Rapid detection of ESBL-producing gram-negative bacteria isolated from blood: a reasonable and reliable tool for middle and low resource countries.

INTRODUCTION Delay in appropriate treatment in patients with bacteraemia can increase morbidity, mortality, and health expenditures. We compared the Rapid Direct Test (RDT) designed to detect ESBL-producing gram-negative bacteria (GNB) directly from positive blood cultures bottles, with two conventional ESBL detection tests: Screening and Confirmatory Disk Diffusion Assay (SC-DDA) and an MIC Screening and ESBL E-test (MIC/ET). MATERIAL AND METHODS We screened all blood cultures in a tertiary care facility from August to December 2005. We only included one positive bottle per patient in which GNB were observed. RDT: Blood from each bottle was inoculated on Mueller-Hinton agar. Ceftazidime and cefotaxime disks with and without clavulanic acid were added and incubated at 35 degrees C +/- 2 degrees C for 24 h. Results were interpreted according to CLSI recommendations for the SC-DDA and MIC/ET. All methods were performed simultaneously. Time for reporting as an ESBL-producer and cost of the tests were recorded. RESULTS We isolated 124 GNB in 114 episodes of bacteraemia, 10 of them (8.8%) polymicrobial; 79 (63.7%) of the GNB were enteric bacteria, 44 (35.5%) glucose non-fermenter GNB and one Haemophilus influenzae. The most common microorganism was Escherichia coli in 56 episodes (45.2%), followed by Pseudomonas aeruginosa in 24 (19.3%), and Klebsiella pneumoniae in 13 (10.5%). Of the 114 episodes, 41 (36%) had at least one GNB resistant to 3rd generation cephalosporins, and 25 (21.9%) were caused by an ESBL-producing GNB. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for the RDT were 96%, 98.9%, 96% and 98.9%, respectively. Agreement by kappa index between RDT and SC-DDA was 0.95 and between the RDT and MIC/ET was 0.92. The RDT detected 24/25 ESBL-producing bacteria. The mean time to detect an isolate as an ESBL producer after a positive blood culture bottle signal was 1.02 +/- 0.19 days when using the RDT, and 3.40 +/- 0.59 days when using any other method. The difference in reporting time was 2.38 +/- 0.63 days (p < 0.0001). Our estimated cost per test was $1.54 for RDT, $2.32 for screening/ confirmatory SC-DDA, and $49.65 for MIC screening and MIC/ET. Conclusions. The RDT is a rapid, reliable and easy analysis to perform, as well as cost-effective.