Resistance among Streptococcus pneumoniae Clinical Isolates by Use of the E Test

Transcription

1 JOURNAL OF CLINICAL MICROBIOLOGY, Jan. 1994, p /94/$ Copyright 1994, American Society for Microbiology Vol. 32, No. 1 Detection of Penicillin and Extended-Spectrum Cephalosporin Resistance among Streptococcus pneumoniae Clinical Isolates by Use of the E Test JAMES H. JORGENSEN,1* MARY JANE FERRARO,2 M. LETICIA McELMEEL,1 JEAN SPARGO,2 JANA M. SWENSON,3 AND FRED C. TENOVER3 Department ofpathology, The University of Texas Health Science Center, San Antonio, Texas Microbiology Laboratory, Massachusetts General Hospital, Boston, Massachusetts ; and Hospital Infections Program, National Centerfor Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia Received 26 July 1993/Returned for modification 9 September 1993/Accepted 12 October 1993 Increasing penicillin resistance and the initial recognition of resistance to extended-spectrum cephalosporins among Streptococcus pneumoniae isolates have placed greater emphasis on accurate methods for susceptibility testing of clinical isolates. This study has evaluated the use of the E test (AB Biodisk NA, Piscataway, N.J.) for the detection of penicillin and cefotaxime resistance among 147 pneumococcal clinical isolates in three geographically separate laboratories. These included 42 penicillin-resistant (MIC,.2 tlg/ml) and 14 cefotaxime-resistant (defined here as an MIC of.2 tig/ml) isolates. E test strips were applied to the surface of Mueller-Hinton sheep blood agar plates and incubated at 35 C in 5% CO2 for 20 to 24 h. E test MICs were compared with MICs determined with lysed horse blood-supplemented Muelier-Hinton broth in a microdilution format as recommended by the National Committee for Clinical Laboratory Standards. Penicillin MICs agreed within one log2 dilution for 136 of 147 (92.5%) isolates, and cefotaxime MICs agreed within one log2 dilution for 142 of 147 (96.6%) isolates. No very major or major interpretive errors occurred with either penicillin or cefotaxime E test MIC results. There were 9.5 and 5.4% minor interpretive category errors with penicillin and cefotaxime E test MICs, respectively. These data indicate that the E test represents a convenient and reliable method for the detection of penicillin or cephalosporin resistance in pneumococci. Penicillin resistance has increased sharply within the past few years among Streptococcus pneumoniae clinical isolates in many areas of the United States (11-13, 18, 23). Because of the concern for potential treatment failures due to penicillin resistance, extended-spectrum cephalosporins have been recommended by some for the empiric therapy of pneumococcal meningitis (9, 17, 24, 25). However, therapeutic failures for meningitis caused by extended-spectrum cephalosporin-resistant strains were reported recently (2, 5, 16). Therefore, clinical microbiology laboratories must be able to recognize antimicrobial resistance in pneumococci quickly and accurately. This problem has been complicated by the fact that well-standardized, practical methods for susceptibility testing of pneumococci have not been widely available. This study has evaluated the use of the E test (AB Biodisk NA, Piscataway, N.J.) for the detection of penicillin and cefotaxime resistance among 147 pneumococcal clinical isolates in three geographically separate laboratories. MATERIALS AND METHODS Test organisms. Approximately 50 unique isolates of S. pneumoniae were tested in each of the three laboratories (i.e., the Centers for Disease Control and Prevention, the Massachusetts General Hospital, and the University of Texas Health Science Center). They were fresh clinical isolates, isolates sent to the laboratories for confirmation of antimicrobial resistance, or isolates selected from among the * Corresponding author. Mailing address: Department of Pathology, University of Texas Health Science Center, San Antonio, TX Phone: (210) Fax: (210) laboratories' stock culture collections to ensure adequate numbers of resistant isolates. Included among the study isolates were 42 penicillin- and 14 cefotaxime-resistant isolates (see definitions of resistance below). S. pneumoniae ATCC (15) and CDC strain CS 101 (multiply-antibiotic resistant) were used as quality control strains in each laboratory. To assess interlaboratory reproducibility, each laboratory tested a coded collection of 10 strains with different levels of penicillin and cephalosporin susceptibilities. Reference broth microdilution susceptibility tests. Penicillin and cefotaxime MICs were determined for each isolate with lysed horse blood-supplemented Mueller-Hinton broth (LHB) in a microdilution format as recommended by the National Committee for Clinical Laboratory Standards (NC- CLS) (14). Two different brands (BBL Microbiology Systems, Cockeysville, Md., and Unipath Oxoid, Ogdensburg, N.Y.) of Mueller-Hinton broth were used to prepare different lots of LHB for the study. Inocula were prepared by suspending growth obtained by 20 to 24 h of incubation in 5% CO2 on sheep blood agar plates. Colonies were suspended to the turbidity of a 0.5 McFarland standard in 0.9% saline and then further diluted within 15 min to provide a final inoculum density of 5 x 105 CFU/ml in the wells of the microdilution trays. The microdilution trays were incubated at 35 C in ambient air for 20 to 24 h prior to the determination of MICs. Penicillin susceptibility was indicated by an MIC of <0.06,g/ml, relative resistance was defined as MICs of 0.12 to 1,g/ml, and resistance was defined as an MIC of 22,g/ml (15). For the purposes of this study, cefotaxime susceptibility was defined as an MIC of <0.25,g/ml, and resistance was defined as an MIC of.2 jg/ml. E tests. MICs were determined with low-concentration

2 160 JORGENSEN ET AL. J. CLIN. MICROBIOL. TABLE 1. Comparison of penicillin and cefotaxime MICs determined by LHB microdilution susceptibility test and E test methods for 10 reference pneumococcal strains tested in all three laboratories MIC Strain and (pg/ml) determined with the indicated method in laboratory: antimicrobial A B C agent LHB E test LHB E test LHB E test 1 Penicillin Cefotaxime Penicillin Cefotaxime Penicillin Cefotaxime Penicillin Cefotaxime S Penicillin Cefotaxime Penicillin Cefotaxime Penicillin Cefotaxime Penicillin Cefotaxime Penicillin Cefotaxime Penicillin Cefotaxime penicillin (gradient range of to 32,ug/ml) and cefotaxime (gradient range of to 32,g/ml) E test strips applied to the surfaces of 150-mm Mueller-Hinton sheep blood agar plates. Three different lots of commercially prepared (BBL) plates were used in the study. The plates were inoculated with a 0.5 McFarland turbidity standard organism suspension prepared in 0.9% saline as described above. The plates were incubated at 35 C in 5% CO2 for 20 to 24 h prior to the determination of MICs. The MIC was defined by the intersection of the growth elipse margin with the E test strip by use of reflected light at an oblique angle to examine the plates. Care was taken to note the margin of pneumococcal growth, not the area of alpha-hemolysis of the medium. Because the E test strips are marked in one-half log2 concentrations, it is possible to record MICs in increments smaller than the usual twofold increments indexed to the base 1. In such instances in this study, E test MICs were rounded to the next higher log2 MICs for purposes of comparison with reference MICs. RESULTS Each of three laboratories tested two S. pneumoniae control strains, 10 coded challenge strains, and approximately 50 unique clinical isolates by using the E test and the NCCLS reference broth microdilution procedure with penicillin and cefotaxime. The expected results were obtained with the control strains in all three laboratories. Table 1 depicts the interlaboratory reproducibility of results obtained with the collection of 10 strains tested in a blinded fashion by each laboratory. These strains spanned a wide range of susceptibility and resistance to both penicillin and cefotaxime. Excellent reproducibility of each test method was demonstrated in this phase of the study; i.e., 100% of penicillin MICs determined by either the E test or broth microdilution agreed within two log2 dilutions, and 96.7% of cefotaxime MICs determined by either method agreed within two log2 dilutions. Thus, the two test methods demonstrated equal interlaboratory reproducibilities with both antibiotics. Table 2 depicts the comparison between penicillin MICs

3 VOL. 32, 1994 E TEST FOR S. PNEUMONIAE DRUG RESISTANCE 161 TABLE 2. Laboratory Comparison of penicillin MICs determined in the three laboratories by the E test and the LHB microdilution test for approximately 50 unique clinical isolates per laboratory No. of E test MICs within indicated log2 dilution of No. of isolates % Interpretive errors' LHB MICs -2-1 ~~~~~~~~~~ ~~~~~~resistant MaoMir -2 -l Same Resistant Relatively major Very Major Minor A B C All a Very major errors = (number of false-susceptible results/number of resistant isolates) x 100. Major errors = (number of false-resistant results/number of susceptible isolates) x 100. Minor errors = (number of S to I, I to S, R to I, or I to R results/total number of isolates tested) x 100, where S = susceptible, I = relatively resistant, and R = resistant. determined by the E test and those determined by broth microdilution for the total collection of 147 unique clinical isolates tested in the three laboratories. Penicillin MICs agreed within one log2 dilution for 136 of 147 (92.5%) isolates when determined by the E test in comparison with broth dilution. Generally speaking, penicillin MICs determined by the E test tended to be the same as or one log2 dilution lower than those determined by broth microdilution. There were no very major or major interpretive category errors with penicillin E test results on the basis of approved NCCLS MIC interpretive categories specific for pneumococcal testing (14). However, there were 9.5% minor interpretive errors with penicillin E test results. The majority of the minor errors (10 of 14) occurred with isolates that were defined as relatively resistant by broth microdilution (usually an MIC of 0.12,tg/ml) but that were categorized as susceptible by the E test (usually an MIC of 0.06 Fg/ml) (data not shown). Table 3 depicts the cefotaxime MICs determined by the E test and the reference method. Cefotaxime E test MICs agreed with broth microdilution MICs within one dilution for 142 of 147 (96.6%) isolates. There were no very major or major interpretive errors with cefotaxime E test results, and there were 5.4% minor errors on the basis of the cefotaxime interpretive criteria invoked for the purposes of this study. Six of eight minor errors occurred with isolates that were defined as relatively resistant by broth microdilution (an MIC of 0.5 pg/ml for five of six strains) but that were categorized as susceptible by the E test (an MIC of 0.25 pg/ml for all six strains). E test MICs were determined with both CO2 and ambient air incubation for the 50 unique isolates in one of the laboratories (Massachusetts General Hospital). Four of 50 isolates (8%) failed to grow when incubated in ambient air but grew under CO2 incubation. Table 4 depicts the penicillin and cefotaxime MICs determined in both incubation atmospheres in comparison with the reference broth microdilution results for the 46 isolates which grew in both atmospheres. E test MICs determined with CO2 incubation were approximately one-half log2 dilution higher than those determined in ambient air, which provided a closer agreement with the reference broth microdilution MICs. DISCUSSION In contrast to the relatively low prevalence of S. pneumoniae clinical isolates resistant to primary agents such as penicillin, chloramphenicol, trimethoprim-sulfamethoxazole, and tetracycline only a few years ago (7, 10, 20), pneumococcal strains resistant to penicillin or to four or more drugs have now been reported from several areas of the United States (11, 13, 18, 22). Perhaps even more troubling are strains which have been found in several locales, that are resistant to various extended-spectrum cephalosporins, and have resulted in treatment failures for meningitis (2, 5, 16, 19). Extended-spectrum cephalosporins have been recommended for the empiric therapy of serious infections due to suspected penicillin-resistant or relatively resistant pneumococci (9, 17, 24, 25). Thus, it is important that clinical microbiology laboratories use routine susceptibility testing procedures which can quickly and accurately recognize resistant pneumococcal isolates. The oxacillin disk test has been found to be reliable for predicting susceptibility to penicillin (11, 21). However, the test cannot distinguish strains that are penicillin resistant (MIC, 22,ug/ml) from those that demonstrate relative resistance to penicillin (MICs, 0.1 to 1,ug/ml). Indeed, some penicillin-susceptible strains of pneumococci produce zones of <20 mm in the oxacillin disk screening test (11, 21, 26). TABLE 3. Laboratory Comparison of cefotaxime MICs determined in the three laboratories by the E test and the LHB microdilution test for approximately 50 unique clinical isolates per laboratory No. of E test MICs within indicated log2 dilution of No. of isolates % Interpretive effors LHB MICs -2-1 Same Resistant Relatively Very Major Minor resistant major Mao Mnr A B C All a See Table 2, footnote a.

4 162 JORGENSEN ET AL. J. CLIN. MICROBIOL. TABLE 4. Comparison of E test MICs determined with CO2 and ambient air incubation and MICs determined by the LHB microdilution test (incubation in ambient air) for 46 pneumococcal isolates tested in one laboratory' Antimicrobial No. of E test MICs within indicated log2 dilution of LHB MICs % Interpretive errors" agent and incubation conditions < Same >+2 Ve*ry Major Minor major Penicillin Air CO Cefotaxime Air CO ' Calculations were based on the 46 isolates which grew in both ambient air and CO2 atmospheres; 4 isolates did not grow in ambient air. b See Table 2, footnote a. Thus, it may be important to define the level of penicillin susceptibility of significant pneumococcal isolates by determining an MIC. Moreover, disk diffusion testing of pneumococci for susceptibility to cephalosporins has not yet been well standardized (4, 22). The MIC procedure recommended by the NCCLS for testing pneumococci involves the use of 2 to 5% LHB or Haemophilus test medium in a broth microdilution test format (15). Lysed horse blood-supplemented medium is tedious to prepare and has not been widely available from commercial sources. Haemophilus test medium may be simpler to prepare or obtain commercially; however, MICs tend to be somewhat lower than those determined in LHB, a fact which can affect the interpretation of results (6), and growth failures may occur with some strains (unpublished observation). The E test is a relatively simple method for determining the MICs of a variety of antimicrobial agents for both common and fastidious or anaerobic bacteria (1, 3, 8). Excellent interlaboratory agreement of MICs was demonstrated in this study with both the E test and the broth microdilution procedure when a coded challenge set of pneumococcal strains was tested independently in the three laboratories. A previous study with pneumococci noted generally favorable, although imperfect, agreement of MICs determined by the E test and NCCLS broth dilution methods (8). However, in that study, the E test involved incubation in ambient air, rather than in a CO2 environment, as in the present study. The results of this study have shown very good agreement between MICs determined by the E test (when incubated in C02) and those determined by the NCCLS broth microdilution test for penicillin and cefotaxime; i.e., 92.5 to 96.6% of MICs were within one log2 dilution. Incubation in CO2 for the pneumococcal E test improves the agreement with MICs determined by the NC- CLS LHB broth microdilution method and avoids the problem caused by the fact that some pneumococcal strains are capneic and do not grow readily under ambient air incubation conditions. No very major or major interpretive category errors were encountered in this study with either penicillin or cefotaxime E test MICs. The majority of the minor errors observed with penicillin occurred with strains for which MICs clustered at the breakpoint separating susceptible and relatively resistant strains, i.e., 0.06 to 0.12,ug/ml. Only 5.4% minor errors occurred among cefotaxime E test results, on the basis of the MIC interpretive criteria selected for use in this study (i.e., susceptible, <0.25,ug/ml; resistant,.2,g/ml). These breakpoints were chosen for use in this study on the basis of reported treatment failures for meningitis caused by strains for which MICs were 2,ug/ml or greater (2, 5, 19). These breakpoints are admittedly conservative and may warrant reevaluation