Penicillinase-Resistant Penicillins and Cephalosporins

JOURNAL OF CLINICAL MICROBIOLOGY, May 1986, p. 832-839 0095-1137/86/050832-08$02.00/0 Copyright C 1986, American Society for Microbiology Vol. 23, No. 5 The Role of 13-Lactamase in Staphylococcal Resistance to Penicillinase-Resistant Penicillins and Cephalosporins LINDA K. McDOUGAL* AND CLYDE THORNSBERRY Antimicrobics and Infection Mechanisms Branch, Hospital Infections Program, Center for Infectious Diseases, Centers for Disease Control, Atlanta, Georgia 30333 Received 4 November 1985/Accepted 13 January 1986 We showed that most Staphylococcus aureus strains that have borderline or intermediate susceptibility to the penicillinase-resistant penicillins (PRPs) react this way because of the activity of their ji-lactamase on these antimicrobial agents. These strains produced large amounts of staphylococcal,i-lactamase that rapidly hydrolyzed penicillin and partially hydrolyzed the PRPs. Susceptibility to hydrolysis was penicillin > oxacillin > cephalothin > methicillin. The borderline results and the hydrolysis could be prevented by the j-lactamase inhibitors clavulanic acid and sulbactam. For intrinsically methicillin-resistant (heteroresistant) S. aureus, the inhibitors reduced the penicillin MICs, but the strains remained resistant to all the I-lactam antimicrobial agents, including penicillin. We conclude that the borderline in vitro susceptibility or resistance to PRPs in most of these S. aureus strains is mediated by j-lactamase and they are not heteroresistant or intrinsically resistant. We do not know whether this in vitro resistance is expressed clinically. We have received several problem isolates of Staphylococcus aureus from different areas of the United States that yielded borderline-susceptible (24-h incubation) or intermediate (48-h incubation) results for oxacillin and methicillin by the usual testing methods (25, 26). The MICs for these strains were 4,ug/ml (methicillin), 2,ug/ml (oxacillin), and 2,ug/ml (cephalothin) after 24 h of incubation at 35 C when tested by the broth microdilution method with cationsupplemented Mueller-Hinton broth (CSMHB) and 2% NaCl (36). After incubation for 48 h, the MICs generally increased 1 dilution. With disk diffusion tests using a 1-,ug oxacillin disk, zone sizes ranged from 6 to 13 mm. These isolates of borderline-susceptible S. aureus produced large amounts of staphylococcal P-lactamase as shown by their high penicillin MICs and immediate strong positive results when tested with nitrocefin. In previous studies, we showed that staphylococcal,-lactamase affects the zone sizes attained with the disk diffusion method, in that S. aureus strains that were P-lactamase positive yielded smaller zones of inhibition than strains that were P-lactamase negative but had the same MIC (24). These observations raised the question of whether these borderline-susceptible S. aureus isolates were truly heteroresistant (often called intrinsically resistant) or whether the apparent in vitro resistance in the MIC test was caused by P-lactamase. Previous reports have shown that staphylococcal,b-lactamase, when present in sufficient quantities for a sufficient duration, slowly hydrolyzes the penicillinase-resistant penicillins (PRPs) and cephalosporins (1, 4, 8, 9, 17-22, 28, 34). In view of this finding, we initiated studies to estimate the amount of P-lactamase produced by each S. aureus isolate by using penicillin MICs and nitrocefin test results. We also studied the effect of two P-lactamase inhibitors (clavulanic acid and sulbactam) on the MICs of penicillin, oxacillin, cephalothin, and methicillin and on the results obtained with oxacillin and methicillin screen plates (36) for these borderline-susceptible S. aureus isolates, as well as for other staphylococci. Finally, we studied the hydrolysis of,-lactam antimicrobial agents by * Corresponding author. 832 staphylococcal P-lactamase and its acid. inhibition by clavulanic MATERIALS AND METHODS Protocol. We grouped the S. aureus isolates by,-lactamase production and suceptibility to 1-lactam antimicrobial agents. To test the theory that staphylococcal P-lactamase was the cause of the decreased susceptibility of the borderlinesusceptible S. aureus to PRPs, we added clavulanic acid to microdilution plates containing,-lactam antimicrobial agents and tested two 3-lactamase-positive strains of S. aureus, a borderline-susceptible one with a high penicillin MIC and a methicillin-susceptible one with a low penicillin MIC. Next, we determined the optimal concentrations of clavulanic acid and sulbactam necessary to reduce penicillin MICs without affecting the MICs of,b-lactamase-negative methicillinresistant and methicillin-susceptible S. aureus. Then we tested penicillin, oxacillin, cephalothin, and methicillin with and without the selected concentrations of clavulanic acid and sulbactam with 66 strains of S. aureus. Methicillin and oxacillin screen plates (36), with and without clavulanic acid, were tested at the same time. Lastly, we determined the stability of penicillin, oxacillin, cephalothin, and methicillin to P-lactamase and the inhibition of staphylococcal P- lactamase by clavulanic acid. Bacteria. S. aureus strains had been isolated from infected patients and were selected from many strains sent to our laboratory for susceptibility studies. The basic groupings are shown in Table 1. The 66 S. aureus isolates were initially selected for their ability to produce staphylococcal P- lactamase and then further grouped on the basis of susceptibility to the P-lactam antimicrobial agents. The P- lactamase-positive methicillin-resistant S. aureus isolates were composed of two groups of organisms, depending on the difficulty with which heteroresistance (intrinsic resistance) could be detected. Three isolates were easily detected methicillin-resistant S. aureus with high penicillin MICs (.128,ug/ml) and were resistant to methicillin, oxacillin, and cephalothin. This resistance could be detected without the addition of NaCl to the medium, longer incubations, or higher inocula. Heteroresistance was more difficult to detect

TABLE 1. ROLE OF P-LACTAMASE IN STAPHYLOCOCCAL RESISTANCE VOL. 23, 1986 833 Distribution of 66 S. aureus strains by 13-lactamase production and susceptibility to,b-lactam antimicrobial agents Antimicrobial.-Lactamase positive MIC (jig/ml) for S. aureus strains P-Lactamase negative agent Methicillin Methicillin Borderline" Partial Methicillin Methicillin Methicillin resistanta resistantb -nd= 12n borderlined susceptible resistante susceptible (n = 3) (n = 14) (n - ) (n = 9) (n = 19) (n = 4) (n = 5) Penicillin -128 C32 2128 <128 C32 16 0.06-0.12 Methicillin >32 >32 4 4 C2 >32 C2 Oxacillin >16 >16 2 <1 S0.5 >16 C0.25 Cephalothin.32 16-.32 2 <1 S0.5.32 C0.25 a Easy to detect; did not require 2% NaCl to detect resistance. b Difficult to detect; required 2% NaCl to detect heteroresistance. c Borderline MICs of penicillinase-resistant,b-lactam antimicrobial agents: methicillin, 4 pg/ml; oxacillin, 2,ug/ml; cephalothin, 2,ug/ml; and penicillin, -128 pg/ml. d Partial borderline MICs of penicillinase-resistant,-lactam antimicrobial agents: methicillin, 4 pg/ml; oxacillin, -1 jig/ml; cephalothin, -1,ug/ml; and penicillin, ru28 2hg/ml. e Did not require 25% NaCl to detect heteroresistance. in 14 isolates that produced less P-lactamase, had lower penicillin MICs (.32,ug/ml), and required the addition of NaCl to demonstrate heteroresistance. Among the borderline-susceptible S. aureus, the MICs for 12 strains were.128,ug of penicillin per ml and the MICs for 9 strains were <128,ug of penicillin per ml. The borderline-susceptible S. aureus isolates were from several nosocomial outbreaks from the continental United States (California, Missouri, Florida, New York) and were isolated mainly in tertiary care centers. Of the 3-lactamase-negative isolates of S. aureus, four were methicillin resistant (detected without the addition of NaCl) and five were methicillin susceptible. S. aureus ATCC 25923 and ATCC 29213 were tested as standard control strains. Antimicrobial agents. Penicillin, methicillin, oxacillin, amoxicillin, cephalothin, gentamicin, vancomycin, rifampin, clavulanic acid, and sulbactam, deemed suitable for susceptibility testing, were obtained from the appropriate pharmaceutical companies. Amoxicillin-clavulanic acid disks were purchased from BBL Microbiology Systems. Nafcillin was excluded from these studies after the preliminary phase because it is not recommended for in vitro testing by the National Committee for Clinical Laboratory Standards (26). Susceptibility tests. MICs were determined by broth microdilution as described in the National Committee for Clinical Laboratory Standards dilution standard M7-T (26), except that 2% NaCl was added to CSMHB (Difco Laboratories) as described previously (36). The trays were incubated at 35 C, and the endpoints were read at 24 and 48 h. A direct inoculum was prepared after 24 h of growth on Trypticase soy agar with 5% sheep blood (BBL) (2). Organisms from the agar culture were added to Mueller-Hinton broth to yield a turbidity equal to a 0.5 McFarland standard, approximately 108 CFU/ml. This inoculum was diluted 1:10, and 5,ul was added to the MIC wells by using the Quick Spense inoculator (Sandy Spring Instrument Co. Inc., Ijamsville, Maryland). The final inoculum was approximately 5 x 105 CFU/ml (5 x 104 CFU per well). The size of this inoculum was confirmed periodically throughout the study by colony counts. In one instance, MICs were determined by the broth macrodilution procedure described by Washington and Sutter (38), except that CSMHB plus 2% NaCl was used. Screening tests. Methicillin (10,ug/ml) or oxacillin (6,ug/ml) and methicillin with clavulanic acid (2 and 4,ug/ml) or oxacillin with clavulanic acid (2 and 4 p.g/ml) were incorporated in molten Mueller-Hinton agar (BBL) (approximately 50 C) containing 4% NaCl (36). The plates were inoculated with a swab from a suspension containing 108 CFU/ml, an inoculum similar to that used for the disk diffusion test. The plates were incubated at 35 C for 24 h. If there was no growth, they were incubated another 24 h. Plates that produced growth at 48 h but not at 24 h were noted. Plates were stored at 4 C and used within 1 week after preparation. -Lactamase tests.,-lactamase tests were performed with nitrocefin (29) by two methods. In method 1, growth from blood agar plates was removed with a swab dipped in nitrocefin (500,ug/ml). In method 2, 50,ul of the nitrocefin solution was added to selected wells of broth microdilution plates (growth control, methicillin [0.5,ug/ml], clavulanic acid [4,ug/ml], and sulbactam [16,ug/ml]). Color change was graded from very faint to dark red (+ to 4+) and was observed at 10, 30, and 60 min. Color change (4+) that occurred in 10 min was considered an immediate positive test. Methicillin (0.5,ug/ml) was used to induce the production of,-lactamase in the S. aureus isolates. j-lactam antimicrobial inactivation test. A borderlinesusceptible S. aureus isolate was subcultured to CSMHB plus 2% NaCl with 1 p.g of methicillin per ml and incubated at 35 C with shaking overnight. Penicillin, oxacillin, and methicillin (each at a concentration of 20,ugIml) were added to equal volumes of the filtered supernatant. The test mixtures were assayed for residual antibacterial activity by a biological assay (35) at 0, 5, 10, 30, and 60 min. Clavulanic acid concentrations (wt/vol) equal to the concentrations of the,-lactam antimicrobial agents were mixed with the supernatant 5 min before the antimicrobial agent was added, and these test mixtures were also assayed for antibacterial activity at 30 and 60 min. Concentrations of,-lactam drugs remaining in the solution were calculated, with appropriate controls, from standard curves by using mean zone diameters calculated from triplicate determinations. RESULTS With the nitrocefin test for,b-lactamase production, all of the,-lactamase-positive S. aureus strains gave a positive result on the swabs from uninduced growth from the blood agar plate in 10 min. When the penicillin MIC was low (<16

834 McDOUGAL AND THORNSBERRY pug/ml), the color change was weak (2+ or less) at 10 min. Swabs from induced growth gave a strong (4+) positive test for all P-lactamase-positive strains. Because of the large amount of growth on the swab, this method was a very sensitive test for P-lactamase production. Although it was difficult to titrate the color change reaction by time (10 to 60 min) and strength (± to 4+ color change), it was possible when testing the smaller, more consistent amount of growth in the broth microdilution wells. When the penicillin MIC was.128,ug/ml and the organism had been induced by growth in 0.5,ug of methicillin per ml, the nitrocefin result was an immediate positive (4+) reaction. If the organism was not induced, the nitrocefin was positive but showed weaker color change (1+) at 30 to 60 min. With most strains with low penicillin MICs (c32,ug/ml), the methicillin-induced growth did not become positive for P-lactamase until incubated 30 to 60 min. Although induced with methicillin, six isolates were negative for P-lactamase at 60 min. Uninduced growth and growth from the sulbactam well gave a very weak (±) color change at 30 to 60 min or were negative at 60 min. When any S. aureus isolate was grown in wells containing 4,ug of clavulanic acid per ml, it was always negative for P-lactamase after 60 min of incubation. On the other hand, S. aureus isolates with penicillin MICs of -64,g/ml that were grown in wells containing 16,ug of sulbactam per ml gave a weak (±) color change at 10 min that became 4+ at 60 min. Uninduced growth from control wells did not give as strong a reaction. In the preliminary phase of this study, several concentrations of clavulanic acid (final concentrations, 0.25, 0.5, and 1.0 pug/ml) were added with the inoculum to microdilution plates containing penicillin, oxacillin, nafcillin, methicillin, cephalothin, gentamicin, vancomycin, or rifampin. With 1.0,ug of clavulanic acid per ml, MICs for a borderlinesusceptible S. aureus were reduced seven twofold dilutions for penicillin, three twofold dilutions for oxacillin, and one twofold dilution for methicillin and cephalothin; the MIC remained the same for nafcillin. The MICs for a,-lactamasepositive methicillin-susceptible S. aureus with a low penicillin MIC were reduced three twofold dilutions for penicillin and one twofold dilution for methicillin, oxacillin, and nafcillin; the MIC remained the same for cephalothin. Clavulanic acid did not affect the MICs of gentamicin, vancomycin, or rifampin. When oxacillin, with and without 1.0,ug of clavulanic acid per ml, was tested by broth macrodilution, the results were the same as those obtained by broth microdilution. In phase 2 of the study, we determined the optimal amounts of clavulanic acid and sulbactam necessary to reduce penicillin MICs without affecting the MICs for,blactamase-negative methicillin-resistant and methicillinsusceptible S. aureus. MIC plates with penicillin and clavulanic acid alone and in a checkerboard configuration were inoculated with 10,-lactamase-positive methicillinsusceptible S. aureus isolates (penicillin MICs ranging from 2.0 to >128,ug/ml), 8 P-lactamase-positive methicillinresistant S. aureus isolates (penicillin MICs from 16 to >128,ug/ml), 2,-lactamase-negative methicillin-resistant S. aureus isolates (penicillin MICs, 16.0,ug/ml), and 2 3- lactamase-negative methicillin-susceptible S. aureus isolates (penicillin MICs, 0.06,ug/ml). The same studies were also performed with penicillin and sulbactam. With the 1Blactamase-positive methicillin-susceptible S. aureus, the growth in the checkerboard area of the MIC plate resembled descending stair steps (decreasing penicillin MICs) as the J. CLIN. MICROBIOL. clavulanic acid (8 to 0.25,ug/ml) and sulbactam (16 to 0.25 jig/ml) concentrations increased. With the 3-lactamasepositive methicillin-resistant S. aureus, penicillin MICs were reduced by small amounts of clavulanic acid (0.5,ug/ml) and sulbactam (2.0,ug/ml), but remained resistant (penicillin MICs of 8 to 16,ug/ml) even with high concentrations of the inhibitors. The penicillin MICs for 13-lactamase-negative methicillin-resistant and methicillin-susceptible S. aureus were essentially not affected by clavulanic acid or sulbactam in that they had the same penicillin MIC (or one dilution lower with high concentrations of inhibitors) as they did in the absence of the inhibitors. Higher concentrations of sulbactam than clavulanic acid were needed to inhibit staphylococcal 1-lactamase. Clavulanic acid alone inhibited most methicillin-susceptible S. aureus at 32,ug/ml, an occasional strain at 16,ug/ml, and 1 strain at 8,ug/ml (1 of 66). The MICs of sulbactam alone for methicillin-susceptible S. aureus were 128 to 256,ug/ml. All methicillin-resistant S. aureus had clavulanic acid MICs of >64,ug/ml and sulbactam MICs of >256,ug/ml. Two concentrations of clavulanic acid (2 and 4,ug/ml) and sulbactam (8 and 16,ug/ml) were selected for further studies with penicillin, two PRPs (methicillin and oxacillin), and cephalothin. The data from this phase of the study are shown in Table 2 as geometric means. Clavulanic acid and sulbactam reduced the MICs for,b-lactamase-positive methicillinsusceptible S. aureus of penicillin, oxacillin, cephalothin, and methicillin as shown by the markedly lower geometric mean MICs. The number of twofold reductions of an MIC depended first on the antimicrobial agent (penicillin > oxacillin > cephalothin > methicillin) and second on the MIC for an organism of that antimicrobial agent. In general, for methicillin- and borderline-susceptible S. aureus, the largest reductions in MICs by clavulanic acid and sulbactam occurred among those organisms with the highest MICs of a,b-lactam antimicrobial agent. Penicillin MICs of -128 jig/ml for borderline-susceptible S. aureus were always reduced seven to eight twofold dilutions in the presence of clavulanic acid (4,ug/ml) and sulbactam (16,ug/ml). Organisms with borderline-susceptible MICs of oxacillin and methicillin had penicillin MICs of.128,ug/ml; the MICs of oxacillin, cephalothin, and methicillin were reduced more for borderline-susceptible than for other methicillin-susceptible S. aureus isolates in the presence of the,b-lactamase inhibitors. Oxacillin and cephalothin MICs for borderline-susceptible S. aureus were reduced twice as much as methicillin MICs in the presence of clavulanic acid and sulbactam. After 24 h of incubation, the MICs of oxacillin, cephalothin, methicillin, and penicillin were only slighty lower with the higher concentration of clavulanic acid and sulbactam. After 48 h of incubation, however, the MICs of all four antimicrobial agents were lower with the higher concentration of clavulanic acid (4,ug/ml). There was little difference between clavulanic acid (4,ug/ml) and sulbactam (16,ug/ml) in their effectiveness in reducing the MICs of 1-lactam antimicrobial agents.,b-lactamase-positive methicillin-resistant S. aureus isolates were grouped according to the degree of difficulty in detecting heteroresistance as described in Materials and Methods. The easy-to-detect strains produced large amounts of,b-lactamase, and the penicillin MICs were.128,jug/ml. To detect heteroresistance in the difficult-to-detect group of methicillin-resistant S. aureus, we added 2% NaCl to CSMHB, especially in tests with methicillin and cephalothin. These strains produced significantly less,b-lactamase

