Comparison of susceptibility test methods to detect penicillin-resistant Streptococcus pneumoniae

213

DIAGN MICROBIOL INFECT DIS 1993;17:213-217

Comparison of Susceptibility Test Methods to Detect Penicillin-Resistant

Streptococcus pneumoniae Richard B. Clark, Olarae Giger, and Joel E. Mortensen

The detection of penicillin-resistant Streptococcus pneumoniae was assessed by six different methods: agar dilution, oxacillin screen by disk diffusion, E-test, and three overnight microdilution test methods that included commercial panels from MicroScan and Micro Media and in-house-made conventional panels using a commercial Haemophilus test medium (HTM) broth. Of the 52 pneumococcal isolates tested, 12 were resistant, 16 were relatively resistant, and 24 were susceptible to penicillin as defined by the reference agar dilution method. The oxacillin screen detected as resistant all 28 resistant and

relatively resistant strains. The percentage of penicillin-resistant isolates detected by each minimum inhibitory concentration (MIC) test method was as follows: E-test (100%), Micro Media (75%), MicroScan (0%), and HTM (0%). With the relatively resistant isolates, the detection percentage was as follows: E-test (88%), Micro Media (94%), MicroScan (69%), and HTM (69%). In conclusion, the E-test and Micro Media MIC tests are acceptable confirmatory tests for detecting penicillin resistance among S. pneumoniae isolates.

INTRODUCTION

mococcal isolates is facililated by the use of the diskdiffusion screening test with 1-~g oxacillin disks. Confirmation of a resistant isolate is desirable, however, because the predictive value of a resistant isolate by the disk-diffusion method has been reported to be low (Doern and Jones, 1991). Furthermore, the screening test does not distinguish between relatively resistant and resistant strains. A number of antibiotic susceptibility methods for quantitative detection of relatively resistant and penicillin-resistant pneumococci have been described. These tests include macrodilution, microdilution, agar dilution, and more recently the E-test (Carroll et al., 1991; D'Amato et al., 1987; Jacobs et al., 1992; Jorgensen et al., 1992; Marshall et al., 1993; Mason et al., 1992; Spangler et al., 1992). Due to the fastidious growth requirements of the pneumococci, most studies have used Mueller-Hinton agar/broth supplemented with either sheep blood or lysed horse blood; however, Haemophilus test medium (HTM) has recently been reported to be an acceptable alternative growth medium for microdilution testing with these organisms (Jorgensen et al., 1992). In this study, we compared the ability of six different susceptibility methods to detect the presence of relatively resistant and penicillin-resistant pneumococci. Included among these methods are two commercial test systems (Mi-

The recovery of Streptococcus pneumoniae with relative resistance [minimum inhibitory concentration (MIC) 0.12-1.0 p,g/ml] and complete resistance to penicillin (MIC ~>2 ~g/ml) has been well documented in many studies and recently reviewed (Klugman, 1990). Penicillin therapy may be sufficient for treating bacteremic or nonbacteremic pneumonia caused by S. pneumoniae relative resistance to penicillin; however, isolates that exhibit complete penicillin resistance require alternative antimicrobial agent therapy such as with cefotaxime, ceftriaxone, and/or vancomycin (Klugman, 1990). Laboratory detection of penicillin-resistant pneuFrom the Department of Microbiology (R.B.C.), American Medical Laboratories, Chantilly, Virginia; and the Department of Pathology (O.G.), Episcopal Hospital; and the Department of Laboratories (J.E.M.), St. Christopher's Hospital, Philadelphia, Pennsylvania, USA. Address reprint requests to Dr. R.B. Clark, Department of Microbiology, American Medical Laboratories, Inc., 14225 Newbrook Drive, PO Box 10841, Chanfilly, VA 22021-0841, USA. Received 26 May 1992; revised and accepted 18 August 1993. © 1993 Elsevier Science Publishing Co., Inc.

