

Table 2.

Signal detection results by SOC and PT using ROR, PRR, IC, and EBGM models applied to GARDASIL 9-related VAERS entries.

SOC	PT	N	ROR(95%Cl)	IC(IC025)	PRR(χ2)	EBGM(EBGM05)
Blood And Lymphatic System Disorders	Lymphadenitis	22	1.76 (1.15–2.68)	0.81 (0.2)	1.76 (7.1)	1.75 (1.23)
	Immune Thrombocytopenic Purpura	17	5.12 (3.15–8.32)	2.31 (1.61)	5.12 (54.09)	4.95 (3.3)
	Mast Cell Activation Syndrome	14	4.09 (2.4–6.96)	1.99 (1.24)	4.09 (31.56)	3.98 (2.55)
	Increased Tendency To Bruise	11	3.82 (2.1–6.97)	1.9 (1.06)	3.82 (22.24)	3.74 (2.26)
	Splenomegaly	11	2.39 (1.31–4.33)	1.24 (0.4)	2.39 (8.68)	2.36 (1.43)
Cardiac Disorders	Postural Orthostatic Tachycardia Syndrome	106	7.89 (6.48–9.6)	2.9 (2.61)	7.88 (598.01)	7.46 (6.33)
	Bradycardia	59	2.02 (1.56–2.61)	1 (0.62)	2.02 (29.78)	2 (1.61)
	Tricuspid Valve Incompetence	10	2.57 (1.37–4.8)	1.34 (0.46)	2.57 (9.37)	2.53 (1.5)
	Sinus Bradycardia	9	2.48 (1.28–4.8)	1.29 (0.37)	2.48 (7.81)	2.45 (1.41)
	Sinus Arrhythmia	7	3.12 (1.47–6.61)	1.62 (0.58)	3.12 (9.83)	3.07 (1.64)
Endocrine Disorders	Autoimmune Thyroiditis	27	3.23 (2.21–4.74)	1.67 (1.11)	3.23 (40.58)	3.18 (2.31)
	Thyroid Mass	9	3.37 (1.74–6.54)	1.72 (0.8)	3.37 (14.59)	3.31 (1.9)
	Thyroid Disorder	9	2.04 (1.05–3.93)	1.01 (0.09)	2.04 (4.66)	2.02 (1.16)
	Thyroid Cyst	5	11.7 (4.67–29.29)	3.43 (2.19)	11.7 (44.63)	10.76 (4.99)
Eye Disorders	Eye Movement Disorder	213	8.68 (7.56–9.98)	3.03 (2.82)	8.66 (1348.13)	8.15 (7.26)
	Vision Blurred	181	1.36 (1.18–1.58)	0.44 (0.23)	1.36 (17.36)	1.36 (1.2)
	Visual Impairment	176	1.62 (1.4–1.88)	0.69 (0.47)	1.62 (41.4)	1.61 (1.42)
	Photophobia	63	1.4 (1.1–1.8)	0.48 (0.12)	1.4 (7.24)	1.4 (1.14)
	Blindness	50	1.91 (1.44–2.52)	0.92 (0.51)	1.91 (21.29)	1.89 (1.5)
Gastrointestinal Disorders	Vomiting	824	1.33 (1.24–1.42)	0.4 (0.3)	1.32 (64.42)	1.32 (1.24)
	Abdominal Pain	264	1.45 (1.28–1.64)	0.53 (0.35)	1.45 (36.44)	1.44 (1.3)
	Lip Swelling	114	1.47 (1.22–1.77)	0.55 (0.28)	1.47 (16.88)	1.46 (1.25)
	Retching	38	1.52 (1.1–2.09)	0.6 (0.13)	1.52 (6.67)	1.51 (1.16)