ROLE OF P-LACTAMASE IN STAPHYLOCOCCAL RESISTANCE VOL. 23, 1986 835 TABLE 2. Geometric mean MICs for penicillin, oxacillin, cephalothin, and methicillin alone and with clavulanic acid (2 and 4,ug/ml) and sulbactam (8 and 16,ug/ml) for 66 strains of S. aureusa Geometric mean MIC (;g/ml) Antimicrobial 13-Lactamase positivec 3-Lactamase negativec agent (qlg/ml)b Methicillin Methicillin Borderline Partial Methicillin Methicillin Methicillin resistant resistant or erme borderline susceptible resistant susceptible (n = 3) (n = 14) (n12) (n = 9) (n = 19) (n = 4) (n = 5) P -128 21.3.128.0 78.3 24.70 16 0.08 P + CA (2) 20 11.1 2.0 1.2 0.31 16 0.07 P + CA (4) 16 10.6 1.0 0.71 0.16 16 0.05 P + SB (8) 16 11.1 1.8 1.5 0.45 16 0.07 P + SB (16) 8 8.0 0.96 0.86 0.30 16 0.06 Ox >16 >16 2.0 0.86 0.52 >16 0.25 Ox + CA (2) >16 >16 0.27 0.28 0.19 >16 0.19 Ox + CA (4) >16 >16 0.25 0.18 0.15 >16 0.25 Ox + SB (8) >16 >16 0.25 0.28 0.18 >16 0.25 Ox + SB (16) >16 >16 0.25 0.18 0.16 >16 0.25 Cf -32 24.5 2.0 1.0 0.52-32 s0.25 Cf + CA (2) -32 22.0 0.53 0.35 0.17.32 50.25 Cf + CA (4).32 21.0 <0.25.0.25 s0.25-32 <0.25 Cf + SB (8) 232 24.5 0.37 0.31 0.16.32 <0.25 Cf + SB (16).32 20.0 s0.25 <0.25.0.25.32 <0.25 Me >32 37 4.0 4.0 2.0 >32 2.3 Me + CA (2) >32 40 2.1 2.0 1.25 >32 2.0 Me + CA (4) >32 40 1.6 1.5 1.0 >32 1.3 Me + SB (8) 20 35 2.0 2.0 1.7 >32 2.0 Me + SB (16) 20 32 1.8 1.7 1.2 >32 2.0 a Tested in CSMHB with 2% NaCI and incubated at 35 C for 24 h. b p, Penicillin; Ox, oxacillin; Cf, cephalothin; Me, methicillin; CA, clavulanic acid; and SB, sulbactam. c See Table 1 for description of groups. (penicillin MIC,.32,ug/ml). In the presence of clavulanic acid or sulbactam, the MICs of methicillin, oxacillin, and cephalothin remained unchanged for methicillin-resistant S. aureus (-32,.16, and 16 to.32,ug/ml, respectively). Penicillin MICs were reduced but remained high enough to be considered resistant, at 8 to 16 jig/ml. None of the MICs of any of the drugs tested against P-lactamase-negative methicillin-resistant and methicillinsusceptible S. aureus were affected by clavulanic acid or sulbactam. The screening tests with oxacillin (6 jxg/ml) alone or methicillin (10,ug/ml) alone and oxacillin or methicillin with clavulanic acid (2 and 4,ug/ml) accurately indicated resistance or susceptibility to methicillin and oxacillin at 24 h for all 66 strains tested. At 48 h, however, most of the borderline-susceptible S. aureus grew on the plates with oxacillin (6,ug/ml) and some grew lightly on the oxacillin with clavulanic acid (2,ug/ml). Oxacillin with clavulanic acid (4,ug/ml) and the methicillin screen plates, with and without inhibitor, were negative for growth of the borderline-susceptible S. aureus at 48 h. The broth supernatant of a borderline-susceptible S. aureus isolate was able to hydrolyze penicillin and to partially hydrolyze oxacillin and methicillin. Penicillin was completely inactivated within 5 min, and more than 75% of oxacillin was inactivated within 60 min. Methicillin was only slightly less active after 60 min of incubation with the enzyme. Inactivation was prevented when clavulanic acid was added to the broth supernatant before the antimicrobial agent. A total of 28,B-lactamase-positive isolates of S. aureus (amoxicillin MICs, 2 to >128 jxg/ml) and 16,-lactamasepositive methicillin-resistant S. aureus isolates were tested against amoxicillin and clavulanic acid in ratios of 2:1 and 4:1 and also against amoxicillin with clavulanic acid and sulbactam at final concentrations of 4 and 16,ug/ml, respectively. S. aureus isolates tested with amoxicillin-clavulanic acid at ratios of 2:1 and 4:1 yielded susceptible amoxicillin MICs of 0.5 to 4,ug/ml for methicillin-susceptible S. aureus, but when tested against amoxicillin with fixed concentrations of clavulanic acid (4,ug/ml) and sulbactam (16 ;ig/ml), the MICs of amoxicillin were one to two twofold dilutions lower (0.25 to 2 jig/ml) than with the 2:1 and 4:1 ratios. Methicillin-resistant S. aureus had resistant MICs of 16 to 32 j,g/ml regardless of the amount of inhibitor. For methicillin-susceptible S. aureus, the zone sizes obtained with the amoxicillin-clavulanic acid combination disk were 22 to 30 mm, which would be interpreted as susceptible, but for methicillin-resistant S. aureus, zones of 8 to 14 mm were obtained, which would be interpreted as resistant to the combination. DISCUSSION Most S. aureus isolates produce an inducible,-lactamase. The proportion of the total,-lactamase liberated into a culture depends on the strain and on the conditions of growth. Isolates endemic in hospitals usually produce large quantities of P-lactamase and release 40 to 60% of it into the medium (10). The composition of the growth medium can have a major effect on the proportion of enzyme that is extracellular. For example, Coles and Gross reported that inorganic phosphate liberated the cell-bound enzyme (7), and Kim and Chipley reported that 10% NaCl (followed

836 McDOUGAL AND THORNSBERRY closely by 5% NaCI) optimally stimulated both the constitutive and induced formation of 13-lactamase (16). When S. aureus cultures were exposed to an inducer during growth in media with 5 to 10% NaCI, they released more 3-lactamase into the media than did those cultures grown in media with NaCl without an inducer. These investigators concluded that since S. aureus can survive in high salt concentrations, the release of P-lactamase was not due to cell damage and that a high salt concentration favored the release and synthesis of P-lactamase (16). Since we recently advocated the use of CSMHB plus 2% NaCl for the detection of methicillin-resistant S. aureus (36), we and others (15) have used this medium for testing S. aureus for susceptibility to methicillin, oxacillin, and cephalothin. Barry and Badel previously recommended the addition of 5% NaCl to the broth (3). However, we believed the addition of 2% NaCl to CSMHB was adequate for discriminating between methicillin-susceptible and heteroresistant strains and that it would not cause errors with the nonheteroresistant strains. We cautioned those who use 5% NaCl that it may make occasional strains appear intermediate to the P-lactam antimicrobial agents (36). Other recommendations that have been made for improving the susceptibility tests for methicillin-resistant S. aureus include lowering the incubation temperature to 30 to 35 C, extending the incubation time to 48 h, and using a larger inoculum. A problem that may arise when all of these recommendations are followed at the same time is that susceptible organisms may not be inhibited or killed by concentrations at the "susceptible" MIC. This may be because all of these recommendations favor the production and release of large amounts of 1-lactamase (7, 16, 21, 23, 39) Ṫhe results of this study show that some S. aureus strains produce large quantities of P-lactamase, an enzyme which is capable of slowly inactivating the so-called PRPs. Our results further show that oxacillin is more susceptible to hydrolysis by staphylococcal P-lactamase than is methicillin. Basker et al., using three types of tests to compare the stabilities staphylococcal,-lactamase, reported that the relative stabilities of the compounds in descending order were methicillin > cloxacillin = dicloxacillin = flucloxacillin > oxacillin = cephalothin = cephradine = cephalexin > cefazolin = cephaloridine (4). Lacey and Stokes reported the relative order of resistance of PRPs to,b-lactamase was