Complete blood count-derived inflammatory indices for discriminating febrile seizures from epilepsy in children: a retrospective comparative study

Abstract Background Inflammation is implicated in febrile seizures and epilepsy, but the clinical value of inflammatory indices derived from the complete blood count (CBC) remains uncertain. We compared the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), systemic immune-inflammation index (SII), systemic inflammation response index (SIRI), and aggregate index of systemic inflammation (AISI) between children with febrile seizures and those with epilepsy and evaluated their discriminative and incremental value. Methods This single-center retrospective study included 100 children with febrile seizures, 98 with epilepsy, and 102 healthy controls evaluated between January 2021 and April 2026. In the patient groups, the first CBC from the first eligible presentation was analyzed. Groups were compared using Kruskal–Wallis and Holm-adjusted Mann–Whitney U tests. Discrimination was assessed by receiver operating characteristic analysis. Log2-transformed indices were entered into separate logistic regression models adjusted for age and sex, and incremental discrimination over an age–sex base model was summarized as the change in the area under the curve (ΔAUC). Results Median age differed among the febrile seizure, epilepsy, and control groups (1.75, 4.00, and 3.50 years, respectively; p < 0.001), whereas sex distribution did not. NLR, SII, SIRI, and AISI differed among the groups, but PLR did not. SIRI and AISI were highest in febrile seizures, intermediate in epilepsy, and lowest in controls. Discrimination between febrile seizures and epilepsy was poor: the AUC was 0.606 (95% confidence interval [CI] 0.527–0.683) for AISI and 0.591 (95% CI 0.510–0.669) for SIRI. After adjustment for age and sex, each doubling of NLR, SII, SIRI, and AISI was associated with higher odds of febrile seizures (adjusted odds ratios 1.35–1.54), but adding these indices to the base model (AUC 0.792) increased the AUC by only 0.012–0.016. Conclusions CBC-derived inflammatory indices, particularly SIRI and AISI, differed across groups but did not distinguish febrile seizures from epilepsy with clinically useful accuracy and added little beyond age and sex. Without a febrile non-seizure control group, seizure-related effects could not be separated from those of fever and infection. These indices are best regarded as exploratory markers of systemic inflammation rather than stand-alone diagnostic tests.