Susceptibility of the Bacteroides fragilis Group in the United

ANTIMICROBIAL AGENTS AND CHEMOTHERPY, Apr. 1983, p. 536-540 0066-4804/83/040536-05$02.0O/0 Copyright C 1983, American Society for Microbiology Vol. 23, No. 4 Susceptibility of the Bacteroides fragilis Group in the United States in 1981 F. P. TALLY,'* G. J. CUCHURAL,' N. V. JACOBUS,' S. L. GORBACH,1 K. E. ALDRIDGE,2 T. J. CLEARY,3 S. M. FINEGOLD,4 G. B. HILL,5 P. B. IANNINI,6 R. V. McCLOSKEY7 J. P. O'KEEFE,8 AND C. L. PIERSON9 Infectious Diseases Section, Department of Medicine, Tufts-New England, Boston, Massachussets 021111; Charity Hospital, Louisiana State University, New Orleans, Louisiana 761122; Jackson Memorial Hospital, Miami, Florida 331363; Wadsworth Veterans Administration, Los Angeles, California 900734; Duke University, Durham, North Carolina 277105; Danbury Hospital, Danbury, Connecticut 068106; Albert Einstein, Philadelphia, Pennsylvania 191477; Loyola University, Chicago, Illinois 601538; and University of Michigan Hospital, Ann Arbor, Michigan 48104' Received 12 November 1982/Accepted 20 January 1983 The minimal inhibitory concentrations of nine antimicrobial agents was determined for over 750 clinical isolates of the Bacteroides fragilis group of anaerobic bacteria collected from nine centers in the United States during 1981. High resistance rates were documented for cefoperazone, cefotaxime, and tetracycline. Cefoxitin had the best activity of the 3-lactam antibiotics, whereas moxalactam and piperacillin had good activities. The resistance rate for clindamycin was 6%. There were no metronidazole- or chloramphenicol-resistant isolates encountered. There were significant differences in susceptibility among the various species of the B. fragilis group, particularly with moxalactam, cefoxitin, and clindamycin. Clustering of clindamycin-, piperacillin-, and cefoxitin-resistant isolates was observed at different hospitals. The variability of resistance rates with the P- lactam antibiotics and clindamycin indicates that susceptibility testing of significant clinical isolates should be performed to define local resistance patterns. The Bacteroides fragilis group of organisms, consisting of Bacteroides fragilis, Bacteroides thetaiotaomicron, Bacteroides distasonis, Bacteroides ovatus, Bacteroides vulgatus, and Bacteroides uniformis, are recognized as important pathogens in suppurative diseases and other infections. These organisms are usually considered together on the basis of taxonomy and because of increased antibiotic resistance as compared with other anaerobic bacteria (1, 6, 7, 13, 24). Of the group, B. fragilis emerges as the most important pathogen; it has the highest frequency of isolation from serious infections and is the most common anaerobic bacterium that invades the bloodstream. However, other species of the group are equally pathogenic once they have invaded deep tissues. The number of useful antimicrobial agents active against these organisms is relatively limited. Most strains of the B. fragilis group are resistant to penicillin G and cephalosporins, including cephalothin, cefazolin, and cefamandole, on the basis of the presence of constitutive,b-lactamases (13, 14, 24, 25). Furthermore, the isolation rate of resistant strains seems to be increasing (4, 17). The change in the resistance rates may be related to genetic mechanisms such as the transferable antimicrobial resistance which has been described for cindamycin-erythromycin and tetracycline and for high levels of penicillin-ampicillin resistance (12). Most clinical microbiology laboratories do not routinely test antimicrobial susceptibility of anaerobic bacteria. The choice of an appropriate therapeutic agent to treat infections involving B. fragilis group is usually based on published susceptibility data, even though these data show wide variations. This variabiilty could be explained by regional differences in the frequency of resistant organisms, the selection of strains to be tested, or differences in the testing methods. Thus, there may be reported falsely high or low resistance rates that could lead to the selection of an ineffective drug or to the use of toxic drug combinations. This study was designed to establish the current susceptibility rates for clinical isolates of the B. fragilis group from nine centers across the United States by standardized susceptibility methods performed in one laboratory. Preliminary results have been previously published (4). 536 MATERILS AND METHODS Bacteral isoates. All clinical isolates of the B. fragilis group collected during the 12-month period of January to December, 1981, were referred from the