following further clinical experience with strains for which cephalosporin MICs are in the range of 0.5 to 2,ug/ml. Indeed, these same MIC interpretive criteria are under consideration presently by the NCCLS (personal communication). The results of this three-laboratory study indicate that the E test represents a reliable method for the detection of penicillin or cephalosporin resistance in pneumococci. Clinical isolates of pneumococci can be tested by the E test in a manner quite similar to that for the disk diffusion method. However, unlike oxacillin disk screening test results, penicillin E test results unambiguously define the level of penicillin susceptibility of pneumococcal isolates. In addition, use of the E test would allow prompt definition of the degree of susceptibility to an extended-spectrum cephalosporin, such as cefotaxime. We recommend that all pneumococcal isolates from patients with meningitis be tested against penicillin and an extended-spectrum cephalosporin as soon as isolated colonies are available because of the serious consequences of inadequate antibiotic treatment (19). Our data suggest that the E test is a convenient and accurate means of performing such testing when microdilution testing with LHB is not readily available. For pneumococcal isolates from patients without meningitis, it may be more cost-effective to screen isolates by the oxacillin disk test and pursue further testing of penicillin and alternative agents only for isolates showing zone sizes of less than 20 mm. ACKNOWLEDGMENT E test strips were kindly provided for the purposes of this study by AB Biodisk NA, Piscataway, N.J. REFERENCES 1. Baker, C. N., S. A. Stocker, D. M. Culver, and C. Thornsberry Comparison of the E-test to agar dilution, broth microdilution, and agar diffusion susceptibility testing techniques by using a special challenge set of bacteria. J. Clin. Microbiol. 29: Bradley, J. S., and J. D. Conner Ceftriaxone failure in meningitis caused by Streptococcus pneumoniae with reduced susceptibility to beta-lactam antibiotics. Pediatr. Infect. Dis. J. 10: Citron, D. M., M. L Ostovari, A. Karlsson, and E. J. C. Goldstein Evaluation of the E test for susceptibility testing of anaerobic bacteria. J. Clin. Microbiol. 29:

5 VOL. 32, 1994 E TEST FOR S. PNEUMONIAE DRUG RESISTANCE Friedland, I. R., S. Shelton, and G. H. McCracken, Jr Screening for cephalosporin-resistant Streptococcus pneumoniae with the Kirby-Bauer disk susceptibility test. J. Clin. Microbiol. 31: Friedland, I. R., S. Shelton, M. Paris, S. Rinderknecht, S. Ehrett, K. Krisher, and G. H. McCracken, Jr Dilemmas in diagnosis and management of cephalosporin-resistant Streptococcus pneumoniae meningitis. Pediatr. Infect. Dis. J. 12: Jorgensen, J. H., G. V. Doern, M. J. Ferraro, C. C. Knapp, J. M. Swenson, and J. A. Washington Multicenter evaluation of the use of Haemophilus test medium for broth microdilution antimicrobial susceptibility testing of Streptococcus pneumoniae and development of quality control limits. J. Clin. Microbiol. 30: Jorgensen, J. H., G. V. Doern, L. A. Maher, A. W. Howell, and J. S. Redding Antimicrobial resistance among respiratory isolates of Haemophilus influenzae, Moraxella catarrhalis, and Streptococcus pneumoniae in the United States. Antimicrob. Agents Chemother. 34: Jorgensen, J. H., A. W. Howell, and L. A. Maher Quantitative antimicrobial susceptibility testing of Haemophilus influenzae and Streptococcus pneumoniae by using the E-test. J. Clin. Microbiol. 