Adverse event reports were retrieved from the Vaccine Adverse Event Reporting System (VAERS) via the CDC WONDER interface, covering the period from January 1, 2013–December 31, 2024 (accessed March 2025). The dataset included patient demographics, vaccine identifiers, Medical Dictionary for Regulatory Activities (MedDRA) preferred terms, and serious event indicators (6).

The study included reports for individuals younger than 18 years of age where an FDA-approved quadrivalent influenza vaccine (QIV) was coded as the primary suspected vaccine. Reports were systematically excluded if they (i) lacked age information, (ii) did not specify the QIV formulation, (iii) were flagged as duplicates through an automated pipeline (collapsing records with identical VAERS_ID and retaining only the most recent RECVDATE), or (iv) were incomplete, defined as lacking both vaccine type and any coded MedDRA preferred term. Data cleaning and handling of missing values were automated using scripted rules in R without manual adjudication or imputation; analyses requiring specific variables (e.g., sex, onset interval) were performed on complete-case subsets.

Six FDA-approved QIVs were evaluated:

Serious adverse events (SAEs) were defined according to FDA criteria, including death, life-threatening conditions, hospitalization or prolonged hospitalization, persistent disability, and congenital anomalies (25). All adverse events were coded using MedDRA (version 27.0).

Categorical variables were summarized as frequencies and proportions, and continuous variables as medians with interquartile ranges (IQRs). Age was stratified into four groups: younger than 6 months, 6 months–5 years, 6–12 years, and 13–17 years.

Disproportionality analyses employed two complementary metrics: (i) reporting odds ratio (ROR), calculated as ROR = (a × d)/(b × c), where a = reports with the event and vaccine of interest, b = other events with the vaccine of interest, c = the event with other vaccines, and d = other events with other vaccines; and (ii) information component (IC), an empirical Bayesian shrinkage measure robust for sparse data, with IC025 indicating the lower 95% credibility bound. A safety signal was identified when both the lower 95% confidence interval of the ROR exceeded 1.0 and IC025>0. To account for multiple comparisons across MedDRA preferred terms, p-values from ROR analyses were adjusted using the Benjamini-Hochberg false discovery rate (FDR) procedure (q < 0.05) (26). All analyses were conducted using R statistical software (version 4.4.2).

This study utilized publicly available and de-identified VAERS data; therefore, institutional review board (IRB) approval was not required. The analysis was conducted in accordance with the principles of the Declaration of Helsinki.

A total of 15,458 pediatric adverse event reports were included, with a median patient age of 7.3 years [interquartile range (IQR): 2.8–11.5 years]. The gender distribution was balanced, with females representing 47.5% of cases. Vaccine-specific report distribution was as follows: Fluzone (39.9%, n = 6,166), Fluarix (16.8%, n = 2,596), FluLaval (16.7%, n = 2,581), FluMist (15.4%, n = 2,373), Flucelvax (7.7%, n = 1,192), and Afluria (3.6%, n = 550).

Serious adverse events (SAEs) were reported in 818 cases (5.29%). The SAE proportions by vaccine were highest for Afluria (6.73%), followed by Fluarix (6.32%), FluMist (5.77%), Fluzone (5.71%), FluLaval (3.87%), and Flucelvax, which exhibited the lowest SAE rate (2.35%).

There were 67 reported deaths, corresponding to a crude mortality rate of 0.43%. Notably, 73% of fatalities occurred among children aged 6 months to 5 years, with 54% involving Fluzone. These associations are temporal, and causality should not be inferred from these reports given the limitations of passive surveillance data (27).

An inverse relationship between age and SAE frequency was observed across all vaccines. Among children aged 6 months to 5 years, SAE rates ranged from 2.37% (Flucelvax) to 10.16% (Afluria). Adolescents aged 13–17 years showed substantially lower SAE rates, ranging from 1.78% to 6.73%. Table 1 presents the detailed age-specific distribution of serious adverse events following quadrivalent influenza vaccination.

A total of 251 MedDRA preferred terms met predefined criteria for disproportionality signals across all QIVs. Notable vaccine-specific safety signals included syncope and seizures associated with Afluria, injection-site reactions linked primarily to egg-based vaccines, and respiratory symptoms observed with FluMist.

Among evaluated rare adverse events, Guillain-Barré syndrome (GBS) was the sole condition meeting both statistical thresholds, with 42 reported cases (ROR = 1.71, 95% CI: 1.25–2.35; IC025 = 0.28). The median time to symptom onset for GBS was 14.5 days (IQR: 4.5–32 days), with distribution across vaccines as follows: Fluzone (n = 16), Fluarix (n = 7), FluMist (n = 6), FluLaval (n = 4), Afluria (n = 3), and Flucelvax (n = 0). Table 2 presents the top five MedDRA preferred terms showing positive disproportionality signals, while Table 3 provides a comprehensive summary of rare adverse events.

For context, the background incidence of Guillain-Barré syndrome in children is estimated at approximately 0.3–1.3 per 100,000 person-years; thus, while GBS met both disproportionality thresholds (ROR and IC), this likely reflects reporting enrichment rather than a true increase in clinical incidence (21, 22, 28, 29).

Annual reporting volume peaked in 2018 with 1,840 reports, followed by a significant reduction ranging from 24.8% to 66.5% during 2021–2024, coinciding with disruptions related to the COVID-19 pandemic. Figure 1 illustrates the annual pediatric VAERS reports by quadrivalent influenza vaccine formulation from 2013 to 2024, clearly demonstrating the impact of the pandemic on reporting patterns.

Among the inactivated vaccines, the median time to adverse event onset was 0 days (IQR: 0–1 day), with 75% of events occurring within 24 h post-vaccination. FluMist demonstrated a slightly broader onset range (IQR: 0–2 days).

To assess potential underreporting during the COVID-19 pandemic, we compared pediatric VAERS reporting volume with national influenza vaccination coverage estimates. Coverage among children aged 6 months-17 years declined from 63.7% (2019–20) to 58.6% (2020–21),and remained relatively stable at 57.8% (2021–22) and 57.4% (2022–23), with a further decline to 55.4% (2023–24) (30–33).By contrast, VAERS reports dropped by 42%–66.5%, a disproportionate decline relative to vaccine uptake, supporting the inference that pandemic-related factors-such as reduced healthcare access, shifting reporting priorities, and system burden-likely contributed to underreporting rather than a true reduction in adverse event incidence.

Figure 2 presents disproportionality analyses for the 30 most frequently reported adverse events following quadrivalent influenza vaccination in children. The forest plots demonstrate vaccine-specific reporting odds ratios (ROR) and Bayesian information components (IC) with 95% confidence intervals for each of the six FDA-approved QIVs. Significant signals were defined by ROR lower bound >1.0 and IC025>0, with notable variations observed across different vaccine formulations.

To further characterize the safety signal landscape, Figure 3 displays volcano plots of disproportionality signals for each QIV. These plots effectively visualize the relationship between statistical significance and effect size, with each point representing an individual MedDRA preferred term. The x-axis represents the log₂-transformed reporting odds ratio (ROR), while the y-axis indicates the log₁₀ of the false discovery rate (FDR)-adjusted p-value. Terms positioned further to the right and higher on the graph represent stronger disproportionality signals, with dot size indicating the number of reports.

To further explore formulation-specific safety profiles, we conducted a sensitivity analysis stratified by vaccine type-inactivated influenza vaccines (IIV4) vs. live attenuated influenza vaccine (LAIV4). This analysis integrated findings from Tables 1, 2 and Figures 2–4. IIV4s (Fluzone, Fluarix, FluLaval, Afluria, and Flucelvax) collectively accounted for 84.6% of reports and were more frequently associated with neurological and injection-site events, such as syncope, seizures, and injection-site erythema. In contrast, LAIV4 (FluMist) was more commonly linked to upper respiratory symptoms including rhinorrhea and nasal congestion. Time-to-onset patterns were broadly similar, with most events occurring within the first day post-vaccination, though LAIV4 showed a slightly wider onset range (IQR: 0–2 days). These trends are consistent with known reactogenicity profiles and vaccine administration routes.