655 Avenue of the Americas, New York, NY 10010. 0732-8893/93/$6.00

214

cro Media and MicroScan) that are reported to have difficulty in identifying these resistant strains (Carrol et al., 1991; Shanholtzer and Peterson, 1986). MATERIALS A N D M E T H O D S Organisms A total of 52 isolates of S. pneumoniae were used for this study. Isolates were recovered from a variety of clinical sites, including blood, cerebrospinal fluid, and respiratory tract. The organisms included 12 penicillin-resistant (MICs ~ 2 ~g/ml), 16 relatively resistant (MICs 0.125-1.0 ~g/ml), and 24 susceptible (MICs ~ 0.06 ~g/ml) isolates as determined by the reference agar dilution method (see below). Included among these isolates was S. pneumoniae ATCC 49619, which shows relative resistance to penicillin (Jorgensen et al., 1992). The remaining penicillinresistant and relatively resistant isolates were recovered from various North American geographic areas: California, Pennsylvania, Virginia, and Utah in the United States and Saskatchewan in Canada. The penicillin-susceptible isolates were recovered from Pennsylvania. All isolates were identified by colony morphology, optochin susceptibility, and bile solubility testing (Facklam and Washington, 1991). Isolates were subcultured twice on blood agar before testing. Pneumococcal test isolates were grown overnight on blood agar before the below test systems were inoculated. All suspensions prepared for inoculation of the various test systems were checked spectrophotometrically to ensure that they were equivalent to a 0.5 McFarland standard. Staphylococcus aureus ATCC 29213 and Enterococcus faecalis ATCC 29212 were used as control organisms as outlined below.

R.B. Clark et al.

McFarland 0.5 standard. Mueller-Hinton agar plates supplemented with 5% sheep blood were inoculated with the test pneumococcal suspension and allowed to air dry for 15 min. E-test strips (AB Biodisk, Culver City, CA, USA) containing penicillin (MIC range, 0.016-256 ~g/ml in incremental concentrations, . . . 1, 1.5, 2.0, 3.0 . . . . ) and oxacillin disks (1 ~g) were placed onto the inoculated plates and incubated overnight at 35°C in amibient air. E-test MICs were read from the intersection of the inhibition ellipse with the test strip. For this study, pneumococcal isolates with an E-test MIC of ~ 1 ~g/ml were considered resistant to penicillin. The oxacillin screen test was interpreted according to the National Committee for Clinical Laboratory Standards (NCCLS, 1991). Pneumococcal isolates with oxacillin zone sizes 1-20 mm were judged to be susceptible to penicillin, whereas isolates with oxacillin zone sizes of 419 mm were considered resistant to penicillin. Microscan Testing Pneumococci were tested for penicillin susceptibility according to the instructions of the manufacturer. Test isolates were emulsified in MicroScan Inoculum Water to match the turbidity of a 0.5 McFarland standard; 0.1 ml of the Inoculum Water suspension was added to 10 ml of Microscan Haemophilus Broth (HB). Then, 0.1 ml of the HB suspension was transferred into the penicillin-containing wells and the growth control well on MicroScan Pos MIC type-6 panels. The final organism concentration was - 1 x 105 CFU/ml. The plates were incubated overnight at 35°C in ambient air and the penicillin MIC was defined as the lowest concentration that inhibited complete growth of the test isolate. Micro Media FOX Testing

Agar D i l u t i o n Penicillin MICs were determined by a previously described agar dilution method for pneumococci (Spangler et al., 1992). Briefly, the test pneumococci were suspended in Mueller-Hinton broth to a turbidity equivalent to a 0.5 McFarland standard. The organisms were spot inoculated [104-colony-forming units (CFU)] with a Steer's replicator onto MuellerHinton agar (BBL, Cockeysville, MD, USA) supplemented with 5% sheep blood. Plates were incubated overnight at 35°C in ambient air. The MIC was defined as the lowest penicillin concentration that inhibited growth of the test isolate. One colony or a fine haze of growth was disregarded.

Micro Media FOX Fastidious MIC panels (Micro Media Systems, Cleveland, OH, USA) were inoculated according to the procedure outlined in the package insert. The final inoculum (2-3 x 10~ CFU/ml) was achieved by preparing a test organism suspension equivalent to a 0.5 McFarland standard in sterile water containing 0.02% Tween 80. The standardized suspension (0.5 ml) was added to 13 ml of sterile water containing 0.02% Tween 80, which was used to inoculate the Micro Media panels by using the pinned Inoculator/Seed Trough. The plates were incubated overnight before MICs were determined. The MIC was defined as the lowest penicillin concentration that inhibited growth of the test isolate.

