(889) Hypo- and Hyperinflammatory Phenotypes in Children With Sepsis/Septic Shock

Introduction: Hypo- and hyper-inflammatory phenotypes of adult and pediatric ARDS have been identified with distinct response to treatment and outcomes. However, the applicability of these biologic phenotypes to other critical illnesses is not known. We hypothesize that hypo- and hyper-inflammatory phenotypes with distinct outcomes would be identified using similar cytokines in children with sepsis as previously reported in ARDS.

Methods: We performed a secondary analysis of 161 patients < 18 years treated for sepsis/septic shock at a single academic children’s hospital who were previously enrolled in an observational study that included serial measurement of cytokines and chemokines.  Using measures of plasma interleukin (IL)-6, IL-8, TNFa, IFNa, sTNFr1, and sTNFr2 collected on day 1-2 following ICU admission for sepsis, we applied latent class analysis (LCA) to identify data-driven subgroups without specification as to the ideal number of groups. Clinical characteristics and hospital mortality were compared between subgroups using Wilcoxon rank sum and Fisher’s exact tests.

Results: LCA using day 1-2 plasma IL-6, IL-8, TNFa, IFNa, sTNFr1, and sTNFr2 identified two distinct phenotypes. Phenotype 1 exhibited lower levels of mean cytokines than phenotype 2, including IL-6 (113 vs 1,844 pg/ml, p< 0.05), IL-8 (222 vs 2,614 pg/ml, p< 0.05), sTNFr1 (3,382 vs 14,672 pg/ml, p< 0.05), and sTNFr2 (20,948 vs 69,318 pg/ml, p< 0.05). TNFa (9.4 vs 5.8 pg/ml, p=0.27) and IFNa (55 vs 65 pg/ml, p=0.26) did not differ significantly between phenotypes.  Age, sex, and race did not differ between phenotypes, but mortality was lower in hypo-inflammatory phenotype 1 compared to hyper-inflammatory phenotype 2 (5% versus 31%, p=0.002), even after controlling for baseline illness severity with PRISM-III (aOR 0.11, 95% CI 0.03-0.44).

Conclusions: Similar to previous reports in ARDS, hypo- and hyper-inflammatory phenotypes are evident in pediatric sepsis with lower mortality in the hypo-inflammatory group. Future research is needed to determine the utility of identifying such phenotypic differences in sepsis and other pediatric critical illnesses.