29: Klass, P. E., and J. 0. Klein Therapy of bacterial sepsis, meningitis, and otitis media in infants and children: 1992 poll of directors of programs in pediatric infectious diseases. Pediatr. Infect. Dis. J. 11: Klugman, K. P Pneumococcal resistance to antibiotics. Clin. Microbiol. Rev. 3: Mason, E. O., Jr., S. L. Kaplan, L. B. Lamberth, and J. Tillman Increased rate of isolation of penicillin-resistant Streptococcus pneumoniae in a children's hospital and in vitro susceptibilities to antibiotics of potential therapeutic use. Antimicrob. Agents Chemother. 36: McDougal, L., R. Facklam, M. Reeves, S. Hunter, J. M. Swenson, B. C. Hill, and F. C. Tenover Analysis of multiply antimicrobial-resistant isolates of Streptococcus pneumoniae from the United States. Antimicrob. Agents Chemother. 36: Munoz, R, T. J. Coffey, M. Daniels, C. G. Dowson, G. Laible, J. Casal, R. Hakenbeck, M. Jacobs, J. M. Musser, B. G. Spratt, and A. Tomasz Intercontinental spread of a multiresistant clone of serotype 23F Streptococcus pneunoniae. J. Infect. Dis. 164: National Committee for Clinical Laboratory Standards Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved standard M7-A2. National Committee for Clinical Laboratory Standards, Villanova, Pa. 15. National Committee for Clinical Laboratory Standards Performance standards for antimicrobial susceptibility testing. Fourth informational supplement M100-S4. National Committee for Clinical Laboratory Standards, Villanova, Pa. 16. O'Keefe, P. T., and R. Bayston Pneumococcal meningitis in a child with ventriculo-peritoneal shunt. J. Infect. 22: Pallares, R., F. Gudiol, J. Unares, J. Ariza, G. Rufi, L. Murguli, J. Dorca, and P. F. Vlldanich Risk factors and response to antibiotic therapy in adults with bacteremic pneumonia caused by penicillin-resistant pneumococci. N. Engl. J. Med. 317: Reichler, M. R., A. A. Aliphin, R. F. Breiman, J. R. Schreiber, J. E. Arnold, L. K. McDougal, R. R. Facklam, B. Boxerbaum, D. May, R. 0. Walton, and M. R. Jacobs The spread of multiply-resistant Streptococcus pneumoniae at a day care center in Ohio. J. Infect. Dis. 166: Sloas, M. M., F. F. Barrett, B. K. English, B. C. Hill, F. C. Tenover, and R. J. IAeadro Cephalosporin treatment failure in penicillin- and cephalosporin-resistant Streptococcus pneumoniae meningitis. Pediatr. Infect. Dis. J. 11: Spika, J. S., R. R. Facklam, B. D. Plikaytis, M. J. Oxtoby, and the Pneumococcal Surveillance Working Group Antimicrobial resistance of Streptococcus pneunoniae in the United States, J. Infect. Dis. 163: Swenson, J., B. C. Hill, and C. Thornsberry Screening pneumococci for penicillin resistance. J. Clin. Microbiol. 24: Tenover, F. C., J. M. Swenson, and L. K. McDougal Screening for extended-spectrum cephalosporin resistance in pneumococci. Lancet 340: Thornsberry, C., J. K. Marler, and T. J. Rich Increased penicillin (Pen) resistance in recent U.S. isolates of Streptococcus pneumoniae, abstr. C-268, p Abstr. 92nd Gen. Meet. Am. Soc. Microbiol Vildanch, P. F., F. Guidol, J. Linares, G. Rufi, J. Ariza, and R. Pallares Characteristics and antibiotic therapy of adult meningitis due to penicillin-resistant pneumococci. Am. J. Med. 84: Weingarten, R. D., Z. Markiewicz, and D. N. Gilbert Meningitis due to penicillin-resistant Streptococcus pneumoniae in adults. Rev. Infect. Dis. 12: Willet, L. D., H. C. Dillon, Jr., and B. M. Gray Penicillinintermediate pneumococci in a children's hospital. Am. J. Dis. Child. 139: