Combining VITEK® 2 with colistin agar dilution screening assist timely reporting of colistin susceptibility.

OBJECTIVES: The rise in carbapenem resistance among Gram-negative bacteria has renewed interest in colistin. Recently, the EUCAST-CLSI Polymyxin Breakpoints Working Group declared that broth microdilution (BMD) is the only valid method for colistin susceptibility testing. BMD is not easily incorporated into the routine work of clinical laboratories, and usually this test is incorporated serially, resulting in delayed susceptibility reporting. We tested a strategy of combining VITEK® 2 with a 2 µg/mL colistin agar dilution (VITEK® 2/AD) screening plate to improve performance and time to reporting of colistin susceptibility. METHODS: Colistin susceptibility for 364 clinical isolates was determined by VITEK® 2/AD and compared with the reference standard BMD according to the ISO 20776-1:2007 and CLSI guidelines. The EUCAST colistin susceptibility breakpoint of ≤2 µg/mL was used. Escherichia coli NCTC 13846 served as quality control strain. Agreement, very major error (VME) and major error rates were determined using ISO 20776-2:2007. RESULTS: The VME rate for VITEK® 2 alone was 30.6% (15/49, 95% CI 18.3-45.4%), and was reduced to 10.2% (5/49, 95% CI 3.4-22.2%) using the VITEK® 2/AD combined testing. The combined testing had categorical agreement with BMD of 97% (354/364, 95% CI 95.0-98.7%), and a major error (ME) rate of 1.6% (5/315, 95% CI 0.5-3.7%). Using the combined testing, even against challenging strains, 349 (95.8%, 95% CI 93.3-97.7%) colistin susceptibility results could be reported, and only 15 isolates required further analysis by BMD. DISCUSSION: Our method is simple to apply and allows rapid reporting of colistin susceptibility.

A case-control study to identify predictors of 14-day mortality following carbapenem-resistant Acinetobacter baumannii bacteraemia.

Carbapenem-resistant Acinetobacter baumannii (CRAB) is an increasingly common nosocomial pathogen. We sought to identify clinical and microbiological predictors of 14-day mortality among patients with CRAB bacteraemia. This case-control study included all adult patients in one Israeli hospital with CRAB on blood culture between July 2008 and June 2011. Cases were defined as patients who died within 14 days of bacteraemia onset and controls as patients who survived over 14 days. Sequence-typing of the blaOXA-51-like gene and REP-PCR identified CRAB clone groups. Logistic regression was performed to analyze predictors of 14-day all-cause mortality. To correct for differences in treatment onset, Cox regression was used to examine the effect of receiving an active antibiotic. Eighty-three cases and 89 controls were included. Six major CRAB clone groups were identified, with 14-day mortality ranging from 17 to 66%. Independent predictors of 14-day mortality were severity of illness (OR = 1.38 for each 1-point increase in Sequential Organ Failure Assessment (SOFA) score; 95% CI, 1.21, 1.56), independence in activities of daily living (ADL) on admission (OR = 3.40; 95% CI, 1.20, 9.67, for fully dependent vs. independent), surgery before bacteraemia (OR = 0.25; 95% CI, 0.11, 0.59) and clone group (OR = 7.76; 95% CI, 2.52, 23.85, for the most virulent group vs. the reference group). In the multivariate Cox model using a propensity score to adjust for SOFA, clone, ADL and surgery, active antibiotic treatment was protective (HR = 0.30; 95% CI, 0.15, 0.60). Differences in virulence between CRAB clones may partly explain heterogeneous results in previous studies of mortality following CRAB infection.

Risk factors for colonization with extended-spectrum beta-lactamase-producing enterobacteriaceae on admission to rehabilitation centres.

Patients newly admitted to rehabilitation centres are at high risk of colonization with multidrug-resistant bacteria because many of them have experienced prolonged stays in other healthcare settings and have had high exposure to antibiotics. We conducted a prospective study to determine the prevalence of and risk factors for colonization with extended-spectrum beta-lactamase-producing Enterobacteriaceae (ESBL-PE) in this population. Subjects were screened by rectal swab for ESBL-PE within 2 days of admission. Swabs were plated on chromagar ESBL plates and the presence of ESBL was verified by a central laboratory. A multilevel mixed effects model was used to identify risk factors for ESBL-PE colonization. Of 2873 patients screened, 748 (26.0%) were positive for ESBL-PE. The variables identified as independently associated with ESBL-PE colonization were: recent stay in an acute-care hospital for over 2 weeks (OR=1.34; 95% CI, 1.12, 1.6), history of colonization with ESBL-PE (OR=2.97; 95% CI, 1.99, 4.43), unconsciousness on admission (OR=2.59; 95% CI, 1.55, 4.34), surgery or invasive procedure in the past year (OR=1.49; 95% CI, 1.2, 1.86) and antibiotic treatment in the past month (OR=1.80; 95% CI, 1.45, 2.22). The predictive accuracy of the model was low (area under the ROC curve 0.656). These results indicate that ESBL-PE colonization is common upon admission to rehabilitation centres. Some risk factors for ESBL-PE colonization are similar to those described previously; however, newly identified factors may be specific to rehabilitation populations. The high prevalence and low ability to stratify by risk factors may guide infection control and empirical treatment strategies in rehabilitation settings.