E-Test and Oxacillin Screen

Haemophilus Test

The test pneumococci were inoculated into MuellerHinton broth to a suspension equivalent with a

MICs were determined by using broth microdilution panels prepared with HTM broth (BBL). Overnight

M e d i u m Microdilution

Susceptibility of Streptococcus pneumoniae

TABLE 1

215

Agreement to Category with Reference Agar Dilution and Other Test Methods for Penicillin

TABLE 2

Geometric Mean Penicillin Minimum Inhibitory Concentrations by Test Method of the Pneumococcal Isolates Category of Penicillin Resistance

Category of Penicillin Resistance Method

Resistant

Relatively Resistant

Susceptible

Method

Resistant

Relatively Resistant

Agar dilution E-test Micro Mediaa MicroScanb HTMc Disk diffusiona

12 14 10 0 0 29

16 14 20 23 19 --

24 24 21 28 29 23

Agar dilution E-test Micro Media MicroScan HTMa

2.2 3.3 1.7 0.44 0.44

0.47 0.58 0.61 0.15 0.18

aOne isolate did not grow in the Micro Media system. bOne isolate did not grow in the MicroScansystem. ~Fourisolates did not grow in Haemophilustest mediumbroth. aNo relative-resistancecategoryavailable. growth of the test isolates was suspended in 0.9% NaC1 to a turbidity equivalent to a 0.5 McFarland standard. Test panels were inoculated with 0.05 ml of a suspension of organism to produce a final concentration of - 5 x 105 CFU/ml. The panels were incubated overnight at 35°C in ambient air and were examined microscopically for evidence of growth. An MIC was defined as the lowest concentration of penicillin that inhibited growth of the test isolate. Statistical M e t h o d s The chi-square and Student's t-test were used to analyze statistical differences (P < 0.05) between results generated by agar dilution and the other test systems.

RESULTS The agreement between the reference agar dilution method and other test systems by category of penicillin resistance is outlined in Table 1. Results generated by the E-test, Micro Media, and disk-diffusion methods agreed most closely with agar dilution. All 12 penicillin-resistant pneumococci were detected by the E-test. In addition, two strains showing relative resistance by agar dilution (both MICs of 1.0 ~g/ml) were penicillin resistant by E-test methodology (MICs of 1.5 and 2.0 p,g/ml). Micro Media detected 10 of 12 penicillin resistant isolates (P --- 0.46). The two resistant isolates not detected by Micro Media (MICs of 1.0 and 0.5 p~g/ml) showed MICs of 2.0 ~g/ml by agar dilution. Furthermore, one strain relatively resistant by agar dilution (MIC 1.0 ~g/ml) was resistant by Micro Media testing (2.0 ~g/ml). In addition, two strains susceptible by both agar dilution and E-test methods were

Susceptible 0.021 0.026 0.061 ~<0.015 ~<0.007

aHaemophilustest medium. relatively resistant by Micro Media (both MICs 0.12 p~g/ml). All penicillin-resistant and relatively resistant pneumococcal isolates were detected by the diskdiffusion method. One additional isolate was peniciUin resistant by the oxaciUin screen (zone size, 17-18 ram); however, the agar dilution, E-test, and Micro Media MICs were all within the susceptible range for this isolate. MICs generated by both the MicroScan and HTM microdilution methods were generally not in agreement with agar dilution MICs. Neither microdilution method detected the 12 penicillin-resistant pneumococcal isolates (P < 0.00001), and all MICs by both methods were in the relative resistance range. In addition, of the 16 relatively resistant isolates, only 11 isolates by MicroScan and HTM methods were detected as such (P = 0.052). The remaining five relatively resistant isolates were classified as susceptible by both methods. Penicillin MICs for the proposed control strain S. pneumoniae ATCC 49619 by agar dilution, E-test, and Micro Media method were all in the recommended range (0.125-0.5 ~,g/ml) as suggested by Jorgensen et al. (1992). In contrast, penicillin MICs by MicroScan and the HTM methods were in the susceptible range (0.03-0.06 ~g/ml). Penicillin MICs for the control strains S. aureus ATCC 29231 and E. faecalis ATCC 29212 were in the NCCLS (1991) recommended ranges for all tests. The geometric mean penicillin MICs by category of penicillin resistance is shown in Table 2. The geometric mean MICs for each resistance category are substantially lower with the MicroScan (penicillinresistant isolates, P < 0.001; and penicillin relatively resistant isolates, P < 0.01) and HTM methods (penicillin-resistant isolates, P < 0.001; and penicillin relatively resistant isolates, P < 0.01). Table 3 shows the comparison of penicillin MICs for the 28 penicillin-resistant and relatively resistant isolates. The penicillin-susceptible data are not in-

216

TABLE 3

R.B. Clark et al.

