JOURNAL OF CLINICAL MICROBIOLOGY, OCt. 1987, p. 1920-19250095-1137/87/101920-06$02.00/0Copyright © 1987, American Society for Microbiology

Vol. 25, No. 10

Optimal Dilution Susceptibility Testing Conditions,Recommendations for MIC Interpretation, and Quality Control

Guidelines for the Ampicillin-Sulbactam CombinationRONALD N. JONES* AND ARTHUR L. BARRY

The Clinical Microbiology Institute, Tualatin, Oregon 97062

Received 4 May 1987/Accepted 29 June 1987

The ampicillin-sulbactam combination was evaluated in vitro to determine the optimal susceptibility testingconditions among five combination ratios and four fixed concentrations of sulbactam. The organisms testedwere markedly resistant to aminopenicillins and most other beta-lactams. The ratio of 2:1 is recommended toassure recognition of the ampicillin-sulbactam spectrum and minimize false-susceptible results among strainsknown to be resistant to this combination. Proposed MIC breakpoint concentrations were compatible withlevels in serum achieved with recommended clinical doses. Cross-resistance analyses comparing ampicillin-sulbactam and amoxicillin-clavulànate showed comparable activity and spectra. However, the major interpre-tive disagreement was sufficient to require separate testing of these aminopenicillin-inhibitor combinations. Therecommended ampicillin-sulbactam MIC susceptibility breakpoints are as follows: (i) <8.0/4.0 ,ug/ml for testsagainst members of the family Enterobacteriaceae, anaerobes, nonenteric gram-negative bacilli, staphylococci,Haemophilus influenzae, and Branhamella catarrhalis; (ii) the ampicillin MICs alone interpreted by NationalCommittee for Clinical Laboratory Standards criteria should predict ampicillin-sulbactam susceptibility forthe enterococci, streptococci, and Listeria monocytogenes. MIC quality control ranges were determined bymultiple laboratory broth microdilution trials for the ampicillin-sulbactam 1:1 and 2:1 ratio tests.

The ampicillin-sulbactam combination is one of severalavailable penicillin-beta-lactamase inhibitor drugs that havea wide spectrum of antimicrobial activity (1, 2, 3, 6, 7, 15, 16,19, 20; R. N. Jones, Drugs, in press). The sulfone beta-lactamase inhibitor sulbactam has an excellent affinity forthe clinically prevalent plasmid-mediated bacterial enzymesthat can readily hydrolyze the aminopenicillins (7, 16, 17;Jones, in press). The combination of ampicillin and sulbac-tam is a particularly good choice since it combines thereliable antimicrobial activity of ampicillin against entero-cocci, streptococci, and beta-lactamase-negative strains ofenteric bacilli with that of the combination, which alsoinhibits penicillinase-producing Staphylococcus spp. andNeisseria gonorrhoeae, beta-lactamase-positive Haemophi-lus influenza and Branhamella catarrhalis, Bacteroidesfragilis group species, and many more members of the familyEnterobacteriaceae (2, 7, 16; Jones, in press).

Ampicillin-sulbactam has been studied clinically as anorally administered ester combination and as a parenteralcombination usually administered in equal doses of eachdrug or as a 2:1 ratio of ampicillin to sulbactam (11). Thepharmacokinetics of each component in humans are verysimilar and the drug ratios in serum or tissue vary between1:1 and 2:1 (ampicillin-sulbactam) depending upon the dosesadministered (5, 8).

In this report, we summarize the results of investigationsto determine the best in vitro dilution susceptibility testconditions for the ampicillin-sulbactam combination. Wealso report the results of a multilaboratory study to establishbroth microdilution MIC quality control (QC) guidelines.

MATERIALS AND METHODSAntimicrobial agents. Ampicillin, amoxicillin, and clavu-

lanic acid were obtained from Beecham Laboratories, Bris-

* Corresponding author.

tol, Tenn., and sulbactam was contributed by Pfizer CentralResearch, Groton, Conn. All other drugs and reagents weresupplied by their principal manufacturers.Organisms tested. Clinical bacterial isolates collected dur-

