Critically ill patients with necrotizing soft tissue infections in the Caribbean area: unsupervised analysis of a retrospective cohort (2014–2023) with identification of factors associated with mortality
Authors' conclusion
Unsupervised analysis of critically ill patients with NSTIs in tropical settings revealed three distinct patient clusters that exhibited unique phenotypic characteristics and clinical outcomes.
Study context
- Population
- Critically ill patients with necrotizing soft tissue infections (NSTIs) admitted to the intensive care unit (ICU) of the University Hospital of Guadeloupe, 2014–2023
- Participants
- 91 patients
- Exposure
- Clusters identified by unsupervised clustering
- Mild NSTIn = 23
- Severe NSTIn = 49
- Fulminant NSTIn = 19
- Outcome
- Mortality
- Follow-up
- 90 days
Methodological detailsobservational, single-centre, unsupervised clustering
- Observational
- Single-centre
- Unsupervised clustering
Main finding
As reported in the abstract
90-day mortality
90-day mortality rate
32%
Observational data: associations, not effects.
Further findings
Unsupervised clustering
Severity and time to surgery
- Median Simplified Acute Physiology Score (SAPS) II (IQR 40-66)
- 45
- Median time from hospital admission to first surgical debridement, h (IQR 6-10 h)
- 8
Organ support in ICU
- Mechanically ventilated
- 65%
- Shock
- 75%
- RRT
- 34%
Unsupervised clustering
- Clusters associated with different ICU courses and outcomes; subcutaneous emphysema and sepsis-associated encephalopathy key components
Factors associated with 90-day mortality (multivariate analysis)
- Mortality associated with SAPS II, subcutaneous emphysema, >8 h from hospital admission to first surgery, and immunocompromised status
Original publication
Critically ill patients with necrotizing soft tissue infections in the Caribbean area: unsupervised analysis of a retrospective cohort (2014–2023) with identification of factors associated with mortality. Ann Intensive Care. Published 2 June 2025. doi:10. 1186/ s13613-025-01488-2
Composed by AI from the published abstract and checked against it. How Briefs are made →