VOL. 23, 1983 nine study group hospitals to the Tufts-New England, Anaerobic Bacteriology Laboratory, Boston, Mass. Only one isolate of the same species per patient was studied. The identification of the strains was confirmed by established methods (8). Antimicrobial agents. Standard powders were obtained as follows: cefoxitin from Merck Sharp & Dohme, Rahway, N.J.; moxalactam from Eli Lilly & Co., Indianapolis, Ind.; cefoperazone from Pfizer Inc., New York, N.Y.; piperacillin from Lederle Laboratories, Pearl River, N.Y.; clindamycin from The Upjohn Co., Kalamazoo, Mich.; cefotaxime from Hoechst-Roussel Inc., Somerville, N.J.; metronidazole from G. D. Searle & Co., Chicago, Ill.; and chloramphenicol and tetracycline from Sigma Chemical Co., St. Louis, Mo. Antibiotic susceptibility testing. The minimal inhibitory concentrations (MICs) were determined by an agar-dilution method with a Steer's replicator by the anaerobic chamber techniques described previously (24). Data analysis. All data analysis was performed on a TRS/80 model I computer (Radio Shack, Tandy Co., Fort Worth, Tex.) with a data base management and statistical package developed by G. J. Cuchural, Tufts-New England, Boston, Mass. RESULTS There was a wide variation in the susceptibility of the Bacteroides strains to the nine antimicrobial agents with regard to the 50 and 90% MICs and to the percentage of resistant strains (Table 1). Two breakpoints for each drug were selected on the basis of achieveable blood levels (the lpwer levels may be more relevant because the drugs must penetrate sites harboring Bacteroides species, namely, abscesses). Although somewhat arbitrary, the lower breakpoints generally agree with previously published data. Higher breakpoints are based on previously published guidelines (9). The,-lactam antibiotics demonstrated variable activity; cefoxitin was the most active, with a 90% MIC of 16,ug/ml. Piperacillin and moxalactam were the next most active 3-lactam antibiotics. There was a high level of resistance with cefoperazone, cefotaxime, and tetracycline. Of the isolates studied, U.S. SURVEY OF B. FRAGILIS SUSCEPTIBILITY 537 6% were resistant to clindamycin at 4,ug/ml. No metronidazole- or chloramphenicol-resistant isolates were encountered. When analyzed by species, B. distasonis was the most resistant to cefoxitin (Table 2). Moxalactam had a strikingly higher rate of resistance among the non-b. fragilis species. B. thetaiotaomicron and B. vulgatus were the most resistant to clindamycin. There was variability in the resistance rates among the nine centers (Table 3). Clindamycin resistance was seen at several centers with the highest rate at the Jackson Memorial Hospital, whereas high cefoxitin resistance rates were encountered at the Danbury Hospital and the Tufts-New England. High resistance rates to piperacillin were seen at several hospitals. DISCUSSION The results of our study indicate that several drugs were active against the B. fragilis group of organisms. Cefoxitin was the most active,blactam antibiotic; the range of resistance rates was 3 to 17%. There was no high-level resistance to cefoxitin (MIC > 128,ug/ml), suggesting that the mechanism of resistance is more likely related to penetration of the compound into the bacterial periplasmic space rather than inactivation of the drug, as has been shown in Bacteroides isolates by investigators in Stockholm (5, 14) Ṫhe next most active,b-lactam antibiotics were piperacillin and moxalactam. Some investigators believe that higher blood levels can be achieved with these two drugs. If 128 and 32,g of drug per ml were used as the respective breakpoints, resistance rates for these two agents would be lower (Table 1). Piperacillin had the lowest incidence of P-lactam resistance other than cefoxitin. This may reflect the substrate specificity of the f-lactamases of Bacteroides species, which are primarily cephalosporinases (14, 25). Some centers had higherrates of piper- TABLE 1. Resistance rates of the B. fragilis group Antimicrobial agent No. of MIC (,Lg/ml) Range % Breakpointb isolates 50%i0 90%o (pg/mi) Resistant' (,g/mi) Cefoxitin 755 8 16 s0.25-128 8 (2) 16 (32) Moxalactam 755 4 64 <0.25->128 22 (12) 16 (32) Cefotaxime 749 32 128 <0.25->128 54 (35) 16 (32) Cefoperazone 729 32 128 50.25->128 57 (33) 16 (32) Piperacillin 754 16 128 <0.25->128 12 (8) 64 (128) Clindamycin 755 <0.25 2 <0.25->32 6 (5) 4 (8) Chloramphenicol 755 4 8 1-8 0 8 (16) Metronidazole 753 0.5 1-0.25-2 0 (0) 8 (16) Tetracycline 705 8 32 <0.25->32 63 (63) 4 (4) a Numbers in parentheses are the percent resistant at higher breakpoint. b Numbers in parentheses are the breakpoints based on data by Kirby et al. (9).

