INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

With antibiotic resistance increasing among many pathogens, accurate antimicrobial susceptibility testing results are vital for both patient care and public health surveillance. In order to monitor drug resistance, clinicians and public health officials depend on accurate performance and appropriate reporting of the results of these susceptibility tests. Disk diffusion (Kirby-Bauer) (3) is one of the most commonly used antimicrobial susceptibility testing (AST) methods among diagnostic laboratories (22). This method is a well-established procedure for which there are accepted standards including those endorsed by the National Committee for Clinical Laboratory Standards (NCCLS) (26). Methods and interpretive criteria for the disk diffusion methodology are published by NCCLS in Performance Standards for Antimicrobial Disk Susceptibility Tests and are revised and updated as needed. These documents include specific guidelines (i.e., medium recommendations, changes in inoculum preparation, incubation conditions, and interpretive criteria) to be followed for organisms designated problem or fastidious organisms by NCCLS (26, 27).

The New York State (NYS) Clinical Laboratory Evaluation Program's Section of Bacteriology is responsible for ensuring the quality of AST and other clinical bacteriologic testing in laboratories throughout the state (~90%) as well as out-of-state laboratories testing specimens originating from NYS patients (~10%). Permitted laboratories are required to meet minimum standards and successfully participate in the NYS Department of Health's proficiency testing program.

Information regarding the appropriateness of the disk diffusion methods used in clinical laboratories is difficult to obtain. Most studies focus on a particular organism or a single step in a method. This study was designed to examine practices in NYS permitted laboratories with regard to disk diffusion testing. The primary objective was to evaluate the level of compliance with NCCLS guidelines for the disk diffusion methodology when testing two organisms designated problem pathogens (Staphylococcus aureus and Enterococcus species) and three organisms designated fastidious pathogens (Streptococcus pneumoniae, Haemophilus influenzae, and Neisseria gonorrhoeae) by NCCLS. We also examined the interpretation and follow-up of screening tests for penicillin resistance in S. pneumoniae and vancomycin resistance in Enterococcus species among NYS permitted laboratories.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Questionnaire. In January 1997, questionnaires were sent to 320 laboratories participating in NYS's proficiency testing program for general bacteriology as part of a routine proficiency testing event. Laboratories in this comprehensive subcategory are approved to perform antimicrobial susceptibility testing with all organisms. The purpose of this survey was to draw attention to the modified methods required for AST of problem and fastidious organisms and to identify topics in which specific continuing education might be beneficial to participating laboratories. The cover letter clearly stated that responses would not affect the laboratories' proficiency testing score. The questionnaire focused on five organisms for which NCCLS guidelines recommend modified methods for AST: S. aureus, Enterococcus species, S. pneumoniae, H. influenzae, and N. gonorrhoeae.

Data entry and analysis. The responses on the questionnaires were entered into a database created with Filemaker Pro, version 3.0, software (Claris Corporation, Santa Clara, Calif.) and were independently verified. Responses for questions relevant to disk diffusion were tabulated with Excel, version 4.0, software (Microsoft Corporation, Redmond, Wash.). For each question and organism, results were tabulated only for participants responding to the specific question resulting in a different total number (n) of responses for each question and each organism analyzed. We did not include missing responses in the analysis of the survey results.

Interpretation. For each organism, compliance was determined as the percentage of respondents whose practices agreed with 1993 (27) or 1997 (26) NCCLS guidelines for each step of the disk diffusion test. Either set of guidelines was considered acceptable, as the 1997 guidelines were released just prior to mailing of the survey. An outline of the responses that we considered in compliance with NCCLS guidelines is shown in Table 1. For temperature of incubation for S. aureus we included choices of 30 to 35°C in room air, 35°C in room air, and 30 to 35°C in 5% CO₂. The 1993 (27) and 1997 (26) NCCLS guidelines suggest 30 to 35°C in room air; however, because of the overlap between the choices, we accepted both 30 to 35°C in room air and 35°C in room air as being in compliance with NCCLS guidelines. For length of incubation, the response choices in our survey included 16 to 18, 20 to 24, and 24 h. Due to the overlap between the choices of 20 to 24 and 24 h, we considered both answers as being in compliance for the survey. We defined overall compliance as compliance with five key steps (inoculum preparation, choice of medium, number of disks per plate, incubation conditions, and length of incubation) for S. aureus and Enterococcus species or as compliance with the first four steps for the fastidious organisms (S. pneumoniae, H. influenzae, and N. gonorrhoeae).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Overview. Two hundred eighty-eight of 320 laboratories (90%) returned a completed survey. The disk diffusion method was used by 58.0, 58.7, 93.0, 68.6, and 37.4% of the laboratories responding to the question regarding the method for testing of S. aureus, Enterococcus species, S. pneumoniae, H. influenzae, and N. gonorrhoeae, respectively (Table 2).

