LETTER
Determining phenotypic antimicrobial susceptibilities of Pseudomonas aeruginosa is particularly valuable due to the complexity of resistance mechanisms this organism can harbor. The Accelerate PhenoTest BC kit (AXDX) provides a fast phenotypic antimicrobial susceptibility testing (AST) method for testing P. aeruginosa directly from positive blood culture. This study evaluated updates to the Accelerate PhenoTest BC kit made in order to improve the performance of beta-lactams when tested against P. aeruginosa (1, 2).
One hundred forty-four P. aeruginosa isolates were used to spike a blood culture bottle containing healthy donor blood and incubated until positivity. Aliquots of positive blood culture were tested on the Accelerate Pheno system (software 1.4.1.25) as previously described (3). AST was also performed in triplicate by CLSI reference broth microdilution (BMD) using isolated colonies (4). MIC results were compared to BMD results to calculate essential agreement (EA), categorical agreement (CA), and rates of very major (susceptible by AXDX, resistant by reference), major (resistant by AXDX, susceptible by reference), and minor (intermediate by one AST method, susceptible or resistant by the other method) errors (5). For EA, BMD results were truncated to the same range as those reported by the Accelerate Pheno system. FDA and CLSI breakpoints were applied (Table 1) (6, 7).
TABLE 1.
Current FDA- and CLSI-designated breakpoints of antipseudomonal beta-lactams
| Beta-lactam antibiotic | Breakpoint (μg/ml) |
||
|---|---|---|---|
| Susceptible | Intermediate | Resistant | |
| Aztreonam (FDA and CLSI) | ≤8 | 16 | ≥32 |
| Cefepime (FDA) | ≤8 | ≥16 | |
| Cefepime (CLSI) | ≤8 | 16 | ≥32 |
| Ceftazidime (FDA) | ≤8 | ≥16 | |
| Ceftazidime (CLSI) | ≤8 | 16 | ≥32 |
| Meropenem (FDA and CLSI) | ≤2 | 4 | ≥8 |
| Piperacillin-tazobactam (FDA and CLSI) | ≤16/4 | 32/4–64/4 | ≥128/4 |
Table 2 provides the EA, CA, and error rates for the isolates tested on both the updated and previous assays. With respect to the updated assay and when interpreted by FDA breakpoints, nine of 11 errors observed for cefepime were within EA, including the single very major error. Cefepime and ceftazidime do not have an intermediate interpretation by FDA breakpoints; therefore, all errors can be classified only as major or very major for these antimicrobials (6). When interpreted with CLSI breakpoints, all cefepime errors were minor and 17/21 errors were within EA. Bias toward a more resistant MIC for cefepime was observed by AXDX (Table 3). High cefepime minor-error rates with P. aeruginosa have been observed in various studies with other automated platforms, such as Vitek2 (9 to 18%), MicroScan WalkAway (32% to 48%), and BD Phoenix (18%) (8–11). When results were interpreted by FDA breakpoints, a total of five errors were observed with ceftazidime, and 2 of the 3 very major errors were within EA, a good case example demonstrating the challenges of interpreting errors when an intermediate breakpoint does not exist. When results were interpreted with CLSI breakpoints, 1 major and 1 very major error remained for ceftazidime, with EA and CA above 90%. Fifteen minor errors (10.4%) were observed with meropenem (Table 2), among which 9 were within EA. Eleven of the minor errors were due to the MIC being interpreted as resistant by AXDX but intermediate by BMD.
TABLE 2.