	Gastrointestinal Disorder	34	1.5 (1.07–2.1)	0.57 (0.08)	1.5 (5.51)	1.49 (1.12)
General Disorders And Administration Site Conditions	No Adverse Event	7408	8.34 (8.14–8.55)	2.86 (2.82)	7.64 (40766.75)	7.25 (7.1)
	Injection Site Pain	1031	1.38 (1.29–1.46)	0.45 (0.36)	1.37 (103.36)	1.37 (1.3)
	Injection Site Erythema	786	1.27 (1.19–1.37)	0.34 (0.24)	1.27 (45.18)	1.27 (1.2)
	Injection Site Swelling	743	1.5 (1.39–1.61)	0.57 (0.46)	1.49 (119.54)	1.49 (1.4)
	Immediate Post-Injection Reaction	584	5.63 (5.18–6.12)	2.43 (2.31)	5.6 (2112.1)	5.4 (5.03)
Hepatobiliary Disorders	Ocular Icterus	7	7.16 (3.34–15.34)	2.77 (1.72)	7.16 (35.02)	6.82 (3.6)
	Liver Injury	6	2.39 (1.06–5.35)	1.24 (0.14)	2.39 (4.74)	2.36 (1.2)
	Gallbladder Disorder	5	3.03 (1.25–7.35)	1.57 (0.38)	3.03 (6.62)	2.98 (1.42)
	Hepatosplenomegaly	4	4.27 (1.58–11.57)	2.06 (0.74)	4.27 (9.67)	4.16 (1.81)
	Hepatic Calcification	3	45.63 (12.11–172)	5.06 (3.4)	45.63 (95.24)	33.46 (11.02)
Immune System Disorders	Hypersensitivity	153	1.23 (1.05–1.44)	0.3 (0.06)	1.23 (6.5)	1.23 (1.07)
	Anaphylactic Shock	45	2.77 (2.06–3.73)	1.45 (1.02)	2.77 (49.81)	2.73 (2.13)
	Autoimmune Disorder	42	2.01 (1.48–2.72)	0.99 (0.55)	2 (20.82)	1.99 (1.54)
	Immune System Disorder	22	1.58 (1.04–2.4)	0.65 (0.05)	1.58 (4.6)	1.57 (1.1)
	Food Allergy	22	2.76 (1.81–4.22)	1.45 (0.84)	2.76 (24.19)	2.72 (1.91)
Infections And Infestations	Cellulitis	91	1.32 (1.08–1.63)	0.4 (0.1)	1.32 (7.12)	1.32 (1.11)
	Papilloma Viral Infection	77	32.68 (25.41–42.03)	4.7 (4.34)	32.64 (1862.43)	25.95 (21.02)
	Injection Site Cellulitis	32	1.65 (1.16–2.34)	0.71 (0.21)	1.65 (8.04)	1.64 (1.22)
	Encephalitis	27	2.31 (1.58–3.38)	1.19 (0.64)	2.31 (19.73)	2.29 (1.66)
	Infectious Mononucleosis	21	5.92 (3.82–9.17)	2.51 (1.88)	5.91 (81.79)	5.69 (3.94)
Metabolism And Nutrition Disorders	Type 1 Diabetes Mellitus	22	2.87 (1.88–4.38)	1.5 (0.89)	2.87 (26.12)	2.82 (1.98)
	Food Intolerance	12	2.94 (1.66–5.22)	1.53 (0.72)	2.94 (15.04)	2.9 (1.79)
	Abnormal Loss Of Weight	11	3.26 (1.79–5.94)	1.68 (0.84)	3.26 (16.83)	3.21 (1.94)
