Evaluation of culture-based approaches for rapid detection of glycopeptide-resistant enterococci: a randomised, investigator-blinded study

Abstract number: O251

Malhotra-Kumar S., Cortiñas Abrahantes J., Lammens C., Molenberghs G., Aerts M., Goossens H.

Background: Rapid detection of gastro-intestinal carriage of glycopeptide-resistant enterococci (GRE) from screening cultures is crucial for an efficient control of their spread. We assessed 4 media – 2 chromogenic, ChromID, (bioMérieux), and CHROMagar (CHROMagar Microbiology), and 2 selective, VRE Selective (Oxoid) and ECCV (BD) – for their ability to detect GRE using well-characterised isolates and stool samples from hospitalised patients at high risk of GRE colonisation.

Methods: Twenty-five isolates consisting of 13 GRE. faecalis/faecium carrying various van genes and 12 non-VRE at concentrations of 106-101 CFU/ml and 106 CFU/ml, respectively, and 37 stool samples were randomised and spiral plated on all media and scored by 5 blinded investigators for characteristic colonies after 24 hrs incubation. Standard confirmatory tests were done on 1 putative GRE colony or on 1 characteristically coloured colony each for E. faecalis/faecium from the selective and chromogenic media, respectively. Detection of van genes, and ddl or sodA based speciation was done on PCR-sequencing. Mean sensitivity (SEN) and specificity (SPEC), and confidence intervals (CIs) were estimated for each medium by a logistic regression model using a penalised likelihood approach based on the reader response for the stool samples and isolates, and additionally on confirmation test results for the stool samples, both at the aggregated (GRE detected) and penalised level (correct species–colony colour correlation).

Results: CHROMagar showed the highest SEN based on reader response at the aggregated and penalised level for both stool samples and isolates (Table). Using confirmation test results at the aggregated level, SEN for ECCV was highest while the two chromogenic media showed a decrease in SEN by at least 11% in comparison to the values obtained based on reader response. SENs for the 2 chromogenic media were even lower (<70%) based on confirmation test results at the penalised level. ECCV and ChromID showed the highest SPECs with both reader response (stool samples) and confirmation test results at the aggregated level, and ChromID also at the penalised level, with narrow CIs indicating a high precision of this parameter estimate. For isolates, SPECs were highest for CHROMagar at both levels.

Conclusions: CHROMagar showed the best overall performance considering both SEN and SPEC estimates. ECCV performed well as a selective medium for GRE detection from stool samples.

Table. Mean sensitivities and specificities of media for detection of GRE after 24 hrs incubation

SamplesVariableMedium forAggregatedPenalised
  GRE detectionSensitivity (%)Specificity (%)Sensitivity (%)Specificity (%)
   Mean95% CIMean95% CIMean95% CIMean95% CI
 responseVRE Selectiveb57.044.6–69.193.786.6–97.5NANANANA
 testsVRE Selectiveb33.523.6–44.698.495.5–99.5NANANANA
 responseVRE Selectiveb65.958.3–72.894.287.8–97.2NANANANA
*Selective media do not differentiate between GRE. faecalis/faecium and thus, the aggregated and penalised responses are the same.
aBD, BE; bOxoid, UK; cBioMérieux, FR; dCHROMagar Microbiology, FR.

Session Details

Date: 16/05/2009
Time: 00:00-00:00
Session name: 19th European Congress of Clinical Microbiology and Infectious Diseases
Location: Helsinki, Finland, 16 - 19 May 2009
Presentation type:
Back to top