ing the previous 5 years were selected for documentedresistance to aminopenicillins and other beta-lactam antibi-otics. These 741 isolates were distributed as follows: 267members of the Enterobacteriaceae (17% susceptible toampicillin, 47% susceptible to cefoperazone, and 0% inhib-ited by 8 ,ug of sulbactam per ml), 88 Staphylococcus spp.isolates (18 were oxacillin resistant, and 34 others werepenicillinase producers), 42 H. influenza isolates (20 beta-lactamase producers), 20 B. catarrhalis isolates (17 beta-lactamase producers), 22 Acinetobacter spp. isolates, 71Pseudomonas spp. isolates (7% susceptible to ampicillin), 97strict anaerobes (63 B. fragilis group isolates resistant toampicillin), and 20 Streptococcus pneumoniae isolates (9relatively resistant and 2 highly resistant to penicillins). Theremaining isolates were generally quite susceptible to ampi-cillin. They included 10 Listeria monocytogenes isolates, 20Streptococcus pyogenes isolates, 20 Streptococcus agalac-tiae isolates, 10 Streptococcus bovis isolates, 10 serogroup Cstreptococci, 10 serogroup G streptococci, 19 Enterococcusfaecalis isolates, 10 Enterococcus faecium isolates, and 5Enterococcus durans isolates. By standardized susceptibil-ity test methods (12-14) or direct beta-lactamase detectionmethods, 63.8% of the strains were ampicillin resistant.These organisms differ greatly from the strains used in ourearlier studies of beta-lactamase inhibitor combinations orthose routinely encountered in clinical practice (3, 4, 9;Jones, in press).

Antimicrobial susceptibility testing and beta-lactamase de-tection methods. The MICs were determined by the brothmicrodilution method described in detail previously (4, 9;Jones, in press) and by the National Committee for ClinicalLaboratory Standards (NCCLS) (13). The inoculum con-

1920

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

AMPIClLLIN-SULBACTAM DILUTION TESTS 1921

TABLE 1. Cumulative percentage of 267 isolates ofEnterobacteriaceae inhibited by various combinations of

ampicillin and sulbactam, ampicillin alone, and sulbactam alone"

Cumulative 57< inhibited by ampicillinTest condition component (ig/!ml):

-0.5 1 2 4 8h 16 32 64

Ampicillin alone 3 8 13 16 17 21 26 37

Ampicillin-sulbactam1:1 2 8 18 25 35 51 70 862:1 3 8 15 23 30 43 57 804:1 2 8 13 21 26 38 51 666.7: 1' 1 7 12 20 24 34 49 598:1 2 7 13 18 25 30 46 58

Fixed sulbactam concn (,ug/ml)1 1 9 16 21 24 29 36 472 1 9 18 23 25 35 42 514 1 12 21 24 29 38 45 538 2 17 23 27 34 42 50 57

Sulbactam alone 1 26 67a Isolates were selected for resistance to beta-lactams. especially amino-

penicillins. The organism collection was 83% resistant to ampicillin.b Results are the percentages susceptible by NCCLS (12, 13) criteria for the

dilution tests.Approximates pediatric formulations.

tained 5 x 105 CFU/ml, and the MIC was read as the lowestantimicrobial concentration completely inhibiting bacterialgrowth in a well after 16 to 20 h of incubation at 35°C.Ampicillin was tested with the enzyme inhibitor sulbactam inratios (ampicillin to sulbactam) of 1:1, 2:1, 4:1, 6.7:1, and8:1. In addition, fixed sulbactam concentrations of 1, 2. 4,and 8 ,ug/ml were tested with ampicillin concentrationsranging from 0.125 to 64 Jg/ml. Dilution test modificationsfor fastidious strains were those advocated by NCCLSstandard M7-A (13). Anaerobes were tested by the brothmicrodilution method (NCCLS standard M17-P) describedpreviously (3, 14).The determination of broth microdilution MIC QC guide-

lines used the study design of Barry et al. (2a). Laboratoriescontributing to this investigation were the Clinical Microbi-ology Institute, Tualatin, Oreg.; The Cleveland Clinic Foun-dation, Cleveland, Ohio; St. Vincent Hospital and MedicalCenter, Portland, Oreg.; St. Francis Regional Medical Cen-ter, Wichita, Kans.; Centers for Disease Control, Atlanta,Ga.; and Ohio State University Hospital, Columbus. Eachlaboratory generated 20 ampicillin-sulbactam (1:1 and 2:1ratios) MICs for each QC strain. The MIC of ampicillin alonewas determined as an internal control. A common lot ofbroth microdilution trays was tested by each investigator(five tests) to assess interlaboratory technical variation (125total MICs per organism). Data were analyzed by methodsdescribed elsewhere (2a).