methicillin (most resistant) > nafcillin > cloxacillin > flucloxacillin (most vulnerable) (22). of P-lactam antimicrobial agents to the action of Thus, we have a paradox: additional NaCl is needed to promote the growth and subsequent detection of the heteroresistant staphylococci (if they are present), but this salt may, at the same time, promote the production and release of 3-lactamase. We and others (1, 4, 8, 9, 17-22, 28, 34) have shown that oxacillin, methicillin, nafcillin, and cephalothin are only partially resistant to hydrolysis by staphylococcal 13-lactamase, and it follows that increased,b-lactamase production could cause increased resistance (whether it is borderline susceptible or borderline resistant). This conclusion was confirmed by the results obtained with clavulanic acid and sulbactam. Clavulanic acid and sulbactam are irreversible P- lactamase inhibitors (11, 31-33) that, by themselves, generally do not inhibit bacteria. When these agents are used in conjunction with other P-lactam antimicrobial agents that J. CLIN. MICROBIOL. would be inactivated by a bacterial 3-lactamase, the bacteria are inhibited by the combination if the 13-lactamase is susceptible to either clavulanic acid or sulbactam. Both clavulanic acid and sulbactam have only weak antibacterial activity against S. aureus when compared with other 3-lactam antimicrobial agents. In the presence of increasing amounts of these inhibitors, the MICs of 1-lactam antimicrobial agents are proportionally reduced (if,b-lactamase is the only reason for resistance) without affecting the MICs of non-,- lactam antimicrobial agents. For methicillin-susceptible S. aureus, the number of twofold reductions in the MIC depends first on the antimicrobial agent (its susceptibility to inactivation by staphylococcal f-lactamase) and second on the MIC for the particular organism. With most strains of methicillin-susceptible S. aureus, the higher the MIC for all P-lactam antimicrobial agents, the greater the reduction of that MIC in the presence of clavulanic acid and sulbactam because the higher MIC is due to the inactivation of the antimicrobial agents by staphylococcal,b-lactamase. If the MIC is high because of intrinsic resistance (heteroresistance), it will not be reduced to the susceptible category in the presence of inhibitors. (With some methicillinresistant S. aureus, the cephalothin MIC may not be in the resistant category to start with, but the MIC will not be reduced.) In tests with,b-lactamase-positive methicillin-susceptible S. aureus in the presence of 4 plg of clavulanic acid or 16,ug of sulbactam per ml, the MICs of penicillin, amoxicillin, oxacillin, cephalothin, and methicillin were reduced. In similar tests with borderline-susceptible S. aureus, MICs of penicillin and amoxicillin, which are very susceptible to r-lactamase hydrolysis, were reduced from -128,ug/ml to 1 to 2,ug/ml. These penicillin and amoxicillin MICs could be reduced even more and would approximate the MICs for P-lactamase-negative methicillin-susceptible S. aureus (0.06 to 0.12,ug/ml) if higher concentrations of clavulanic acid and sulbactam were used, but these concentrations would affect the MICs for P-lactamase-negative methicillin-susceptible S. aureus and some S. aureus strains with lower susceptibility (MICs of 8 to 16,ug/ml) to clavulanic acid alone. Penicillin MICs of c32,ug/ml were reduced to 0.25 to 0.12,g/ml for isolates of methicillin-susceptible S. aureus. Oxacillin MICs were reduced from borderline-susceptible or intermediate (2 to 4,ug/ml) to susceptible (.0.25,ug/ml). Methicillin appears to be more resistant to staphylococcal,-lactamase than the other drugs tested, and its MICs in the presence of the inhibitors were reduced from 4,ug/ml to 1 to 2,ug/ml. MICs of oxacillin, cephalothin, and methicillin for methicillinresistant S. aureus were not affected by clavulanic acid and sulbactam at the concentrations we tested. Penicillin MICs for methicillin-resistant S. aureus were reduced but remained in the resistant category at 8 to 16,ug/ml. Thus, with,-lactamase inhibitors, the borderline or intermediate oxacillin-susceptible strain became susceptible, but methicillinresistant S. aureus remained resistant. Although sulbactam inhibits staphylococcal,-lactamase, it appears to also act as an inducer. S. aureus strains (those isolates with penicillin MICs of.128,ug/ml) when grown in sulbactam gave a stronger, faster positive nitrocefin test than when grown without sulbactam. Clavulanic acid, however, did not appear to induce staphylococcal 3-lactamase. When we previously described an S. aureus strain with intermediate MICs of methicillin, oxacillin, and nafcillin (24, 36), we stated that a real intermediate result was rare. The susceptibility of this strain and three similar strains to the PRPs remained the same in the presence of clavulanic acid and sulbactam. These strains have low penicillin MICs (1 to 8,ug/ml), and these MICs are not reduced more than one

VOL. 23, 1986 ROLE OF P-LACTAMASE IN STAPHYLOCOCCAL RESISTANCE 837 dilution with P-lactamase inhibitors. This resistance appears, therefore, to be due to factors other than the 1- lactamase inactivation of these antimicrobial agents. Tests on screening plates (Mueller-Hinton agar plus 4% NaCl with 6 jig of oxacillin per ml) are effective in discriminating between methicillin-susceptible and methicillinresistant S. aureus after 24 h of incubation (36). Further incubation for 48 h or more of a plate with no growth, however, may permit inactivation of the drugs by,blactamase and thus make them appear oxacillin resistant. If clavulanic acid (4,ug/ml) or sulbactam (16,ug/ml) is included in the screen test medium, the inactivation of oxacillin by 13-lactamase is averted and the test remains negative, i.e., susceptible after 48 h of incubation. Screening plates containing methicillin (10,ug/ml) are not as affected by inactivation by,b-lactamase, because methicillin is more resistant to this 1-lactamase than is oxacillin. These screen plates should be incubated for only 24 h because methicillin-resistant S. aureus grows in 24 h, and strains that grow after 24 h are not heteroresistant but instead resistant because of 1-lactamase. The 24-h screen test should be a test for heteroresistant staphylococci. In our experience, there is an inverse relationship between the penicillin MIC and the degree of difficulty in detecting heteroresistance by in vitro tests. If the penicillin MICs for methicillin-resistant S. aureus are.128,ug/ml, heteroresistance is easier to detect in vitro than in those strains that produce less 1-lactamase and have penicillin MICs of c32,ug/ml. A possible reason for these characteristics is that there may be a relationship between the phenotypic