Abstract
Current National Committee for Clinical Laboratory Standards (NCCLS) susceptibility guidelines for quality control testing with Haemophilus influenzae do not include a β-lactamase-producing strain that could detect the deterioration of the β-lactamase inhibitor components of amoxicillin-clavulanic acid, ampicillin-sulbactam, and piperacillin-tazobactam. The objective of the study was to determine if comparable quality control results for Escherichia coli ATCC 35218, a β-lactamase-producing strain, would be produced for the three β-lactam–β-lactamase inhibitor agents with Haemophilus test medium and Mueller-Hinton medium. The criteria used in this study to determine if Haemophilus test medium was acceptable for quality control testing of E. coli ATCC 35218 was that 100% of the results obtained with an antimicrobial agent-methodology combination needed to be within the acceptable NCCLS ranges established with Mueller-Hinton medium. The MIC testing results obtained by the broth microdilution and E-test methods with amoxicillin-clavulanic acid and piperacillin-tazobactam were all within the NCCLS ranges; however, the results obtained with ampicillin-sulbactam by both methods were not within the NCCLS ranges. Acceptable results were obtained by the disk diffusion methodology with ampicillin-sulbactam and piperacillin-tazobactam but not with amoxicillin-clavulanic acid. When performing susceptibility testing with H. influenzae with the β-lactam–β-lactamase inhibitors, in addition to quality control testing with H. influenzae ATCC 49247, testing of E. coli ATCC 35218 on Haemophilus test medium is an effective way to monitor the β-lactamase inhibitors in some antimicrobial agent-methodology combinations.
The current quality control isolate recommended by the National Committee for Clinical Laboratory Standards (NCCLS) for use when performing susceptibility testing with Haemophilus influenzae is a non-β-lactamase-producing strain, H. influenzae ATCC 49247 (6, 7). NCCLS guidelines for quality control testing of H. influenzae susceptibility do not include a β-lactamase-producing strain and therefore do not provide a method for the detection of the presence of adequate amounts of the β-lactamase inhibitor components of amoxicillin-clavulanic acid, ampicillin-sulbactam, and piperacillin-tazobactam. As the percentage of β-lactamase-producing H. influenzae isolates increases, appropriate quality control for susceptibility testing becomes important. Among countries in Europe, variable percentages of β-lactamase-producing H. influenzae isolates have been reported (3, 8, 9). In a U.S. multicenter surveillance study performed in 1994 and 1995, 36.4% of 1,537 clinical isolates of H. influenzae produced β-lactamase (2). A U.S. surveillance study of respiratory tract pathogens performed in 1996 and 1997 reported 33.4% β-lactamase producers among 1,558 H. influenzae isolates (10). Another U.S. study performed in 1997 showed that among 1,676 H. influenzae isolates, 41.6% produced β-lactamase (4). With these percentages of β-lactamase producers, it is critical that quality control be performed to ensure that adequate amounts of the β-lactamase inhibitors are present.
The β-lactamase-producing strains of H. influenzae are not suitable for use as quality control organisms in evaluations of the clavulanic acid content of amoxicillin-clavulanic acid. This is due to the high level of inhibition of the TEM-1 and ROB-1 β-lactamases produced by H. influenzae provided by clavulanic acid. This was demonstrated in a study in which 43 β-lactamase-producing strains of H. influenzae for which amoxicillin MICs ranged from 1 to >16 μg/ml were tested. Only 0.12 μg of clavulanic acid per ml was needed to restore amoxicillin activity to the NCCLS susceptibility breakpoint for amoxicillin-clavulanic acid (≤4 μg/ml) for all 43 isolates (1).
Escherichia coli ATCC 35218 is recommended by NCCLS (6, 7) as the quality control organism for the β-lactam–β-lactamase inhibitor agents. However, current guidelines recommend the testing of E. coli ATCC 35218 with Mueller-Hinton (MH) medium. The primary objective of this study was to determine if comparable quality control results for E. coli ATCC 35218 would be produced for the three β-lactam–β-lactamase inhibitor agents with MH medium and Haemophilus test medium (HTM). It would be preferable to test E. coli ATCC 35218 with HTM, since the test conditions for quality control isolates should be the same as those for patient isolates.
Another objective of this study was to determine if comparable results could be obtained with Mueller-Hinton chocolate (MHC) agar and MH agar when E. coli ATCC 35218 was tested with all three β-lactam–β-lactamase inhibitor agents by the E-test and the disk diffusion methodologies. Prior to 1990, NCCLS recommended the use of MHC agar for disk diffusion tests with H. influenzae (5). Isolates were tested in triplicate with each medium type on each day of testing for 10 consecutive days by the broth microdilution, E-test, and disk diffusion methods. The results obtained with HTM were compared to the results obtained with MH medium and to the acceptable NCCLS ranges for E. coli ATCC 35218.