Performance of antipseudomonal beta-lactams tested against P. aeruginosa isolates on the Accelerate PhenoTest BC kit compared with BMD
| Beta-lactam antibiotica | No. of isolatesb
|
No. (%) with agreement |
No. (%) of errors |
|||||
|---|---|---|---|---|---|---|---|---|
| Total | S | R | CA | EA | Very major | Major | Minor | |
| Aztreonam (FDA and CLSI)* | 144 | 105 | 35 | 134 (93.1) | 135 (93.8) | 0 (0) | 1 (1.0) | 9 (6.2) |
| Aztreonam (FDA and CLSI) | 144 | 105 | 35 | 122 (84.7) | 124 (86.1) | 0 (0) | 1 (1.0) | 21 (14.6) |
| Cefepime (FDA)* | 143 | 107 | 36 | 132 (92.3) | 136 (95.1) | 1 (2.8) | 10 (9.3) | |
| Cefepime (FDA) | 144 | 108 | 36 | 84 (58.3) | 81 (56.2) | 0 (0) | 60 (55.6) | |
| Cefepime (CLSI)* | 143 | 107 | 29 | 122 (85.3) | 136 (95.1) | 0 (0) | 0 (0) | 21 (14.7) |
| Cefepime (CLSI) | 144 | 108 | 29 | 76 (52.8) | 81 (56.2) | 0 (0) | 2 (1.9) | 66 (45.8) |
| Ceftazidime (FDA)* | 141 | 103 | 38 | 136 (96.5) | 136 (96.5) | 3 (7.9) | 2 (1.9) | |
| Ceftazidime (FDA) | 144 | 104 | 40 | 46 (31.9) | 47 (32.6) | 0 (0) | 98 (94.2) | |
| Ceftazidime (CLSI)* | 141 | 103 | 31 | 132 (93.6) | 136 (96.5) | 1 (3.2) | 1 (1.0) | 7 (5.0) |
| Ceftazidime (CLSI) | 144 | 104 | 33 | 40 (27.8) | 47 (32.6) | 0 (0) | 20 (19.2) | 84 (58.3) |
| Meropenem (CLSI and FDA)* | 144 | 102 | 25 | 127 (88.2) | 136 (94.4) | 0 (0) | 2 (2.0) | 15 (10.4) |
| Meropenem (CLSI and FDA) | 144 | 102 | 25 | 98 (68.1) | 107 (74.3) | 0 (0) | 2 (2.0) | 44 (30.6) |
| Piperacillin-tazobactam (CLSI and FDA)* | 138 | 101 | 30 | 130 (94.2) | 133 (96.4) | 0 (0) | 0 (0) | 8 (5.8) |
| Piperacillin-tazobactam (CLSI and FDA) | 144 | 106 | 31 | 45 (31.2) | 52 (36.1) | 0 (0) | 12 (11.3) | 52 (36.1) |
An asterisk indicates that the improved software was used.
S, susceptible; R, resistant.
TABLE 3.
Error trends of beta-lactam antibiotics tested against P. aeruginosa isolates on Accelerate PhenoTest BC kit compared with BMD
| Beta-lactam antibiotic (no. of errors) | No. of results |
||
|---|---|---|---|
| More susceptible | More resistant | Within EA | |
| Aztreonam (10) | 6 | 4 | 9 |
| Cefepime (11) | 1 | 10 | 10 |
| Cefepime (CLSI) (21) | 9 | 12 | 17 |
| Ceftazidime (5) | 3 | 2 | 2 |
| Ceftazidime (CLSI) (9) | 3 | 6 | 6 |
| Meropenem (17) | 1 | 16 | 9 |
| Piperacillin-tazobactam (8) | 1 | 7 | 6 |
Overall, the most notable improvements with the updated assay are within the major and minor error rates. In the original clinical trial data set for the Accelerate Pheno system, a total of 43 major errors were observed among the Gram-negative organisms, with 26% of them being for beta-lactams tested against P. aeruginosa. This resulted in major-error limitations imposed by the FDA and the aim for the updates to the assay described herein (1). The data presented here are from a different population of isolates than those used in the original clinical trial. Specifically, the current data set was enriched to include approximately 20% of isolates with MICs at the breakpoint, allowing a robust evaluation of performance postimprovement. Furthermore, the population described here is approximately 10% less susceptible than what is likely to be observed in clinical laboratories based on U.S. surveillance of P. aeruginosa bloodstream infections (12). This is important, as differences in MIC distributions impact the propensity of errors. Therefore, direct comparisons between two different isolate sets, such as the present data and that described by Pancholi et al. (1), cannot be directly made. Nonetheless, the improvements described herein led to the removal of major-error limitations for piperacillin-tazobactam, meropenem, ceftazidime, and cefepime.
P. aeruginosa susceptibility testing is known to be challenging (8–11). As technologies for susceptibility testing advance, development of assays for these difficult-to-test organisms is prudent and likely an ongoing necessity. Moreover, clinical microbiology labs should seek to understand their local epidemiology when evaluating an assay, as performance can vary among different populations of isolates. These data demonstrate markedly improved performance, particularly with respect to major errors, of beta-lactams against P. aeruginosa on the Accelerate Pheno system compared with previous versions of the assay.