RESULTS

Ampicillin-sulbactam tests against members of the Entero-bacteriaceae. The ampicillin-sulbactam dilution tests wereperformed with five ratios and four fixed sulbactam concen-trations against 267 members of the Enterobacteriaceae(Table 1). The organism collection was only 17% susceptibleto ampicillin alone, and none of the isolates was inhibited by8.0 ,ug of sulbactam per ml. The widest spectrum of activityfor ampicillin-sulbactam was achieved by testing at the 1:1

ratio or at a fixed sulbactam concentration of 8.0 p.g/ml.Under each of these test conditions, the susceptible break-point concentration was 8.0 p.g of ampicillin and sulbactamper ml. Similarly, the 2:1 ratio and the fixed sulbactamconcentration of 4 F.g/mI had the same breakpoint concen-trations of 8 ,ug of ampicillin and 4 ,ug of sulbactam per ml.However, the ratio tests yielded more moderately suscepti-ble MICs (16 ,tg of ampicillin per ml).The expanded spectrum produced by the 1:1 ratio or the

8.0 <tg/ml sulbactam fixed ratio tests were among the Entero-bacter spp. (6 of 73 strains), Klebsiella pneunmoniae (5 of 22strains), Proi'idencia sluartii (1 of 19 strains), and Proteuss'ulgaris (5 of 10 strains). The ampicillin MICs for nearly allof these strains were 8.0 ,ug/mI associated with 8.0 ji.g ofsulbactam per ml. The presence of sulbactam at 8 p.g/ml inratio or fixed concentration tests doubled the ampicillin invitro efficacy (17 to 35%) against these very resistant entericbacilli.

Ampicillin-sulbactam tests against staphylococci, H. influ-enzae, and B. catarrhalis. The results of dilution testing ofstrains of Staphylococcus aureus, coagulase-negative Staph-ylococcus spp., H. influenzae, and B. catarrhalis are shownin Table 2. The ampicillin-sulbactam MICs for staphylococcithat produce a penicillinase were routinely higher than thosefor enzyme-negative strains. The 4.0 or 8.0 Fsg/mI fixedconcentration tests produced lower MICs for 90% of thestrains tested (MIC90s) than did the 1:1 or 2:1 ratio tests withpenicillin-resistant staphylococci; however, all ampicillin-sulbactam MICs were c4.0/4.0 and '4.0/2.0 ,ug/ml in the 1:1and 2:1 ratio tests, respectively. This represents a -16-foldreduction in the ampicillin MIC.The ampicillin MICs for the beta-lactamase positive H.

influenzae isolates were generally reduced at least eightfoldin the 1:1 or 2:1 ratio tests of ampicillin-sulbactam. Again,fixed 2.0, 4.0, or 8.0 p.g/ml sulbactam tests produced lowerampicillin MtCs, but al strains were susceptible to -8.0 Ftgof ampicillin per ml when tested with at least 4.0 p.g ofsulbactam per ml.

B. catarrhalis isolates were generally ampicillin resistantby a beta-lactamase mechanism. All ratio and fixed sulbac-tam concentration tests reduced the MICs for B. catarrhalisto c0.5 p.g/ml. Sulbactam was directly active as an antimi-crobial agent against B. catcarrhalis (MIC90, 16 ,ug/ml). Thus,the combination acts as a synergistic antimicrobial pair, aswell as via the inhibitory effect of sulbactam on the beta-lactamase of the organism.The testing of all ratios and fixed sulbactam concentra-

tions produced MICs identical to those of ampicillin alonefor the beta-lactamase-deficient strains of these threegroups. For oxacillin-resistant Staphylococcus spp., ampi-cillin MICs ranged from 4.0 to >64 ,ug/ml (MIC90, >64,ug/ml). Nearly all ampicillin-sulbactam MICs for the oxacil-lin-resistant staphylococci (ORS) were 8.0/8.0 or 8.0/4.0,ug/ml, a borderline level of activity.

Ampicillin-sulbactam tests against enterococci, Streptococ-cus spp., and L. monocytogenes. The ampicillin MICs for the34 enterococci and six species or groups of streptococciwere unaffected by the presence of sulbactam. The sulbac-tam MICs for these organisms were 32 to >64 ,ug/ml. Severalpneumococci relatively or highly resistant to penicillin weretested and found to be resistant to the ampicillin-sulbactamcombinations. For the two highly resistant S. pneumoniaestrains, the ampicillin-sulbactam (2:1 ratio) MICs were 2/1and 4/2 ,ug/ml. L. monocytogenes strains were directlyinhibited by ampicillin alone, and there was no furtherinhibition after sulbactam was added.

VOL. 25, 1987

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

TABLE 2. Ampicillin-sulbactam susceptibility testing results for beta-lactamase-producing and -deficient isolates of Staphylococcus spp.,H. influenza, and B. catarrhalis with five ratios and four different fixed sulbactam concentrations

MIC (range) foray:Test condition B. catarrhalis H. influenzae Staphylococcus spp.