Comparison of Penicillin Minimum Inhibitory Concentrations (MICs) for 28 Resistant and Relatively Resistant Pneumococcal Strains Number of MICs Within the Indicated log2 Concentration of Agar Dilution

Method

4-3

E-test Micro Media MicroScan 6 HTMa 10

- 2 - 1 -0.5 0 +0.5 +1 +2 ~ + 3 3 1 16 10

4 4 4

12 18 2 1

5

8 5 2

"Haemophilustest medium.

cluded because most of the MicroScan and HTM MICs were off scale (40.015 and 40.007 ~g/ml, respectively). E-test MICs were within one doubling dilution of the reference agar dilution MICs for all 28 resistant and relatively resistant isolates; 27 of 28 Micro Media MICs were within one doubling dilution of agar dilution (P = 1). However, less agreement was observed with the MicroScan and HTM methods. MicroScan MICs were within one doubling dilution of reference MICs for only six isolates (21%) (P < 0.00001), within two doubling dilutions for 22 (79%) isolates, and within three doubling dilutions for all 28 resistant and relatively resistant isolates. For the HTM method, MICs were within one doubling dilution of reference MICs for only seven isolates (26%) (P < 0.00001), within two doubling dilutions for 18 isolates (67%), and within three doubling dilutions for 24 isolates (89%).

DISCUSSION The accurate detection of penicillin-resistant S. pneumoniae is of paramount importance in the clinical microbiology laboratory. In this study, we compared six different test methods for their ability to detect the presence of these resistant isolates. Our results indicate that both the E-test and the Micro Media commercial microdilution system are acceptable confirmation methods for detecting penicillin-resistant pneumococci. At this time, we cannot recommend the use of the MicroScan conventional microdilution panels or commercial HTM broth prepared in microdilution panels to detect these penicillin-resistant isolates. Others have also reported a good correlation between E-test MICs and reference agar dilution MICs when assessing pneumococci susceptibility to penicillin (Jacobs et al., 1992). The E-test is simple to perform and the MICs are relatively easy to read.

Furthermore, one may run the E-test and oxacillin screen simultaneously on the same agar plate. One recent publication reported that the Micro Media microdilution system failed to detect penicillin-resistant pneumococci accurately (Carroll et al., 1991). In this earlier study, all 11 relatively resistant pneumococci were falsely susceptible by the Micro Media method, and one resistant strain was judged to be relatively resistant. In our hands, however, the Micro Media commercial panels performed extremely well in detecting the penicillin-resistant pneumococci. The reason(s) for the conflicting results in both studies is unclear; however, strain differences and possible lot-to-lot variations in the preparation of Micro Media plates may help explain the discrepancies in our studies. MicroScan commercial MIC plates have also been previously reported to be unreliable for detecting penicillin-resistant S. pneumoniae (Shanholtzer and Peterson, 1986). In this earlier study, two relatively resistant and three resistant isolates were falsely susceptible by MicroScan testing. We agree with these authors, at this time, that the MicroScan conventional MIC panels should not be utilized for susceptibility testing with pneumococci. The Haemophilus inoculum broth provided by the manufacturer may not support adequate growth of the pneumococci since one isolate in the previous study and one of our strains did not grow in the MicroScan panel wells. Furthermore, the proposed control strain S. pneumoniae ATCC 49619 showed false susceptibility to penicillin when tested by the MicroScan method. In our experience, as recently described by others (Marshall et al., 1993), HTM broth was unreliable for detecting penicillin-resistant pneumococci. Four pneumococcal isolates (8%) in our study did not grow in the commercial HTM broth and penicillin MICs (including the candidate control strain S. pneumoniae ATCC 49619) were well below the reference agar dilution MICs. The lower penicillin MICs may indicate suboptimal growth of the pneumococci in the commercial HTM broth. Different preparations of HTM broth may be important since a previous report (Jorgensen et al., 1992) showed that 97.6% of penicillin MICs to the test pneumococci were within 1 log2 dilution of the reference microdilution method. However, others (Mendelman et al., 1990) studying Haemophilus influenzae have also reported problems with HTM broth due to strain-dependent growth variability and unreliable detection of ampicillinresistant, nonqB-lactamase-producing H. influenzae. In conclusion, we recommend that the E-test or Micro Media commercial MIC plates be utilized to confirm penicillin resistance among pneumococci resistant by the oxacillin screen test. The E-test