	Gluten Sensitivity	11	8.75 (4.74–16.13)	3.04 (2.18)	8.75 (70.42)	8.23 (4.93)
	Vitamin D Deficiency	11	3.23 (1.77–5.87)	1.66 (0.82)	3.23 (16.45)	3.17 (1.92)
Musculoskeletal And Connective Tissue Disorders	Muscular Weakness	241	1.25 (1.1–1.42)	0.31 (0.13)	1.25 (11.6)	1.24 (1.12)
	Musculoskeletal Stiffness	230	1.59 (1.39–1.81)	0.66 (0.47)	1.59 (49.14)	1.58 (1.41)
	Posture Abnormal	143	13.34 (11.22–15.85)	3.6 (3.34)	13.31 (1468.08)	12.1 (10.47)
	Muscle Twitching	131	2.63 (2.22–3.13)	1.38 (1.12)	2.63 (129.85)	2.6 (2.25)
	Muscle Rigidity	65	7.48 (5.82–9.61)	2.83 (2.46)	7.48 (343.53)	7.1 (5.76)
Neoplasms Benign, Malignant And Unspecified (Incl Cysts And Polyps)	Skin Papilloma	32	19.98 (13.74–29.03)	4.11 (3.57)	19.97 (495.29)	17.29 (12.65)
	Anogenital Warts	23	33.73 (21.25–53.53)	4.73 (4.08)	33.72 (571.76)	26.62 (18.09)
	Cervix Carcinoma	18	14.7 (9.01–23.98)	3.72 (3.03)	14.7 (205.04)	13.22 (8.78)
	Neoplasm Malignant	15	1.98 (1.19–3.29)	0.97 (0.24)	1.98 (7.13)	1.96 (1.28)
	Leukaemia	8	3.94 (1.95–7.97)	1.94 (0.97)	3.94 (17.01)	3.85 (2.14)
Nervous System Disorders	Dizziness	2086	1.73 (1.66–1.81)	0.77 (0.7)	1.71 (620.53)	1.7 (1.64)
	Syncope	2070	5.46 (5.23–5.71)	2.37 (2.3)	5.35 (7040.83)	5.16 (4.97)
	Loss Of Consciousness	1354	5.67 (5.37–5.99)	2.43 (2.35)	5.59 (4898.27)	5.39 (5.15)
	Seizure	486	3.3 (3.02–3.61)	1.69 (1.56)	3.29 (753.95)	3.23 (2.99)
	Tremor	411	1.61 (1.46–1.78)	0.68 (0.54)	1.61 (93.88)	1.6 (1.48)
Pregnancy, Puerperium And Perinatal Conditions	Pregnancy	10	2.76 (1.47–5.16)	1.44 (0.56)	2.76 (10.97)	2.72 (1.61)
	Abortion	8	7 (3.43–14.28)	2.74 (1.75)	7 (38.93)	6.68 (3.68)
	Foetal Cardiac Disorder	4	13.91 (4.94–39.13)	3.65 (2.28)	13.91 (43)	12.58 (5.29)
Psychiatric Disorders	Anxiety	169	1.19 (1.02–1.39)	0.25 (0.03)	1.19 (5.21)	1.19 (1.05)
	Disorientation	84	1.93 (1.55–2.39)	0.93 (0.62)	1.93 (36.79)	1.91 (1.6)
	Depression	66	1.7 (1.34–2.17)	0.76 (0.41)	1.7 (18.92)	1.69 (1.38)
	Tic	37	7.15 (5.13–9.96)	2.77 (2.29)	7.15 (184.75)	6.81 (5.16)
	Staring	36	4.36 (3.13–6.08)	2.08 (1.6)	4.36 (89.96)	4.24 (3.21)
Renal And Urinary Disorders	Urinary Incontinence	91	4.22 (3.43–5.2)	2.04 (1.73)	4.22 (216)	4.11 (3.45)
	Lupus Nephritis	4	7.73 (2.81–21.23)	2.87 (1.54)	7.73 (22.02)	7.32 (3.14)
	Renal Vein Compression	3	182.52 (30.5–1092.37)	6.2 (4.34)	182.51 (216.62)	73.61 (16.47)
	Bladder Dysfunction	3	3.44 (1.09–10.85)	1.76 (0.29)	3.44 (5.06)	3.38 (1.29)
	Cystitis Interstitial	3	3.88 (1.23–12.26)	1.92 (0.45)	3.88 (6.22)	3.79 (1.45)
Reproductive System And Breast Disorders	Amenorrhoea	57	1.7 (1.31–2.21)	0.76 (0.37)	1.7 (16.18)	1.69 (1.36)
	Cervical Dysplasia	33	20.71 (14.31–29.95)	4.16 (3.63)	20.7 (528.69)	17.83 (13.1)
	Ovarian Cyst	24	5.23 (3.48–7.88)	2.34 (1.75)	5.23 (78.8)	5.06 (3.59)
	Premature Menopause	20	12.36 (7.8–19.57)	3.5 (2.84)	12.35 (189.45)	11.31 (7.69)
	Infertility Female	17	14.27 (8.63–23.58)	3.69 (2.97)	14.27 (187.7)	12.87 (8.45)
Respiratory, Thoracic And Mediastinal Disorders	Hyperventilation	41	2.42 (1.78–3.3)	1.26 (0.81)	2.42 (33.52)	2.39 (1.85)
	Pharyngeal Oedema	23	2.58 (1.71–3.91)	1.35 (0.76)	2.58 (21.87)	2.55 (1.81)
	Pulmonary Mass	13	1.79 (1.04–3.1)	0.83 (0.05)	1.79 (4.48)	1.78 (1.13)
	Snoring	12	5.64 (3.16–10.06)	2.44 (1.62)	5.64 (43.75)	5.43 (3.35)
	Grunting	8	5.66 (2.79–11.5)	2.45 (1.46)	5.66 (29.33)	5.45 (3.01)
Skin And Subcutaneous Tissue Disorders	Erythema	618	1.19 (1.1–1.29)	0.25 (0.13)	1.19 (18.49)	1.19 (1.11)
	Urticaria	552	1.39 (1.27–1.51)	0.46 (0.34)	1.38 (58.26)	1.38 (1.29)
	Hyperhidrosis	497	1.45 (1.32–1.58)	0.52 (0.39)	1.44 (67)	1.44 (1.33)
	Skin Warm	225	1.28 (1.12–1.46)	0.35 (0.16)	1.28 (13.71)	1.28 (1.14)
	Cold Sweat	199	2.63 (2.29–3.03)	1.38 (1.17)	2.63 (196.91)	2.6 (2.31)
Vascular Disorders	Pallor	1098	9 (8.47–9.58)	3.06 (2.97)	8.89 (7178.59)	8.35 (7.93)
	Hypotension	147	1.45 (1.24–1.71)	0.53 (0.29)	1.45 (20.51)	1.45 (1.26)
	Peripheral Coldness	54	1.44 (1.1–1.88)	0.52 (0.13)	1.44 (7.18)	1.43 (1.15)
	Cyanosis	47	1.55 (1.16–2.07)	0.63 (0.21)	1.55 (9.15)	1.55 (1.22)
	Orthostatic Hypotension	20	3.76 (2.41–5.86)	1.88 (1.24)	3.76 (39.23)	3.67 (2.53)

