Invasive group A streptococcus infections in the intensive care unit: an unsupervised cluster analysis of a multicentric retrospective cohort
Authors' conclusion
Based on simple and readily available clinical admission characteristics of critically ill patients with iGAS, unsupervised clustering analysis identified three specific patient populations that differed regarding ICU management.
Study context
- Population
- Patients admitted to 9 ICUs of Paris University Hospitals for invasive group A streptococcus (iGAS) infection, 01/03/2018–01/08/20237 unclassified
- Participants
- 148 patients
- Exposure
- Clinical phenotypes of iGAS infection in ICU
- Cluster 1n = 73
- Cluster 2n = 42
- Cluster 3n = 26
- Outcome
- Clinical phenotypes (clusters)
Methodological detailsmulticentre, unsupervised cluster analysis (k-prototypes)
- Multicentre
- Unsupervised cluster analysis (k-prototypes)
Main finding
As reported in the abstract
- Patients grouped in 3 clusters (Silhouette criterion); 7 patients unclassified
Observational data: associations, not effects.
Further findings
Antitoxin antibiotics — Cluster 1 vs Cluster 2 vs Cluster 3 · less frequent in cluster 2
P < 0.001
ECMO support — Cluster 1 vs Cluster 2 vs Cluster 3 · trend to more frequent in cluster 2
P = 0.07
Invasive mechanical ventilation — Cluster 1 vs Cluster 2 vs Cluster 3 · less likely in cluster 1
P = 0.005
ICU mortality — Cluster 1 vs Cluster 2 vs Cluster 3 · no difference between clusters
P = 0.32
Days
| Outcome | Cluster 1 | Cluster 2 | Cluster 3 | P value |
|---|---|---|---|---|
| Time to source controlCluster 1 vs Cluster 2 vs Cluster 3 · no significant difference across clusters | 0 (IQR 0–0) | 0 (IQR 0–0) | 0 (IQR 0–1) | P = 0.12 |
Original publication
Composed by AI from the published abstract and checked against it. How Briefs are made →