Introduction: It is not known whether there are underlying physiologic or immunologic differences between febrile seizures (FS) triggered by vaccines versus other causes. Furthermore, while secular and individual-level factors have been associated with FS risk, they are rarely evaluated simultaneously.

Methods: Subjects included members of Kaiser Permanente Southern California aged 6 months to 3 years from July 1, 2003-December 31, 2011. Primary outcome was first diagnosis of FS. Vaccine-associated (VA) FS were defined as those occurring from day 0 to day 15 following any vaccine; non-vaccine associated (NVA) FS were those outside this period. We compared incidence rates of VA-FS versus NVA-FS. Poisson regression was used to assess the association between FS and secular and individual-level factors. We also evaluated interactions between vaccine exposure and each model covariate on the risk of FS.

Results: Among 265,275 children, 3348 FS were identified; 383(11%) were VA-FS, and 2965(89%) were NVA-FS. Incidence rates were 2.73 and 2.05 per 100,000 person-days for VA-FS and NVA-FS, respectively. Multivariable analyses confirmed previously reported increased risk of FS by age, low gestational age, and winter months. Increased risk was also associated with VA exposure (RR=1.63[95% CI: 1.27-2.11]), non-White race/ethnicity vs. White (African-American RR=1.41[1.22-1.63]; Asian RR=1.58[1.40-1.79]; Hispanic RR=1.60[1.47-1.75]), and maternal age 29 years or less vs. 40+ years (≤ 19 years RR=1.28[1.00-1.65]; 20-29 years RR=1.21[1.02-1.42]). Females were at lower risk of NVA-FS (RR=0.77[0.72-0.83]), but were similar to males for VA-FS (RR=0.97[0.79-1.19]). Children with low 1 min Apgar scores (≤ 3) had increased risk of VA-FS (RR=3.40[1.86-6.22]), but no increased risk for NVA-FS (RR=1.05[0.69-1.60]) compared to children with normal Apgar scores (≥ 7).

Discussion: This study suggests that there may be immunogenetic differences underlying VA-FSs compared with other FSs. However, further studies are needed. An understanding of the mechanisms behind these findings may help improve vaccine design or policies.