Figure 4 presents Kaplan–Meier curves depicting the cumulative incidence of adverse events over time for each quadrivalent influenza vaccine formulation. Across all vaccines, the median onset was 0 days, indicating an acute response profile. Over 95% of adverse events occurred within the first 7 days, with most reported on the day of vaccination. These rapid onset patterns support the acute nature of most reported events and align with known immunization response timelines.

This large-scale post-marketing surveillance study provides robust evidence supporting the overall safety of quadrivalent influenza vaccines (QIVs) in pediatric populations. The observed 5.29% rate of serious adverse events (SAEs) reflects the inherent characteristics of passive surveillance systems. Large-scale active surveillance studies conducted through the Vaccine Safety Datalink demonstrate substantially lower rates of confirmed serious adverse events, with one study of over 590,000 live attenuated vaccine doses finding only rare confirmed cases of anaphylaxis (1.7 per million doses) and syncope (8.5 per million doses) (34), while another study of over 91,000 inactivated vaccine doses in young children found no evidence of serious medically attended events (35). However, interpretation must consider the inherent limitations of passive surveillance systems, which typically capture only 1%–10% of actual adverse events and are more likely to detect serious outcomes (36).

Among the rare adverse events assessed, Guillain-Barré syndrome (GBS) was the only one that met both signal detection thresholds (ROR=1.71), consistent with selected epidemiological studies (37). While some meta-analyses have found no confirmed risk of vaccine-associated GBS (38), CDC surveillance has reported modest relative risk increases within expected post-vaccination intervals (17). The median onset of 14.5 days is biologically plausible and supports a potential immune-mediated mechanism.

Importantly, the absolute risk of GBS following QIV administration remains exceedingly low. Even at the upper limit of the confidence interval, the estimated excess risk is 1–2 additional cases per million doses. This must be balanced against the substantially higher GBS incidence following natural influenza infection, which is estimated to be 4–7 times greater (37, 39). Recent pediatric surveillance data indicate a background incidence of Guillain-Barré syndrome of approximately 0.34–0.69 per 100,000 person-years in children under 16 years (22). Therefore, although GBS met both disproportionality thresholds (ROR and IC), this signal most likely reflects nhanced reporting rather than a clinically meaningful increase in incidence.

The elevated SAE rates observed in children aged 6 months to 5 years are clinically noteworthy. This group exhibited 2- to 4-fold higher SAE rates compared to adolescents, which may reflect age-related differences in immune response and heightened parental or provider vigilance in reporting (40, 41). The immature immune system characteristics in early childhood, including differences in innate and adaptive immune responses, may contribute to varying reactogenicity profiles across pediatric age groups (40, 42).

Flucelvax demonstrated the lowest SAE rate (2.35%), suggesting potential safety advantages associated with cell culture-based platforms, including the absence of egg-adaptive mutations (43) and possibly lower innate immunogenicity due to production processes. Conversely, the elevated frequency of neurological adverse events associated with Afluria is noteworthy, especially given prior reports linking this formulation to febrile seizures (44, 45). These findings support the hypothesis that manufacturing methods and antigenic composition may influence vaccine reactogenicity.

The marked decline in annual adverse event reporting (24.8%–66.5%) during 2021–2024 likely reflects the multifactorial consequences of the COVID-19 pandemic, including reduced pediatric immunization visits, shifts in healthcare delivery, and reporting fatigue among healthcare providers (24, 46, 47). Additionally, changes in public health messaging, vaccine formulation preferences, and seasonal uptake may have further confounded reporting trends.

These findings may inform tailored vaccine recommendations, particularly in pediatric subgroups. For instance, cell-based QIVs may offer preferable safety profiles for immunocompromised children or those with a history of seizures (48, 49). Continued surveillance and stratified safety assessments are essential for optimizing vaccine selection across diverse populations (50).

This study is subject to several limitations inherent to the VAERS system, including underreporting, reporting and selection bias, absence of denominator data, and the inability to determine causality without an unvaccinated comparator group. These constraints highlight the importance of corroborating findings with active surveillance systems and well-controlled epidemiological studies.