Exploring disease axes as an alternative to distinct clusters for characterizing sepsis heterogeneity
Authors' conclusion
Cluster analysis of sepsis patients across various cohorts showed modest reproducibility. Sepsis heterogeneity is better characterized through continuous disease axes that coexist to varying degrees within the same individual instead of mutually exclusive subtypes.
Study context
- Population
- Sepsis patients from six distinct cohorts with similar sepsis-related characteristics
- Participants
- 63,547 patients · 6 cohorts
- Exposure
- Top-down vs Bottom-up
- Outcome
- Reproducibility of sepsis subtypes across cohorts (normalized mutual information, 0 to 1, higher = better concordance)
Methodological detailssix cohorts, 27 clustering schemes, principal component analysis
- Six cohorts
- 27 clustering schemes
- Principal component analysis
Main finding
As reported in the abstract
Reproducibility of clustering subtypes (NMI)
Top-down
0.64
median · IQR 0.49–0.78
Bottom-up
0.23
median · IQR 0.2–0.31
P < 0.001
- Reproducibility across studies modest (median normalized mutual information (NMI) range)
- 0.08 to 0.54
Observational data: associations, not effects.
Further findings
Reproducibility within United States (US) cohorts
- Greater reproducibility when transfer performed within US cohorts
Disease axes (principal component analysis (PCA))
- Correlation pattern between variables consistent across all cohorts; first two axes were "shock axis" and "systemic inflammatory response syndrome (SIRS) axis"
Original publication
Exploring disease axes as an alternative to distinct clusters for characterizing sepsis heterogeneity. Intensive Care Med. Published 4 October 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →