Reassessment of susceptibility test interpretive criteria for ticarcillin and ticarcillin-clavulanic acid.

Abstract

There are at least four different existing or proposed interpretive criteria for the disk diffusion susceptibility testing of ticarcillin and ticarcillin plus clavulanic acid (T/C). To assess these criteria, 570 gram-negative bacillary isolates were tested for susceptibility to ticarcillin and T/C by both disk diffusion and broth microdilution methods. These included 53 strains of the family Enterobacteriaceae selected for ticarcillin resistance and high-level beta-lactamase production. The broth microdilution test results were more influenced by increased beta-lactamase production than were disk diffusion results. In the absence of published data indicating which of the two standardized test methods better predicts clinical response, we conclude that until such data are available the more conservative National Committee for Clinical Laboratory Standards tentative criteria for tests with members of the Enterobacteriaceae are appropriate. Our data do not support the use of separate T/C interpretive criteria for Pseudomonas spp. and members of the Enterobacteriaceae. The appropriateness of different interpretive criteria needs further evaluation.

Full text

Page 1

Vol. 27, No. 11
JOURNAL OF CLINICAL MICROBIOLOGY, Nov. 1989, p. 2475-2481
0095-1137/89/112475-07$02.00/0
Copyright © 1989, American Society for Microbiology
Reassessment of Susceptibility Test Interpretive Criteria for
Ticarcillin and Ticarcillin-Clavulanic Acid
PETER C. FUCHS,'* RONALD N. JONES,2 AND ARTHUR L. BARRY2
St. Vincent Hospital and Medical Center, Portland, Oregon 97225,1 and The Clinical Microbiology Institute,
Tualatin, Oregon 970622
Received 25 May 1989/Accepted 31 July 1989
There are at least four different existing or proposed interpretive criteria for the disk diffusion susceptibility
testing of ticarcillin and ticarcillin plus clavulanic acid (T/C). To assess these criteria, 570 gram-negative
bacillary isolates were tested for susceptibility to ticarcillin and T/C by both disk diffusion and broth
microdilution methods. These included 53 strains of the family Enterobacteriaceae selected for ticarcillin
resistance and high-level P-lactamaseproduction. The broth microdilution test results were more influenced by
increased0-lactamaseproduction than were disk diffusion results. In the absence of published data indicating
which of the two standardized test methods better predicts clinical response, we conclude that until such data
are available the more conservative National Committee for Clinical Laboratory Standards tentative criteria
for tests with members of the Enterobacteriaceae are appropriate. Our data do not support the use of separate
T/C interpretive criteria for Pseudomonas spp. and members of the Enterobacteriaceae. The appropriateness of
different interpretive criteria needs further evaluation.
The combination of clavulanic acid (CA), a potent ,-
lactamase inhibitor, with ticarcillin has proven to be an
effective therapeutic agent (4, 11). However, there is some
recent controversy regarding the in vitro susceptibility test-
ing interpretive criteria for ticarcillin plus CA (T/C), as well
as ticarcillin alone. In 1984, we recommended the 75/10-,ug
T/C disk for disk diffusion susceptibility testing and tenta-
tively suggested the same interpretive zone diameter break-
points that were recommended for ticarcillin alone (2). The
National Committee for Clinical Laboratory Standards
(NCCLS) adopted these recommendations for disk suscep-
tibility testing of T/C (5). At least two problems exist with
these NCCLS criteria. (i) The MIC breakpoints for dilution
susceptibility testing of T/C, ticarcillin, and related penicil-
lins found in one NCCLS document (6) differ from the MIC
correlates of the disk diffusion breakpoints found in another
NCCLS document (5).
frequency of discrepancies between disk diffusion and broth
dilution susceptibility test results with T/C when members of
the family Enterobacteriaceae that produce high levels of
plasmid-mediated ,B-lactamases are tested (10).
The NCCLS recently published new tentative interpretive
criteria for this group of drugs which are designed to
minimize the discrepancies between the disk diffusion and
dilution susceptibility test documents (7, 8). These tentative
criteria provide different breakpoints for Pseudomonas spp.
and non-Pseudomonas gram-negative bacteria (NPGNB).
Sanders et al. proposed yet another set of criteria for T/C
disk diffusion tests designed to reduce discrepant categori-
zation ofmembers ofthe Enterobacteriaceae producing high
levels ofP-lactamase(10). The purpose of the present study
was to evaluate the cited interpretive criteria by simulta-
neously testing a large number of clinical isolates by disk
diffusion and broth microdilution methods. To further eval-
uate the criteria proposed by Sanders et al. (10), additional
tests were performed with strains of Enterobacteriaceae
selected because of their increased ,B-lactamase production
and resistance to pseudomonas-active penicillins.
(ii) There appears to be a high
*Corresponding author.
MATERIALS AND METHODS
Organisms. A total of 570 gram-negative bacterial isolates,
consisting of96 Pseudomonas spp. and 474 NPGNB (includ-
ing 53
ticarcillin-resistant members of the Enterobac-
teriaceae selected for high-level production of 1-lactamase),
were tested.
Antibiotic susceptibility testing. MICs were determined by
the broth microdilution method described by the NCCLS
(6). Ticarcillin and potassium clavulanate were provided as
standardized powders by Beecham Laboratories, Bristol,
Tenn. Each broth microdilution tray contained three series
of ticarcillin twofold dilutions with concentrations ranging
from 512 to 1 ,ug/ml in cation-supplemented Mueller-Hinton
broth. In one ofthese was 2 ,ug ofCA per ml in each well; the
second series contained 4 ,ug ofCA per ml in each well; and
the third series of dilutions contained no CA. ,B-Lactamase-
producing isolates of Salmonella and Shigella spp. were
collected and tested by the laboratory directed by C. Thorns-
berry, Centers for Disease Control, Atlanta, Ga. Disk diffu-
sion susceptibility tests were conducted according to the
recommendations of the NCCLS (5). Commercial 75-,ug
ticarcillin and 75/10-,ug T/C disks manufactured by Difco
Laboratories, Detroit, Mich., were used. Standard quality
control strains were included on each day of testing.
RESULTS
The MICs of ticarcillin and T/C for the 517 clinical isolates
not selected for high-level production of ,-lactamase are
summarized in Table 1. The susceptibility patterns for the
different species were similar to those previously reported
(1, 3). When the disk diffusion inhibitory zone diameters
were plotted against the MICs for this group of organisms,
the correlation coefficients for ticarcillin were 0.87 and 0.76
for members of the Enterobacteriaceae and Pseudomonas
spp., respectively; corresponding values for T/C were 0.83
and 0.90.
Table 2 summarizes the effect of 2 versus 4,ugof CAper
ml in the susceptibility testingofT/Cagainstmembers of the
Enterobacteriaceae. No significant difference was observed
2475

Page 2

TABLE 1. Ticarcillin and T/C MICs for 517 gram-negative isolates
MIC (1Lg/ml)b
T/C (2,ug/mI)'
Organism
(no. of isolates)
Ticarcillin
T/C (4 ,ug/ml)
Range
50%
Range
50%
Range
50%
Acinetobacter anitratus (30)
Citrobacter diversus (14)
Citrobacterfreundii (13)
Enterobacter aerogenes (31)
Enterobacter agglomerans (10)
Enterobacter cloacae (32)
Escherichia coli (50)
Klebsiella spp. (32)
Morganella morganii (20)
Proteus mirabilis (31)
Proteus vulgaris (20)
Providencia spp. (54)
Pseudomonas acidovorans (5)
Pseudomonas aeruginosa (55)
Pseudomonas cepacia (8)
Pseudomonasfluorescens (8)
Pseudomonas maltophilia (10)
Pseudomonas putida (5)
Pseudomonas stutzeri (5)
Serratia spp. (33)
Shigella spp. (20)
Yersinia enterocolitica (10)
Other NPGNBd (21)
aMICs of T/C reflect the concentration of ticarcillin in the combination.
b50%, MIC for 50% of isolates tested. When fewer than six isolates were tested, individual MICs are shown. The inferior number is the number of isolates
with the indicated MIC.
CConcentration of CA.
dOtherNPGNB included: fourAchromobacter xylosoxidans, threeAeromonas spp., two Cedecea lapagei, three Citrobacter amalonaticus, three Enterobacter
sakazafi, two Flavobacterium spp., and three Hafnia alvei isolates and one Salmonella typhi isolate.
2.0-16
64-128
2.0->512
2.0-256
2.0-512
<1.0->512
'1.0->512
32->512
-1.0-64
s1.0-128
2.0-64
sl1.0->512
321, 641, 1282, 256f
8.0->512
2.0->512
64->512
2.0-64
1282, 2562,512,
-11, 42, 82
2.0->512
<1.0->512
4.0-512
'1.0->512
8.0
'1.0-16
2.0-4.0
2.0->512
2.0-256
2.0-512
2.0->512
'1.0-512
2.0-512
l1.0-128
'1.0
<1.0-2.0
'<1.0-32
'14, 21
8.0->512
4.0->512
64->512
2.0-64
1281, 2564
<11, 22, 41, 81
2.0->512
'1.0-64
l1.0-8.0
'1.0->512
4.0
4.0
8.0
4.0
8.0
4.0
8.0
4.0
16
'1.0
sl.0
sl.0
'1.0-8.0
2.0-4.0
2.0->512
2.0-256
4.0-512
2.0->512
'1.0-512
2.0-512
'1.0-64
'1.0
'1.0
'<1.0-16
<14, 21
8.0->512
4.0->512
64->512
4.0-128
1282, 2562, 512,
.12, 23
2.0->512
<1.0-64
'1.0-8.0
'1.0->512
2.0
2.0
8.0
4.0
8.0
4.0
4.0
4.0
8.0
128
64
4.0
128
4.0
8.0
256
4.0
'1.0
16
'1.0
'1.0
'1.0
'.
32
16
32
16
32
16
256
16
256
512
4.0
8.0
8.0
2.0
8.0
2.0
4.0
8.0
16
2.0
4.0
8.0
256
16
with routine clinical isolates, but a significant (fourfold)
reduction in T/C MICs was seen with 4 ,ug of CA per ml in
the test system when producers ofhigh levels of P-lactamase
were tested.
Scattergrams for ticarcillin MICs plotted against inhibitory
zone diameters are displayed in Fig. 1 for these organisms,
illustrating the two NCCLS breakpoints. Table 3 provides
the calculated error rates for each subset of organisms when
each of the sets of interpretive criteria was applied: criteria
A, original NCCLS-recommended criteria (5); criteria B,
tentative NCCLS criteria for Pseudomonas spp. (7); criteria
C, tentative NCCLS criteria for NPGNB (7); and criteria D,
proposed criteria of Sanders et al. for members of the
Enterobacteriaceae (10).
Application of criteria A to ticarcillin yielded high very
major error rates (false-susceptibility disk diffusion results)
for NPGNB (2.6%) and Pseudomonas spp. (3.1%). These
were reduced to acceptable levels with the application of the
tentative NCCLS criteria, i.e., to 0.5% for NPGNB and to
1.0% for Pseudomonas spp. Application ofcriteria B yielded
the lowest error rates for both NPGNB and Pseudomonas
spp.
The corresponding scattergrams for T/C are given in Fig.
2. NPGNB (Fig. 2A and B) had an unacceptably high very
major error rate with criteria A (1.4%), but these errors were
eliminated by the other criteria. However, the total error
rate was highest with criteria D (10.0%), owing exclusively
to minor discrepancies.
