Photo Credit: Freepik

The following is a summary of “Creating a prediction model for invasive candidiasis in the intensive care unit using a case control design: a European multicentre approach,” published in the May 2025 issue of BMC Infectious Diseases by Benders et al. 

Invasive candidiasis (IC) in intensive care unit (ICU) remains challenging to diagnose early, with existing tools lacking sensitivity (Sens) and multinational validation despite the importance of timely treatment.  

Researchers conducted a retrospective study to determine the factors linked with IC in ICU settings and to develop an alternative prediction model using a large international dataset.  

They used ICU-acquired IC as the primary endpoint and collected data on 285 cases and 285 matched controls from the EUCANDICU database. Information on comorbidities, illness severity, and established IC risk factors was available. Univariate analysis identified 31 independent risk factors. A random 80% subset of the data was used to construct the optimal prediction model. Predictor selection was performed using the Least Absolute Shrinkage and Selection Operator (LASSO) method with a regularization parameter (λ = 1SE), which allowed slight precision loss to enhance external validity. The remaining 20% of cases were used to evaluate the model’s predictive performance.  

The results showed that factors such as the Simplified Acute Physiology Score II (SAPS II), Sequential Organ Failure Assessment (SOFA) score, past infections, renal impairment, and multiple Candida colonization sites were independently related with an increased risk of developing IC. A LASSO regression analysis incorporated 22 of the 31 identified variables, yielding an Area Under the Receiver Operating Characteristic (AUROC) of 0.7433.  

Investigators concluded that predicting which patient admitted to ICU would develop IC remained challenging, and the performance of their prediction model was insufficient for clinical practice, even with an alternative methodology applied to a large multinational database. 

Source: bmcinfectdis.biomedcentral.com/articles/10.1186/s12879-025-10644-9