Evaluation of the Phoenix™ Automated Microbiology System for Detection of (i) Extended-Spectrum Beta-Lactamase-Producing Escherichia coli and Klebsiella spp., and of (ii) Ciprofloxacin-Resistant Acinetobacter spp. and Pseudomonas aeruinosa European Clinic

S. BRISSE, A. C. FLUIT, K. KUSTERS, M. KLOOTWIJK, S. DE VAAL, M. A. LEVERSTEIN-VAN HALL, J. VERHOEF AND D. MILATOVIC University Medical Centre Utrech, Utrech the Netherlands Contact. sbrissc@tabazunl

INTRODUCTION Reliable automated detection of antimicrobial resistance is of the utmost importance for clinical microbiology laboratories. The Phoenix™ Automated Microbiology System (BD Biosciences, Sparks, MD, USA), a new instrument combining identification and susceptibility testing, as evaluated for detection of two increasingly important antimicrobial resistance phenotypes: extended-spectrum beta-lactamase (ESBL) production in gram negative organisms and ciprofloxacin resistance in non-fermenters. MATERIALS AND METHODS

  1. Bacterial isolates and Antimicrobial Susceptibility Testing (AST) reference methods.
    • 74 ESBL-producing isolates formerly identified using the E-test (ceftazidime/ceftazidime+clavulanate and cefotaxime/cefotaxime+clavulanate) (AB Biodisk, Solna, Sweden).
    • Seventy-five Pseudomonas aeruginosa isolates, formerly analyzed by broth microdilution following NCCLS standard method, were selected to be equally distributed in minimum inhibitory concentrations (MIC) classes of 15 isolates each (MIC of ciprofloxacin<0.5mg/L. MIC=0.5 or 1 mg/L=2mg/L. MIC=4 or 8 mg/L and MIC = 16 mg/L) (see Table 2).
    • 205 Acinetobacter isolates, similary distributed into MIC classes (see Table 3).
  2. AST using the Phoenix™ instrument
    • The manufacturer’s recommendations were followed. The priniciples of AST testing and ESBL screening by the Phoenix™ are indicated in Figures 1 and 2, respectively.

 

Figure 1. AST features of the Phoenix™
  • Full doubling dilution MIC method.
  • Growth detection via metabolic indicator and turbidity.
  • Advanced algorithms to give accurate prediction of resistance in the shortest possible time
  • Integrated ESBL Screening Test (Gram-negative)
  • Integrated β-Lactamase Tests on both ID and AST side of the panel (Gram-positive bacteria)

 

Figure 2. Extended Spectrum β-Lactamaxe (ESBL) Screen
  • Evaluation of growth in wells containing low concentrations of cefpodoxime or ceftazidime triggers a check of cephalosporin/clavulanate growth response in the Phoenix™ System.
  • Inhibition in both wells is considered a negative result.
  • Inhibition in ceftazidime/clavulanate, ceftriaxone/clavulanate, or cefotaxime/clavulanate can indicate an ESBL.
  • Growth in all wells indicate a ESBL negative result.

RESULTS Table 1. ESBL screen

Phoenix™ AST Results – ESBL Screening E-test Phoenix™ Results
ESBL not ESBL
Escherichia coli n=76 ESBL (n=26) 23 3*
NON-ESBL (n=42) 1 41
Klebsiella oxytoca n=26 ESBL (n=13) 13 0
NON-ESBL (n=13) 6** 7
Klebsiella pneumoniae n=50 ESBL (n=39) 38 1*
NON-ESBL (n=11) 1 10
Proteus mirabilis n=10 ESBL (n=3) 3 0
NON-ESBL (n=7) 0 7
Total n=162 ESBL (n=81) 77 (95%) 4 (5%)
NON-ESBL (n=81) 8 (10%)** 73 (90%)

* Identical result was obtained with VITEK 2 ** Overestimation due to a biased choice of test strains Table 2. P. aeruginosa – ciprofloxacin MICs PHOENIX™ AST – Pseudomonas aeruginosa ciprofloxacin

Tested Strains MIC as determined by Phoenix™
MIC (mg/L) No. <=0.25 0.5 1 2 4 >4
<=0.25 15 15          
0.5 5   1 1 1    
1 10 1*   9      
2 15     7 2 6  
4 8         6 2
8 7         3 4
>=16 14           14
* Not retested  
Very Major Error: 0% Major Error: 0%
Minor Error: 9% Essential Concordance: 97%

Table 3., Acinetobacter spp. – ciprofloxacin MICs PHOENIX™ AST– Acinetobacter spp. ciprofloxacin

Tested Strains MIC as determined by Phoenix™
MIC (mg/L) No. <=0.25 0.5 1 2 4 >4
<=0.25 16 15 1        
1 16 4 3 7     2
2 15 1   7 6   1
4 8     3* 5    
8 7         1 2
>=16 146 2*       1 143
* Not retested  
Very Major Error: 3% Major Error: 1%
Minor Error: 7% Essential Concordance: 94%

CONCLUSIONS The ESBL screen of the Phoenix™ instrument appeared very reliable. The determination of Minimum Inhibitory Concentrations of ciprofloxacin appeared very reliable in both non-fermenter taxa tested.