expression of heteroresistance, the production and release of,b-lactamase, and the higher salt concentration tolerated by these organisms. The higher salt environment would provide a more favorable environment for the production and release of 1-lactamase, which in turn could protect the organism against penicillin until the mechanism for intrinsic resistance is in place and can provide protection. The higher salt concentration may also permit the heteroresistant cells to survive as protoplasts long enough to permit the development of the mechanism for intrinsic resistance. Recent studies have shown that the development of intrinsic resistance may be related to the synthesis of a penicillin-binding protein (PBP) called PBP 2a or PBP 2' (probably the same PBP) and that this PBP can be induced by,b-lactam antimicrobial agents (14, 37). This process would be compatible with the roles we have postulated for the,b-lactamase and higher salt concentration, in that they would protect the heteroresistant cells until they could be induced by the 1-lactam antimicrobial agent to produce enough PBP 2a or 2' to express intrinsic resistance. A probable reason that these difficult-to-detect methicillin-resistant S. aureus strains are even more difficult to detect with methicillin is that methicillin is less susceptible to the action of 1-lactamase and the cells thus have less protection. This lends support to the recommendation that oxacillin be used as the test agent for this group of antimicrobial agents (36). On the other hand, rare strains of methicillin-resistant S. aureus are,b-lactamase negative but relatively easy to detect without the use of additional salt in the medium. Although this appears to contradict the role we have postulated for 1-lactamase and salt, we have observed and others have reported (9) that these strains express lesser degrees of heterogeneity and have a larger subpopulation of resistant cells than do other methicillin-resistant S. aureus. They appear, therefore, to have already been induced to express their heteroresistance and do not need the aid of the,b-lactamase or salt. It is likely that if an S. aureus cell has the gene for heteroresistance, the usual chain of events leading to phenotypic expression will be the addition of the inducing 1-lactam antimicrobial agent; early protection by 1-lactamase, which is increased by the salt; early protection against lysis by salt; and, finally, induced heteroresistance. Therefore, cells that have been fully induced, as with the 1-lactamase-negative methicillinresistant S. aureus, may not need the intervention of,blactamase and salt. The ability of some S. aureus strains to be resistant to penicillinase-resistant 1-lactam antimicrobial agents because they are able to partially inactivate these agents with their 1-lactamase is a resistance mechanism that is distinctly different from that of intrinsic resistance, and only the latter should be called heteroresistance. 1-Lactamase, however, could play a protective role in heteroresistant strains with small resistant populations or intrinsic resistance, and,blactamase production may act additively or even synergistically (30). Whether,B-lactamase produced by some methicillin-susceptible S. aureus strains can cause clinical failure with oxacillin, nafcillin, methicillin, or cephalothin cannot be predicted on the basis of these in vitro data, but the possibility exists. Lacey and colleagues reported that flucloxacillin was unable to eliminate,b-lactamase-positive S. aureus in two patients and concluded that the therapeutic failure was due to destruction of flucloxacillin in vivo by staphylococcal,b-lactamase (21, 22). Norden and Dickens, from their studies of the treatment of staphylococcal osteomyelitis in rabbits, reported that a significant amount of cephaloridine was inactivated by induced 1-lactamase produced by large numbers of S. aureus (27). They stated that the number of organisms in the diseased bone (approximately 105 to 106 CFU/ml) was large enough, at least theoretically, for the production of enough,b-lactamase to inactivate cephaloridine. Burgess and Evans reported that apparently adequate doses of cephaloridine did not cure staphylococcal endocarditis, presumably because of the production of staphylococcal,b-lactamase (6). In general, the PRPs and cephalothin can be used successfully to treat infections caused by 1-lactamase-producing S. aureus, even though the,b-lactamase can be shown in vitro to slowly inactivate the drug. The assumption is that the rate of hydrolysis is not rapid enough to be translated into clinical resistance. In light of the vulnerability of the cephalosporins and the isoxazolyl penicillins to 1-lactamase hydrolysis, infections caused by nonheteroresistant S. aureus in severely ill patients might preferably be treated with methicillin or nafcillin. There is no information, however, on the efficacy of these drugs if the strain produces enough 13- lactamase to cause an in vitro borderline result. Another factor to be considered is the role of 1-lactamase inhibitors used in combination with other,b-lactam antimicrobial agents to treat infections caused by S. aureus with borderline susceptibility or resistance. Clavulanic acid in combination with amoxicillin (Augmentin) and with ticarcillin (Timentin) is now commercially available, and combinations of sulbactam and ampicillin and of sulbactam and cefoperazone are under development. Amoxicillin, ticarcillin, ampicillin, and cefoperazone are clinically active against nonheteroresistant staphylococci in the absence of - lactamase production. Since clavulanic acid and sulbactam inhibit staphylococcal 1-lactamase, theoretically these combinations could be used to treat staphylococcal infections, although the possibility of induction of the 13-lactamase by sulbactam needs further study. There is limited information on the use of 1-lactamase inhibitors in combination with