Beta-lac+ Beta-lac- Beta-lac+ Beta-lac- Beta-lac+ Beta-lac-

Ampicillin alone <0.12-16 <0. 12-O0S 16->64 <0.12-0.25 0.25->64 -0.12-0.25

Ampicillin-sulbactam1:1 <0.12-0.25 O0.12 0.5-4.0 <0.12-0.25 <0. 12-4.0 <0.12-0.252:1 <0.12-0.25 sO.12 1.0-8.0 s0.12-0.25 <0.12-4.0 <0.12-0.254:1 O0.12-0.5 sO.12 1.0-16 <0.12-0.25 <O. 12-8.0 sO.12-0.256.7:1 <O.12-0.5 <0.12-0.25 2.0-16 <0.12-0.5 <0.12-8.0 sO.12-0.258:1 O0.12-0.5 <0. 12-0.25 2.0-16 <O.12-0.25 O.12-16 sO.12-0.25

Fixed sulbactam concn (,ug/ml)1 <0.12 s0.12 0.5-16 <0.12-0.25 sO.12-8.0 sO.12-0.252 <0.12 sO.12 0.25-4.0 <0.12-0.25 <O. 12-8.0 0.12-0.254 O0.12 sO.12 sO.12-2.0 s0.12-0.25 <0. 12-2.O sO.12-0.258 <O.12 0.12 -0.12-1.0 <0.12-0.25 s0.12-1.0 s0.12

Sulbactam alone 1.0-32 1.0-8.0 32->64 32->64 >64 64-64

a Beta-lac+, Beta-lactamase-producing isolates; Beta-lac-, beta-lactamase-deficient isolates.

Ampicillin-sulbactam tests against anaerobic bacteria. Fourtesting conditions for the ampicillin-sulbactam combinationwere assessed by broth microdilution (15) methods against97 strict anaerobic isolates (Table 3). Ampicillin MICs weremarkedly reduced by the addition of sulbactam under all testconditions. The overall ampicillin spectrum was increasedfrom 41.2 to 99% by the sulbactam inhibition of beta-lactamases produced by the Bacteroides spp. The only strainnot susceptible to this combination was a beta-lactamase-negative Peptococcus asaccharolyticus isolate for which theampicillin MIC was 16 ,ug/ml, i.e., moderately susceptible.The ampicillin MIC50s for the B. fragilis group strains weresO.5 iig/ml under all test conditions.

Ampicillin-sulbactam tests against Acinetobacter spp. andpseudomonads. The results of testing ampicillin, sulbactam,and nine ampicillin-sulbactam combinations against 22 Aci-netobacter strains are shown in Table 4. Sulbactam alonewas very active against these organisms, with an MIC90 of4.0 ,ug/ml. The results of the combination testing reflectedthis direct sulbactam action. Therefore, the 1:1 or 2:1 ratiotests and the 4.0 or 8.0 ,ug/ml sulbactam fixed concentrationtests more accurately assessed the value of the combinationagainst the Acinetobacter spp. isolates.

Sulbactam had little effect on the pseudomonads in pro-

TABLE 3. Susceptibility testing results for ampicillin alone andin combination with sulbactam against 97 anaerobic isolates"

Test condition' MIC (pg/mI) %

50% 90% Range g/mI

Ampicillin alone 16 >32 <0.5->32 41.2

Ampicillin-sulbactam1:1 s0.5 4.0 c0.5-16 99.02:1 cO.5 4.0 <0.5-16 99.0

Fixed sulbactam concn (,ug/ml)4 sO.5 1.0 <0.5-16 99.08 -O.5 1.0 -0.5-16 99.0

a Including 25 anaerobic gram-positive strains (8 species), 63 strains fromthe B. fragilis group, and 9 other Bacteroides spp.

b Only four ampicillin-sulbactam combinations were tested.

ducing greater susceptibility to ampicillin. Only two addi-tional strains (P. acidovorans and P. maltophilia) becamesusceptible to ampicillin when .4.0 ,ug of sulbactam per mlwas added. This increased the ampicillin spectrum only from7.0 to 9.9% among the 71 Pseudomonas isolates.

Ampicillin-sulbactam MICs compared with amoxicillin-clavulanic acid MICs. The beta-lactamase inhibitors sulbac-tam and clavulanic acid have both been combined withaminopenicillins for clinical use. The MICs for the 2:1 ratiosof amoxicillin-clavulanate and ampicillin-sulbactam, ratioseither used clinically or demonstrating the best correlationwith clinical pharmacokinetics or both, are shown in Table 5.Ampicillin-sulbactam possessed comparable activity againstmembers of the Enterobacteriaceae and most other orga-nisms. The four enteric organisms susceptible to ampicillin-sulbactam and resistant to amoxicillin-clavulanate werestrains of Enterobacter spp. and P. stuartii. However, fiveEscherichia coli isolates were susceptible to amoxicillin-clavulanate but resistant to ampicillin-sulbactam.

TABLE 4. Antimicrobial activity of ampicillin, sulbactam, andnine different ampicillin-sulbactam combinations against 22 strains

of Acinetobacter spp.

