Journal of Antimicrobial Chemotherapy (1992) 30, 811-820
Antimicrobial susceptibility of anaerobic bacteria in Australia Sharon C A. Chen, Thomas Gottlieb, Jennifer M. Palmer, GabrkOe Morris and Gwendolyn L. Gilbert
Department of Clinical Microbiology, Centre for Infectious Diseases and Microbiology, Westmead Hospital, Westmead NSW 2145, Australia The susceptibilities of 900 clinical isolates of anaerobic bacteria to 14 antimicrobial agents were determined by an agar dilution technique. Chlorampbcnicol, imipenem and metronidazole were found to be active against virtually all of the strains; only a single Bacteroides fragilis isolate was resistant to both imipenem and metronidazole. The addition of clavulanic acid to amoxytillin and ticarcillin potentiated the activities of these agents against all anaerobes including members of the B. fragilis group. Ampkallin/sulbactam and clindamycin were the next most active agents, 91 and 89% of isolates respectively being susceptible. Seventy-three per cent of the bacteria tested were susceptible to cefoxitin and 65% to cefotetan, with the MICs of almost 50% of the isolates clustering between 16 and 32 mg/L. There was also clustering around the breakpoint (64 mg/L) of piperacillin. Azithromycin exhibited poor activity against the B. fragilis group; only 18% of isolates were susceptible to < 4 mg/L. However, 92% of non-2), fragilis Bacteroides group strains were susceptible to this agent We conclude that imipenem, metronidazole, chloramphenicoL, ticarcillin/clavulanate, co-amoxiclav and, to a lesser extent, ampicillin/sulbactam are suitable as empirical therapy for infections caused by anaerobic bacteria.
Introduction
Antimicrobial susceptibility testing of anaerobic bacteria is difficult to standardize and is therefore not routinely performed. There is also disagreement about the ability of in-vitro tests to predict responses in vivo, particularly with the newer cephalosporins. However, occasional surveys provide guidelines for effective empirical treatment, enable patterns of bacterial resistance to be monitored and are necessary to evaluate the activities of new agents. Regional differences in the antimicrobial resistance patterns of anaerobes, as well as intracentre variations with time, are well documented (Cuchural et al., 1984; Tally et al., 1985). Recent studies have demonstrated increasing resistance to /Mactams, not only among the Bacteroides fragilis group but also among non-i?. fragilis Bacteroides and Fusobacteritan spp. (Cuchural et al., 1984; Appelbaum, Spangler & Jacobs, 1990). The addition of /Mactamase inhibitors such as clavulanic acid and sulbactam reduces the MICs of some /Mactam antibiotics to susceptible concentrations but this effect is not invariable (Appelbaum et al., 1990; Jacobs, Spangler & Appelbaum, 1990). Moreover, resistance to both imipenem and metronidazole amongst clinical isolates Correspondence to: Profeoor G. L. Gilbert. 811 0305-7453/92/120811 + 1 0 S08.00/0
© 1992 The Brituh Society for Antimicrobial Chemotherapy
812
S. C A. Chen et aL
belonging to the B.fragilis group has recently been noted in the USA and Europe (Breuiil et al., 1989; Jacobs et al., 1990). At Westmead Hospital antibiotic susceptibility testing of anaerobic bacteria has been performed periodically. Earlier results (Collignon, Munro & Morris, 1988) showed that metronidazole had the most predictable in-vitro activity and that resistance to /Mactams (imipenem was not tested) was frequent and clinically important. We report here additional data of the susceptibilities of recent anaerobic isolates to 14 antimicrobial agents, including several not included in the previous report. Materials and methods
Bacterial strains Nine hundred recent clinical isolates of anaerobic bacteria were included in this study which was carried out between 1986 and 1991. Repeat isolates from the same patients were excluded. Organisms were identified by conventional biochemical tests and gasliquid chromatography as specified in the Virginia Polytechnic Institute (Holdeman, Cato & Moore, 1977) and Wadsworth Anaerobic Laboratory Manuals (Suttcr et al., 1979). To facilitate comparison with previous results the old classification of non-B. fragilis group Bacteroides spp. was used (Shah & Collins, 1988). Antimicrobial agents The following antibiotics (and their respective breakpoint MICs) were tested: penicillin G (4 mg/L), metronidazole (16 mg/L), clindamycin (4 mg/L), cefotetan (32 mg/L), cefoxitin (32 mg/L), piperacillin (64 mg/L), ampicillin (4 mg/L), ampicillin/ sulbactam (16/8 mg/L), co-amoxiclav (8/4 mg/L), ticarcillin/clavulanate (64/2 mg/L), imipenem (8 mg/L), chloramphenicol (16 mg/L), erythromycin (4 mg/L) and azithromycin (4 mg/L). The breakpoints are in accordance with the current recommendations of the National Committee for Clinical Laboratory Standards (1990), with the exception of those for erythromycin and azithromycin which were based on previously established, achievable serum or tissue concentrations (Barry & Jones, 1988). The MICs of the /Mactam//Mactamase inhibitor combinations were expressed in terms of the concentrations of the /Mactam component. Ticarcillin was combined with a fixed concentration (2 mg/L) of clavulanic acid and co-amoxiclav consisted of amoxycillin and clavulanic acid in a 2:1 ratio. Sulbactam and ampicillin were combined in a 1 : 1 ratio according to the manufacturer's instructions. Penicillin, metronidazole, clindamycin and cefoxitin were evaluated throughout the study period. Imipenem was tested against 634 isolates, cefotetan against 402, ampicillin and ampicillin/sulbactam against 266 and co-amoxiclav, ticarcillin/clavulanate, piperacillin, chloramphenicol, erythromycin and azithromycin against 232. MIC determinations MICs were determined by an agar dilution method. Portions of five or more colonies from 48-72 h cultures growing on supplemented blood agar (Dowell & Hawkins, 1979) were inoculated into the thioglycollate broth supplemented with haemin and vitamin K. The broths were incubated anaerobically for 48 h. After the turbidity was adjusted to a 0-S McFarland standard the organisms were inoculated with a Denley
Susceptibility testing of anaerobic bacteria
813
replicator on to Wilkins-Chalgren agar containing doubling dilutions of antibiotic (range 0-03-128 mg/L). The final inoculum density was approximately 103 cfu per spot. Antibiotic-free control plates were inoculated at the beginning and end of each series. One set of control plates was incubated aerobically and the other was incubated with the test plates at 37°C for 48 h in an anaerobic chamber containing 80% nitrogen, 10% hydrogen and 10% carbon dioxide. The MICs for each organism were recorded as the lowest concentration yielding no growth, a fine haze, one discreet colony or multiple tiny colonies (NCCLS, 1990). Organisms which were non-viable or where the control plates were contaminated were excluded. B. fragilis ATCC 25285, Bacteroides thetcdotomicron ATCC 29741, Clostridium pefringens ATCC 13124 and Peptostreptococcus asaccharolyticus ATCC 29743 were included as controls with each batch tested. Results The susceptibilities of the isolates to the 14 antibiotics are shown in the Table. Bacteroides spp. predominated, accounting for 466 (52%) of the 900 organisms studied; 310 of these belonged to the B. fragilis group and included 144 B. fragilis strains. The next commonest bacteria were the non-sporing, Gram-positive bacilli, including 130 Propionibacterium acnes isolates. Isolates of the B. fragilis group were the least susceptible overall. Within the group, B. fragilis isolates were the most susceptible, followed by members of the indolenegative non-fragilis B. fragilis group. The indole-positive group strains were the least susceptible to /7-lactam antibiotics. Cefotetan and cefoxitin exhibited similar activities against the B. fragilis group (55% and 56% of strains respectively were susceptible to < 16 mg/L) but cefotetan was slightly more active against B. fragilis (88% vs 79% of strains). Raising the breakpoint to 32 mg/L increased the percentages of B. fragilis group isolates which were inhibited to 73% and 63% for cefoxitin and cefotetan, respectively. The cefoxitin MICs for 95 of 274 (35%) of these isolates were clustered between 16 and 32 mg/L. Similar clustering of 49 of 108 (45%) isolates was observed for cefotetan. Addition of a /Mactamase inhibitor lowered the MICs of the penicillins, rendering most of the Bacteroides spp. susceptible. However, ampicillin/sulbactam was active against only 79 of 87 (91%) of the B. fragilis group strains. The majority of isolates were susceptible to piperacillin but the MIC^s for all of the strains in the B. fragilis group were clustered around the breakpoint (64 mg/L). Resistance to chloramphenicol, metronidazole and imipenem was rare. Only a single strain of B. fragilis was resistant to imipenem (MIC 8 mg/L) and metronidazole (MIC 16 mg/L). Metronidazole had predictably low activity against the Gram-positive, nonsporing bacilli. Little resistance to clindamycin was detected amongst the isolates with the exception of the indole-negative non-fragilis B fragilis group; nine of 50 (18%) isolates, including seven of the 24 Bacteroides distasonis strains tested, were resistant to this agent. Sixty-three per cent and 56% of B. distasonis isolates were also resistant to cefotetan and cefoxitin, respectively. /?-Lactam antibiotics in general were highly active against Fusobacterium spp.; only two strains demonstrated reduced susceptibility to penicillin and imipenem. Azithromycin, like erythromycin, exhibited poor activity against B. fragilis group isolates; only 18 of 102 (18%) isolates were susceptible to < 4 mg/L. However, 23 of 25 (92%) isolates of the non-fi. fragilis Bacteroides species, 91% of anaerobic
814
S.CA. CbtaetaL Table. MICs* of 14 antimicrobial agents for 900 anaerobic isolates
Organism
Antimicrobial
MIC a MIC*
B. fragUis group (all spedes)
penicillin metronidazole clindamycin cefotetan cefoxitin ampicillin ampitiUin/sulbactam azithromycin erythromycin co-amoxidav ticarcillin/clavulanate piperacillin imipenem chloramphenicol penicillin metronidazole clindamycin cefotetan cefoxitin ampicillin ampidlHn/sulbactam azithromycin erythromydn co-amoxidav ticarcillin/clavulanate piperacillin imipenem chloramphenicol
16 05 025 4 16 32 2 16 8 05 012 8 006 2 16 05 025 4 8 32 1 16 8 025 056 4 003 2 16 05 012 32 8 32 2 16 8 025 006 4 003 2 32 05 05 32 16 32
B. fragUis
non-fragUis B. fragUis
penicillin
group (indole—ve)
metronidazole dindamycin cefotetan cefoxitin ampicillin ampicillin/sulbactam azithromycin erythromydn co-amoxidav ticarcillin/clavulanate piperacillin imipenem chloramphenicol
non-fragUis B. fragUis
penicillin
group (indole+ ve)
metronidazole dindamycin cefotetan cefoxitin ampicillin
128 1 8 164 32 128 8 128 128 4 4 128 05 8 128 1 1 32 16 128 4 128 32 1 025 32 025 8 128 1 8 64 32 128 16 128 128 4 4 64 005 4 128 1 2 64 32 128
Range
% of strain's susceptible at breakpoint concentrations*
003-128 003-16 003-16 003-64 003-128 025-128 025-16 05-128 025-128 003-16 003-128 003-128 O03-8 O03-8 O12-128 006-4 003-16 025-64 025-128 4-128 05-8 05-128 1-128 012-128 003-128 003-128 O03-8 025-4 012-128 006-16 003-16 1-128 O25-128 025-128 025-16 05-128 1-128 003-16 003-128 003-128 O03-1 025-4 012-128 006-16 003-16 003-128 003-128 4-128
16 99 89 63 73 2 91 18 42 100 99 91 99 100 7 100 98 95 84 0 100 17 42 100 97 97 99 100 21 100 82 75 78 5 88 4 23 100 100 90 100 100 6 100 96 16 52 0
Stnceptibffity testing of anaerobic bacteria
815
Table.—continued
Organism
Antimicrobial
non-fragilis B. fragilis ampitillin/sulbactam group (indole+ve) azithromycin erythromycin co-amoxiclav ticarrillin/davulanate piperacillin nxupcncm chloramphenicol
M I C , MIC* 2 4 16 128 4 64 0-5 4 0-5 4 32 128 0-12 05 4 4
Range
% of strain's susceptible at breakpoint concentrations*
05-16 1-128 025-128 003-^t 003-4 003-128 003-1 003-4
97 32 68 100 100 80 100 100
non-B. fragilis group Bacteroides spp.
penicillin metronidazole clindamycin cefotetan cefoxitin ampicillin ampicillin/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanatc piperacillin umpenem chloramphenicol
025 0-12 0-5 0-5 0-5 0-5 0-25 0-5 1 006 003 2-0 003 1
8 1 1 8 4 32 2 2 8 O5 1 8 006 2
003-128 003-4 003-16 003-64 003-128 003-128 003-16 003-128 003-128 003-8 003-16 003-128 003-05 003-8
74 100 94 93 97 67 97 92 88 100 100 90 100 100
Fusobacterium spp.
penicillin metronidazole clindamycin cefotetan cefoxitin ampiciUin ampicillin/sulbactam azithromycin erythromycin co-amoxiclav ticarcLUin/davulanate piperacillin imipenem chloramphenicol
003 006 012 006 012 006 006 16 128
003-16 003-1 003-4 O03-2 003-8 003-8 003-2 025-16 4-128 —
92 100 (100) 100 100 (83) (100) (50) (50) —
003 —
025 025 012 2 4 8 2 16 128 — — — 4 —
— O03-4 —
— (89) —
peniciUin metronidazole clindamycin cefotetan cefoxitin ampicillin ampicilUn/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanate piperacillin
025 050 006 025 05 025 025 4 16 006 05 8
8 2 012 8 1 1 05 16 64 1 128 128
003-32 003-8 003-2 003-8 003-128 012-1 012-05 006-64 003-64 003-1 003-128 1-128
51 95 100 100 95 100 100 64 36 100 89 78
Veillonella spp.
c
— —
816
S. C. A. Cben et aL Table.—continued
Organism
Antimicrobial
Veillonella spp.
imipencm chloramphcnicol
Anaerobic Grampositive cocci
Propionibacterium acnes
% of strain's susceptible at breakpoint concentrations*
MIC* MIQo
Range
006 2
006 8
003-05 005-8
100 100
penicillin metronidazole clindamycin cefotetan ccfoxitin ampicilUn ampicillin/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanate pipcracillin imipenem chloramphenicol
0-12 0-50 0-25 0-12 0-12 0-12 012 025 1 003 003 006 003 1
2 2 05 4 1 8 1 16 32 1 2 16 006 1
003-64 003-8 003-4 003-64 003-64 003-32 003-32 003-128 003-128 003-2 003-64 003-128 003-05 003-4
94 98 96 96 98 89 93 91 91 100 100 93 100 100
penicillin metronidazole clindamycin cefotetan ccfoxitin ampidllin ampicillin/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanate piperacillin imipencm chloramphenicol
003 16 006 006 006 012 006 006 003 003 003 025 003 05
006 003-^ 32 4-32 012 003-1 0-25 003-32 025 003-128 025 006-05 025 003-025 012 0O3-8 0-12 003-64 006 003-0-12 003 003-1 025 003-1 003 003 1 006-1
99 19 100 96 99 100 100 97 97 100 100 100 100 100
Clostridium perfringens penicillin metronidazole clindamycin cefotetan cefoxitin ampicillin ampicillin/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanate piperacillin
012 025 025 025 003 025 006 2 1 003 006 006 003 2
chloramphenicol Anaerobic Grampositive bacilli (non-sporing)
penicillin metronidazole clindamycin cefotetan
006 16 006 012
05 2 2 025 2 2 1 2 2 025 16 1 006 4
003-16 006-2 003-32 003-2 006-64 003-2 003-1 05-2 1-2 003-025 003-16 003-1 003-0-5 1-4
97 100 96 100 100 (100) (100) (100) (100) (100) (100) (100) 100 (100)
05 32 05 4
003-64 003-32 003-4 003-128
96 15 99 100
Sosctfrtibflfty testing of anaerobic bacteria
817
Table.—continued
Oroflnigin
Antimicrobial
M I C , MIC*
Anaerobic Grampositive bacilli (non-sporing)
ccfoxitin ampidllin ampirilHn/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanate piperacillin mupenem chloramphenicol
012 0-12 0-12 0-06 003 003 003 025 003 O5
1 05 05 1 1 O06 003 1 O03 1
Range
% of strain's susceptible at breakpoint concentrations*
003-128 O03-O5 O03-O5 O03-8 O03-8 003-012 O03 O03-4 003-2 006-1
98 93 100 98 94 100 100 100 100 100
*MICf expressed in mg/L. *See matf-riiili and method* for breakpoint concentrations; where < 10 (trains were tested, numbers are within brackets. *Not tested.
Gram-positive cocci, 98% of Gram-positive bacilli and all eight isolates of C. perfringens were susceptible. Discussion The results of this investigation revealed that there have not been any significant changes in the overall susceptibility patterns of most of the antibiotics tested since an earlier study undertaken by this laboratory (Collignon et al., 1988). Chloramphenicol, co-amoxiclav, ticarcillin/clavulanate, imipenem and metronidazole remain highly active against almost all anaerobes. Predictably, the non-sporing, Gram-positive bacilli, particularly P. acnes, were not susceptible to metronidazole. The latter has become increasingly important as a pathogen in neurosurgical patients; penicillin G remains the drug of choice for use against this organism. Of the /J-lactams, imipenem, ticarcillin/clavulanate and co-amoxiclav were the most active agents against the B. fragilis group. The single strain resistant to imipenem was identified as B. fragilis; it was also resistant to metronidazole. Previous studies in Australia failed to detect resistance to imipenem (Downes & Andrew, 1988); however recent international experience suggests that resistance to this drug is emerging (Breuil et al., 1989; Cuchural et al., 1990; Jacobs et ai., 1990). Metronidazole resistance has not previously been documented in Australia. Broad-spectrum /Mactam antibiotics such as the cephamycins are attractive as single agents for both the prevention and treatment of mixed aerobic and anaerobic infections. Our results demonstrate substantial resistance to both cefoxitin and cefotetan amongst the B. fragilis group. These results are similar to those found elsewhere (Downes & Andrew, 1988; Cuchural et al., 1990). However, the MIQQS of cefoxitin and cefotetan were 32 and 64 mg/L, respectively, concentrations that are achievable in vivo when high doses of these agents are used. It was noteworthy that the MICs of
818
S. C. A. Chen et aL
both antibiotics clustered around 16-32 mg/L, highlighting the difficulty in selecting a suitable breakpoint for the cephalosporins. This problem has been identified by other investigators (Wexler et al., 1986; Collignon et al., 1988). Cefotctan has a spectrum of activity similar to that of cefoxitin, although it is less active against the indole-positive non-fragilis B.fragilis group. Cefotetan has the advantage of a longer half-life and can be admininstered twice daily. However, activity against Bacteroides spp. is not predictable and there is a risk of hypoprothrombinaemia. Neither cephamycin can currently be considered a reliable choice as empirical therapy for serious anaerobic infections in Australia. Piperacillin remains active against the majority of anaerobes although there was clustering of the MICs around the breakpoint (64 mg/L); this has been noted by others (Barry & Fuchs, 1991). Piperacillin MICs were particularly high amongst isolates which were also resistant to cefoxitin. /J-Lactamase production is commonplace amongst all Bacteroides spp. as well as Fusobacterium spp. (Appelbaum et al., 1990; Jacobs et al., 1990). The addition of /J-lactamase inhibitors such as clavulanic acid and sulbactam renders most Bacteroides isolates susceptible to /Mactams in vitro, with a more than five-fold reduction in the MICJOS (Appelbaum et al., 1990; Cuchural et al., 1990; Wexler, Molitoris & Finegold, 1991). Our results confirm this observation, although sulbactam was less effective than clavulanic acid. Sulbactam was combined with ampicillin in a 1 : 1 ratio which provides adequate sulbactam to inhibit the /Mactamase activity of Bfragilis (Lang & Thomas, 1985). However, previous investigators have used this combination in a 2 : 1 ratio (Appelbaum, Spangler & Jacobs, 1991; Wexler et al., 1991) and found that the activity against all Bacteroides spp. was comparable to that of co-amoxiclav. The variation in technique and breakpoint concentrations in different centres makes comparison of results from different studies difficult (Brown, 1991). The incidence of clindamycin resistance among members of the B. fragilis group was similar to that observed by Collignon et al. (1988); resistance was largely confined to the indole-negative non-fragilis B. fragilis group which tends to be more susceptible to antibiotics in general. Within this group nearly 30% of B. distasonis isolates were resistant to clindamycin. Resistance amongst B. distasonis isolates to the cephamycins was also high, lending support to the earlier observation that it is the most resistant member of the B.fragilis group (Downes & Andrew, 1988). Azithromycin, a new macrolide which achieves high tissue concentrations rapidly after an oral dose (Girard et al., 1987), was found to have acitivity against anaerobes which was similar to that of erythromycin. The MIGgS for most isolates, especially the Bacteroides spp., were > 16 mg/L. However, because of its high tissue penetration, in-vitro data may underestimate its in-vivo activity. For infections caused by anaerobic Gram-positive cocci and C. perfringens this drug may be an alternative for patients intolerant of other agents. Our results are similar to those described elsewhere (Kitris et al., 1990). In conclusion, the susceptibility profile of antimicrobial agents previously tested in our institution against anaerobic bacteria remains largely unchanged. However, occasional resistance to metronidazole and imipenem was noted. Cefoxitin and cefotetan cannot be considered to be reliably active in vitro against the B.fragilis group. This inability to predict the efficacy of the cephalosporins indicates that the selection of these agents for the treatment of serious anaerobic infections should not be based on collective institutional or even national data but on the results of the susceptibility testing of individual isolates. This could be facilitated by alternative methods such as
Sosceptfbfflty testing of anaerobic bacteria
819
the E-tcst (Citron et al., 1991). Preliminary studies in our laboratory confirm that there is a close correlation between the results of the E-tcst and the agar dilution method (unpublished data). Because of patterns of resistance, particularly among the B. fragUis group, and the reliance of clinicians on drugs with less than optimal activity against this important group of pathogens, it would seem that periodic testing of all anaerobic isolates to commonly-used antimicrobial agents remains necessary. Acknowledgements The authors wish to thank Mrs Judith Rogers and Ms Claire Greenall for their excellent secretarial assistance and Pfizer Australia for financial support. References Appelbaum, P. D., Spangler, S. K. & Jacobs, M. R. (1990). /7-Lactamase production and susceptibilities to amoxicillin, amoxicillin-clavulanate, ticarcillin, ticarcilin-clavulanate, ccfoxitin, imipenem, and metronidazole of 320 non-Bacteroides fragUis Bacteroides isolates and 129 fusobacteria from 28 U.S. centres. Antimicrobial Agents and Chemotherapy 34, 1546-50. Appelbaum, P. C , Spangler, S. -K. & Jacobs, M. R. (1991). Susceptibilities of 394 Bacteroides fragUis, non-B. fragUis group Bacteroides species, and Fusobacterium species to newer antimicrobial agents. Antimicrobial Agents and Chemotherapy 35, 1214-8. Barry, A. L. & Fuchs, P. C. (1991). In-vitro activity of four penicillin /Mactamase inhibitor combinations against cefoxitin-susceptible and cefoxitin-resistant Bacteroides fragUis isolates. Journal of Antimicrobial Chemotherapy 27, 243-4. Barry, A. L. & Jones, R. N. (1988). Interpretative criteria for the agar diffusion susceptibility test with azithromycin. Journal of Antimicrobial Chemotherapy 22, 637-41. Breuil, J., Burnat, C , Patcy, O. & Dublanchet, A. (1989). Survey of Bacteroides fragUis susceptibility patterns in France. Journal of Antimicrobial Chemotherapy 24, 69-75. Brown, E. M. (1991). Lack of standardization of in vitro susceptibility testing of the Bacteroides fragUis group to amoxydllin-clavulanic acid. Journal of Antimicrobial Chemotherapy 28, 147-9. Citron, D. M., Ostovari, M. I., Karlsson, A. & Goldstein, E. J. C. (1991). Evaluation of the E-test for susceptibility testing of anaerobic bacteria. Journal of Clinical Microbiology 29, 2197-203. Collignon, P. J., Munro, R. & Morris, G. (1988). Susceptibility of anaerobic bacteria to antimicrobial agents. Pathology 20, 48-52. Cuchural, G. J., Tally, F. P., Jacobus, N. V., Cleary, T., Finegold, S. M., Hill, G. et al. (1990). Comparative activities of newer /J-lactam agents against members of the Bacteroides fragUis group. Antimicrobial Agents and Chemotherapy 34, 479-80. Cuchand, G. J., Tally, F. P., Jacobus, N. V., Gorbach, S. L., Aldridge, K., Cleary, T. et al. (1984). Antimicrobial susceptibilities of 1292 isolates of the BacteroidesfragUisgroup in the United States: comparison of 1981 with 1982. Antimicrobial Agents and Chemotherapy 26, 145-8. DowelL V. R. & Hawkins, T. M. (1979). Laboratory Methods. In Anaerobic Bacteriology: CDC Laboratory Manual. US Department of Health, Education and Welfare, Public Health Service, Centre for Disease Control, Atlanta, Georgia. Downes, J. & Andrew, J. H. (1988). Susceptibility of 114 isolates of the Bacteroides fragUis group to imipenem and eight other antimicrobial agents. Pathology 20, 260-3. Girard, A. E., Girard, D., English, A. R., Gootz, T. D., Cimochowski, C. R., Faiella, J. A. et al. (1987). Pharmacokinetic and in vivo studies with azithromycin (CP-62, 993), a new macrolide with an extended half-life and excellent tissue distribution. Antimicrobial Agents and Chemotherapy 31, 1948-54. Holdeman, L. V., Cato, E. P. & Moore, W. E. C. (Eds) (1977). Anaerobe Laboratory Manual. 4th edn. Virginia Polytechnic Institute and State University. Blacksburg, VA.
820
S. C. A. Chen et aL
Jacobs, M. R., Spangler, S. K. & Appelbaum, P. C. (1990). 0-Lactamase production, /Mactam sensitivity and resistance to synergy with clavulanate of 737 Bacterouks fragilis group organisms from thirty-three US centres. Journal of Antimicrobial Chemotherapy 26, 361-70. Kitzis, M. D., Goldstein, F. W., Miegi, M. & Acar, J. F. (1990). In-vitro activity of azithromycin against various Gram-negative bacilli and anaerobic bacteria. Journal of Antimicrobial Chemotherapy 25, Suppl. A, 15-8. Lang, S. D. & Thomas, M. G. (1985). A comparison of the in vitro activity of ampicillinsulbactam and amoxycillin-clavulanic acid. New Zealand Journal of Medical Laboratory Technology 39, 100-2. National Committee for Clinical Laboratory Standards. (1990). Methods for Antimicrobial Susceptibility Testing of Anaerobic Bacteria, 2nd edn; Approved Standard M11-A2. NCCLS, Villanova, PA. Shah, H. N. & Collins, M. D. (1988). Proposals for ^classification of Bacteroiaes asaccharolyticus, Bacteroiaes gingivalis, and Bacteroiaes endodontalis in a new genus, Porphyromonas. International Journal of Systematic Bacteriology 38, 128-31. Sutler, V. L., Barry, A. L., Wilkins, T. D. & Zabransky, R. J. (1979). Collaborative evaluation of a proposed reference dilution method of susceptibility testing of anaerobic bacteria. Antimicrobial Agents and Chemotherapy 16, 495-502. Tally, F. P., Cuchural, G. J., Jacobus, N. V., Gorbach, S. L., Aldridge, K., Cleary, T. et al. (1985). Nationwide study of the susceptibility of the Bacteroiaes fragilis group in the United States. Antimicrobial Agents and Chemotherapy 28, 675-7. Wexler, H. M., Harris, B., Carter, W. T. & Finegold, S. M. (1986). Six-year retrospective survey of the resistance of Bacteroiaes fragilis group species to clindamycin and cefoxitin. Diagnostic Microbiology and Infectious Disease 4, 247-53. Wexler, H. M., Molitoris, E. & Finegold, S. M. (1991). Effect of /?-lactamase inhibitors on the activities of various /Mactam agents against anaerobic bacteria. Antimicrobial Agents and Chemotherapy 35, 1219-24. {Received 18 February 1992; accepted 14 July 1992)
Antimicrobial susceptibility of anaerobic bacteria in Australia Sharon C A. Chen, Thomas Gottlieb, Jennifer M. Palmer, GabrkOe Morris and Gwendolyn L. Gilbert
Department of Clinical Microbiology, Centre for Infectious Diseases and Microbiology, Westmead Hospital, Westmead NSW 2145, Australia The susceptibilities of 900 clinical isolates of anaerobic bacteria to 14 antimicrobial agents were determined by an agar dilution technique. Chlorampbcnicol, imipenem and metronidazole were found to be active against virtually all of the strains; only a single Bacteroides fragilis isolate was resistant to both imipenem and metronidazole. The addition of clavulanic acid to amoxytillin and ticarcillin potentiated the activities of these agents against all anaerobes including members of the B. fragilis group. Ampkallin/sulbactam and clindamycin were the next most active agents, 91 and 89% of isolates respectively being susceptible. Seventy-three per cent of the bacteria tested were susceptible to cefoxitin and 65% to cefotetan, with the MICs of almost 50% of the isolates clustering between 16 and 32 mg/L. There was also clustering around the breakpoint (64 mg/L) of piperacillin. Azithromycin exhibited poor activity against the B. fragilis group; only 18% of isolates were susceptible to < 4 mg/L. However, 92% of non-2), fragilis Bacteroides group strains were susceptible to this agent We conclude that imipenem, metronidazole, chloramphenicoL, ticarcillin/clavulanate, co-amoxiclav and, to a lesser extent, ampicillin/sulbactam are suitable as empirical therapy for infections caused by anaerobic bacteria.
Introduction
Antimicrobial susceptibility testing of anaerobic bacteria is difficult to standardize and is therefore not routinely performed. There is also disagreement about the ability of in-vitro tests to predict responses in vivo, particularly with the newer cephalosporins. However, occasional surveys provide guidelines for effective empirical treatment, enable patterns of bacterial resistance to be monitored and are necessary to evaluate the activities of new agents. Regional differences in the antimicrobial resistance patterns of anaerobes, as well as intracentre variations with time, are well documented (Cuchural et al., 1984; Tally et al., 1985). Recent studies have demonstrated increasing resistance to /Mactams, not only among the Bacteroides fragilis group but also among non-i?. fragilis Bacteroides and Fusobacteritan spp. (Cuchural et al., 1984; Appelbaum, Spangler & Jacobs, 1990). The addition of /Mactamase inhibitors such as clavulanic acid and sulbactam reduces the MICs of some /Mactam antibiotics to susceptible concentrations but this effect is not invariable (Appelbaum et al., 1990; Jacobs, Spangler & Appelbaum, 1990). Moreover, resistance to both imipenem and metronidazole amongst clinical isolates Correspondence to: Profeoor G. L. Gilbert. 811 0305-7453/92/120811 + 1 0 S08.00/0
© 1992 The Brituh Society for Antimicrobial Chemotherapy
812
S. C A. Chen et aL
belonging to the B.fragilis group has recently been noted in the USA and Europe (Breuiil et al., 1989; Jacobs et al., 1990). At Westmead Hospital antibiotic susceptibility testing of anaerobic bacteria has been performed periodically. Earlier results (Collignon, Munro & Morris, 1988) showed that metronidazole had the most predictable in-vitro activity and that resistance to /Mactams (imipenem was not tested) was frequent and clinically important. We report here additional data of the susceptibilities of recent anaerobic isolates to 14 antimicrobial agents, including several not included in the previous report. Materials and methods
Bacterial strains Nine hundred recent clinical isolates of anaerobic bacteria were included in this study which was carried out between 1986 and 1991. Repeat isolates from the same patients were excluded. Organisms were identified by conventional biochemical tests and gasliquid chromatography as specified in the Virginia Polytechnic Institute (Holdeman, Cato & Moore, 1977) and Wadsworth Anaerobic Laboratory Manuals (Suttcr et al., 1979). To facilitate comparison with previous results the old classification of non-B. fragilis group Bacteroides spp. was used (Shah & Collins, 1988). Antimicrobial agents The following antibiotics (and their respective breakpoint MICs) were tested: penicillin G (4 mg/L), metronidazole (16 mg/L), clindamycin (4 mg/L), cefotetan (32 mg/L), cefoxitin (32 mg/L), piperacillin (64 mg/L), ampicillin (4 mg/L), ampicillin/ sulbactam (16/8 mg/L), co-amoxiclav (8/4 mg/L), ticarcillin/clavulanate (64/2 mg/L), imipenem (8 mg/L), chloramphenicol (16 mg/L), erythromycin (4 mg/L) and azithromycin (4 mg/L). The breakpoints are in accordance with the current recommendations of the National Committee for Clinical Laboratory Standards (1990), with the exception of those for erythromycin and azithromycin which were based on previously established, achievable serum or tissue concentrations (Barry & Jones, 1988). The MICs of the /Mactam//Mactamase inhibitor combinations were expressed in terms of the concentrations of the /Mactam component. Ticarcillin was combined with a fixed concentration (2 mg/L) of clavulanic acid and co-amoxiclav consisted of amoxycillin and clavulanic acid in a 2:1 ratio. Sulbactam and ampicillin were combined in a 1 : 1 ratio according to the manufacturer's instructions. Penicillin, metronidazole, clindamycin and cefoxitin were evaluated throughout the study period. Imipenem was tested against 634 isolates, cefotetan against 402, ampicillin and ampicillin/sulbactam against 266 and co-amoxiclav, ticarcillin/clavulanate, piperacillin, chloramphenicol, erythromycin and azithromycin against 232. MIC determinations MICs were determined by an agar dilution method. Portions of five or more colonies from 48-72 h cultures growing on supplemented blood agar (Dowell & Hawkins, 1979) were inoculated into the thioglycollate broth supplemented with haemin and vitamin K. The broths were incubated anaerobically for 48 h. After the turbidity was adjusted to a 0-S McFarland standard the organisms were inoculated with a Denley
Susceptibility testing of anaerobic bacteria
813
replicator on to Wilkins-Chalgren agar containing doubling dilutions of antibiotic (range 0-03-128 mg/L). The final inoculum density was approximately 103 cfu per spot. Antibiotic-free control plates were inoculated at the beginning and end of each series. One set of control plates was incubated aerobically and the other was incubated with the test plates at 37°C for 48 h in an anaerobic chamber containing 80% nitrogen, 10% hydrogen and 10% carbon dioxide. The MICs for each organism were recorded as the lowest concentration yielding no growth, a fine haze, one discreet colony or multiple tiny colonies (NCCLS, 1990). Organisms which were non-viable or where the control plates were contaminated were excluded. B. fragilis ATCC 25285, Bacteroides thetcdotomicron ATCC 29741, Clostridium pefringens ATCC 13124 and Peptostreptococcus asaccharolyticus ATCC 29743 were included as controls with each batch tested. Results The susceptibilities of the isolates to the 14 antibiotics are shown in the Table. Bacteroides spp. predominated, accounting for 466 (52%) of the 900 organisms studied; 310 of these belonged to the B. fragilis group and included 144 B. fragilis strains. The next commonest bacteria were the non-sporing, Gram-positive bacilli, including 130 Propionibacterium acnes isolates. Isolates of the B. fragilis group were the least susceptible overall. Within the group, B. fragilis isolates were the most susceptible, followed by members of the indolenegative non-fragilis B. fragilis group. The indole-positive group strains were the least susceptible to /7-lactam antibiotics. Cefotetan and cefoxitin exhibited similar activities against the B. fragilis group (55% and 56% of strains respectively were susceptible to < 16 mg/L) but cefotetan was slightly more active against B. fragilis (88% vs 79% of strains). Raising the breakpoint to 32 mg/L increased the percentages of B. fragilis group isolates which were inhibited to 73% and 63% for cefoxitin and cefotetan, respectively. The cefoxitin MICs for 95 of 274 (35%) of these isolates were clustered between 16 and 32 mg/L. Similar clustering of 49 of 108 (45%) isolates was observed for cefotetan. Addition of a /Mactamase inhibitor lowered the MICs of the penicillins, rendering most of the Bacteroides spp. susceptible. However, ampicillin/sulbactam was active against only 79 of 87 (91%) of the B. fragilis group strains. The majority of isolates were susceptible to piperacillin but the MIC^s for all of the strains in the B. fragilis group were clustered around the breakpoint (64 mg/L). Resistance to chloramphenicol, metronidazole and imipenem was rare. Only a single strain of B. fragilis was resistant to imipenem (MIC 8 mg/L) and metronidazole (MIC 16 mg/L). Metronidazole had predictably low activity against the Gram-positive, nonsporing bacilli. Little resistance to clindamycin was detected amongst the isolates with the exception of the indole-negative non-fragilis B fragilis group; nine of 50 (18%) isolates, including seven of the 24 Bacteroides distasonis strains tested, were resistant to this agent. Sixty-three per cent and 56% of B. distasonis isolates were also resistant to cefotetan and cefoxitin, respectively. /?-Lactam antibiotics in general were highly active against Fusobacterium spp.; only two strains demonstrated reduced susceptibility to penicillin and imipenem. Azithromycin, like erythromycin, exhibited poor activity against B. fragilis group isolates; only 18 of 102 (18%) isolates were susceptible to < 4 mg/L. However, 23 of 25 (92%) isolates of the non-fi. fragilis Bacteroides species, 91% of anaerobic
814
S.CA. CbtaetaL Table. MICs* of 14 antimicrobial agents for 900 anaerobic isolates
Organism
Antimicrobial
MIC a MIC*
B. fragUis group (all spedes)
penicillin metronidazole clindamycin cefotetan cefoxitin ampicillin ampitiUin/sulbactam azithromycin erythromycin co-amoxidav ticarcillin/clavulanate piperacillin imipenem chloramphenicol penicillin metronidazole clindamycin cefotetan cefoxitin ampicillin ampidlHn/sulbactam azithromycin erythromydn co-amoxidav ticarcillin/clavulanate piperacillin imipenem chloramphenicol
16 05 025 4 16 32 2 16 8 05 012 8 006 2 16 05 025 4 8 32 1 16 8 025 056 4 003 2 16 05 012 32 8 32 2 16 8 025 006 4 003 2 32 05 05 32 16 32
B. fragUis
non-fragUis B. fragUis
penicillin
group (indole—ve)
metronidazole dindamycin cefotetan cefoxitin ampicillin ampicillin/sulbactam azithromycin erythromydn co-amoxidav ticarcillin/clavulanate piperacillin imipenem chloramphenicol
non-fragUis B. fragUis
penicillin
group (indole+ ve)
metronidazole dindamycin cefotetan cefoxitin ampicillin
128 1 8 164 32 128 8 128 128 4 4 128 05 8 128 1 1 32 16 128 4 128 32 1 025 32 025 8 128 1 8 64 32 128 16 128 128 4 4 64 005 4 128 1 2 64 32 128
Range
% of strain's susceptible at breakpoint concentrations*
003-128 003-16 003-16 003-64 003-128 025-128 025-16 05-128 025-128 003-16 003-128 003-128 O03-8 O03-8 O12-128 006-4 003-16 025-64 025-128 4-128 05-8 05-128 1-128 012-128 003-128 003-128 O03-8 025-4 012-128 006-16 003-16 1-128 O25-128 025-128 025-16 05-128 1-128 003-16 003-128 003-128 O03-1 025-4 012-128 006-16 003-16 003-128 003-128 4-128
16 99 89 63 73 2 91 18 42 100 99 91 99 100 7 100 98 95 84 0 100 17 42 100 97 97 99 100 21 100 82 75 78 5 88 4 23 100 100 90 100 100 6 100 96 16 52 0
Stnceptibffity testing of anaerobic bacteria
815
Table.—continued
Organism
Antimicrobial
non-fragilis B. fragilis ampitillin/sulbactam group (indole+ve) azithromycin erythromycin co-amoxiclav ticarrillin/davulanate piperacillin nxupcncm chloramphenicol
M I C , MIC* 2 4 16 128 4 64 0-5 4 0-5 4 32 128 0-12 05 4 4
Range
% of strain's susceptible at breakpoint concentrations*
05-16 1-128 025-128 003-^t 003-4 003-128 003-1 003-4
97 32 68 100 100 80 100 100
non-B. fragilis group Bacteroides spp.
penicillin metronidazole clindamycin cefotetan cefoxitin ampicillin ampicillin/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanatc piperacillin umpenem chloramphenicol
025 0-12 0-5 0-5 0-5 0-5 0-25 0-5 1 006 003 2-0 003 1
8 1 1 8 4 32 2 2 8 O5 1 8 006 2
003-128 003-4 003-16 003-64 003-128 003-128 003-16 003-128 003-128 003-8 003-16 003-128 003-05 003-8
74 100 94 93 97 67 97 92 88 100 100 90 100 100
Fusobacterium spp.
penicillin metronidazole clindamycin cefotetan cefoxitin ampiciUin ampicillin/sulbactam azithromycin erythromycin co-amoxiclav ticarcLUin/davulanate piperacillin imipenem chloramphenicol
003 006 012 006 012 006 006 16 128
003-16 003-1 003-4 O03-2 003-8 003-8 003-2 025-16 4-128 —
92 100 (100) 100 100 (83) (100) (50) (50) —
003 —
025 025 012 2 4 8 2 16 128 — — — 4 —
— O03-4 —
— (89) —
peniciUin metronidazole clindamycin cefotetan cefoxitin ampicillin ampicilUn/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanate piperacillin
025 050 006 025 05 025 025 4 16 006 05 8
8 2 012 8 1 1 05 16 64 1 128 128
003-32 003-8 003-2 003-8 003-128 012-1 012-05 006-64 003-64 003-1 003-128 1-128
51 95 100 100 95 100 100 64 36 100 89 78
Veillonella spp.
c
— —
816
S. C. A. Cben et aL Table.—continued
Organism
Antimicrobial
Veillonella spp.
imipencm chloramphcnicol
Anaerobic Grampositive cocci
Propionibacterium acnes
% of strain's susceptible at breakpoint concentrations*
MIC* MIQo
Range
006 2
006 8
003-05 005-8
100 100
penicillin metronidazole clindamycin cefotetan ccfoxitin ampicilUn ampicillin/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanate pipcracillin imipenem chloramphenicol
0-12 0-50 0-25 0-12 0-12 0-12 012 025 1 003 003 006 003 1
2 2 05 4 1 8 1 16 32 1 2 16 006 1
003-64 003-8 003-4 003-64 003-64 003-32 003-32 003-128 003-128 003-2 003-64 003-128 003-05 003-4
94 98 96 96 98 89 93 91 91 100 100 93 100 100
penicillin metronidazole clindamycin cefotetan ccfoxitin ampidllin ampicillin/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanate piperacillin imipencm chloramphenicol
003 16 006 006 006 012 006 006 003 003 003 025 003 05
006 003-^ 32 4-32 012 003-1 0-25 003-32 025 003-128 025 006-05 025 003-025 012 0O3-8 0-12 003-64 006 003-0-12 003 003-1 025 003-1 003 003 1 006-1
99 19 100 96 99 100 100 97 97 100 100 100 100 100
Clostridium perfringens penicillin metronidazole clindamycin cefotetan cefoxitin ampicillin ampicillin/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanate piperacillin
012 025 025 025 003 025 006 2 1 003 006 006 003 2
chloramphenicol Anaerobic Grampositive bacilli (non-sporing)
penicillin metronidazole clindamycin cefotetan
006 16 006 012
05 2 2 025 2 2 1 2 2 025 16 1 006 4
003-16 006-2 003-32 003-2 006-64 003-2 003-1 05-2 1-2 003-025 003-16 003-1 003-0-5 1-4
97 100 96 100 100 (100) (100) (100) (100) (100) (100) (100) 100 (100)
05 32 05 4
003-64 003-32 003-4 003-128
96 15 99 100
Sosctfrtibflfty testing of anaerobic bacteria
817
Table.—continued
Oroflnigin
Antimicrobial
M I C , MIC*
Anaerobic Grampositive bacilli (non-sporing)
ccfoxitin ampidllin ampirilHn/sulbactam azithromycin erythromycin co-amoxiclav ticarcillin/clavulanate piperacillin mupenem chloramphenicol
012 0-12 0-12 0-06 003 003 003 025 003 O5
1 05 05 1 1 O06 003 1 O03 1
Range
% of strain's susceptible at breakpoint concentrations*
003-128 O03-O5 O03-O5 O03-8 O03-8 003-012 O03 O03-4 003-2 006-1
98 93 100 98 94 100 100 100 100 100
*MICf expressed in mg/L. *See matf-riiili and method* for breakpoint concentrations; where < 10 (trains were tested, numbers are within brackets. *Not tested.
Gram-positive cocci, 98% of Gram-positive bacilli and all eight isolates of C. perfringens were susceptible. Discussion The results of this investigation revealed that there have not been any significant changes in the overall susceptibility patterns of most of the antibiotics tested since an earlier study undertaken by this laboratory (Collignon et al., 1988). Chloramphenicol, co-amoxiclav, ticarcillin/clavulanate, imipenem and metronidazole remain highly active against almost all anaerobes. Predictably, the non-sporing, Gram-positive bacilli, particularly P. acnes, were not susceptible to metronidazole. The latter has become increasingly important as a pathogen in neurosurgical patients; penicillin G remains the drug of choice for use against this organism. Of the /J-lactams, imipenem, ticarcillin/clavulanate and co-amoxiclav were the most active agents against the B. fragilis group. The single strain resistant to imipenem was identified as B. fragilis; it was also resistant to metronidazole. Previous studies in Australia failed to detect resistance to imipenem (Downes & Andrew, 1988); however recent international experience suggests that resistance to this drug is emerging (Breuil et al., 1989; Cuchural et al., 1990; Jacobs et ai., 1990). Metronidazole resistance has not previously been documented in Australia. Broad-spectrum /Mactam antibiotics such as the cephamycins are attractive as single agents for both the prevention and treatment of mixed aerobic and anaerobic infections. Our results demonstrate substantial resistance to both cefoxitin and cefotetan amongst the B. fragilis group. These results are similar to those found elsewhere (Downes & Andrew, 1988; Cuchural et al., 1990). However, the MIQQS of cefoxitin and cefotetan were 32 and 64 mg/L, respectively, concentrations that are achievable in vivo when high doses of these agents are used. It was noteworthy that the MICs of
818
S. C. A. Chen et aL
both antibiotics clustered around 16-32 mg/L, highlighting the difficulty in selecting a suitable breakpoint for the cephalosporins. This problem has been identified by other investigators (Wexler et al., 1986; Collignon et al., 1988). Cefotctan has a spectrum of activity similar to that of cefoxitin, although it is less active against the indole-positive non-fragilis B.fragilis group. Cefotetan has the advantage of a longer half-life and can be admininstered twice daily. However, activity against Bacteroides spp. is not predictable and there is a risk of hypoprothrombinaemia. Neither cephamycin can currently be considered a reliable choice as empirical therapy for serious anaerobic infections in Australia. Piperacillin remains active against the majority of anaerobes although there was clustering of the MICs around the breakpoint (64 mg/L); this has been noted by others (Barry & Fuchs, 1991). Piperacillin MICs were particularly high amongst isolates which were also resistant to cefoxitin. /J-Lactamase production is commonplace amongst all Bacteroides spp. as well as Fusobacterium spp. (Appelbaum et al., 1990; Jacobs et al., 1990). The addition of /J-lactamase inhibitors such as clavulanic acid and sulbactam renders most Bacteroides isolates susceptible to /Mactams in vitro, with a more than five-fold reduction in the MICJOS (Appelbaum et al., 1990; Cuchural et al., 1990; Wexler, Molitoris & Finegold, 1991). Our results confirm this observation, although sulbactam was less effective than clavulanic acid. Sulbactam was combined with ampicillin in a 1 : 1 ratio which provides adequate sulbactam to inhibit the /Mactamase activity of Bfragilis (Lang & Thomas, 1985). However, previous investigators have used this combination in a 2 : 1 ratio (Appelbaum, Spangler & Jacobs, 1991; Wexler et al., 1991) and found that the activity against all Bacteroides spp. was comparable to that of co-amoxiclav. The variation in technique and breakpoint concentrations in different centres makes comparison of results from different studies difficult (Brown, 1991). The incidence of clindamycin resistance among members of the B. fragilis group was similar to that observed by Collignon et al. (1988); resistance was largely confined to the indole-negative non-fragilis B. fragilis group which tends to be more susceptible to antibiotics in general. Within this group nearly 30% of B. distasonis isolates were resistant to clindamycin. Resistance amongst B. distasonis isolates to the cephamycins was also high, lending support to the earlier observation that it is the most resistant member of the B.fragilis group (Downes & Andrew, 1988). Azithromycin, a new macrolide which achieves high tissue concentrations rapidly after an oral dose (Girard et al., 1987), was found to have acitivity against anaerobes which was similar to that of erythromycin. The MIGgS for most isolates, especially the Bacteroides spp., were > 16 mg/L. However, because of its high tissue penetration, in-vitro data may underestimate its in-vivo activity. For infections caused by anaerobic Gram-positive cocci and C. perfringens this drug may be an alternative for patients intolerant of other agents. Our results are similar to those described elsewhere (Kitris et al., 1990). In conclusion, the susceptibility profile of antimicrobial agents previously tested in our institution against anaerobic bacteria remains largely unchanged. However, occasional resistance to metronidazole and imipenem was noted. Cefoxitin and cefotetan cannot be considered to be reliably active in vitro against the B.fragilis group. This inability to predict the efficacy of the cephalosporins indicates that the selection of these agents for the treatment of serious anaerobic infections should not be based on collective institutional or even national data but on the results of the susceptibility testing of individual isolates. This could be facilitated by alternative methods such as
Sosceptfbfflty testing of anaerobic bacteria
819
the E-tcst (Citron et al., 1991). Preliminary studies in our laboratory confirm that there is a close correlation between the results of the E-tcst and the agar dilution method (unpublished data). Because of patterns of resistance, particularly among the B. fragUis group, and the reliance of clinicians on drugs with less than optimal activity against this important group of pathogens, it would seem that periodic testing of all anaerobic isolates to commonly-used antimicrobial agents remains necessary. Acknowledgements The authors wish to thank Mrs Judith Rogers and Ms Claire Greenall for their excellent secretarial assistance and Pfizer Australia for financial support. References Appelbaum, P. D., Spangler, S. K. & Jacobs, M. R. (1990). /7-Lactamase production and susceptibilities to amoxicillin, amoxicillin-clavulanate, ticarcillin, ticarcilin-clavulanate, ccfoxitin, imipenem, and metronidazole of 320 non-Bacteroides fragUis Bacteroides isolates and 129 fusobacteria from 28 U.S. centres. Antimicrobial Agents and Chemotherapy 34, 1546-50. Appelbaum, P. C , Spangler, S. -K. & Jacobs, M. R. (1991). Susceptibilities of 394 Bacteroides fragUis, non-B. fragUis group Bacteroides species, and Fusobacterium species to newer antimicrobial agents. Antimicrobial Agents and Chemotherapy 35, 1214-8. Barry, A. L. & Fuchs, P. C. (1991). In-vitro activity of four penicillin /Mactamase inhibitor combinations against cefoxitin-susceptible and cefoxitin-resistant Bacteroides fragUis isolates. Journal of Antimicrobial Chemotherapy 27, 243-4. Barry, A. L. & Jones, R. N. (1988). Interpretative criteria for the agar diffusion susceptibility test with azithromycin. Journal of Antimicrobial Chemotherapy 22, 637-41. Breuil, J., Burnat, C , Patcy, O. & Dublanchet, A. (1989). Survey of Bacteroides fragUis susceptibility patterns in France. Journal of Antimicrobial Chemotherapy 24, 69-75. Brown, E. M. (1991). Lack of standardization of in vitro susceptibility testing of the Bacteroides fragUis group to amoxydllin-clavulanic acid. Journal of Antimicrobial Chemotherapy 28, 147-9. Citron, D. M., Ostovari, M. I., Karlsson, A. & Goldstein, E. J. C. (1991). Evaluation of the E-test for susceptibility testing of anaerobic bacteria. Journal of Clinical Microbiology 29, 2197-203. Collignon, P. J., Munro, R. & Morris, G. (1988). Susceptibility of anaerobic bacteria to antimicrobial agents. Pathology 20, 48-52. Cuchural, G. J., Tally, F. P., Jacobus, N. V., Cleary, T., Finegold, S. M., Hill, G. et al. (1990). Comparative activities of newer /J-lactam agents against members of the Bacteroides fragUis group. Antimicrobial Agents and Chemotherapy 34, 479-80. Cuchand, G. J., Tally, F. P., Jacobus, N. V., Gorbach, S. L., Aldridge, K., Cleary, T. et al. (1984). Antimicrobial susceptibilities of 1292 isolates of the BacteroidesfragUisgroup in the United States: comparison of 1981 with 1982. Antimicrobial Agents and Chemotherapy 26, 145-8. DowelL V. R. & Hawkins, T. M. (1979). Laboratory Methods. In Anaerobic Bacteriology: CDC Laboratory Manual. US Department of Health, Education and Welfare, Public Health Service, Centre for Disease Control, Atlanta, Georgia. Downes, J. & Andrew, J. H. (1988). Susceptibility of 114 isolates of the Bacteroides fragUis group to imipenem and eight other antimicrobial agents. Pathology 20, 260-3. Girard, A. E., Girard, D., English, A. R., Gootz, T. D., Cimochowski, C. R., Faiella, J. A. et al. (1987). Pharmacokinetic and in vivo studies with azithromycin (CP-62, 993), a new macrolide with an extended half-life and excellent tissue distribution. Antimicrobial Agents and Chemotherapy 31, 1948-54. Holdeman, L. V., Cato, E. P. & Moore, W. E. C. (Eds) (1977). Anaerobe Laboratory Manual. 4th edn. Virginia Polytechnic Institute and State University. Blacksburg, VA.
820
S. C. A. Chen et aL
Jacobs, M. R., Spangler, S. K. & Appelbaum, P. C. (1990). 0-Lactamase production, /Mactam sensitivity and resistance to synergy with clavulanate of 737 Bacterouks fragilis group organisms from thirty-three US centres. Journal of Antimicrobial Chemotherapy 26, 361-70. Kitzis, M. D., Goldstein, F. W., Miegi, M. & Acar, J. F. (1990). In-vitro activity of azithromycin against various Gram-negative bacilli and anaerobic bacteria. Journal of Antimicrobial Chemotherapy 25, Suppl. A, 15-8. Lang, S. D. & Thomas, M. G. (1985). A comparison of the in vitro activity of ampicillinsulbactam and amoxycillin-clavulanic acid. New Zealand Journal of Medical Laboratory Technology 39, 100-2. National Committee for Clinical Laboratory Standards. (1990). Methods for Antimicrobial Susceptibility Testing of Anaerobic Bacteria, 2nd edn; Approved Standard M11-A2. NCCLS, Villanova, PA. Shah, H. N. & Collins, M. D. (1988). Proposals for ^classification of Bacteroiaes asaccharolyticus, Bacteroiaes gingivalis, and Bacteroiaes endodontalis in a new genus, Porphyromonas. International Journal of Systematic Bacteriology 38, 128-31. Sutler, V. L., Barry, A. L., Wilkins, T. D. & Zabransky, R. J. (1979). Collaborative evaluation of a proposed reference dilution method of susceptibility testing of anaerobic bacteria. Antimicrobial Agents and Chemotherapy 16, 495-502. Tally, F. P., Cuchural, G. J., Jacobus, N. V., Gorbach, S. L., Aldridge, K., Cleary, T. et al. (1985). Nationwide study of the susceptibility of the Bacteroiaes fragilis group in the United States. Antimicrobial Agents and Chemotherapy 28, 675-7. Wexler, H. M., Harris, B., Carter, W. T. & Finegold, S. M. (1986). Six-year retrospective survey of the resistance of Bacteroiaes fragilis group species to clindamycin and cefoxitin. Diagnostic Microbiology and Infectious Disease 4, 247-53. Wexler, H. M., Molitoris, E. & Finegold, S. M. (1991). Effect of /?-lactamase inhibitors on the activities of various /Mactam agents against anaerobic bacteria. Antimicrobial Agents and Chemotherapy 35, 1219-24. {Received 18 February 1992; accepted 14 July 1992)
