-
1
The authors argue that clustering provides quantitative, reproducible candidemia phenotypes and identified enrichment of Candida glabrata in Phenotype 2 and Candida parapsilosis in Phenotype 3.
-
2
Phenotypes 1 and 2 had comparably high mortality despite differing clinical profiles characterized by immunosuppression and cirrhosis, respectively.
-
3
After SAPS II was removed, Phenotype 1 remained identifiable, with mortality of 70.1% versus 72.9% in the original model.
-
4
Without SAPS II, Phenotypes 2 and 3 became less distinct, patients shifted between groups, and Candida glabrata enrichment in the cirrhotic/digestive cluster persisted.
-
5
The authors acknowledge the study’s retrospective design and lack of external validation and propose future clustering that separates biological signatures from severity stratification.