AE adverse event, PT preferred term, SOC system organ classification.

In the subgroup analysis of individuals under 18 years of age, who accounted for 73.7% of all evaluable cases, the most frequently reported adverse events included syncope (4.77%), dizziness (4.66%), loss of consciousness (3.11%), headache (2.96%), and pallor (2.90%). Disproportionality analysis confirmed that these terms also yielded elevated ROR values, notably for loss of consciousness (ROR: 3.39, 95% CI: 3.16–3.63), syncope (ROR: 3.12, 95% CI: 2.95–3.30), and dizziness (ROR: 2.69, 95% CI: 2.54–2.85). These events are characteristic of transient neurocardiogenic responses, such as vasovagal syncope, which is well-documented among adolescent populations receiving intramuscular vaccines. Other signals in this age group included nausea, injection site pain, and fatigue, all of which were generally non-serious and aligned with the established safety profile of GARDASIL 9. (Supplementary Figure S2–S3 and Supplementary Table S1).

Serious reports following GARDASIL 9 alone

To further clarify potential safety signals attributable solely to GARDASIL 9, a subanalysis was conducted using serious adverse events (SAEs) reported in cases where the vaccine was administered without concomitant immunization. A total of 1,179 serious reports were retrieved for this subset, enabling isolation of potential signals specifically linked to the monovalent effect of GARDASIL 9.

Disproportionality analysis using the ROR method revealed several PTs with significantly elevated risk estimates. POTS emerged as the most prominent signal, with an ROR of 10.39 (95% CI: 7.75–13.93), reinforcing findings from the overall dataset and supporting prior concerns regarding its temporal association with HPV vaccines. Other notable neurologic PTs included seizure (ROR: 3.06), syncope (ROR: 2.76), loss of consciousness (ROR: 2.37), migraine (ROR: 2.07), and anxiety (ROR: 2.25), all of which presented RORs exceeding the threshold and narrow confidence intervals (Figure 7).

Figure 7.

Predominant serious adverse events showing significant ROR elevations in VAERS reports attributed to GARDASIL 9.

In addition to neurologic and psychiatric categories, a set of musculoskeletal and dermatological signals were also identified, including muscular weakness (ROR: 1.87), rash (ROR: 1.88), and erythema (ROR: 2.34). These suggest that systemic inflammatory and cutaneous manifestations, although less frequent, may be involved in serious post-vaccination presentations. Of note, gastrointestinal symptoms such as abdominal pain (ROR: 1.74) and general disorders like gait disturbance and peripheral swelling also met signal detection criteria.

Death reports

Among the total VAERS records reviewed, 57 reports involved death following GARDASIL 9 vaccination, comprising 36 females, 12 males, and 9 cases without gender documentation. Of these, 36 deaths occurred following GARDASIL 9 administration alone, while the remaining 21 involved co-administration with other vaccines. Ages were available for 18 cases under 18 years, while the rest were either adults or lacked age information. Temporal analysis showed that onset within 30 days post-vaccination was recorded in 19 cases, whereas the majority of remaining reports lacked a clear timeline.

Disproportionality analysis revealed several Preferred Terms statistically overrepresented among fatal cases, most notably “Unevaluable Event,” which yielded a markedly elevated ROR (44.15; 95% CI: 15.84–123.04), suggesting substantial uncertainty or incomplete documentation within these fatal reports. Other PTs associated with fatal outcomes included “Brain Oedema” (ROR: 8.67, 95% CI: 2.76–27.24), “Rash,” “Syncope,” and “Headache,” all of which demonstrated signal strength across multiple algorithms (Table 3). Notably, many of these PTs are neurologically or immunologically mediated symptoms, potentially indicating central nervous system involvement or systemic hypersensitivity responses. However, it should be emphasized that these signals derive from small sample sizes and should be interpreted with caution in light of VAERS’s passive reporting nature and inherent limitations in establishing causality.

Table 3.

Preferred terms exhibiting signal positivity in death-related reports after GARDASIL 9 administration, as identified by at least one statistical method.

PT	N，(%)	Soc_name	ROR(95%Cl)	PRR(χ2)	EBGM(EBGM05)	IC(IC025)
HEADACHE	9(15.7)	Nervous system disorders	4.16 (2.15–8.07)	4.1 (21.03)	4.08 (2.34)	2.03 (1.1)
SYNCOPE	4(7.0)	Nervous system disorders	3.23 (1.21–8.68)	3.22 (6.08)	3.2 (1.4)	1.68 (0.38)
UNEVALUABLE EVENT	4(7.0)	General disorders and administration site conditions	44.15 (15.84–123.04)	43.78 (154.13)	40.42 (17.15)	5.34 (3.98)
BRAIN OEDEMA	3(5.3)	Nervous system disorders	8.67 (2.76–27.24)	8.62 (19.89)	8.5 (3.26)	3.09 (1.62)
RASH	3(5.3)	Skin and subcutaneous tissue disorders	7.02 (2.24–22.02)	6.98 (15.19)	6.9 (2.65)	2.79 (1.33)

Reports in pregnant individuals

A total of 18 reports involving pregnant individuals were identified, including one case in a minor and 10 in adults, while the remaining cases lacked age documentation. All reported events were classified as non-serious, and no fatal outcomes were recorded.

Although preliminary signal detection analyses suggested elevated reporting odds for certain terms, these were primarily associated with non-clinical classifications such as pregnancy exposure or product use errors. After excluding terms unrelated to actual clinical symptoms, no positive signals were identified. These findings indicate that, based on the available data, there is no evidence of increased risk for adverse events following GARDASIL 9 administration during pregnancy.

Discussion

This study provides a ten-year overview of GARDASIL 9-associated adverse event reports submitted to the VAERS system, with a focus on disproportionality signal detection and subgroup-specific outcomes. The overwhelming majority of reported events were non-serious in nature, with dizziness, syncope, headache, and various injection site reactions being the most frequently documented PT, consistent with the safety profile described in pre-licensure trials and official product labeling. These outcomes also align with prior passive and active surveillance reports for both GARDASIL 9 and its predecessor vaccines, reinforcing the vaccine’s favorable tolerability in routine immunization practice. Of note, individuals under 18 years of age accounted for over 70% of all reports, and age-stratified signal analysis revealed that syncope, dizziness, and loss of consciousness were particularly enriched in this subgroup. These PTs, all indicative of transient autonomic or neurocardiogenic responses, are well-documented in adolescent vaccine recipients and likely reflect heightened vasovagal reactivity rather than vaccine-specific toxicity.

Among serious events, signals associated with transient neurologic or autonomic dysfunction were prominent. In particular, elevated ROR values were observed for syncope, pallor, and loss of consciousness, likely reflecting neurocardiogenic responses such as vasovagal syncope, especially in adolescent populations – a well-characterized phenomenon across multiple vaccines. Especially, a strong signal was identified for postural orthostatic tachycardia syndrome (POTS), a condition of autonomic dysregulation previously flagged in registry-based and anecdotal reports involving HPV vaccination, though no regulatory authority has thus far established a causal relationship.^20,21 The detection of POTS primarily in serious reports following GARDASIL 9 monotherapy may warrant renewed prospective evaluation, particularly given the reproducibility of this signal across multiple disproportionality models.

This study also uncovered a set of signals not previously emphasized in labeling or earlier reviews, including eye movement disorders, posture abnormalities, muscular weakness, and autoimmune thyroiditis. While these outcomes remain infrequent in absolute terms, their consistently elevated RORs and corresponding IC and EBGM estimates suggest possible mechanistic pathways involving neuroimmune or autoimmune modulation. Several of these signals – such as autoimmune thyroiditis and muscular rigidity – are biologically plausible in the context of post-immunization syndromes and may reflect molecular mimicry or immune system perturbation in genetically susceptible hosts. Importantly, these signals were not observed exclusively among adolescent recipients, indicating the potential relevance of monitoring across broader age ranges.

Compared to GARDASIL 4, GARDASIL 9 introduces five additional high-risk HPV types and contains a higher total antigenic load and adjuvant concentration. These compositional changes have been hypothesized to contribute to the slight increase in local reactogenicity and systemic complaints observed in clinical trials.¹ Our findings extend this observation by showing that although the overall safety signal remains reassuring, the range of detectable signals in post-marketing data for GARDASIL 9 appears broader than for GARDASIL 4, particularly in relation to dysautonomia-related PTs and autoimmune manifestations.²² Nonetheless, definitive comparisons are limited by differences in usage periods, reporting bias, and temporal distance from initial licensure.

