Journal of Antimicrobial Chemotherapy (1991) 27, Suppl. C, 1-7 An evaluation of the susceptibility patterns of Gram-negative organisms isolated in cancer centres with aminoglycoside usage J. J. Muscato", D. W. Wilbur*, J. J. Stout* and R. A. Fab-lender* "Columbia Regional Hospital and Boone Hospital Center, Columbia, MO 65201; b Loma Linda University School of Medicine, Loma Linda, CA 92350; 'Bristol-Myers Squibb Company, Pharmaceutical Group, Evansville, IN 47721, USA This study in 12 cancer treatment centres across the United States was designed to evaluate the potential for increased resistance to with unrestricted use. An initial 3-month baseline period during which the use of was restricted and that of and unrestricted was followed by a period of at least 12 months when was the primary aminoglycoside. Resistance of Gramnegative bacilli to these aminoglycosides from hospitalized patients was monitored and compared for the two periods. Amikacin usage increased from a mean of 20-1% to a mean of 83-9% of aminoglycoside patient-days. A reduction in the use of and were observed with means of 66-1 and 10%, and 13-9 and 6-1%, respectively for the two periods. Resistance to was 0-85% at baseline and 1-3% at end-point which was not clinically significant (P = 0-614). Baseline resistance was 6-5 and 7-6%, while final resistance was 2-6 and 4-8%, respectively for (P = 0-1) and (P = 52). Introduction Hospitalized, immunocompromised cancer patients are often subject to serious nosocomial infections. The most common Gram-negative pathogens are Pseudomonas spp., Providencia spp., Serratia spp., and indole-positive Proteus. They are susceptible to a relatively limited number of antibiotics. The aminoglycoside antibiotics, combined with /Mactam antibiotics, have provided the mainstay of empirical therapy for these infections (Young et al., 1977). While the aminoglycosides have been in use for many years, increasing resistance to and has occurred among Gram-negative bacilli (Gerding et al., 1979). Resistance has not been limited to the aminoglycosides but included more recent agents such as the quinolones and imipenem (Ogle, Reller & Vasil, 1988). Therefore, the combination of an aminoglycoside and /Mactam antibiotic remains the principal regimen for the empirical treatment of infections in the immunocompromised cancer patient (Young et al., 1977). Several studies have prospectively investigated the emergence of resistance to specific aminoglycosides in relationship to their usage (Moody et al., 1982; Wielunsky et al., 1983; Betts et al, 1984; Gerding & Larson 1985; Levine et al., 1985; Cunha, 1987). The Corresponding address: Dr J. J. Muscato, Hematology-Oncology Association of Columbia, 31 Berrywood Drive, Columbia, MO 65201, USA. O305-7453/91/27CO01 +07 $02./0 1 1991 The British Society for Antimicrobial Chemotherapy
2 J. J. Muscato et al. majority of these studies were carried out in hospitals. Moody et al. (1982) evaluated the emergence of resistance to aminoglycosides at the University of Maryland Cancer Center over a 13-month period, during a time when served as the primary aminoglycoside. A total of 29 strains was isolated from 315 patients among which resistance was significantly less than in an earlier baseline study period, during which the use of had been very low. This observation led to the conclusion that unrestricted use of did not lead to an increase in resistance among Gram-negative bacilli. Betts et al. (1984) carried out a prospective study over 5 years. While there was no significant difference in the percentage of resistance in the two study periods (baseline and usage periods), there was a significant decrease in resistance during the period when served as the primary aminoglycoside. Similar results were shown by Wielunsky et al., (1983), Gerding & Larson (1985), and Cunha (1987) in other studies. However, Levine et al., (1985) observed an increase in resistance over an 18 month period coincident with an increase in usage. The current study was designed to evaluate the incidence of resistance in 12 cancer centres across the United States. Materials and methods A prospective survey of aminoglycoside resistance was conducted at 12 cancer centres located in Alabama, California, Colorado, Illinois, Indiana, Missouri, Oklahoma, Tennessee