S. aureus. Of the 166 laboratories that reported use of disk diffusion testing for S. aureus, 85.0% reported testing oxacillin disks rather than methicillin (10.2%) or nafcillin (0.6%) disks. Table 3 shows that many laboratories (50.6%) reported that their protocol was in compliance with NCCLS guidelines for five key steps (preparation of inoculum, choice of medium, number of disks per plate, incubation conditions, and length of incubation) of the disk diffusion procedure. The two most common areas of noncompliance were the use of the growth method to prepare the inoculum rather than the direct suspension method recommended by NCCLS and the failure to incubate the test for a full 24 h to accurately detect methicillin or oxacillin resistance (Table 3). Most (89.0%) laboratories complied with NCCLS protocols by using the strain (ATCC 25923) recommended for quality control testing of the disk diffusion procedure, although some (18.9%) laboratories reported using the strain (ATCC 29213) recommended for MIC testing (Table 3).

Enterococcus species. Table 3 shows that 47.7% of the laboratories reported compliance with the five key steps evaluated (preparation of inoculum, choice of medium, number of disks per plate, incubation conditions, and length of incubation). The most common deviation was the use of media other than unsupplemented Mueller-Hinton agar (Table 3). Laboratories often reported using sheep blood-supplemented Mueller-Hinton agar (31.6%) or Trypticase soy agar (2.3%). Laboratories (16.2%) also reported using a strain for quality control (ATCC 51299) that is not designated for quality control of disk diffusion testing (Table 3).

S. pneumoniae. Overall, 41.6% of respondents (Table 3) reported following NCCLS guidelines for the four key steps analyzed for this organism (preparation of inoculum, choice of medium, number of disks per plate, and incubation conditions). Common departures from the NCCLS guidelines (Table 3) included placement of too many disks for the size of plate chosen (40.8%) and incubation in room air rather than 5% CO₂ (22.6%). NCCLS provides quality control zone diameter limits for only one strain of S. pneumoniae, ATCC 49619; however, only 68.6% of laboratories reported using this strain for quality control (Table 3). We found that 70.9% of laboratories reported that they interpreted zone sizes for S. pneumoniae using the table specifically designated for S. pneumoniae (Table 3). The 1993 NCCLS guidelines (27) differed from the 1997 NCCLS guidelines (26) by changing the age of inoculum from 18 to 20 h to 16 to 18 h. If the results of the survey are analyzed by the guidelines reportedly in use in the participants' laboratories, then 37.9% of laboratories were in compliance; if either answer was considered in compliance, then 52.4% of the laboratories were in compliance.

H. influenzae. Overall compliance with the four key steps (preparation of inoculum, choice of medium, number of disks per plate, and incubation conditions) was 30.1% (Table 3). Although compliance with NCCLS guidelines was evident as 65.2% chose Haemophilus test medium, others chose either chocolate or Mueller-Hinton agar. Many (53.6%) reported placing no more than four disks per 100-mm plate or nine disks per 150-mm plate, and 87.6% correctly chose to incubate their plates in CO₂ (Table 3). Similar to the results for S. pneumoniae, only 68.4% of laboratories reported using the table specifically designated for interpretation of the results for H. influenzae (Table 3).

N. gonorrhoeae. Only 96 of 257 (37.4%) laboratories reported performing disk diffusion testing with this organism. Many (61.9%) use only beta-lactamase tests, which do not detect chromosomally mediated resistance or resistance to other classes of antimicrobial agents. Table 3 shows that 14.3% of the laboratories reporting disk diffusion susceptibility testing for N. gonorrhoeae were in compliance with NCCLS guidelines for four key steps (preparation of inoculum, choice of medium, number of disks per plate, and incubation conditions). Many laboratories used chocolate agar (44.6%) or Mueller-Hinton chocolate agar (19.6%) rather than the recommended GC agar containing 1% growth supplement (29.3%). Many (57.8%) reported using no more than the recommended number of disks per plate, and only 69.7% of laboratories used strain ATCC 49226 for antimicrobial susceptibility testing of N. gonorrhoeae (Table 3). Also of concern, as shown in Table 3, only 58.2% of laboratories reported choosing the specific table for interpretation of the zone sizes obtained when testing N. gonorrhoeae.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This survey was designed to assess laboratory compliance with NCCLS guidelines for AST by the disk diffusion method for problem and fastidious organisms and to determine areas where further education was needed. The results presented here are based on the results specifically reported by NYS permitted laboratories; however, it is likely that a similar lack of compliance occurs in clinical laboratories in other locales as well. This is supported by the fact that the specific results found in this study are similar to those previously reported by the American College of Pathologists and other investigators (17-20).

