Authors defend candidemia ICU phenotypes after SAPS II sensitivity analysis

In a published response to methodological criticism, the authors report that one of three previously described candidemia phenotypes remained identifiable after SAPS II was removed from the clustering variables. They argue that severity scores capture relevant clinical heterogeneity, while acknowledging the original study’s retrospective design and lack of external validation.

One phenotype persisted after SAPS II was excluded.
Reported mortality was 70.1% versus 72.9% originally.
Other phenotype distinctions blurred without severity scoring.
Original study was retrospective and lacked external validation.

The authors respond to a critique of their earlier retrospective study, which used unsupervised clustering to classify intensive care patients with candidemia into three phenotypes. The critique questioned whether including the SAPS II severity score drove the phenotype separation and associated outcome differences. The response does not report the cohort size or new patient recruitment; instead, it presents a post-hoc sensitivity analysis that repeated clustering without SAPS II.

The authors say the phenotype characterized by hematological malignancy, neutropenia, fever, and low leukocyte counts remained identifiable, with mortality of 70.1% compared with 72.9% in the original model. Without SAPS II, the distinction between the other two phenotypes became less clear, and patients were redistributed. The authors also report that microbiological associations persisted, including Candida glabrata enrichment in the cirrhotic/digestive cluster (31.5% in the sensitivity analysis versus 24.2% originally).

They argue that severity scores can be legitimate descriptors in clustering intended to characterize the full clinical syndrome, while suggesting future studies use biological clustering followed by severity stratification. The response acknowledges that the original study was retrospective and lacked external validation; the findings do not establish that phenotype-guided treatment improves outcomes.

AI summary · Not yet editor-reviewed

Is this summary clinically accurate?

Help fellow clinicians: your rating sends inaccurate summaries straight to our editors.

Source

Critical Care: Unsupervised clustering to classify patients with candidemia in intensive care: algorithmic tautology or the generation of highly anticipated new data? ↗

This is an automated AI-condensed summary that has not yet been reviewed by an editor. Always consult the full item at the original source.