Test condition MIC (tg/mi) % s8501% 90% Range p.g/mI

Ampicillin alone 32 64 4.0->64 18.1

Ampicillin-sulbactam1:1 1.0 8.0 0.5->64 95.52:1 2.0 4.0 1.0->64 95.54:1 4.0 16 2.0->64 86.46.7:1 4.0 16 2.0->64 63.68:1 8.0 16 2.0->64 63.6

Fixed sulbactam concn (,u.g/ml)1 1.0 32 <0.12->64 63.62 0.12 16 <0.12->64 77.34 <0.12 <O.12 <0.12->64 95.58 <0.12 sO.12 s0.12->64 95.5

Sulbactam alone 2.0 4.0 1.0->64 95.5

1922 JONES AND BARRY J. CLIN. MICROBIOL.

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

AMPICILLIN-SULBACTAM DILUTION TESTS 1923

TABLE 5. Cross-resistance and -susceptibility to ampicillin-sulbactam and amoxicillin-clavulanic acid tested at

2:1 ratios (12, 13)

No. of isolateswith ampicillin-

Organisms (no. tested) Amoxicillin-clavulanate sulbactam (2:1)(2:1) MIC (p.g/ml) MIC (p.g/ml):s8 16 -32

Enterobacteriaceae (209) 8 69 10 5a17 4 8 13.32 4h 16 80

Acinetobacter and <8 17 1 6`Pseudomonas spp. (78) 16 1 3

.32 50a Five E. coli strains.b Including one strain of Enterobacter aerogenes, two strains of P. stuartii,

and one strain of Enterobacter cloacae.C Including four strains of P. maltop/hilia and two strains of P. acidovorans.

Comparison of these two inhibitor combinations againstthe 78 non-Enterobacteriaceae gram-negative bacilli indi-cated greater activity for amoxicillin-clavulanate. Threestrains of P. maltophilia and two strains of P. acidovoranswere susceptible only to amoxicillin-clavulanate. The abso-lute agreement between these drug combinations was 88.2%,with 9.8% minor and 2.0% major (false-susceptible or -resis-tant) discrepancies. However, since many of the majorerrors were among the E. coli isolates (nearly 50% of routineenteric isolates), the small major error rate may be greatlyamplified in some Enterobacteriaceae populations.Comparison of the two combination MICs for the

enterococci (r = 0.93) and streptococci (r = 0.98) showednearly identical results indistinguishable from the compari-son of activity of their active aminopenicillin components.Similarly, the staphylococcal, B. catarrhalis, and H. influ-enzae correlation coefficients were >0.95, with 100% inter-pretive agreement using the previously published susceptibleamoxicillin-clavulanate breakpoints for both combinations(4, 9, 12, 13; Jones, in press).

Ampicillin-sulbactam MlC QC. The results of a prospec-

tive multilaboratory investigation to determine ampicillin-sulbactam QC guidelines for broth microdilution MIC testsare shown in Table 6. The five centers reported 20 MICs foreach QC strain and five additional MICs for each organismdetermined with a common lot of broth microdilution panels.The control drug was ampicillin, and all MICs were withinthe QC range recommended by the NCCLS (14). The twotested ampicillin-sulbactam ratios each gave very similarresults. The proposed QC limits for the 1:1 and 2:1 ratios are

as follows: (i) for E. coli ATCC 25922, 2.0 to 8.0 ,ug/ml; (ii)for E. coli ATCC 35218, 4.0 to 16 ,xg/ml; (iii) for E. faecalisATCC 29213, 0.5 to 2.0 p.g/ml; and (iv) for S. aureus ATCC29213, 0.25 to 1.0 ,ug/ml. These are identical to those forampicillin alone, except for E. coli ATCC 35218, for whichthe mode was >32 ,ug/ml. All QC organism MICs (36 MICsfor three organisms) obtained during the ampicillin-sul-bactam combination experiments described above were

within these guidelines.

DISCUSSION

Several beta-lactamase inhibitors have been developedthat enhance the antimicrobial activity and spectra of beta-lactam drugs used in combination with them (1, 2, 7, 16).Among these inhibitors, sulbactam has received extensive invitro and in vivo evaluations combined with ampicillin andcefoperazone (15; Jones, in press). Sulbactam restores theefficacy of ampicillin which existed before the era of more

prevalent beta-lactamase production by bacterial speciessuch as members of the Enterobacteriaceae, H. influenza,N. gonorrhoeae, staphylococci, and gram-negative an-

aerobes (1, 2, 3, 4, 6, 7, 15, 16, 19, 20; Jones, in press).Sulfone beta-lactamase inhibitors have excellent affinity forthe plasmid-mediated enzymes found to be most prevalent inclinically important organisms (17) and also have marginalinhibitory activity against some type I and IV beta-lactamases (2, 15; Jones, in press). Thus, sulbactam be-comes a potent irreversible enzyme inhibitor with excellentbactericidal activity when combined with ampicillin against a

