Background
Comparative evidence on antiviral treatment in pediatric influenza remains limited. We compared the clinical features of influenza A and B and assessed the comparative outcomes of oseltamivir, baloxavir marboxil, and peramivir in children.

Methods
This retrospective single-center cohort included 3,718 children with nucleic acid-confirmed influenza A or B who presented to a tertiary pediatric emergency department between November 2023 and April 2024 and initiated antiviral treatment within 48 h of fever onset. Primary endpoints were time to fever resolution and time to full symptom resolution. Propensity score matching was performed separately within the influenza A and influenza B cohorts for prespecified pairwise comparisons, followed by Kaplan–Meier and Cox regression analyses.

Results
Among 3,718 children, 1,717 had influenza A and 2,001 had influenza B. Influenza A was more often associated with higher fever and neurologic complications, whereas influenza B more commonly presented with gastrointestinal symptoms. After propensity score matching, baloxavir and peramivir were associated with shorter times to fever resolution than oseltamivir in both cohorts (influenza A: baloxavir 1.57 days vs. oseltamivir 2.11 days, peramivir 1.72 days vs. oseltamivir 2.23 days; influenza B: baloxavir 1.50 days vs. oseltamivir 2.29 days, peramivir 1.71 days vs. oseltamivir 2.25 days). Differences in time to full symptom resolution were smaller and less consistent across comparisons. Overall adverse events were documented in 16.5% of children receiving oseltamivir, compared with 5.0% receiving peramivir and 3.6% receiving baloxavir; notably, no serious adverse events were observed across the cohorts.

Conclusions
Influenza A and B showed distinct clinical phenotypes in this pediatric cohort. In propensity score-matched analyses, baloxavir and peramivir were associated with modestly faster fever resolution and lower recorded rates of adverse events than oseltamivir. Given the retrospective, non-randomized design and the modest absolute differences observed, these findings should be interpreted as associative and hypothesis-generating rather than as evidence of causal superiority. Prospective multicenter studies with standardized outcome assessment are needed.