Rapid detection of extended-spectrum lactamase-producing Enterobacteriaceae: arandomised, investigator-blinded evaluation of culture-based approaches

Abstract number: P2057

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

Background: Rapid and accurate detection of extended-spectrum b-lactamase-producing Enterobacteriaceae (ESBL-En) is crucial for effective infection control. We assessed 4 chromogenic media-ChromID, (bioMérieux), CHROMagar (CHROMagar Microbiology), Amber (AES Chemunex), and a yet to be introduced formulation, Chromogenic-ESBL (Oxoid) – and 1 selective medium – EbSA (Alpha-Omega) – for their ability to correctly identify ESBL-En using well-characterised isolates and spiked stool samples.

Methods: Eighty-four samples consisting of 16 ESBL-En (E. coli, K. pneumoniae, Enterobacter spp., P. mirabilis, P. aeruginosa harbouring CTX-M, SHV, TEM or PER), and 5 non-ESBL-En (E. coli, K. pneumoniae, Enterobacter spp., P. mirabilis) at concentrations of 101 CFU/ml and 106 CFU/ml, respectively, and each of the 21 isolates spiked into stools at 3 concentrations (106, 103, 101 CFU/ml) were randomised and spiral plated on the 5 media. Media were read by 5 blinded investigators for characteristic colonies after 24 and 48 hrs incubation. One putative ESBL-En colony from the selective medium and 1 colony of each colour/type from the chromogenic media for each plated sample was confirmed for species identification on biochemical tests and for presence of ESBL by double-disk synergy test. Mean sensitivity (SEN) and specificity (SPEC), and confidence intervals (CIs) were estimated for each medium by logistic regression model based on reader response for both incubation times, and both at the aggregated (any ESBL-En detected) and penalised level (correct species-colony colour correlation), using the penalised likelihood approach.

Results: Chromogenic-ESBL showed almost equal to 100% mean SEN and SPEC at both 24 and 48 hrs with the aggregated reader response and narrow CIs indicating a high precision of these parameter estimates (Table). Although, Chromogenic-ESBL also showed the highest SEN and a high SPEC with the penalised reader response for both incubation times, these values were lower than the aggregated response primarily due to misclassifications of E. aerogenes (TEM) and P. aeruginosa (PER) based on colony colour. Mean SENs for the other 4 media increased on average by 6.5% from 24 to 48 hrs. EbSA and ChromID showed almost equal to 100% mean SPECs at both incubation times, and the latter also with both reader responses.

Conclusions: Chromogenic-ESBL showed the best performance overall irrespective of sample concentration, reader or incubation time.

Table. Mean sensitivities and specificities of media for ESBL-En detection

Media for ESBL-En detection24 hours48 hours
 Mean95% CIMean95% CIMean95% CIMean95% CI
Reader response aggregated
EbSA* (Alpha-omega, BE)79.6%75.3, 83.399.6%94.6, 100.086.0%82.6, 88.899.4%91.8, 100.0
ChromID (bioMérieux, FR)84.1%80.2, 87.499.6%94.7, 100.089.3%86.3, 91.799.4%91.9, 100.0
CHROMagar (CHROMagar Microbiology, FR)77.5%73.0, 81.496.2%91.4, 98.484.4%80.8, 87.594.1%87.1, 97.4
Amber (AES Chemunex, FR)54.4%47.6, 61.074.0%62.5, 83.065.3%58.8, 71.264.4%51.4, 75.6
Chromogenic ESBL (Oxoid, UK)99.4%97.6, 99.899.2%95.0, 99.999.6%98.5, 99.998.7%92.3, 99.8
Reader response penalised
ChromID (bioMérieux, FR)75.9%71.5, 79.799.6%94.4, 100.081.8%78.1, 85.099.5%92.1, 100.0
CHROMagar (CHROMagar Microbiology, FR)49.7%44.8, 54.696.3%91.5, 98.458.6%53.7, 63.394.7%88.3, 97.7
Amber (AES Chemunex, FR)26.4%22.2, 31.167.6%57.2, 76.533.9%29.0, 39.159.3%48.4, 69.4
Chromogenic ESBL (Oxoid, UK)82.2%78.3, 85.698.1%93.7, 99.486.9%83.6, 89.597.3%91.2, 99.2
*EbSA is a selective medium that does not differentiate the ESBL-En and thus the aggregated and penalised responses are the same.

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