(886) The Stability of the Transcriptome in Childhood Sepsis

Introduction: Whole blood gene expression analyses (transcriptomics) have been used to identify groups of patients with sepsis that share biological features (endotypes) that may respond to similar treatments. Endotypes have been validated across adult cohorts, but less so in child cohorts, where the immune response to infection is demonstrably different to adults. In critical care, childhood mortality from sepsis is lower than in adults, and most mortality occurs in the first 24 hours following presentation to hospital. This censors the most informative patients and predisposes false inferences from sampled cohorts. We sought to define early gene expression changes in childhood sepsis, examine the stability of the transcriptome, and describe associations with clinical trajectory.

Methods: We prospectively sampled children with sepsis at first contact with paediatric critical care (PCC) retrieval physicians (T1), and at 48 hours (T2). We selected 19 children with septic shock and with paired T1 and T2 samples for further analysis. Whole blood was sampled into PAXgene (Qiagen, UK) tubes, and RNA was extracted using standard methods, globin and DNA depleted, and sequenced (Illumina NextSeq2000). Detailed clinical data from the cohort were extracted from prospectively collected data. Bioinformatic analyses using differential gene expression analysis and hierarchical clustering are ongoing.

Results: Of 19 children analysed, 5 (26%) were < 1 year of age, 8 (42%) were 1–4 years of age and 6 (32%) were ≥5 years of age; 12 (63%) were male. By clinical syndrome, 3 (16%) had respiratory tract infection, 3 had meningitis, 3 had unknown source, 1 had soft tissue infection (5%) and 5 (26%) had other sources: all met criteria for septic shock. Children had a median of 42 (IQR 15–64) hours of vasoactive therapy and 125 (IQR 87–198) hours of mechanical ventilation; 9 (47%) had steroids for refractory septic shock, ARDS or meningitis. All children survived and had a median of 22 (16–24) PCC-free days at day 30.

Conclusions: This cohort represents the trajectory of severe sepsis in children. Our analyses will describe early gene expression changes, inform the stability of endotypes in childhood sepsis, and may highlight treatment effects on the dynamic transcriptome.