Critical care
Are there hidden biological types of septic shock in children?
Open access · cc by · source: Europe PMC
Clustering children with septic shock by the genes active in their blood revealed three subgroups, and the one with switched-off immune and zinc-related genes was the sickest and most likely to die.
Study at a glance
- Design
- Other — Observational discovery study: whole-blood microarray profiles taken within 24 hours of PICU admission were clustered without using outcome data (hierarchical clustering, then ANOVA-filtered genes, principal components and K-means), and the resulting subclasses were compared on clinical features; leave-one-out cross-validation tested how well a reduced gene set reassigned patients.
- N
- N=98 · 98 children with septic shock from 11 institutions, compared against 32 healthy control children for normalisation; 67 patients and all controls had appeared in earlier analyses of the same dataset.
- Population
- Children aged 10 years or younger admitted to paediatric intensive care units in the United States with septic shock.
- Outcome
- Gene-expression-defined subclasses and their association with illness severity (PRISM III), organ failure and mortality.
Structured fields used in claim comparison tables when every cited study has a complete layer.
Key findings
Three subclasses (A, B and C) emerged, separated by thousands of differently expressed genes. Subclass A had higher illness-severity scores, more organ failure and higher mortality (36%) than B and C, and its patients showed broad repression of genes for adaptive immunity, glucocorticoid-receptor signalling and zinc biology, without differing in lymphocyte counts. A support vector machine using 307 pathway genes reassigned 89 of 98 patients (91%) to their subclass in leave-one-out cross-validation.
Methodology
The team took blood samples from 98 children in septic shock within a day of intensive care admission and measured the activity of tens of thousands of genes using microarrays. Without telling the analysis who survived, they clustered patients by gene-expression pattern, checked the clusters with principal component analysis, and then compared the resulting groups on illness severity, organ failure and death. They also tested whether a smaller set of pathway genes could reassign patients to the right group.
Limitations
This is a discovery analysis in only 98 patients with no independent validation cohort, so the three subclasses might not replicate; the cross-validation reused the same data that defined the groups. The clustering choices (fold-change filters, where to cut the tree) were somewhat arbitrary, and whole-blood RNA could partly reflect different mixes of white cells rather than changes within cells. The study shows associations with outcome, not that repressed immune or zinc genes cause death or that targeting them would help.
How this study connects
Role on claims
Each row is a claim on a concept or method page where this paper supports, challenges, or qualifies the statement. Roles are hand-checked — not a model guess.
Not yet placed on a claim. This paper has study layers, but no concept page yet cites it as support, challenge, or qualifier.
Related papers in this topic
Same topic cluster — not a recommendation engine.