Catheter-Associated Candiduria: Aggregates, Microscopy, and CFU Variability

TL;DR

This study examines how to better diagnose yeast infections in catheterized patients by comparing different lab methods and evaluating visible urine aggregates.

Contribution

The study introduces a standardized quantitative culture workflow and evaluates its performance against existing methods for diagnosing catheter-associated candiduria.

Findings

Standardized quantitative culture (QM) showed poor agreement with routine semi-quantitative culture (SQM) for CFU detection.

Visible urine aggregates were not reliable indicators of Candida spp. presence.

Intra-patient CFU variability was observed in 17.6% of patients, with 6.8% fluctuating over 10³ CFU/mL.

Abstract

Diagnosis of catheter-associated candiduria is limited by absent colony-forming unit (CFU) thresholds and heterogeneous laboratory workflows, complicating patient management. Within a quality-improvement (QI) initiative we aimed to (1) implement and evaluate a standardised quantitative culture (QM) workflow for catheter urine, (2) compare QM with routine semi-quantitative culture (SQM) for CFU detection and reproducibility, and (3) assess whether visible urine aggregates predict Candida spp. presence. From February 2024 to February 2025, we analysed 222 yeast-positive urine samples from 74 catheterised patients. This pilot combined process standardisation, parallel SQM and QM testing, targeted microscopy of aggregate-containing samples, and species identification by CHROMID® Candida agar and MALDI-TOF. Agreement between methods, intra-patient CFU variability over three daily samples, and the predictive value of aggregates were assessed. Aggregates were present in 29/222 samples (13.1%), of which 13 (44.8%) contained Candida spp., mainly C. albicans. No reliable macroscopic features distinguished Candida-positive from -negative aggregates. QM and SQM showed poor agreement (Bowker p < 0.001, κ = 0.17). QM detected growth in samples reported as “sterile” by SQM and provided continuous CFU counts. Cohort-level median QM counts were stable, but 13 patients (17.6%) showed notable variability, with 5 (6.8%) fluctuating > 103 CFU/mL. Process data suggested batching and variable pre-analytics as contributors to discordance. A standardised QM workflow revealed substantial discordance with SQM and highlighted pre-analytical variability. Visible aggregates are unreliable indicators for Candida spp. presence. Adoption of standardised quantitative culture, systematic microscopy, and structured lab-clinician communication may improve diagnostic consistency; prospective evaluation is warranted.

Figures