External Validation of Risk Prediction Score for Candidemia in Critically Ill Patients: A Retrospective Observational Study.

Abstract

Candidemia is associated with high mortality in critically ill patients. Early diagnosis of candidemia is imperative for starting treatment. Therefore, this study was designed to externally validate the candidemia risk prediction scores. This single-center, retrospective observational study included all critically ill patients admitted to the intensive care unit at a tertiary referral center from January 2018 to December 2023. The risks and patient outcomes were analyzed using bivariate and multivariate regression analysis. A total of 500 patients were analyzed with 2 dropouts due to incomplete data. Candidemia incidence was 8.86%, with parenteral nutrition and candida colonization identified as independent risk factors. Compared to an established risk prediction score, this study demonstrated a sensitivity of 75.0% [59.7-86.8], a specificity of 65.4% [60.8-69.8], a negative predictive value of 96.4% [94.2-97.8], and a positive predictive value of 17.3% [14.5-20.5]. The candidemia group had a significantly higher mean SOFA score, longer time in ICU, longer hospital length of stay, and higher rates of both ICU and in-hospital mortality. This study shows that the risk prediction score is more effective as a tool for excluding rather than predicting candidemia. We recommend against using it as the sole diagnostic guide.