S. aureus. Temperature and length of incubation are critically important in the detection of methicillin-resistant S. aureus (4, 6), and minor changes in these protocols may have a significant impact on susceptibility testing results. We found that 50.6% of our laboratories were in compliance with all five steps, which is comparable to other similar studies (20, 29). Some (18.9%) of the reporting laboratories chose to prepare their inocula by the growth method rather than the recommended direct suspension method. This is comparable to the rate reported by Pfaller et al. (29). In addition, laboratories also reported incubation times of less than the critical 24 h required for optimum detection of methicillin (4, 21, 25) and vancomycin (34) resistance. Pfaller et al. (29) also reported that laboratories did not incubate plates for the recommended full 24 h.

Enterococcus species. Because multiple mechanisms of resistance in enterococci exist, it may be necessary to use more than one antimicrobial susceptibility testing method for accurate analysis (33). Overall, the laboratories in this study adhered to most of the steps in the procedure for Enterococcus sp. except that 31.6% of laboratories selected Mueller-Hinton agar supplemented with sheep blood as the test medium. Mueller-Hinton agar containing 5% sheep blood was considered acceptable in 1979 and 1983, but NCCLS subsequently recommended Mueller-Hinton agar without supplements when studies demonstrated that the blood in the agar interacted with aminoglycoside disks to yield false-positive results for susceptibility (16). The continued use of this medium by some laboratories raises concerns that outdated practices still occur in clinical laboratories.

S. pneumoniae. The progressive increase in the United States of pneumococci resistant to penicillin (see reference 12 for a review) underscores a need for accurate susceptibility testing so that therapeutic and epidemiologic decisions are appropriate. The overall rate of compliance with the four key steps in the disk diffusion technique for AST of S. pneumoniae was 41.6%. A similar survey of testing for pneumococcal antibiotic resistance in Tennessee found that only 27% of hospitals properly tested isolates according to NCCLS guidelines (8). Although penicillin resistance was less common in 1987, Brummitt et al. (5) found a lower percentage (23%) of 111 laboratories surveyed statewide in Minnesota performed the recommended test with a 1-µg oxacillin disk, whereas 80.1% of the laboratories in this study performed the test.

H. influenzae. There have been numerous revisions of NCCLS guidelines with respect to application of the disk diffusion method for H. influenzae (see reference 9 for a review). Recent NCCLS guidelines recommend that inoculum suspensions be made from colony growth 20 to 24 h old because of the importance of an accurate inoculum concentration (26, 27). Care must be taken in preparing the inoculum of this organism for disk diffusion testing because it may demonstrate an "inoculum effect" in that inaccurate zone size determinations result if the number of CFU exceeds that recommended. Overall compliance with NCCLS guidelines for the four key steps noted was 30.1% for this organism, which is similar to that reported by Scriver et al. (31). One-quarter of laboratories described preparation of the inoculum from a source with 16 to 18 h of growth, which is younger than recommended. The majority of respondents (87.0%) to this survey correctly chose the direct colony method to prepare the inoculum for this organism.

N. gonorrhoeae. Our survey found that antimicrobial susceptibility testing of N. gonorrhoeae was usually limited to checking of isolates for the presence of a beta-lactamase enzyme. While 37.4% of the respondents in the current study used the disk diffusion method for in vitro susceptibility testing, the remainder performed only a beta-lactamase test. A positive beta-lactamase test result predicts resistance to penicillin, ampicillin, and amoxicillin by plasmid-mediated mechanisms but fails to detect chromosomally mediated resistance to penicillin and to select cephalosporins (see reference 13 for a review). Altaie et al. (2) found that of 51 penicillin-resistant isolates detected by the disk diffusion method, 17 (33%) were beta-lactamase negative and therefore would have been missed by beta-lactamase testing alone (2). Fox et al. (14) showed a lower but increasing level of chromosomally mediated resistance to penicillin in the United States between 1988 and 1994. In addition, beta-lactamase testing would not detect resistance to non-beta-lactam classes of antimicrobial agents such as the fluoroquinolones or the macrolides to which resistance may be emerging (14, 24).

Follow-up. As our study showed that the rate of compliance with NCCLS guidelines for susceptibility testing of these problem and fastidious organisms was suboptimal in some categories, we implemented several different follow-up approaches to improve susceptibility testing of these organisms. After the survey was evaluated, written results from the questionnaire and abbreviated guidelines outlining the correct methodology for disk diffusion testing of each organism were distributed to all participants. Likewise, on three occasions after this survey was taken, educational information was provided to laboratories regarding AST of problem and fastidious organisms along with proficiency testing materials. In addition, specific guidelines regarding the susceptibility testing of problem and fastidious organisms were incorporated into the Laboratory Standards for New York State Department of Health in September 1998 (28a). Lastly, the importance of reporting of antimicrobial resistance for several of these pathogens (S. pneumoniae, H. influenzae, N. gonorrhoeae, as well as vancomycin-resistant S. aureus) was emphasized in an educational update on Laboratory Reporting of Communicable Diseases 1999 (28b) provided by the NYS Department of Health to all permitted laboratories. Although we believe that it is too early to assess the impact of these approaches, we plan to continue our emphasis on the need to use accurate and reliable susceptibility testing methods.