TABLE 6. Repeated microdilution tests with ampicillin, ampicillin with equal concentrations of sulbactam (1:1 ratio), and ampicillin withhalf the recommended amount of sulbactam (2:1 ratio)a

Antimicrobial agent(s) and No. of times MIC (,ug/ml) recorded'control strain 0.25 0.5 1.0 2.0 4.0 8.0 16 32 >32

AmpicillinE. coli ATCC 25922 (8 100 17)`E. coli ATCC 35218 (125)E. faecalis ATCC 29212 (0 118 7)S. aureus ATCC 29213 (0 55 45) 25

Ampicillin-sulbactam (1:1)E. coli ATCC 25922 (79 46 0)E. coli ATCC 35218 (8 113 4)E. faecalis ATCC 29212 (0 120 5)S. aureus ATCC 29213 (61 64 0)

Ampicillin-sulbactam (2:1)E. coli ATCC 25922 (43 80 2)E. coli ATCC 35218 (1 54 70)E. faecalis ATCC 29212 (0 120 5)S. aureus ATCC 29213 (41 71 13)a Each test was done with four standard control strains in five separate institutions.b MICs expressed as the concentration of ampicillin alone or in the presence of sulbactam; 125 MICs were recorded for each drug-microorganism combination.Values in parentheses are concentrations that are within the established control limits for ampicillin (13) or within the proposed limits for ampicillin-sulbactam

combined in a 1:1 or 2:1 ratio.

VOL. 25, 1987

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

1924 JONES AND BARRY

wide variety of gram-positive cocci, gram-negative entericbacilli, and anaerobic organisms (2, 16). Sulbactam also hasdirect antimicrobial activity against Acinetobacter spp.(MIC90, 4.0 itg/ml), B. catarrhalis (MIC9(, 16 ptg/ml), andpathogenic Neisseria spp. (MIC9(s, 0.31 to 3.2 ,ug/ml) (2).The synergistic interaction of sulbactam and ampicillin has

been documented for Ent>robact>r spp., Serratia spp., andindole-positive members of the tribe Proteîue, all dominatedby chromosomally mediated beta-lactamases (2, 16). Syn-ergy was also found in the absence of beta-lactamase pro-duction, indicating an alternate mode of sulbactam interac-tion with beta-lactam co-drugs (2). This finding indicates thatampicillin-sulbactam has a slightly wider spectrum of activ-ity against some members of the Enterobacteriaceae com-pared with amoxicillin-clavulanic acid (Jones, in press).ORS are also inhibited by ampicillin-sulbactam at clini-

cally achievable drug levels (8, 15; Jones, in press). It ispostulated that the beta-lactamase inhibitor "unmasks" theusable activity of ampicillin directly on the ORS, an activitythat can be two- to eightfold greater than that of penicillin-ase-resistant penicillins (15). Aminopenicillins may havegreater affinity for the altered penicillin-binding proteins thandoes oxacillin or methicillin (10). Ampicillin-sulbactam ther-apy has been successful in the rabbit endocarditis model andother animal models for infections caused by ORS (6, 15, 16,18). Therefore, ampicillin-sulbactam MICs for these ORSisolates may truly reflect the potential clinical application ofthis combination as an alternative to vancomycin.

Ampicillin-sulbactam inhibited nearly all anaerobes at the1:1 or 2:1 ratio breakpoint. Wexler et al. (19) found that99.6% of anaerobic strains were inhibited by 16/8.0 p.g/ml ofampicillin-sulbactam. These findings are similar to thosepreviously reported by our laboratory (3) for B. fragilisgroup isolates (>96%Hs) and reported here (99%7) for a largernumber of strict gram-positive and gram-negative anaerobicbacteria. Comparable in vitro efficacy has been reported forother beta-lactamase inhibitor combinations tested againstanaerobic organisms (1). Some of these comparison studieshave shown ampicillin-sulbactam to be superior to wellknown antianaerobe drugs such as cefoxitin, clindamycin,chloramphenicol, metronidazole, and moxalactam (3, 19).Human pharmacokinetic studies with parenteral ampicil-

lin-sulbactam assure levels of the active beta-lactam andinhibitor that allow susceptible breakpoint concentrations,including s4 ,ug of sulbactam per ml (5, 8). The concentra-tions of each component in body fluids and tissues usuallyvary by about an average of a 1:1 to 2:1 ratio with therecommended 2:1 dosing ratio (5). At the lowest anticipatedclinical dose, 0.5 g of sulbactam, delivered by intramuscularinjection, a peak of only 13 ,ug/mi was observed (8). Simi-larly, Brown et al. (5) found sulbactam blister fluid peakswith a 500-mg intravenous administration to be 19.2 Fg/ml, avalue comparable to concentrations in serum. Therefore, theconservative level of sulbactam for use in the in vitrosusceptibility test system appears to be c4.0 Ftg of sulbac-tam per ml. The half-lives of these two drugs in serum arevery similar, 1 to 1.5 h.

In susceptibility testing with any beta-lactamase inhibitor,it must first be considered whether the inhibitor can bedelivered at a sufficient concentration in vivo to preventbeta-lactamase hydrolysis of the co-drug. This has beenachieved for the ampicillin-sulbactam combination and at theclinically recommended dosage (2 g of ampicillin and 1 g ofsulbactam every 6 to 8 h), favors a dilution testing schemausing a 2:1 ampicillin-sulbactam ratio or a fixed concentra-tion of c4 1xg of sulbactam per ml. A 2:1 ratio test is also

used for the orally administered amoxicillin-clavulanatecombination (9). Ampicillin-sulbactam in vitro tests havepreviously used a 1:1 ratio of the components as disks or indilution test systems (4; Jones, in press). Results of studiespresented here and elsewhere indicate very little differencebetween these test ratios, but some species-specific differ-ences have been identified (7, 15, 16, 20; Jones, in press). Wesuggest the use of the 2:1 ratio for dilution test methods,which approximates the clinically used drug formulationsand maximizes the clinical spectrum of ampicillin-sulbactamas it correlates with clinical outcome statistics and humanpharmacokinetics (11; data on file, Pfizer, Inc.). This corre-lation seems most important for the staphylococci and theenteric bacilli that harbor chromosomally encoded type Icephalosporinases. We have a great concern about thetreatment of pediatric H. influenzae infections with 1:1 or 2:1ratio tests as a guide to therapy. Data presented here indicatethat the 6.7:1 pediatric formulation may not inhibit thebeta-lactamase-producing isolates, and higher sulbactamdoses could be required.The recommendations for 2:1 ratio ampicillin-sulbactam

testing by dilution methods are as follows: (i) for members oftheEEnter obacteriaceae, anaerobes ,Ac inetobaeîter spp., Pseu-domonas spp., staphylococci, H. influenzae, and B. catar-rhalis, -8.0/4.0 1tg/ml as susceptible, 16/8.0 Fxg/ml as mod-erately susceptible, and >16/16 ,ug/ml as resistant; (ii) forenterococci, streptococci, and L. monocytogenes, use thesusceptibility test results for ampicillin interpreted inaccordance with NCCLS criteria (9, 12, 13); and (iii) cross-resistance to ampicillin-sulbactam and amoxicillin-clavulan-ic acid is not sufficiently complete to allow a spectrum classfor these drug combinations, especially for members of theEnterobacteriaceae. These criteria do not greatly differ fromthose originally suggested by Barry et al. (4; Jones, in press).These proposed criteria can be amended as more informa-tion is developed regarding the efficacy of ampicillin-sulbactam against ORS and beta-lactamase-negative ampi-cillin-resistant Haemophilus and Branhamella spp.

ACKNOWLEDGMENTS

We thank R. R. Packer and J. McClung for suggestions andassistance in preparation of the manuscript. Also, QC data werecontributed by the following principal investigators: Thomas L.Gavan (The Cleveland Clinic Foundation), Peter C. Fuchs (St.Vincent Hospital and Medical Center), E. Hugh Gerlach (St.Francis Regional Medical Center), Clyde Thornsberry (Centers forDisease Control), and Leona W. Ayers (Ohio State UniversityHospital).

LITERATURE CITED

1. Appelbaum, P. C., M. R. Jacobs, S. K. Spangler, and S.Yamabe. 1986. Comparative activity of ,3-lactamase inhibitorsYTR 830, clavulanate, and sulbactam combined with ,B-lactamsagainst 3-lactamase-producing anaerobes. Antimicrob. AgentsChemother. 30:789-791.

2. Aswapokee, N., and H. C. Neu. 1978. A sulfone beta-lactamcompound which acts as a beta-lactamase inhibitor. J. Antibiot.31:1238-1243.

2a.Barry, A. L., T. L. Gavan, R. N. Jones, L. W. Ayers, P. C.Fuchs, E. H. Gerlach, and C. Thornsberry. 1987. Quality controllimits for microdilution susceptibility tests with aztreonam,imipenem, ceftriaxone, ceftazidime, ceftizoxime, cefuroximeand cefonicid. Diagn. Microbiol. Infect. Dis. 7:83-87.

3. Barry, A. L., R. N. Jones, and R. R. Packer. 1986. In vitrosusceptibility of the Bacieroidesfragilis group to cefoperazone,

J. CLIN. MICROBIOL.

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from

AMPICILLlN-SULBACTAM DILUTION TESTS 1925

ampicillin, ticarcillin and amoxycillin combined with beta-lactamase inhibitors. J. Antimicrob. Chemother. 17:125-127.