Additionally, in many clinical laboratories, quality control isolates are repeatedly subcultured and used for quality control testing. To determine if this practice would have an effect on the susceptibility results, an isolate was passaged consecutively on each of the 10 test days and was tested in triplicate on each day. The results were compared to the results obtained with a fresh isolate by the broth microdilution, E-test, and disk diffusion methods.
MATERIALS AND METHODS
Test organisms.
E. coli ATCC 35218 and H. influenzae ATCC 49247 were used.
Isolate preparation.
Prior to testing, E. coli ATCC 35218 and H. influenzae ATCC 49247 were subcultured twice on sheep blood agar plates and chocolate agar plates, respectively, from CULTI-LOOPS (Lake Charles, La.). The passaged E. coli ATCC 35218 isolate was subcultured three times. The passaged isolate was then subcultured consecutively on each day of the study up to day 10.
Antimicrobial agents.
Powders of antimicrobial agents with known potencies were obtained as follows: amoxicillin and clavulanic acid, SmithKline Beecham, Collegeville, Pa.; ampicillin and piperacillin, Sigma Chemical Co., St. Louis, Mo.; sulbactam, Pfizer Inc., Groton, Conn.; and tazobactam, Wyeth-Ayerst, Pearl River, N.Y. Amoxicillin-clavulanic acid and ampicillin-sulbactam were tested in 2:1 ratios. Tazobactam was tested with a fixed 4-μg/ml concentration in each well.
Media.
HTM and MHC medium from three different manufacturers were used. Prepared media were obtained as follows: MH broth and agar, Becton Dickinson; HTM broth and agar, Becton Dickinson, Remel (Lenexa, Kans.), PML Microbiologicals (Wilsonville, Oreg.); and MHC agar, Becton Dickinson, Remel, and Northeast Laboratories (Waterville, Mass.). One lot of each medium type from each manufacturer was tested.
Broth microdilution MIC assays.
Broth microdilution MIC assays were performed by the NCCLS methodology (6) in cation-adjusted MH broth (CAMHB) and HTM broth. MICs were determined visually.
E-test MIC testing.
E-test strips impregnated with amoxicillin-clavulanic acid, ampicillin-sulbactam, and piperacillin-tazobactam were obtained from AB Biodisk (Piscataway, N.J.). The E test was performed according to the recommendations of the manufacturer on MH, HTM, and MHC agars.
Disk diffusion susceptibility testing.
Antibiotic-impregnated disks containing 20 μg of amoxicillin with 10 μg of clavulanic acid, 10 μg of ampicillin with 10 μg of sulbactam, and 100 μg of piperacillin with 10 μg of tazobactam were obtained from Becton Dickinson Microbiology Systems, Sparks, Md. Disk diffusion testing was performed by the NCCLS methodology (7) on MH, HTM, and MHC agars.
Analysis of results.
Susceptibility testing with E. coli ATCC 35218 was performed on HTM broth and agar and MHC agar media from three different manufacturers of each medium. The isolates were tested in triplicate for 10 consecutive days. This generated a total of 90 test results for each medium by each methodology for each antimicrobial agent. Susceptibility testing with E. coli ATCC 35218 was performed on MH media in triplicate on each day of testing. The results of the broth microdilution, E-test, and disk diffusion methods with E. coli ATCC 35218 on MH media and HTM were compared for amoxicillin-clavulanic acid, ampicillin-sulbactam, and piperacillin-tazobactam. E-test and disk diffusion testing of E. coli ATCC 35218 with all three antimicrobial agents was performed on MHC agar. All results obtained with HTM and MHC media were compared to the acceptable ranges established by NCCLS for E. coli ATCC 35218 with MH media. The number of dilution differences between HTM and MH media was also compared for the broth microdilution and E-test MIC testing methods. E-test results that fell between the doubling dilutions seen by the broth microdilution method were rounded to the next higher dilution. The zone diameter size was measured in millimeters for disk diffusion testing. The diameter sizes obtained with HTM and MHC agars were compared to the diameter sizes obtained with MH agar. The results for matched E. coli ATCC 35218 and passaged E. coli ATCC 35218 isolates were compared for all of the media and antibiotic agents by each methodology. For the disk tests, straight-line regression analysis was used to evaluate the paired differences. For the MIC tests, confidence intervals based on a binomial model were used for the statistical assessment.
RESULTS AND DISCUSSION
Quality control.
A total of 90 test results for each antimicrobial agent for each broth and agar were generated for E. coli ATCC 35218. Quality control results with CAMHB and MH agar for E. coli ATCC 35218 were in range each of the 10 days by all three methodologies. Quality control results for H. influenzae ATCC 49247 with HTM were in range for each day of testing with one exception. On day 4, there was contamination by the broth microdilution method only. Day 4 results were not included in the study for the broth microdilution method; therefore, the total number of tests for the microbroth dilution method was reduced to 81.
Broth microdilution method with HTM broth.
The broth microdilution test results for E. coli ATCC 35218 with HTM broth were compared to the NCCLS reference ranges, which were established with CAMHB. Of the results for amoxicillin-clavulanic acid and piperacillin-tazobactam with HTM, 100% were within the acceptable NCCLS range. A total of 86% (74 of 81) of the results for ampicillin-sulbactam were within the NCCLS range. The seven results that were not acceptable for ampicillin-sulbactam were out of range by 1 dilution (Table 1). When the broth microdilution test results with HTM were compared to the results obtained with CAMHB, all results were within 1 dilution for all of the antimicrobial agents tested. When a difference was observed, the results were generally 1 doubling concentration higher in HTM broth than in CAMHB (Table 2).
TABLE 1.
Distribution of MICs for E. coli ATCC 35218 obtained by the broth microdilution method and E test with HTM and MHC agar
| MICa (μg/ml) | No. of resultsb:
|
||||||||
|---|---|---|---|---|---|---|---|---|---|
| Broth microdilution method with HTM broth
|
E test with HTM agar
|
E test with MHC agar
|
|||||||
| A/C | A/S | P/T | A/C | A/S | P/T | A/C | A/S | P/T | |
| 128 | 1 | ||||||||
| 64 | 23 | ||||||||
| 32 | 7 | 21 | 42 | ||||||
| 16 | 74 | 8 | 69 | 30 | 24 | ||||
| 8 | 72 | 82 | 56 | ||||||
| 4 | 9 | 4 | |||||||
| 2 | 8 | 48 | 71 | ||||||
| 1 | 71 | 42 | 19 | ||||||
| 0.5 | 2 | ||||||||
MIC refers to that of the β-lactam component only.
Boldface numbers indicate the acceptable NCCLS range. A/C, amoxicillin/clavulanic acid (2:1); A/S, ampicillin/sulbactam (2:1); P/T, piperacillin/tazobactam (4 μg of tazobactam per ml).
TABLE 2.
Difference between HTM and MH medium by test method
| Differencea | No. of resultsb
|
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| HTM broth, broth microdi-lution method
|
HTM agar E testc
|
HTM agar disk diffusion method
|
|||||||||
| A/C | A/S | P/T | A/C | A/S | P/T | A/C | A/S | P/T | |||
| 3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||
| 2 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | ||
| 1 | 67 | 13 | 11 | 8 | 21 | 37 | 5 | 13 | 9 | ||
| 0 | 13 | 68 | 68 | 82 | 69 | 52 | 10 | 35 | 36 | ||
| −1 | 1 | 0 | 2 | 0 | 0 | 1 | 42 | 36 | 29 | ||
| −2 | 0 | 0 | 0 | 0 | 0 | 0 | 28 | 5 | 14 | ||
| −3 | 0 | 0 | 0 | 0 | 0 | 0 | 4 | 1 | 2 | ||
The difference between HTM and MH for the broth microdilution method and the E test was measured by MIC doubling dilution; the difference for the disk diffusion method was measured in millimeters.
See footnote b of Table 1 for definitions of abbreviations.
For MIC results obtained between doubling dilutions of the MIC, the results were rounded up to the next higher MIC.
E test with HTM agar.
When the results obtained by the E test with HTM agar for E. coli ATCC 35218 were compared with the NCCLS established reference range, 100% of the results for amoxicillin-clavulanic acid and piperacillin-tazobactam were within the acceptable range. However, only 77% (69 of 90) of the results for ampicillin-sulbactam were in range. All 21 results that were out of range for ampicillin-sulbactam were within 1 dilution of the acceptable range (Table 1). When E-test results with HTM agar were compared to the results obtained with MH agar, all results were within 1 doubling dilution for all of the antimicrobial agents tested. When a difference was observed, the results were generally 1 doubling concentration higher with HTM agar than with MH agar (Table 2).
Disk diffusion testing with HTM agar.
The disk diffusion testing results for E. coli ATCC 35218 with HTM agar were compared to the NCCLS reference range. Of the results for ampicillin-sulbactam and piperacillin-tazobactam with HTM agar, 100% were within the acceptable NCCLS range. For amoxicillin-clavulanic acid, 82% (74 of 90) of the results were in range. All 16 results that were out of range for amoxicillin-clavulanic acid were out of range by 1 mm (Table 3). Ninety-six percent (86 of 90) of the results for amoxicillin-clavulanic acid, 99% (89 of 90) of the results for ampicillin-sulbactam, and 98% (88 of 90) of the results for piperacillin-tazobactam with HTM agar were within 2 mm of the results with MH agar. When a difference was observed between the agars, zone sizes were generally smaller with HTM agar than with MH agar (Table 2).
TABLE 3.
Disk diffusion zone diameter distribution for E. coli ATCC 35218 with HTM agar and MHC agar
| Diam (mm) | No. of resultsa
|
|||||
|---|---|---|---|---|---|---|
| HTM agar
|
MHC agar
|
|||||
| A/C | A/S | P/T | A/C | A/S | P/T | |
| 10 | 9 | |||||
| 11 | 23 | |||||
| 12 | 20 | |||||
| 13 | 9 | 10 | ||||
| 14 | 46 | 1 | 19 | |||
| 15 | 34 | 3 | 9 | |||
| 16 | 1 | 18 | ||||
| 17 | 16 | 26 | ||||
| 18 | 46 | 27 | ||||
| 19 | 23 | 13 | ||||
| 20 | 5 | 2 | ||||
| 21 | ||||||
| 22 | ||||||
| 23 | 2 | |||||
| 24 | 17 | |||||
| 25 | 4 | 24 | ||||
| 26 | 42 | 26 | ||||
| 27 | 31 | 16 | ||||
| 28 | 12 | 4 | ||||
| 29 | 1 | 1 | ||||
| 30 | ||||||
| 31 | ||||||
Boldface numbers indicate the acceptable NCCLS range. See footnote b to Table 1 for definitions of abbreviations.
E test with MHC agar.
E-test results for E. coli ATCC 35218 with MHC agar were compared to the NCCLS established reference range for E. coli ATCC 35218 with MH agar. Of the results for amoxicillin-clavulanic acid and piperacillin-tazobactam with MHC agar, 100% were within the acceptable NCCLS range. However, only 27% (24 of 90) of the results for ampicillin-sulbactam were within range. The results for 66 tests on MHC agar that were out of the acceptable NCCLS range for ampicillin-sulbactam were 1 to 4 doubling dilutions higher than the results for tests on MH agar (Table 1). When the results of the E test with MHC agar were compared to the results of the E test with MH agar, all results were within 1 dilution for amoxicillin-clavulanic acid and piperacillin-tazobactam. Of the results for ampicillin-sulbactam with MHC agar, a 2-doubling dilution difference was seen for 28% (25 of 90) of the results. When a difference was observed, the results were generally higher with MHC agar.
Disk diffusion testing with MHC agar.
The results obtained by disk diffusion testing of E. coli ATCC 35218 with MHC agar were compared to the NCCLS reference range. Complete agreement with established NCCLS ranges was not seen for any of the antimicrobial agents. Of the results for piperacillin-tazobactam with MHC agar, 98% were within the acceptable NCCLS range. However, only 47% of the results for amoxicillin-clavulanic acid agar with MHC and 42% of the results for ampicillin-sulbactam with MHC agar were within the acceptable NCCLS range (Table 3). When the disk diffusion testing results obtained with MHC agar were compared to those obtained with MH agar, up to a 5-mm difference was observed for all antimicrobial agents tested. When a difference was observed between agars, the zone sizes were generally smaller with MHC agar than with MH agar. Theoretically, the differences observed between the media may be due to better growth in the richer medium. Another reason for the difference could be that the components in the enriched medium may be affecting the antimicrobial agents.
Passaged isolates.
The same procedures that were performed with the E. coli ATCC 35218 isolates were performed with the passaged E. coli ATCC 35218 isolates. Quality control results for the passaged E. coli ATCC 35218 isolates were in range on each of the 10 days on CAMHB and MH agar by all three methodologies and with all three antimicrobial agents. There were no statistically significant differences between the results obtained with the fresh and passaged isolates for any of the media, antimicrobial agents, or methodologies tested. P values were greater than 0.05 in all cases.
The primary objective of the study was to determine if comparable quality control results for E. coli ATCC 35218 would be produced for the three β-lactam–β-lactamase inhibitors with MH medium and HTM. Acceptable antimicrobial agent-methodology combinations are highlighted in Table 4.
TABLE 4.