and Texas. Surveillance of resistance patterns of all Gram-negative bacilli obtained from hospitalized patients was performed using standard susceptibility tests on Mueller-Hinton agar, supplemented with calcium and magnesium, and standard antibiotic discs at each individual hospital. All isolates were identified using standard laboratory methods. All laboratories were certified by the College of American Pathologists. The study was divided into two separate data collection periods. The first was a baseline period of at least 3 months when data were collected prospectively by monitoring the resistance patterns and or were used as the first line aminoglycoside. This period was followed by at least a 12-month data collection period during which was the primary aminoglycoside (^51%). Other aminoglycosides could be used in either study period if clinical situations so dictated. All aerobic and facultatively anaerobic Gram-negative bacilli isolated from hospitallized patients were screened for resistance. Only single source isolates were reported, and each resistant species from a single patient was counted as a single resistant organism. Similarly, each susceptible species from a single patient was counted as a single susceptible organism. Resistance to was confirmed at each centre by measuring MICs, to, and. Resistance was defined as an MIC > 16 mg/1 for and > 4 mg/1 each for and. Duplicate samples of -resistant organisms were sent to the Microbiology Research Laboratory, Mechanisms of Resistance Service, Bristol-Myers Squibb Company Industrial Division, Syracuse, NY for verification. Fisher's exact test (twotailed) was used to compare resistance rates. The total number of aminoglycoside patient-days (AP-Ds) was recorded in each study period. An AP-D was defined as the number of days a patient received an aminoglycoside at any time during that day. Comparison of AP-Ds was made by drug, study period, and incidence of resistance.
Susceptibility and aminoglycoside usage 3 If a periodic review of resistance rates (usually monthly) indicated an alarming increase, the matter was investigated further to determine whether it was necessary to discontinue the study. Aminoglycoside utilization Results The mean duration of the baseline study-period was 3-6 months, while the mean duration of the second period of usage was 23-1 months. During the baseline period (Table I) the average use of was 2% of the AP-Ds (range 1-2- -1%). Tobramycin was the most commonly used aminoglycoside during the baseline period with a mean of 661% (range 41-6 to 92-6%) of the recorded AP-Ds; was used in 13-9%, (range 3-9 to 47-2%). The pattern of aminoglycoside usage was radically different during the second study-period (Table I): was used in 83-9% (range 69-7 to 1%), in 61% (range 0 to 12-2%) and in 10% (range 0 to 21 1%) of the recorded AP-Ds. The AP-Ds per month ranged from 58-7 to 275-3 at the 12 centres and totalled 5053 for the entire baseline period. The usage period induced a total of 22,665 AP-Ds, with a range of 50-8 to 8-4 per month between institutions. Nine centres reported a decrease in mean AP-Ds per month during the usage period. Four of these decreases were statistically significant (P < 1). Gram-negative isolates The five most commonly isolated organisms in decreasing order were Escherichia coli, Pseudomonas aeruginosa, Klebsiella spp., Enterobacter spp. and Proteus mirabilis. Although the protocol called for complete screening of all isolates, there was some variability in the standard screening procedures at the 12 centres, and not all of the aminoglycosides were always included in the susceptibility screening. This accounts for the slight variation in the number of isolates tested against each of the three aminoglycosides. Table II displays the total resistance pattern at the participating centres. The baseline data reveal that was tested against 352 Gram-negative organisms, while and were tested against 302 and 339 organisms, Table I. Aminoglycoside usage during the baseline period and second period of predominant usage Study period/antibiotic Percentage AP-D" Total AP-D" Baseline Second 2 13-9 661 83-9 61 1 10 7 3342 19,018 137V 2270 5053 5053 5053 22,665 22,665 22,665 "Aminoglycoside patient-day.