538 TALLY ET AL. TABLE 2. Resistance rates of species of the B. fragilis group ANTIMICROB. AGENTS CHEMOTHER. (no. of % Resistant isolates at lower breakpoints with&: Species isolates)' Cfx Mox CIn Ctx Cpz Pip Tet B. distasonis (57) 17 50 2 43 56 22 50 B. fragilis (463) 7 9 5 56 59 12 65 B. ovatus (150) 8 42 6 58 57 6 57 B. thetaiotaomicron (63) 7 46 17 70 68 11 62 B. vulgatus (60) 7 18 12 22 39 13 71 a Metronidazole and chloramphenicol are not included because there were no resistant strains. B. uniformis and Bacteroides species were not included because of a low number of isolates. b Cfx, Cefoxitin; Mox, moxalactam; Cln, clindamycin; Ctx, cefotaxime; Cpz, cefoperazone; Pip, piperacillin; Tet, tetracycline. acillin resistance, suggesting the presence of a new P-lactamase in Bacteroides species or the fact that the gene(s) coding for resistance have been amplified (12, 19) or both. On the basis of previously published data, piperacillin appears to be more active than ticarcillin and carbenicillin against the B. fragilis group of organisms (3, 28) Ċhloramphenicol and metronidazole were the most active of the non-p-lactam antimicrobial agents, and there were no resistant isolates encountered. The lack of resistance to chloramphenicol is puzzling since there are at least two mechanisms by which Bacteroides species can inactivate the drug, acetylation and nitroreduction (2, 15). Although there are scattered reports of metronidazole-resistant Bacteroides species this phenomenon has not been documented in the United States. All putative metronidazoleresistant anaerobic bacteria referred to the Tufts Anaerobic Laboratory were actually susceptible or were contaminated with a microaerophilic or TABLE 3. aerobic organism or Propionibacterium acnes. Clindamycin remains a highly active agent; 94% of the isolates were inhibited at <4 plg/ml. Further analysis of the in vitro activity of this drug showed three classes of organisms: highly susceptible (MIC, :0.25,ug/ml); intermediate (MIC, 0.5 to 4,ug/ml); and resistant (MIC, >4 ILg/ml), which confirms previous observations (21). Most of the resistant group had MICs of 128 to 512,ug/ml. The transfer of high-level clindamycin resistance has been demonstrated to be mediated by at least two transfer factors, pbftm10 and pbf4 (pip410) (12, 17, 26, 27, 29). These two transfer factors have common clindamycin resistance gene(s) as determined by DNA-DNA hybridization studies (20). These genes also code for erythromycin and streptogramins resistance (11, 17). Thus, the clinical use of clindamycin, erythromycin, or the streptogramins may increase the frequency of clindamycin resistance in Bacteroides species. The 6% resistance rate in this survey may reflect a real Resistance rates of the B. fragilis group by referral center Center (no. of % Resistant isolates at lower breakpoints with': isolates)a Cfx Mox Cln Ctx Cpz Pip Tet Albert Einstein (50) 12 24 8 50 70 8 58 Danbury Hospital (47) 17 23 0 53 60 9 59 Duke University (136) 7 24 5 58 62 14 65 Jackson Memorial (86) 5 23 13 62 62 17 59 Hospital Louisiana State (89) 11 19 9 58 53 12 65 University Loyola University (100) 5 12 5 43 58 5 73 University of (101) 6 27 8 63 51 16 56 Michigan Tufts-New England (%) 14 23 0 51 52 13 55 Wadsworth Medical (67) 7 19 10 43 42 22 75 Center a Metronidazole and chloramphenicol are not included because there were no resistant strains. b Cfx, Cefoxitin; Mox, moxalactam; Cln, clindamycin; Ctx, cefotaxime; Cpz, cefoperazone; Pip, piperacillin; Tet, tetracycline.