No very major errors occurred with T/C testing of Pseu-
domonas spp. by any of the criteria evaluated. Criteria A
and C yielded the lowest total error rates with this genus
(13.5%), but criteria A and B gave a 2.1% major error rate
compared with no major errors with criteria C (Table 3).
Table 4 summarizes the T/C error rates obtained when the
same interpretive criteria were applied to the 382 isolates of
Enterobacteriaceae among the 517 bacteria tested, as well as
the 53 members of the Enterobacteriaceae producing high
levels of ,-lactamase. Scattergrams for these organisms with
TABLE 2. Effects of 2 and 4 p.g of CA per ml on susceptibility to ticarcillin of isolates of Enterobacteriaceae
Agent
(CA concn;
(CAml)
Cumulative % susceptible at ticarcillin concn (p.g/ml) of:
4.0
8.0
16
32
Enterobacteriaceae
(no. of isolates)
1.0
2.0
64
128
256
512
>512
Unselected (382)
Ticarcillin
T/C (2)
T/C (4)
18
23
23
29
36
43
46
57
59
54
72
73
58
80
82
63
83
86
68
87
92
74
90
95
81
96
97
84
98
98
100
100
100
High-level-p-lactamaseproducers (53)
Ticarcillin
T/C (2)
T/C (4)
0
0
0
0
0
0
0
0
0
0
0
4
0
0
8
0
0
0
0
4
8
100
100
13
51
34
75
70
96
92
100
21
J. CLIN. MICROBIOL.
2476
FUCHS ET AL.

Page 3

TICARCILLIN AND T/C SUSCEPTIBILITY CRITERIA
2477
61S
A
214
3 6432
3
11
1
111
231112
~~~12822212311
6 115936 1 2 1
1
22
.
1 434914544
2562141181374
25
1
10
à
30
>N
2
C
1
1
22 21
1
1
2
1
13232
1
1
368421
1~~~~
11
.
411
41
1
2 1
>52-
se-
256-
la-
64-
32-
16-
8-
1
zm-
se-
la.
641
32-
S1-
8-
4.
2-
D
l
a
25
3
>a
9
2
8453321
I1
2
361
214
3 6432
3
1
B
1
231112
12622212311
61159 836121
216155464
1434914544
1
11
2
21
1
1
2562141M81374
il
- 0--
2D
25
35
2
.
1
D
I
1
1
22 21
1
1
11
1
1
21
13
232 2
11
368421
11
I
411
I
1
I
2 1
10
i
mi
25
30 >à
FIG. 1. Ticarcillin scattergrams. (A and B) NPGNB with former (A; criteria A) and proposed (B; criteria C) NCCLS breakpoints; (C and
D) Pseudomonas spp. with former (C; criteria A) and proposed (D; criteria B) NCCLS breakpoints. x Axis, Zone diameters (millimeters); y
axis, MICs (micrograms per milliliter).
T/C are shown in Fig. 3. No major errors occurred with any
of the criteria. The very major error rates were <l% for ail
criteria with the 382 unselected clinical strains. For the 53
producers ofhigh levels of ,-lactamase, the very major error
rates were extremely high for criteria A, B, and C (77, 59,
and 9%, respectively), but no very major errors occurred
with criteria D. The minor and total error rates were more
than twice as high for criteria D (10.5%) than for any of the
TABLE 3. Ticarcillin and T/C disk susceptibility error rates with 517 gram-negative isolates using four interpretive criteria
No. of interpretive errors (% of total tested)b
Criteria
Organisms
(no. of isolates)
Ticarcillin
T/C
V.
Major
Minor
Total
Mj
Major
Major
Minor
Total
Major
A (S = C64 ,ug/ml, 215 mm;
I = 12-14 mm; R = 2128
,ug/ml,
Pseudomonas spp. (96)
NPGNB (421)
Total (517)
3 (3.1)
11 (2.6)
14 (2.7)
4 (4.2)
0
4 (0.8)
6 (6.3)
16 (3.8)
22 (4.2)
13 (13.5)
27 (6.4)
40 (7.7)
0
6 (1.4)
6 (1.2)
2 (2.1)
0
2 (0.4)
11 (11.5)
14 (3.3)
25 (4.8)
13 (13.5)
20 (4.8)
33 (6.4)
u11 mm)
B (S = C64 ,ug/ml, 215 mm;
MS = 128ktg/ml,12-14
mm; R = 2256 ,ug/ml,
u11 mm)
Pseudomonas spp. (96)
NPGNB (421)
Total (517)
1 (1.0)
1 (0.2)
2 (0.4)
4 (4.2)
0
4 (0.8)
11 (11.5)
21 (5.0)
32 (6.2)
16 (16.7)
22 (5.2)
38 (7.4)
0
0
0
2 (2.1)
0
2 (0.4)
12 (12.5)
19(4.5)
31 (6.0)
14 (14.6)
19 (4.5)
33 (6.4)
C (S = -'16 ,ug/ml, 220 mm;
MS = 32-64 ,Ug/ml, 15-19
mm; R =2128 ,ug/ml,
<14 mm)
Pseudomonas spp. (96)
NPGNB (421)
Total (517)
1 (1.0)
2(0.5)
3 (0.6)
1 (1.0)
0
1 (0.2)
22 (22.9)
29 (6.8)
51 (9.9)
24 (25.0)
31 (7.4)
55 (10.6)
0
1 (0.2)
1 (0.2)
0
0
0
13 (13.5)
21 (5.0)
34 (6.6)
13 (13.5)
22 (5.2)
35 (6.8)
D (S= <16 ,ug/ml, -23 mm;
MS = 32-64 ,Ug/ml, 19-22
mm; R = 2128 p.g/ml,
s18 mm)
Pseudomonas spp. (96)
NPGNB (421)
Total (517)
0
1 (0.2)
1 (0.2)
3 (3.1)
0
3 (0.6)
35 (36.5)
52 (12.4)
87 (16.8)
38 (39.6)
53 (12.6)
91 (17.6)
0
0
0
0
0
0
33 (34.4)
42 (10.0)
75 (14.5)
33 (34.4)
4 (10.0)
75 (14.5)
aCriteria are described in Results. S, susceptible; I, intermediate; R, resistant; MS, moderately susceptible.
bV. Major, Susceptibility zone diameters with resistance MICs; Major,resistance zone diameters withsusceptibility MICs; Minor,zone diameter or MIC (but
not both) indicating moderate susceptibility or intermediate susceptibility.
9
945332
2
2
3
>512
256.
64-
32~
1.
4.
2-
O1-
,m-
>512-
256-
64-
8-
4.
2
<-1
VOL. 27, 1989
1
1

Page 4

2478
FUCHS ET AL.
2
2
1 3 2 1
1 2
2 1 2
3 1li 3 1
111
A
231
111211311
174 0 1 8 332
369181366
2
2
22
1
34
8 212151111 24
2539875542
22
19081711386
25
1
10 15
N2a
3
?5
S
7
1
121
211
1 2
121
1
1
631122
246641
~~~~~1
1
11111I
1
2
11
S
«1
12
11
2
3
>512
56-
lu-
64-
32-
1-
8-
4.
2-
>,m-
512-
S3-
la.
64-
32-
16-
8:
4-
2-
<-1-
6
2
2
212
1 3 2 1 2 1 2
1
3 1 1 3 1
B
r11
231
111211311
174 8 1 8 332
36911366
2 s 821215111
2
22
1
34
2 4
2509875542
22
19817113986
25
1
~~I
I
3
20
3
10
15
30
>35
8
1
7
D
1
121
211
1
121
1 12
1
631122
246641
1
11
1
2
1
~~~~~1
111
1 2
il1
2
3
b
15
20
25
3N
>N m
o
i5
FIG. 2. T/C scattergrams. See the legend to Fig. 1.
other criteria with the 382 unselected strains. With the 53
producers of high levels ofP-lactamase, however, the total
error rate for criteria D (32%) was less than half that of any
of the other criteria.
Antagonism with the combination of ticarcillin and CA
(defined as a fourfold-or-greater increase in ticarcillin MIC)
was observed with 12 ofthe 517 isolates studied (Table 5), all
members of the Enterobacteriaceae. Only with two organ-
isms did the disk diffusion test demonstrate a significantly
smaller zone diameter with T/C.
20
2s
;-
DISCUSSION
None of the cited tentative or proposed interpretive crite-
ria for the disk diffusion susceptibility testing of T/C is
strongly supported by the data in this study. On the other
hand, these data do raise a number of questions that should
be resolved before definitive interpretive criteria for T/C and
related drugs are selected.
Microbiologically, our data provide no support for the
concept of different interpretive criteria for Pseudomonas
TABLE 4. T/C disk diffusion susceptibility error rates with unselected and ticarcillin-resistant members of the Enterobacteriaceae with
four different interpretive criteria
Criteria
Organisms
(no. of isolates)
No. of T/C errors(rate)b
Major
V. Major
Minor
Total
A (S = <64 jtg/ml, .15 mm; I =
12-14 mm; R = .128 ,ug/ml,
'11 mm)
Unselected (382)
High-level-p-lactamaseproducers (53)
4 (1.0)
41 (77.4)
0
0
14 (3.7)
4 (7.5)
18 (4.7)
45 (84.9)
B (S = c64 ,ug/ml, .15 mm; MS
= 128 p.g/ml, 12-14 mm; R =
.256 jtg/ml, '11 mm)
Unselected (382)
High-level-p-lactamase producers (53)
0
0
0
17 (4.5)
13 (24.5)
17 (4.5)
44 (83.0)
31 (58.5)
C (S = <16 ,ug/ml, .20 mm; MS
= 32-64 ptg/ml, 15-19 mm; R
=-128 ,tg/ml, <14 mm)
Unselected (382)
High-level-,-lactamase producers (53)
1 (0.3)
5 (9.4)
0
0
18 (4.7)
41 (77.4)
19 (5.0)
46 (86.8)
D (S = a<16 p.g/ml, .23 mm; MS
= 32-64 ,ug/ml, 19-22 mm; R
= >128 ,u.g/ml, .18 mm)
aSee footnote a to Table 3.
bSee footnote b to Table 3.
Unselected (382)
High-level-p-lactamaseproducers (53)
0
0
0
0
40 (10.5)
17 (32.1)
40 (10.5)
17 (32.1)
6
2
>52 -
512-
256-
64-
32-
1 -
8-
4-
2,-
512 -
256-
lau-
64-
32-
16-
8-
4-"
2-
<1 -
J. CLIN. MICROBIOL.
1
1
1
111
1

Page 5

TICARCILLIN AND T/C SUSCEPTIBILITY CRITERIA
2479
5 1
22
1 3 2 1
1 2
2 1 2
31i
11 1
A
i1 1
1
131
221
1 1 121 1 21 1
1 S 3 8 1 7 3 2 2
36511266
2 5 8 1717149 9 2 4
2
2478854432
34
252
42343
1
c
2
1 3 31 1 1
ilI
il
...-.~~~
>512-
512-
m6
lu-
84-
32-
1.
8 2
4.
2
5121
1u
52~
84.
1.
4.
