Letter to the Editor: Implementation of the Rapid Polymyxin Acinetobacter Test to Detect Colistin-Resistant Acinetobacter baumannii

Dear Editor:

Acinetobacter baumannii is among the most predominant nosocomial pathogens worldwide. ¹ The intrinsic resistance of the microorganism to several antimicrobials combined with the acquisition of additional mechanisms (carbapenems, aminoglycosides, etc.) leads to limited therapeutic options. ² Polymyxins (polymyxin B and colistin) remain the “last resort” antimicrobial agents for the treatment of serious infections caused by multidrug-resistant A. baumannii. ³

The broth microdilution method (BMD) proposed by EUCAST for testing susceptibility to colistin is time consuming, with a turnaround time of ∼24 hr. ⁴ The Rapid Polymyxin™ Acinetobacter test, developed by Elitech Microbiology, has been recently evaluated by Lescat et al. for rapid detection of colistin-resistant A. baumannii and the results were in concordance with those obtained by BMD. ⁴

From June to October 2018, the Rapid Polymyxin Acinetobacter test was introduced and evaluated in the routine microbiology laboratory of the University Hospital of Larissa (Central Greece), as an additional tool for detecting colistin resistance rapidly and accurately. According to the protocol, isolates that were Gram-stain negative, oxidase negative, no-mobile, and had MacConkey colony morphology like Acinetobacter were immediately tested by the Rapid Polymyxin Acinetobacter test. In parallel, the isolates were also tested according to the traditional protocol, which included identification to species level plus susceptibility testing by the Vitek 2 automated system (BioMerieux) and determination of minimum inhibitory concentration (MIC) to colistin by the BMD, according to EUCAST guidelines. The colistin-susceptible strain Pseudomonas aeruginosa ATCC 27853 and the colistin-resistant Escherichia coli R2739 were used as negative and positive controls, respectively, for the determination of MIC of colistin by the BMD method. The interpretation of the susceptibility testing results was based on EUCAST criteria. ⁴

A total of 75 A. baumannii isolates were collected. Duplicate isolates from the same patient such as isolates that belonged to other species based on Vitek 2 identification were excluded. All isolates were resistant to ampicillin–sulbactam, piperacillin, piperacillin–tazobactam, cefotaxime, ceftazidime, cefepime, imipenem, meropenem, aztreonam, amikacin, gentamicin, tetracycline, and ciprofloxacin. According to MICs of colistin, 33 of them were susceptible (MICs ranged from 0.25 to 2 mg/L; MIC geometric mean value: 0.715 mg/L, MIC_50/90: 0.25/0.5 mg/L, respectively) and 42 were resistant (MICs ranged from 4 to 32 mg/L; MIC geometric mean value: 19.185 mg/L, MIC_50/90: 8/16 mg/L, respectively).

The Rapid Polymyxin Acinetobacter test identified correctly 39 out of the 42 colistin-resistant isolates and 28 out of the 33 colistin-susceptible isolates. The test failed to detect three colistin-resistant isolates that had MICs 4, 8, and 16 mg/L, respectively, whereas it has falsely characterized as resistant five colistin-susceptible isolates; three of them had MIC 2 mg/L, one 1 mg/L, and one 0.5 mg/L. These eight isolates were retested by both BMD and Rapid Polymyxin Acinetobacter test, and the results remained the same. The sensitivity, specificity, the positive predictive value (PPV), and the negative predictive value (NPV) of the test were 93.3%, 86.8%, 89.4%, and 91.7%, respectively. The test gave results much sooner than BMD (maximum time 4 hr vs. 24 hr by BMD).

Very major errors (VMEs) and major errors (MEs) corresponding to false-susceptible and false-resistant results, respectively, were calculated as described previously. ⁵ Comparing our results with those of the study of Lescat et al., three VMEs and five MEs were observed versus 0 VME and 0 ME, respectively. Lescat et al. have evaluated the test using a collection of 21 isolates from different countries. In our study, the test was evaluated in real time and in a clinical setting with a high rate of colistin resistance. The rapidity of the results, combined with its high NPV (91.7%), helps the clinician to initiate therapy with colistin, specifically in clinical settings with high prevalence of colistin-resistant bacteria. Despite the test showing good performance for the detection of colistin-resistant A. baumannii isolates, it revealed the presence of some false negative and false positive results. To investigate false results, further experiments are planned for examining the clonality and colistin-resistant mechanisms of the respective isolates.