Antimicrobial agent-methodology results for quality control testing of the β-lactam–β-lactamase inhibitors with E. coli ATCC 35218 with HTM and MHC agar
| Antibiotic | Acceptability of resulta
|
||||
|---|---|---|---|---|---|
| HTM
|
MHC agar
|
||||
| Broth microdilution method | E test | Disk diffusion method | E test | Disk diffusion method | |
| Amoxicillin-clavulanic acid | A | A | U | A | U |
| Ampicillin-sulbactam | U | U | A | U | U |
| Piperacillin-tazobactam | A | A | A | A | U |
A, acceptable (100% of results within the NCCLS ranges for E. coli ATCC 35218); U, unacceptable (<100% of results within the NCCLS range).
When performing Haemophilus susceptibility testing with the β-lactam–β-lactamase inhibitor agents, in addition to the testing of H. influenzae ATCC 49247, quality control should be performed with an isolate capable of detecting deterioration in the β-lactamase inhibitors. E. coli ATCC 35218 detects deterioration in the β-lactamase inhibitors, and NCCLS has established quality control ranges for this organism with MH media. The results of this study indicate that quality control testing with E. coli ATCC 35218 and HTM was acceptable for the broth microdilution and E-test methods with amoxicillin-clavulanic acid and piperacillin-tazobactam and for disk diffusion testing with ampicillin-sulbactam and piperacillin-tazobactam.
ACKNOWLEDGMENT
We thank Paul R. McAllister, SmithKline Beecham SCS/Biometrics/Statistical Sciences, for statistical analysis.
REFERENCES
- 1.Cooper C E, Slocombe B, White A R. Effect of low concentrations of clavulanic acid on the in-vitro activity of amoxicillin against β-lactamase-producing Branhamella catarrhalis and Haemophilus influenzae. J Antimicrob Chemother. 1990;26:371–380. doi: 10.1093/jac/26.3.371. [DOI] [PubMed] [Google Scholar]
- 2.Doern G V, Brueggemann A B, Pierce G, Holley H P, Jr, Rauch A. Antibiotic resistance among clinical isolates of Haemophilus influenzae in the United States in 1994 and 1995 and detection of β-lactamase-positive strains resistant to amoxicillin-clavulanate: results of a national multicenter surveillance study. Antimicrob Agents Chemother. 1997;41:292–297. doi: 10.1128/aac.41.2.292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Doern, G. V., and The Alexander Project Collaborative Group. 1996. Antimicrobial resistance among lower respiratory tract isolates of Haemophilus influenzae: results of a 1992–93 Western Europe and USA collaborative surveillance study. J. Antimicrob. Chemother. 38(Suppl. A.):59–69. [DOI] [PubMed]
- 4.Jacobs M R, Bajaksouzian S, Lin G, Appelbaum P C. abstracts of the 98th General Meeting of the American Society for Microbiology 1998. Washington, D.C: American Society for Microbiology; 1998. Susceptibility of Streptococcus pneumoniae and Haemophilus influenzae to oral agents: results of a 1997 epidemiological survey, abstr. A-31; p. 43. [Google Scholar]
- 5.National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial disk susceptibility tests. NCCLS publication no. M2-A3. Wayne, Pa: National Committee for Clinical Laboratory Standards; 1983. [Google Scholar]
- 6.National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. NCCLS publication no. M7-A4. Wayne, Pa: National Committee for Clinical Laboratory Standards; 1997. [Google Scholar]
- 7.National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial disk susceptibility tests. NCCLS publication no. M2-A6. Wayne, Pa: National Committee for Clinical Laboratory Standards; 1997. [Google Scholar]
- 8.Richard M P, Gil Aguado A, Mattina R, Marre R the SPAR Study Group. Sensitivity to sparfloxacin and other antibiotics, of Streptococcus pneumoniae, Haemophilus influenzae and Moraxella catarrhalis strains isolated from adult patients with community-acquired lower respiratory tract infections: a European multicentre study. J Antimicrob Chemother. 1998;41:207–214. doi: 10.1093/jac/41.2.207. [DOI] [PubMed] [Google Scholar]
- 9.Schito, G. C., S. Mannelli, A. Pesce, and The Alexander Project Group. 1997. Trends in the activity of macrolide and β-lactam antibiotics and resistance development. J. Chemother. 9(Suppl. 3.):18–27. [PubMed]
- 10.Thornsberry C, Ogilvie P, Kahn J, Mauriz Y the Laboratory Investigator Group. Surveillance of antimicrobial resistance in Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis in the United States in 1996–1997 respiratory season. Diagn Microbiol Infect Dis. 1997;29:249–257. doi: 10.1016/s0732-8893(97)00195-8. [DOI] [PubMed] [Google Scholar]