2-
5
1
2
13 21 2 1 2
212
2
1
31i1 1
11 1
B
131
2 22
I
1 1j1 2 1 1 2 1 1
i15:,3817322
36:5151266
2 581717149 2 4
2: 2479854432
_:''11
34
1881811986
12
~~252
4 4234 3
2
.
r:î
D
2
1,3
1 1 1II
Il
AA
ieis
253
>a
à
iD
i
à
2
>I
FIG. 3. T/C scattergrams for members of the Enterobacteriaceae. (A and B) All 382 strains ofEnterobacteriaceae from current study with
former (A) and proposed (B) NCCLS breakpoints. Zone diameter breakpoints (broken lines) in panels B and D are those proposed by Sanders
et al. (10). (C and D) Strains of Enterobacteriaceae (n = 53) selected for high-level 13-lactamase production with the same breakpoints as in
panels A and B, respectively. x Axis, Zone diameters (millimeters); y axis, MICs (micrograms per milliliter).
spp. and NPGNB (Table 3). We recognize that the MICs of
the drugs tend to be higher for Pseudomonas spp. than for
NPGNB and that in the absence of susceptibility data for the
specific infecting organisms it is appropriate to treat with
higher doses when Pseudomonas spp. are present or sus-
pected. However, we are unaware ofany data to support the
assumption that Pseudomonas aeruginosa is more suscepti-
ble than Escherichia coli when the MIC for each is the same;
TABLE 5. MICs and zone diameters of ticarcillin and T/C
against 12 members of the Enterobacteriaceae with which
the combination drug gave antagonistiC results
MIC
(plg/ml)
Zone diam
(mm)
Organism and
isolate no.
Ticar
T/C-2T/C-4
TicarT/C
Enterobacter aerogenes 86
E. aerogenes K75
E. agglomerans 365-82
E. cloacae 69
E. sakazakii 125
Morganella morganii
179
180
184
286-82
F46
F47
Serratia marcescens 442-82
4.0
16
4.0
'1.0
32
16
64
16
32
64
32
25
22
28
28
23
24
14
29
28
2.0
8.0
128128
6
'1.0
'1.0
2.0
2.0
'1.0
2.0
4.0
4.0
8.0
8.0
16
16
8.0
16
8.0
8.0
4.0
8.0
4.0
8.0
16
28
27
27
28
28
28
26
28
30
27
28
28
27
26
aTicar, Ticarcillin with no CA; T/C-2, ticarcillin with 2ktgof CA per ml;
T/C-4, ticarcillin with 4k.gof CA per ml.
e.g., with a T/C MIC of 64 ,ug/ml, P. aeruginosa would be
considered susceptible but E. coli would be moderately
susceptible according to the latest NCCLS proposal (7). This
double standard is potentially confusing to both the micro-
biologist and the clinician and, in our opinion, should be
avoided unless and until there are clinical outcome data to
support it.
Production of high levels of P-lactamase by certain mem-
bers ofthe Enterobacteriaceae may prove to be a significant
problem for T/C susceptibility testing. Both microdilution
and disk diffusion methods correctly recognized all 53 strains
producing high levels ofg-lactamase to be resistant to
ticarcillin. On the other hand, significant interpretive dis-
crepancies occurred with those strains between the two
methods of testing when T/C was tested, which confirms the
findings of Sanders et al. (10). That the increasedP-lacta-
mase production is a major contributor to the discrepant
results is supported by the fact that ticarcillin MICs for the
strains producing high levels of ,B-lactamase were fourfold
lower when ticarcillin was combined with 4 ,ug of CA per ml
than when it was combined with the currently recommended
2 ,ug of CA per ml (Table 2). A similar effect of higher
concentrations of CA on ticarcillin MICs against ticarcillin-
resistant enteric bacilli was previously reported (9). By
contrast, the T/C MICs for the unselected members of the
Enterobacteriaceae were reduced less than one-quarterof a
twofold dilution when 4 ,ug ofCA was used in lieu of 2,ug/ml
(Tables 1 and 2). It appears that increased ,3-lactamase
production has a greater effect on the broth microdilution
susceptibility test than on the disk diffusion test under the
512
2%-
32-
84
2
<1
>52-
512-
2S
la-
64-
32-
1s-
8-
4-
2-
<-1
VOL. 27, 1989
.il