838 McDOUGAL AND THORNSBERRY P-lactam antimicrobial agents for treating staphylococcal infections. Lacey has indicated that the role of these drugs is uncertain (19), but the successful use of clavulanic acid and amoxicillin in treating staphylococcal soft tissue infections has been reported for animal models (5) and for children (12). The other side of the question on therapy for infections caused by these borderline (nonheteroresistant) S. aureus is whether they have to be treated with vancomycin, the drug that is recommended to treat methicillin-resistant S. aureus infections in the United States. This question is important for these borderline S. aureus because of the potential toxicity and cost of vancomycin. We emphasize, however, that vancomycin is the drug of choice for heteroresistant S. aureus infections. Obviously, these questions of therapy cannot be answered until appropriate clinical trials are done to compare the efficacy of different regimens of therapy. Another important consideration is whether to place hospitalized patients who are infected or colonized with these borderline, nonheteroresistant S. aureus on isolation precautions. Until more data are accumulated, the conservative approach is to consider the isolates of these patients as methicillin-resistant S. aureus epidemiologically and to implement the isolation precautions described in the Centers for Disease Control guidelines (13). In addition to the implications for therapy and infection control, these findings also have implications for the clinical microbiology laboratory. We think that these borderline strains should be reported as resistant to PRPs if the in vitro test results are interpreted as resistant according to established interpretive guidelines. The more difficult question is whether such strains should be reported as heteroresistant or *nonheteroresistant. A report of a methicillin-resistant, nonheteroresistant isolate would probably have little meaning to most clinicians and would probably create confusion. Until many of the therapeutic and epidemiological questions are answered, the most conservative approach is to treat the isolates as intrinsically resistant methicillin-resistant S. aureus and not be concerned about the mechanisms. If a different decision is to be made, it will be arbitrary; it should be made because of a high incidence of the P- lactamase-mediated oxacillin-resistant strains, and it should be made by the appropriate hospital personnel, e.g., microbiologist, infectious disease physician, and epidemiologist. If laboratory workers wish to determine whether a borderline culture is heteroresistant, they could do an MIC test that includes a P-lactamase inhibitor in combination with the P-lactam antimicrobial agent, they could do a disk diffusion test with amoxicillin-clavulanic acid (25), or they could use the 24-h agar screen test (36). If the isolate is borderline resistant because of P-lactamase (i.e., not intrinsically resistant), the MIC will be reduced to a susceptible level, the amoxicillin-clavulanic acid zone size will be interpreted as susceptible, and no growth will be seen on the screen plate at 24 h; if the isolate is intrinsically resistant to oxacillin, the results will be exactly the opposite. The recognition of these P-lactamase-mediated oxacillinresistant strains further complicates the terminology associated with isolates of S. aureus resistant to the PRPs. The terms methicillin resistant, methicillin-resistant S. aureus, intrinsically resistant, occult resistant (15), and heteroresistant are used interchangeably to refer to those strains with chromosomally mediated resistance not only to methicillin but also to oxacillin, nafcillin, cloxacillin, flucloxacillin, cephalosporins, and probably all P-lactams. This resistance is related to altered or low-affinity targets, since the PBPs appear to be involved (14, 37). On the other hand, I- lactamase-mediated oxacillin resistance is probably always plasmid mediated (although this was not a part of this study), but it also refers to resistance to other P-lactams depending on their susceptibility to hydrolysis by this enzyme. Undoubtedly, terms will be coined to separate these mechanistically different resistant staphylococci, adding to a list of names that is already too long. One way to circumvent some of the confusion linked to the use of methicillin, or another antimicrobial agent, or the use of heteroresistant in the name would be to call the isolates intrinsically resistant S. aureus and acquired-resistant S. aureus. Intrinsically resistant S. aureus would be those strains that are resistant to PRPs because of chromosomally mediated changes in PBPs (14, 37), and acquired-resistant S. aureus would be those strains that are resistant to a PRP because of the P-lactamase they produce. Before these designations are used, the organism must be shown to be resistant to a PRP. If an S. aureus isolate is found to be resistant to oxacillin, for example, it could be further tested as described above and further designated as either intrinsically or acquired-resistant S. aureus. It should be remembered that production of,blactamase is not synonymous with resistance to PRPs. In summary, some S. aureus strains that produce large amounts of,-lactamase may show borderline susceptibility or resistance to PRPs because the staphylococcal 13- lactamase can hydrolyze these antimicrobial agents. These strains are not heteroresistant. If these strains are tested against these,b-lactam antimicrobial agents in combination with clavulanic acid or sulbactam (P-lactamase inhibitors), they become susceptible to the 13-lactam antimicrobial agents. Additional studies are needed to determine the therapeutic and epidemiological implications of these findings. LITERATURE CITED J. CLIN. MICROBIOL. 1. Ayliffe, G. A. J., and M. Barber. 1963. Inactivation of benzylpenicillin and methicillin by hospital staphylococci. Br. Med. J. 2:202-205. 2. Baker, C. N., C. Thornsberry, and R. W. Hawkinson. 1983. Inoculum standardization in antimicrobial susceptibility testing: evaluation of overnight agar cultures and the rapid inoculum standardization system. J. Clin. Microbiol. 17:450-457. 3. Barry, A. L., and R. E. Badel. 1977. Reliability of the microdilution technique for detection of methicillin-resistant strains of Staphylococcus aureus. Am. J. Clin. Pathol. 67:489-495. 4. Basker, M. J., R. A. Edmondson, and R. Sutherland. 1980. Comparative stabilities of penicillins and cephalosporins to staphylococcal,b-lactamase and activities against Staphylococcus aureus. J. Antimicrob. Chemother. 6:333-341. 5. Brook, I., J. C. Coolbaugh, and R. I. Walker. 1983. Antibiotic and clavulanic acid treatment of subcutaneous abscesses caused by Bacteroides fragilis alone or in combination with aerobic bacteria. J. Infect. Dis. 148:156-159. 6. Burgess, H. A., and R. J. Evans. 1966. Failure of cephaloridine in a case of staphylococcal endocarditis. Br. Med. J. 2:1244. 7. Coles, N. W., and R. Gross. 1967. Liberation of surface-located penicillinase from Staphylococcus aureus. Biochem. J. 102: 742-747. 8. Dyke, K. G. H. 1966. Penicillinase production and intrinsic resistance to penicillins in Staphylococcus aureus. Lancet i:835-838. 9. Dyke, K. G. H. 1969. Penicillinase production and intrinsic resistance to penicillins in methicillin-resistant cultures of Staphylococcus aureus. J. Med. Microbiol. 2:261-278. 10. Dyke, K. G. H. 1979.,-Lactamase of Staphylococcus aureus, p. 291-310. In J. M. T. Hamilton-Miller and J. T. Smith (ed.), Beta-lactamase. Academic Press, Inc., New York.