REFERENCES
- 1.Pancholi P, Carroll KC, Buchan BW, Chan RC, Dhiman N, Ford B, Granato PA, Harrington AT, Hernandez DR, Humphries RM, Jindra MR, Ledeboer NA, Miller SA, Mochon AB, Morgan MA, Patel R, Schreckenberger PC, Stamper PD, Simner PJ, Tucci NE, Zimmerman C, Wolk DM. 2018. Multicenter evaluation of the Accelerate PhenoTest BC kit for rapid identification and phenotypic antimicrobial susceptibility testing using morphokinetic cellular analysis. J Clin Microbiol 56:e01329-17. doi: 10.1128/JCM.01329-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Calderaro A, Buttrini M, Martinelli M, Covan S, Montecchini S, Ruggeri A, Arcangeletti MC, De Conto F, Chezzi C. 2020. Rapid microbial identification and phenotypic antimicrobial susceptibility testing directly from positive blood cultures: a new platform compared to routine laboratory methods. Diagn Microbiol Infect Dis 96:114955. doi: 10.1016/j.diagmicrobio.2019.114955. [DOI] [PubMed] [Google Scholar]
- 3.Bowler SL, Towne JW, Humphries RM, Doi Y. 2019. Evaluation of the Accelerate Pheno system for identification of Acinetobacter clinical isolates and minocycline susceptibility testing. J Clin Microbiol 57:e01711-18. doi: 10.1128/JCM.01711-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Clinical and Laboratory Standards Institute. 2018. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically; approved standard, 10th ed. M07-A11. Clinical and Laboratory Standards Institute, Wayne, PA. [Google Scholar]
- 5.Humphries RM, Ambler J, Mitchell SL, Castanheira M, Dingle T, Hindler JA, Koeth L, Sei K, Hardy D, Zimmer B, Butler-Wu S, Dien Bard J, Brasso B, Shawar R, Dingle T, Humphries R, Sei K, Koeth L, CLSI Methods Development and Standardization Working Group of the Subcommittee on Antimicrobial Susceptibility Testing. 2018. CLSI Methods Development and Standardization Working Group best practices for evaluation of antimicrobial susceptibility tests. J Clin Microbiol 56:e01934-17. doi: 10.1128/JCM.01934-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Food and Drug Administration. 2017. FDA-recognized antimicrobial susceptibility test interpretive criteria. Food and Drug Administration, Washington, DC. https://www.fda.gov/Drugs/DevelopmentApprovalProcess/DevelopmentResources/ucm410971.htm Accessed 30 May 2020. [Google Scholar]
- 7.Clinical and Laboratory Standards Institute. 2020. Performance standards for antimicrobial susceptibility testing; 30th informational supplement. M100-S30. Clinical and Laboratory Standards Institute, Wayne, PA. [Google Scholar]
- 8.Jekarl DW, Han SB, Kim YJ, Shin SH, Park KG, Park JJ, Han K, Park Y-J. 2010. Evaluation of Vitek2 and BD Phoenix in antimicrobial susceptibility testing of Acinetobacter baumannii and Pseudomonas aeruginosa. Diagn Microbiol Infect Dis 67:384–386. doi: 10.1016/j.diagmicrobio.2010.02.024. [DOI] [PubMed] [Google Scholar]
- 9.Bobenchik AM, Deak E, Hindler JA, Charlton CL, Humphries RM. 2017. Performance of Vitek 2 for antimicrobial susceptibility testing of Acinetobacter baumannii, Pseudomonas aeruginosa, and Stenotrophomonas maltophilia with Vitek 2 (2009 FDA) and CLSI M100S 26th edition breakpoints. J Clin Microbiol 55:450–456. doi: 10.1128/JCM.01859-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sader HS, Fritsche TR, Jones RN. 2006. Accuracy of three automated systems (MicroScan WalkAway, VITEK, and VITEK 2) for susceptibility testing of Pseudomonas aeruginosa against five broad-spectrum beta-lactam agents. J Clin Microbiol 44:1101–1104. doi: 10.1128/JCM.44.3.1101-1104.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Juretschko S, Labombardi VJ, Lerner SA, Schreckenberger PC, Pseudomonas AST Study Group. 2007. Accuracies of beta-lactam susceptibility test results for Pseudomonas aeruginosa with four automated systems (BD Phoenix, MicroScan WalkAway, Vitek, and Vitek 2). J Clin Microbiol 45:1339–1342. doi: 10.1128/JCM.01716-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Shortridge D, Arends SJR, Duncan LR, Streit JM, Flamm RK. 2019. Ceftolozane-tazobactam activity against difficult-to-treat resistance in Pseudomonas aeruginosa from bloodstream infections in US hospitals. Open Forum Infect Dis 6(Suppl 2):S580. doi: 10.1093/ofid/ofz360.1453. [DOI] [Google Scholar]