Susceptibility of Streptococcus pneumoniae

could even be u s e d routinely instead of/or with the oxacillin screen test to detect p n e u m o c o c c i with decreased susceptibility to penicillin.

217

The authors thank the following individuals for supplying some of the isolates used in this study: E. Chan, P. Coudron, D. Halstead, and K. McGowan.

REFERENCES Carrol K, Cohen S, Saxon B, Hamilton L (1991) Failure of the Micro Media FOX microdilution system to detect decreased susceptibility of Streptococcus pneumoniae to penicillin. ] Clin Microbiol 29:2651-2652. D'Amato RF, Swenson JA, McKinley GA, Hochstein L, Wallman AA, Cleri DJ, Mastellone AJ, Fredericks L, Gonzalez L, Pincus DH, Thornsberry C (1987) Quantitative antimicrobial susceptibility test for Streptococcus pneumoniae using inoculum supplemented with whole defibrinated sheep blood. J Clin Microbio125:1753-1756. Doern GV, Jones RN (1991) Antimicrobial susceptibility tests: fastidious and unusual bacteria. In Manual of Clinical Microbiology. Eds, A Balows, WJ Hausler Jr, KL Herrmann, HD Isenberg, and HJ Shadomy. Washington, DC: American Society for Microbiology, pp 11261132. Facklam RR, Washington JA Jr (1991) Streptococcus and related catalase-negative Gram-positive cocci. In Manual of Clinical Microbiology. Eds, A Balows, WJ Hausler Jr, KL Herrmann, HD Isenberg, and HJ Shadomy. Washington, DC: American Society for Microbiology, pp 238-257. Jacobs MR, Bajaksouzian S, Appelbaum PC, Bolmstrom A (1992) Evaluation of the E-test for susceptibility testing of pneumococci. Diagn Microbiol Infect Dis 15:473478. Jorgensen JH, Doern GV, Ferraro MJ, Knapp CC, Swenson JM, Washington JA (1992) Multicenter evaluation of the use of Haemophilus test medium for broth microdilution antimicrobial susceptibility testing of Streptococcus pneu-

moniae and development of quality control limits. J Clin Microbiol 30:961-966. Klugman KP (1990) Pneumococcal resistance to antibiotics. Clin Microbiol Rev 3:171-196. Marshall KJ, Musher DM, Watson D, Mason EO (1993) Testing of Streptococcus pneumoniae for resistance to penicillin. J Clin Microbiol 31:1246-1250. Mason EO, Kaplan SL, Lamberth LB, Tillman J (1992) Increased rate of penicillin-resistant Streptococcus pneumoniae in a children's hospital and in vitro susceptibility to antibiotics of potential therapeutic use. Antimicrob Agents Chemother 36:1703-1707. Mendelman PM, Wiley EA, Stull TL, Clausen C, Chaffin DO, ONay O (1990) Problems with current recommendations for susceptibility testing of Haemophilus influenzae. Antimicrob Agents Chemother 34:1480-1484. National Committee for Clinical Laboratory Standards (NCCLS) (1991) Third informational supplement M100$3: performance standards for antimicrobial susceptibility testing. Villanova, PA: NCCLS. Shanholtzer CJ, Peterson LR (1986) False susceptible penicillin G MICs for Streptococcus pneumoniae with a commercial microdilution system. Am J Clin Pathol 85:626629. Spangler SK, Jacobs MR, Appelbaum PC (1992) Susceptibilities of penicillin-susceptible and -resistant strains of Streptococcus pneumoniae to RP 59500, vancomycin, erythromycin, PD 131628, sparfloxacin, temafloxacin, Win 57273, ofloxacin, and ciprofloxacin. AntimicrobAgents Chemother 36:856-859.