Introduction: Which of the two leading infectious killers of children—lower respiratory infections (LRIs) or diarrheal diseases—ends first should order child-survival investments, yet whether their transitions are symmetric is unknown. We compared the systems on a single counterfactual scale—a cross-system death-ratio series with formal uncertainty—at ages 0–19 years in the Global Burden of Disease Study (GBD) 2023. Methods: We extracted deaths and incident episodes, 1990–2023, from the GBD 2023 Results Tool, aggregating the under-5 band plus ages 5–9, 10–14 and 15–19. We computed cross-system death-ratio series, estimated annual percentage changes (EAPCs), pandemic-window changes with Monte Carlo uncertainty, an episode-fatality ratio proxy (EFR; deaths per 1,000 episodes, for 204 countries in 2023), and cross-system Escherichia coli trend correlations. Results: In 1990 the systems were indistinguishable (2.04 million LRI versus 2.00 million diarrhea deaths; ratio 0.98, 95% uncertainty interval 0.74–1.31). By 2023, LRI deaths (711,713) exceeded diarrhea deaths (379,958) 1.87-fold (ratio 0.53, 0.36–0.79; parity first excluded in 2013 and in every year thereafter except the 2020–2021 pandemic window). Improvement was age-synchronous for diarrhea but age-asymmetric for LRI (5–19-year EAPC −4.48% versus −1.74% per year). The pandemic window suppressed then released LRI deaths (year-on-year −18.4%, −9.0%, +12.8%, +8.1%; rebound in all five Socio-demographic Index quintiles) while diarrhea declined monotonically. Fatality gradients widened twice over: the 2023 LRI EFR was 49 times the diarrhea EFR; the low-to-high-SDI gradient steepened from 2.58 to 7.29 (LRI) and 6.40 to 22.50 (diarrhea); and country-level diarrhea EFRs tracked GDP per capita (ρ=−0.76) and LRI EFRs (ρ=0.85). E. coli mortality, coupled across systems until 2010 (ρ=1.000), decoupled thereafter (ρ=0.371): enteric deaths halved, respiratory deaths plateaued. Conclusion: The two systems share a starting point but no longer a trajectory. Pediatric infection frameworks should treat fatality, not incidence, as the binding constraint for LRI and extend case management and surveillance to the school-age children the transition has left behind.