(339) Data mining for adverse events of influenza vaccine in children using a tree-temporal scan statistic analysis

Background: Influenza vaccines demand continuous safety surveillance due to the annual variation in recommended vaccine strains. Additionally, the increased use of quadrivalent vaccines in children may alter the current safety profile from what was previously known.

Objectives: To identify potential adverse events (AEs) following influenza vaccination in children from 2016/17 to 2021/22 flu seasons using the tree-temporal scan statistic data-mining method

Methods: The study population consisted of influenza vaccinees aged 12 or less during six flu seasons (2016/17 to 2021/22) from a nationwide linked database of the Korea Disease Control and Prevention Agency immunization registry and the National Health Insurance Service claims data. Children receiving multiple flu vaccinations per season were excluded. The self-controlled tree‐temporal scan statistic was used to identify statistically significant clustering within the hierarchy of ICD-10 code with 5 levels. We excluded some diagnosis categories indicating the conditions such as neoplasms and congenital or hereditary conditions from the tree as implausible vaccine-related. Incident diagnoses recorded in the inpatient setting within 1–84 days post-vaccination were included. We evaluated all temporal risk windows between 2 and 42 days long, started 1–28 days and ended 2–42 days after vaccination. The control window consisted of the days within the follow-up period that were not in the risk window. Patients diagnosed with same code for the first 3 characters in the year prior to vaccination were excluded. The specified p-value cut-off for statistical significance was 0.01.

Results: Eligible patients numbered 83,122 in 2016/17; 94,862 in 2017/18; 123,188 in 2018/19; 136,229 in 2019/20; 50,961 in 2020/21; and 67,670 in 2021/22 flu seasons. Each season detected between 71 and 120 clusters. While clusters varied by season, common clusters included (1) unspecified adverse effects such as infection following immunization (Days 1–2, p < .001); (2) conditions suggestive of influenza such as respiratory disease, including pneumonia (Days 2–42, p ≤ .002), or hematologic disease, including anemia (Days 3–39, p ≤ .005); (3) local or systemic adverse effects known to be associated with some vaccines, such as cellulitis (Days 1–19, p ≤ .001) and convulsions (Days 1–3, p < .001); (4) genitourinary disorders (Days 2–39, p ≤ .007); and (5) injuries caused by external factors (Days 1–41, p ≤ .009).

Conclusions: The study confirmed existing adverse reactions to influenza vaccines and noted unrelated events, like injuries from external factors. Nonetheless, the potential indirect link between these injuries and vaccine-related events requires further study to better understand their relationship.