VOL. 23, 1986 ROLE OF 1-LACTAMASE IN STAPHYLOCOCCAL RESISTANCE 839 11. English, A. R., J. A. Retsema, A. E. Girard, J. E. Lynch, and W. E. Barth. 1978. CP-45,899, a beta-lactamase inhibitor that extends the antibacterial spectrum of beta-lactams: initial bacteriological characterization. Antimicrob. Agents Chemother. 14:414-419. 12. Fleisher, G. R., C. M. Wilmott, and J. M. Campos. 1983. Amoxicillin combined with clavulanic acid for the treatment of soft tissue infections in children. Antimicrob. Agents Chemother. 24:679-681. 13. Garner, J. S., and B. P. Simmons. 1983. CDC guidelines for isolation precautions in hospitals. Infect. Control 4:249-325. 14. Hartman, B. J., and A. Tomasz. 1984. Low-affinity penicillinbinding protein associated with,b-lactam resistance in Staphylococcus aureus. J. Bacteriol. 158:513-516. 15. Hindler, J. A., and C. B. Inderlied. 1985. Effect of the source of Mueller-Hinton agar and resistance frequency on the detection of methicillin-resistant Staphylococcus aureus. J. Clin. Microbiol. 21:205-210. 16. Kim, T. K., and J. R. Chipley. 1974. Effect of salts on penicillinase release by Staphylococcus aureus. Microbios 1OA:55-63. 17. Lacey, R. W. 1975. Antibiotic resistance plasmids of Staphylococcus aureus and their clinical importance. Bacteriol. Rev. 39:1-32. 18. Lacey, R. W. 1982. Stability of the isoxazolyl penicillins to staphylococcal P-lactamase. J. Antimicrob. Chemother. 9:239-243. 19. Lacey, R. W. 1983. Treatment of staphylococcal infections. J. Antimicrob. Chemother. 11:3-6. 20. Lacey, R. W. 1984. Mechanisms of resistance to beta-lactam antibiotics in Staphylococcus aureus. Scand. J. Infect. Dis. 42(Suppl.):64-71. 21. Lacey, R. W., and E. L. Lewis. 1975. Further evolution of a strain of Staphylococcus aureus in vivo: evidence for significant inactivation of flucloxacillin by penicillinase. J. Med. Microbiol. 8:337-347. 22. Lacey, R. W., and A. Stokes. 1977. Susceptibility of the 'penicillinase-resistant' penicillins and cephalosporins to penicillinase of Staphylococcus aureus. J. Clin. Pathol. 30:35-39. 23. Leggate, J., and W. H. Holms. 1968. Gratuitous synthesis of P-lactamase in Staphylococcus auireus. J. Bacteriol. 96: 2110-2117. 24. McDougal, L. K., and C. Thornsberry. 1984. New recommendations for disk diffusion antimicrobial susceptibility tests for methicillin-resistant (heteroresistant) staphylococci. J. Clin. Microbiol. 19:482-488. 25. National Committee for Clinical Laboratory Standards. 1984. Performance standards for antimicrobic disc susceptibility test, 3rd ed. Approved standard M2-A3. National Committee for Clinical Laboratory Standards, Villanova, Pa. 26. National Committee for Clinical Laboratory Standards. 1983. Standard methods for dilution antimicrobial tests for bacteria which grow aerobically. Tentative standard M7-T. National Committee for Clinical Laboratory Standards, Villanova, Pa. 27. Norden, C. W., and D. R. Dickens. 1973. Experimental osteomyelitis. lii. Treatment with cephaloridine. J. Infect. Dis. 127:525-528. 28. Novick, R. P., and M. H. Richmond. 1965. Nature and interactions of the genetic elements governing penicillinase synthesis in Staphylococcus aureus. J. Bacteriol. 90:467-480. 29. O'Callaghan, C. H., A. Morris, S. M. Kirby, and A. H. Shingler. 1972. Novel method for detection of P-lactamases by using a chromogenic cephalosporin substrate. Antimicrob. Agents Chemother. 1:283-288. 30. Parker, M. T., and J. H. Hewitt. 1970. Methicillin-resistance in Staphylococcus aureus. Lancet i:800-804. 31. Reading, C., and M. Cole. 1977. Clavulanic acid: a betalactamase-inhibiting beta-lactam from Streptomyces ciavuligerus. Antimicrob. Agents Chemother. 11:852-857. 32. Reading, C., and P. Hepburn. 1979. The inhibition of staphylococcal P-lactamase by clavulanic acid. Biochem. J. 179:67-76. 33. Retsema, J. A., A. R. English, and A. E. Girard. 1980. CP-45,899 in combination with penicillin or ampicillin against penicillinresistant Staphylococcus, Haemophilus influenzae, and Bacteroides. Antimicrob. Agents Chemother. 17:615-622. 34. Richmond, M. H. 1963. Purification and properties of the exopenicillinase from Staphylococcus aureus. Biochem. J. 88:452-459. 35. Thornsberry, C., and L. A. Kirven. 1978.,B-lactamase of the Legionnaires' bacterium. Curr. Microbiol. 1:51-54. 36. Thornsberry, C., and L. K. McDougal. 1983. Successful use of broth microdilution in susceptibility tests for methicillinresistant (heteroresistant) staphylococci. J. Clin. Microbiol. 18:1084-1091. 37. Utsui, Y., and T. Yokota. 1985. Role of an altered penicillinbinding protein in methicillin- and cephem-resistant Staphylococcus aureus. Antimicrob. Agents Chemother. 28:397-403. 38. Washington, J. A., II, and V. L. Sutter. 1980. Dilution susceptibility test: agar and macro-broth dilution proce,dures, p. 453-458. In E. H. Lennette, A. Balows, W. J. Hausfer, Jr., and J. P. Truant (ed.), Manual of clinical microbiology, 3rd ed. American Society for Microbiology, Washington, D.C. 39. Watanakunakorn, C. 1985. Effect of inoculum size on in vitro susceptibility of methicillin-resistant Staphylococcus aureus to eighteen antimicrobial agents. Eur. J. Clin. Microbiol. 4:68-69.