VOL. 23, 1983 increase over that reported in the early 1970s, when lower resistance rates were reported (1, 6, 10, 13, 16, 22, 23). Our study confirms the widespread resistance to tetracycline in Bacteroides species that was first recognized in the 1960s (6, 10, 13). This high resistance rate may be explained by an efficient tetracycline resistance transfer system which is stimulated by the presence of low levels of the drug (12, 17, 25, 27). Tetracycline resistance has also been encountered in anaerobic cocci and clostridia. These observations relegate tetracycline to a drug of primarily historical interest in the treatment of anaerobic infections. Analysis of resistance patterns at the referring hospitals indicated regional clustering of resistance to cefoxitin, clindamycin, and piperacillin. A mechanism for the clustering of clindamycin and piperacillin could be the presence of transferable genes coding for these resistances. These elements are known to exist in Bacteroides species. In addition, these organisms displayed a close relationship between tetracycline and clindamycin resistance. Previous studies have demonstrated that the transfer of these two resistances is associated (17, 27). There were striking differences in susceptibility to moxalactam among B. fragilis and the other species. B. fragilis was the most susceptible, whereas high-level resistance was found in the other species. Some variation in species susceptibility was also encountered with cefoxitin, piperacillin, cefotaxime, and clindamycin. Although the non-b. fragilis species are less common in clinical specimens, indicating less virulence, once these organisms have invaded deep tissues, they should be considered pathogens, thus requiring appropriate treatment. It may be important to speciate the B. fragilis group as a guide to therapy and for epidemiological studies, as advocated by Rolfe and Finegold (18). The regional differences in resistance patterns indicate that antimicrobial susceptibility studies should be performed on selected clinical isolates of the B. fragilis group. These include organisms isolated from blood cultures, from unusual sites such as bone, joint spaces, and spinal fluid, from very sick patients, and from other sites in patients who have failed to respond to appropriate therapy. This study underscores the fact that anaerobic bacteria, like aerobic bacteria, have changed their susceptibility to a number of antimicrobial agents. The clustering of resistances at various hospitals indicates that we can no longer rely on published data to choose antimicrobial therapy. Microbiological laboratories will have to develop their own data base from which clinicians can make rational therapeutic decisions. U.S. SURVEY OF B. FRAGILIS SUSCEPTIBILITY 539 ACKNOWLEDGMENTS F.P.T. is the recipient of Research Career Development Award AI00393 from the National Institute of Allergy and Infectious Diseases. LITERATURE CITED 1. Appelbaum, P., and S. Chatteto. 1978. Susceptibility of anaerobic bacteria to ten antimicrobial agents. Antimicrob. Agents Chemother. 14:371-376. 2. Brltz, M. L., and R. G. Wimon. 1978. Chloramphenicol acetyltransferase of Bacteroides fragilis. Antimicrob. Agents Chemother. 14:105-111. 3. Brown, J. E., V. E. DelBene, nd C. D. CollIn. 1981. In vitro activity of N-formimidoyl thienamycin, moxalactam, and other new beta-lactam agents against Bacteroides fragilis: contribution of beta-lactamase to resistance. Antimicrob. Agents Chemother. 19:248-252. 4. Cuchural, G., N. Jacobus, S. L. Gorbach, and F. P. Taly. 1981. A survey of Bacteroides susceptibility in the United States. J. Antimicrob. Chemother. 8(Suppl. D):27-31. 5. Dornbusch, K., B. Olsson-Lilequlst, and C. E. Nord. 1980. Antibacterial activity of new,-lactam antibiotics on cefoxitin resistant strains of Bacteroides fragilis. J. Antimicrob. Chemother. 6:207-216. 6. Dubois, J., and J. C. Pechere. 1978. Activity of ten antimicrobial agents against anaerobic bacteria. J. Antimicrob. Chemother. 4:329-334. 