Death reports accounted for a small proportion (0.62%) of the total VAERS entries but continue to attract considerable public and regulatory attention. This study identified 57 death reports following GARDASIL 9 vaccination, 36 of which occurred after administration of GARDASIL 9 alone. While multiple PTs – including syncope, brain edema, headache, rash, and unevaluable event – displayed positive disproportionality signals across algorithms, most lacked clinical narratives sufficient to ascertain underlying etiologies. For example, “Unevaluable Event,” which yielded the highest ROR, reflects incomplete case documentation rather than a specific pathological process. Brain edema, one of the few PTs plausibly associated with central nervous system pathology, was reported in only four cases and lacked confirmatory diagnostic data. Similarly, terms such as syncope and headache, while flagged statistically, are nonspecific and insufficient to infer causality in isolation. No consistent temporal pattern or pathophysiologic mechanism emerged across the fatal reports, and a large proportion of cases were missing autopsy data or detailed clinical follow-up. These findings remain aligned with the conclusions of previous regulatory evaluations, which have not identified any causal link between GARDASIL 9 and mortality events.

Reports involving pregnant recipients were notably limited, and no serious outcomes were observed after filtering out gestational exposure codes and product-related terms. This observation is consistent with existing cohort data showing no increase in adverse maternal or neonatal outcomes associated with inadvertent HPV vaccination during pregnancy.²³ However, the small sample size and lack of systematic pregnancy surveillance preclude firm conclusions, underscoring the need for prospective registries in this population.

While many of the signals detected herein align with established pharmacovigilance findings, the emergence of atypical or unlisted events – particularly those involving neurological or autoimmune symptoms – raises important questions about individual vulnerability, delayed-onset syndromes, and signal persistence. For instance, conditions such as eye movement disorder, thyroid mass, and posture abnormality – though absent from the package insert – were among those with significant disproportionality metrics across all four algorithms, suggesting that signal robustness was not method-dependent.¹⁶ Given the rarity of these events, their reproducibility merits further exploration through longitudinal observational studies and biological modeling rather than immediate changes to regulatory status.

Ultimately, our analysis affirms the overall safety of GARDASIL 9 while identifying a small subset of disproportionate PTs that warrant closer scrutiny. Although passive surveillance tools like VAERS are inherently limited by underreporting, stimulated bias, and the absence of denominators, they play an indispensable role in hypothesis generation.¹⁰ Moreover, the lack of a defined population denominator and absence of an internal control group preclude reliable incidence estimation or definitive attribution of causality. These constraints highlight the necessity of interpreting VAERS signals in conjunction with data from active surveillance systems. Triangulation of VAERS-derived signals through active monitoring platforms such as the Vaccine Safety Datalink (VSD) and post-marketing registries will be crucial in contextualizing these observations and guiding evidence-based policy decisions.

Conclusion

This ten-year VAERS-based analysis reaffirms the overall safety profile of GARDASIL 9, with most reported adverse events being non-serious and consistent with pre-licensure data. Several neurological and autonomic symptoms, including postural orthostatic tachycardia syndrome, showed elevated disproportionality signals and warrant continued monitoring. No new safety concerns were identified in reports involving pregnancy, and death cases lacked sufficient detail to establish causality. These findings support the favorable benefit-risk balance of GARDASIL 9 while underscoring the importance of ongoing pharmacovigilance to detect rare or unexpected events and sustain public confidence in HPV vaccination programs.

Biographies

Guojun Liang, specializing in the field of surgery, the analysis and research of public health databases, proficient in big data mining and statistical analysis.

Yang Song, Visiting Scholar at University of California, Davis Joint Orthopaedic Center. Be proficient in designing multi-center clinical trials to address specific diseases among different populations, and at the same time evaluate the impact of health policies on resource allocation.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. During the preparation of this work the author(s) used DeepL (https://www.deepl.com/en/write) in order to improve readability. After using this tool/service, the author(s) reviewed and edited the content as needed. The author(s) take(s) full responsibility for the content of the publication.

Data availability statement

The data related to this study has been stored in a publicly accessible repository (https://vaers.hhs.gov/data/datasets.html). All data produced or examined in this article are available within the published paper and its supplementary information files.

Ethics statement

This study involves a secondary analysis of publicly accessible summary statistics and does not necessitate explicit ethical approval.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/21645515.2025.2530831