4. Barry, A. L., R. N. Jones, and C. Thornsberry. 1984. Interpre-tive standards and quality control limits for susceptibility testswith ampicillin-sulbactam combination disks. J. Clin. Microbiol.19:134-139.

5. Brown, R. M., R. Wise, J. M. Andrews, and J. Hancox. 1982.Comparative pharmacokinetics and tissue penetration of sulbac-tam and ampicillin after concurrent intravenous administration.Antimicrob. Agents Chemother. 21:565-567.

6. English, A. R., D. Girard, C. Cimochowski, J. Faiella, J. A.Retsema, and J. E. Lynch. 1986. Activity of sulbactam/ampicil-lin in screening and in discriminative animal models of infec-tions. Rev. Infect. Dis. 8(Suppl. 5):S535-S542.

7. English, A. R., J. A. Retsema, A. E. Girard, J. E. Lynch, andW. E. Barth. 1978. CP-45,899, a beta-lactamase inhibitor thatextends the antibacterial spectrum of beta-lactams: initial bac-teriological characterization. Antimicrob. Agents Chemother.14:414-419.

8. Foulds, G. 1986. Pharmacokinetics of sulbactam/ampicillin inhumans: a review. Rev. Infect. Dis. 8(Suppl. 5):S503-S511.

9. Fuchs, P. C., A. L. Barry, C. Thornsberry, T. L. Gavan, andR. N. Jones. 1983. In vitro evaluation of Augmentin by brothmicrodilution and disk diffusion susceptibility testing: regres-sion analysis, tentative interpretive criteria, and quality controllimits. Antimicrob. Agents Chemother. 24:31-38.

10. Hartman, B. J., and A. Tomasz. 1984. Low-affinity penicillin-binding protein associated with 1-lactam resistance in Staphy-lococcus aureus. J. Bacteriol. 158:513-516.

11. Lees, L., J. A. Milson, A. K. Knirsch, and K. Greenhalgh. 1986.Sulbactam plus ampicillin: interim review of efficacy and safetyfor therapeutic and prophylactic use. Rev. Infect. Dis. 8(Suppl.5):S644-S650.

12. National Committee for Clinical Laboratory Standards. 1984.Performance standards for antimicrobic disk susceptibility

tests. Approved standard M2-A3, 3rd ed. National Committeefor Clinical Laboratory Standards, Villanova, Pa.

13. National Committee for Clinical Laboratory Standards. 1985.Standard methods for dilution antimicrobial susceptibility testsfor bacteria which grow aerobically. Approved standard M7-A.National Committee for Clinical Laboratory Standards, Vil-lanova, Pa.

14. National Committee for Clinical Laboratory Standards. 1985.Proposed guideline M17-P. Alternative methods for antimicro-bial susceptibility testing of anaerobic bacteria. National Com-mittee for Clinical Laboratory Standards, Villanova, Pa.

15. Retsema, J. A., A. R. English, A. Girard, J. E. Lynch, M.Anderson, L. Brennan, C. Cimochowski, J. Faiella, W. Norcia,and P. Sawyer. 1986. Sulbactam/ampicillin: in vitro spectrum,potency, and activity in models of acute infection. Rev. Infect.Dis. 8(Suppl. 5):S528-S534.

16. Retsema, J. A., A. R. English, and A. E. Girard. 1980.CP-45,899 in combination with penicillin or ampicillin againstpenicillin-resistant Staphylococcus, Haemophilus influenzae,and Bacieroides. Antimicrob. Agents Chemother. 17:615-622.

17. Roy, A., A. Foz, and C. Segura. 1980. Plasmid-mediated beta-lactamases identified in a group of 204 ampicillin-resistantEnterobacteriaceae. J. Antimicrob. Chemother. 12:507-510.

18. Washburn, R. G., and D. T. Durack. 1981. Efficacy of ampicillinplus a beta-lactamase inhibitor (CP-45,899) in experimentalendocarditis due to SiaphvIococcus aureus. J. Infect. Dis.144:237-243.

19. Wexler, H. M., B. Harris, W. T. Carter, and S. M. Finegold.1985. In vitro efficacy of sulbactam combined with ampicillinagainst anaerobic bacteria. Antimicrob. Agents Chemother.27:876-878.

20. Wise, R., J. M. Andrews, and K. A. Bedford. 1980. Clavulanicacid and CP-45,899; a comparison of their in vitro activity incombination with penicillins. J. Antimicrob. Chemother. 6:197-206.

VOL. 25, 1987

on June 2, 2018 by guesthttp://jcm

.asm.org/

Dow

nloaded from