7. Henderson, D. K., A. W. Chow, and L. B. Guze. 1977. Comparative susceptibility of anaerobic bacteria to ticarcillin, cefoxitin, metronidazole, and related antimicrobial agents. Antimicrob. Agents Chemother. 11:679482. 8. Hden, L. V., E. P. Cato, and W. E. Moore. 1977. Anaerobic laboratory manual, 4th ed. Virginia Polytechnic Institute, Blacksburg, Va. 9. Kirby, B. D., W. L. George, V. L. Sutter, D. M. Citron, and S. M. Flnegold. 1980. Gram-negative anaerobic bacilli, their role in infection and patterns of susceptibility to antimicrobial agents. Little known Bacteroides species. Rev. Infect. Dis. 2:914-951. 10. Kislak, J. W. 1972. The susceptibility of Bacteroides fragilis to 24 antibiotics. J. Infect. Dis. 125:295-299. 11. Magot, M., F. Fayofle, G. Prlvtera, and M. Sebald. 1981. Transposon-like structures in the Bacteroides fragilis MLS plasmid pip410. Mol. Gen. Genet. 181:559-561. 12. Malamy, M. H., and F. P. Tally. 1981. Mechanisms of drug resistance transfer in Bacteroidesfragilis. J. Antimicrob. Chemother. 8(Suppl. D):59-75. 13. Martin, W. J., M. Gardner, and J. HI. Washington. 1972. In vitro antimicrobial susceptibility of anaerobic bacteria isolated from clinical specimens. Antimicrob. Agents Chemother. 1:148-158. 14. Olsson, B., C.-E. Nord, and T. Walstrum. 1976. Formation of beta-lactamase in Bacteroides fragilis: cell-bound and extracellular activity. Antimicrob. Agents Chemother. 9:727-735. 15. Onderdonk, A. B., D. L. Kasper, B. J. Manshelm, T. J. Loule, S. L. Gorbach, and J. G. Bartlett. 1979. Experimental animal models for anaerobic infections. Rev. Infect. Dis. 1:291-301. 16. PbIllIps, I. 1979. Antibiotic sensitivity of anaerobic bacteria. Ann. Ist. Super. Sanita 15:115-122. 17. Privltera, G., F. FayoUe, and M. Sebald. 1981. Resistance to tetracycline, erythromycin and clindamycin in the Bacteroides fragilis group. Induceable versus constructive tetracycline resistance. Nature (London) 20:314-320. 18. Rolfe, R. D., and S. M. Flnegold. 1981. Comparative in vitro activity of new beta-lactam antibiotics against anaerobic bacteria. Antimicrob. Agents Chemother. 20:600-609. 19. Sato, K., Y. Matsuura, M. Inoue, and S. Mabuhashi. 1982. Properties of a new penicillinase type produced by Bacteroides fragilis. Antimicrob. Agents Chemother. 22:579-584. 20. Shimell, M. J., C. J. Smith, F. P. Tally, F. L. Macrlna,

540 TALLY ET AL. ad M. H. MNa y. 1982. Hybsidizaton studies reveal homologies between pbf4 and pbfia1o, two clindamycin erhronycin resistace trander pasmids of Bacteroidesfragilis. J. Dacteriol. 10-953. 21. Son, A., F. P. Taly, N. V. Jacoba, ad S. L. Gerba*. 1982. Bacteroides fragis esistance to clindamycin in vitro. Antimicrob. Agents Chemother. 22:771-774. 22. Simek, J. L., ad J. A. Waml_ H. 1974. Antmicrobial sueptibits of anuaerobic bacteria: recent clinical isolates. Anmicrob. Agents Chemother. 6:311-315. 23. Sa*r, V. L. 1977. In vitro susceptibility of anacrobes: comparison of clindamycin and other antimicrobial agents. J. Infect. Dis. 135(Suppi):57-512. 24. Taly, F. P., N. V. Jaobv, J. G. BardS, and S. L. Gwbach. 1975. Susceptibility of anerobes to cefoxitin and cephalosporins. Antimicrob. Agents Chemother. 7:128-132. ANTIMICROB. AGENTS CHEMOTHER. 25. Taft, F. P., J. P. O'Kmf., N. M. Sula, an S. L. Gorbada. 1979. Inactivation of cephalospoins by Bacteroides. Atimicrob. Agents Chemother. 16:565-571. 26. Tally F. P., D. R. Snyda, S. L. Gorbac, dmm. H. NaM_. 1979. Plasmid-mediatd tbansfrk resistance to clindamycin and erythromycin in Bacteroidesfragilis. J. Infect. Dis. 139:83-88. 27. Tall, F. P., D. R. Snydma, M. J. Shini, ad M. H. Maamy. 1982. Chacterzation of pbptml0, a cdidamycin-erythromycin resistance transfer factor from Bacteroidesfragilis. J. Bacteiol. 151:686-1. 28. Trelsm, I., D. Kay, and M. E. LevI.. 1979. Activity of semisynthetic penicillins and synergism with mecillinam against Bacteroides species. Antimicrob. Agents Chemother. 16283-286. 29. Wleb, R. A., ad F. L. MNiw 1981. Physical characterization of Bacteroides fragilis R plasmid pbf4. J. Bacteriol. 145:867-872. Downloaded from http://aac.asm.org/ on April 25, 2018 by guest