Cross-Canada survey of resistance of 2747 aerobic blood culture isolates to piperacillin/tazobactam and other antibiotics

Division of Microbiology, Department of Pathology and Laboratory Medicine, Queen Elizabeth II Health Sciences Centre, Halifax, Nova Scotia; Department of Microbiology and Infectious Diseases, Hôpital Maisoneuve-Rosemont, Montréal, Québec; Microbiology and Public Health, University of Alberta Hospital, Edmonton, Alberta; Wyeth-Ayerst Canada Inc; Department of Microbiology, Mount Sinai & Princess Margaret Hospitals; Department of Microbiology, Sunnybrook Health Science Centre, Toronto, Ontario Correspondence and reprints: Dr Kevin Forward, Service Chief, Division of Microbiology, Queen Elizabeth II Health Sciences Centre, 5788 University Avenue, Halifax, Nova Scotia B3H 1V8. Telephone 902-473-4109, fax 902-473-4432, e-mail Kevin.Forward@dal.ca Accepted for publication January 28, 1997. Accepted April 1, 1997 KR Forward, PA Franks, DE Low, R Rennie, AE Simor, and the Canadian Piperacillin/Tazobactam Study Group. Cross-Canada survey of resistance of 2747 aerobic blood culture isolates to piperacillin/tazobactam and other antibiotics. Can J Infect Dis 1998;9(1):33-44.

P iperacillin/tazobactam is the most recent of a number of beta-lactam/beta-lactamase inhibitor combinations to become available to Canadian physicians. The combination offers a number of advantages over previously marketed compounds. Piperacillin has a broader spectrum of activity than either the amino or carboxy penicillins, and tazobactam is a potent inhibitor of both chromosomally mediated and plasmid-mediated beta-lactamases (1,2,3).
Numerous clinical trials have demonstrated the efficacy of piperacillin/tazobactam in the treatment of a variety of infections, including intra-abdominal infections, skin and skin structure infections, pelvic infections in women, pneumonia and fever in neutropenic patients undergoing chemotherapy for both solid and hematological malignancies (4)(5)(6)(7)(8)(9)(10)(11)(12)(13). Noncomparative trials have also evaluated the efficacy of piperacillin/tazobactam with or without an aminoglycoside for the treatment of bacteremia (14,15).
We conducted a prospective study of susceptibilities of blood culture isolates collected in 1995 from 58 sites across Canada to determine the prevalence of resistance to commonly used broad spectrum antibiotics, including piperacillin/tazobactam. Such data may help to establish the degree of confidence with which single agents may be used to treat serious bacteremic infections.

MATERIALS AND METHODS
Strains: Fifty-eight teaching and community hospitals from nine Canadian provinces participated in the evaluation. In most cases, 50 consecutive aerobic cultures were collected, beginning in February 1995. A few smaller centres fell short of this goal. Methicillin-resistant Staphylococcus species and anaerobes were not included. Potential contaminants, such as coagulase-negative Staphylococcus species and viridans streptococci, were submitted only if recovered on three different occasions from the same patient within one week. Only one strain from each species was submitted per patient. Methods: Broth microdilution susceptibility studies were performed with 96-well microtitre plates containing serial twofold dilutions of study antibiotics. Panels were prepared by Sensititre (AccuMed International Inc, Ohio) and distributed to each of five regional testing centres. Organisms were speciated in their originating laboratory by using standard methods. Nonfastidious organisms were tested in accordance with National Committee for Clinical Laboratory Standards (NCCLS) guidelines (NCCLS M7-A3 and M100-S5) (16,17). Inocula were prepared by directly inoculating four to five colonies of an overnight culture of the organism incubated at 35 to 37°C into demineralized water to produce a suspension corresponding to a 0.5 McFarland standard. Ten microlitres of inoculum was suspended in 10 mL of cation-supplemented Mueller Hinton broth. Fifty microlitres of the resulting suspension was transferred to each well of the microtitre plates, which were then incubated aerobically at 34 to 36°C for 18 h.
For susceptibility testing of Streptococcus pneumoniae, the inoculum was prepared in Mueller Hinton broth from overnight growth on 5% sheep blood agar. After mixing, 100 µL of inoculum was added to 10 mL of cation-supplemented Mueller Hinton broth with added 2% to 5% lysed horse blood (PML Microbiological, Oregon). One hundred microlitres was inoculated to each well of the microtitre panel with an autoinoculator. Plates were incubated aerobically at 35°C for 20 to 24 h. Susceptibility testing for streptococci other than S pneumoniae was performed as for S pneumoniae except that only 10 µL was added to 10 mL of Mueller Hinton broth and 50 µL was added to each well.
For Haemophilus influenzae, the inoculum was prepared in

RESULTS
Of the 2747 blood culture isolates tested, piperacillin/tazobactam was most active, inhibiting 98% of strains, followed closely by imipenem, which inhibited 97.8% of strains (Table 1). Ceftazidime, ceftriaxone, ciprofloxacin and ticarcillinclavulanate each inhibited over 90% of strains. Overall, the proportion of strains susceptible to third-generation cephalosporins and ciprofloxacin was lower than this percentage, primarily because of their intrinsic inactivity against enterococci, which comprised 160 of 2747 isolates. Against the 1611 strains of Enterobacteriaceae species tested, piperacillin/tazobactam was second only to imipenem in terms of the percentage of strains susceptible; only 0.2% of strains were resistant ( Table 2). The minimum inhibitory concentration of 90% of isolates (MIC 90 ) of Enterobacteriaceae species was 2 mg/L compared with 64 mg/L for piperacillin alone. Seventeen per cent of Enterobacteriaceae species were susceptible to piperacillin/tazobactam but resistant to piperacillin. Enterobacteriaceae species remained highly susceptible to all agents tested, with the exception of piperacillin and cefoxitin, which inhibited 82% and 81% of strains, respectively. Figure 1 compares piperacillin/tazobactam with piperacillin MICs against strains of Enterobacteriaceae species, E coli, Enterobacter cloacae and Acinetobacter species. The majority of strains had lower piperacillin MICs in the presence of tazobactam than with piperacillin alone.        strains were susceptible to ciprofloxacin. Canadian strains remain highly susceptible to aminoglycosides, the MIC 90 for tobramycin being only 1 mg/L. Twenty-five strains of Acinetobacter species were tested and 96% of these were susceptible to piperacillin/tazobactam compared with only 76% of strains susceptible to piperacillin. Only 76% of strains were susceptible to aminoglycosides, and only 68% were susceptible to ceftriaxone. Ceftazidime was somewhat more active than ceftriaxone (MIC 90 8 mg/L versus 64 mg/L) against Acinetobacter species.
All 361 strains of S pneumoniae were susceptible to piperacillin/tazobactam, piperacillin and imipenem. The highest observed piperacillin/tazobactam MIC was 1 mg/L. The highest imipenem MIC was 0.25 mg/L. Only two ceftriaxone-resistant strains (MIC 2 mg/L or greater) were identified. The highest ceftriaxone MIC observed was 8 mg/L. Of the S pneumoniae strains tested, 11.3% had decreased susceptibility to ciprofloxacin (MIC 2 mg/L or greater).
Only three vancomycin-resistant E faecalis strains were identified, and only two of 29 Enterococcus faecium strains 42 Can J Infect Dis Vol 9 No 1 January/February 1998

DISCUSSION
The antibiotic armamentarium offers a number of choices for the initial empirical treatment of severely ill patients with bacteremia. The choice of agents in this setting should take into account the observed susceptibility patterns of isolates most frequently encountered. In this study, we compared the activity of a number of broad spectrum antibiotics with Canadian strains collected from both tertiary centres, and community and regional hospitals. With the exception of anaerobes and methicillin-resistant staphylococci, the numbers and proportion of strains tested should be quite representative of the Canadian experience. We have previously made the observation that Canadian strains are often more susceptible to antibiotics than those in Europe and the United States, and, therefore, it is important to use Canadian susceptibility data when making therapeutic choices (18).
Of the agents tested, piperacillin/tazobactam and imipenem had the broadest spectrums of activity. We would also expect these agents to be active against the anaerobes that were recovered during the same period of time but were not tested (19). Of the strains tested, only E faecium and Stenotrophomonas maltophilia were frequently resistant to these agents -E faecium, by virtue of altered penicillin binding proteins and S maltophilia because of its broad spectrum metalloenzyme and decreased permeability.
Piperacillin/tazobactam was much more active against E coli than piperacillin (MIC 90 1 mg/L versus 128 mg/L). Twenty-two per cent more strains were susceptible to piperacillin/tazobactam than to piperacillin alone. This presumably reflects the prevalence of TEM1 beta-lactamases in E coli and their susceptibility to tazobactam (1). For all Enterobacteriaceae species tested, the MIC 90 was lower for piperacillin/tazobactam than for piperacillin. Possibly, this represents a small amount of intrinsic activity of tazobactam, as well as its activity against beta-lactamases from a number of classes. We also observed that E cloacae strains usually had lower MICs to piperacillin/tazobactam than to piperacillin. This has been observed by others and is likely a reflection of the weak activity of tazobactam against class 1 beta-lactamases (20). When piperacillin/tazobactam was compared with ticarcillin-clavulanate, only S maltophilia was more susceptible to ticarcillinclavulanate (MIC 90 64 mg/L versus 256 mg/L). This has been noted by others (21)(22)(23)(24).
A number of agents showed reduced activity against Acinetobacter species. Only 76% of strains were susceptible to gentamicin and tobramycin, and only 71% were susceptible to cefotaxime. Imipenem remained highly active against Acinetobacter species, the MIC 90 being 0.5 mg/L. Ninety-six per cent of strains were susceptible to imipenem. Piperacillin/tazobactam was more active than piperacillin (MIC 90 8 mg/L versus 256 mg/L), and 20% more strains were more susceptible to piperacillin/tazobactam. These findings likely reflect both the intrinsic