J. J. Muscato et al. Table II. Total aminoglycoside resistance during the baseline period and second period of predominant usage Study period/antibiotic No. of isolates No. resistant (%) Baseline Second P = 0-614. b P = 52. 'P = 0-1. 352 302 339 2512 2479 2481 3(0-85) 23 (7-6) 22 (6-5) 33 (1-3)" 120 (4-8)* 64 (2.6)' respectively. During the second period of usage, 2512 organisms were tested against and 2479 and 2481 against and, respectively. The mean number of Gram-negative organisms isolated per month did not change significantly between the baseline and usage periods. Tobramycin resistance Total baseline resistance to among the participating centres was 6-5% (Table II). Table III depicts the response of the individual species. Pseudomonas spp. (25%), Enterobacter spp. (20%), Citrobacter spp. (91%) and P.aeruginosa (8-6%) demonstrated significant resistance during the baseline period. While resistance to Pseudomonas spp. increased during the period to -4%, total resistance, as well as resistance to a number of the individual Gram-negative species, decreased dramatically. Total resistance for all Gram-negative bacilli was significantly lower (P = 01) following the period (2-6 vs 6-5%, Table II). Gentamicin resistance During the usage period, resistance to among Gram-negative isolates at the participating institutions decreased markedly from a mean of 7-6% during the baseline period to a mean of 4-8%, a decrease of 36-8% (P = 52). Table III reveals that Pseudomonas spp. (37-5%), P.aeruginosa (15-4%), Acinetobacter spp. (14-3%) and Enterobacter spp. (120%) had the highest resistance rates during the baseline period. Resistance of all species except Pseudomonas spp. and P. aeruginosa remained stable during the usage period; resistance declined in all other species (Table III). Amikacin resistance There was essentially no change in the incidence of resistance when comparing the baseline data with the period. Among the 12 participating
Susceptibility and aminoglycoside usage 5 Table III. Percentage of resistant species during the baseline period and second period of predominant usage Baseline Second Organism Antibiotic No. of isolates Percentage resistant No. of isolates Percentage resistant Providencia spp. Pseudomonas spp. P. aeruginosa E. coli Citrobacter spp. Klebsiella spp. Enterobacter spp. Serratia spp. P. mirabilis Proteus (indole+) spp. Acinetobacter spp. 3 3 3OO 93 52 81 82 78 84 59 59 59 25 25 25 9 36 39 12 12 12 8 7 8 12-5 37-5 250 OO 15-4 8-6 1-2 3-8 3-6 91 91 0.0 3-4 51 12-0 2 2-8 2-6 8-3 14-3 13 13 13 53 49 47 549 533 536 879 871 882 54 50 53 424 422 418 231 235 231 64 64 63 162 161 159 42 41 41 39 7-7 26-4 34-7 -4 1-6 15-4 3-5 01 0-6 0-6 1-9 0-7 0-7 1-7 4-7 3-9 1-6 1-6 7-9 1-2 2-4 2-4 2-5 51 institutions, resistance in the Gram-negative isolates increased slightly from 0-85% to 1-3% (P - 0-614). Resistance among Pseudomonas spp. and P. aeruginosa both demonstrated increases from 12-5 to 26-4% and from 0 to 1-6%, respectively (Table III). A slight decrease in resistance among E. coli and Serratia spp. was observed. However, no changes within individual species were statistically significant (Table III).
6 J. J. Muscato et al. Discussion In the present study usage was associated with a slight rise in the incidence of resistance, and a significant decrease in Gram-negative resistance to and/or. These results are strikingly similar to those of earlier studies (Moody etal., 1982; Wielunsky et al., 1983; Gerding & Larson, 1985; Levine et al., 1985; Cunha, 1987). Surveillance studies such as these cannot accurately define the reason for the decrease in Gram-negative resistance. It may be postulated that it is probably due to decreased usage of and/or. This explanation becomes less likely when it is shown that the usage of or was very low in the baseline period and did not differ significantly in the usage period in some of these studies and yet the incidence of resistance declined. An alternative explanation is that kills organisms which contain enzyme systems that inactivate and/or, but do not affect. The data generated by this multi-centre study in cancer treatment centres as well as the results seen in the general hospital surveillance studies, suggest a causal relationship between increased usage and decreased Gram-negative resistance to other aminoglycosides. Another interesting point brought to light by these surveillance studies is the fact that even though very little was used before or during the baseline phase, there was already some resistance present among the Gram-negative organisms. Perhaps the use of and provided the selective pressure which leads to the emergence of resistance. Mechanisms of resistance have a significant impact on both the incidence of resistance and its spread. However, they were not evaluated in this study. No major changes in infection control procedures were reported at any of the participating centres during the study and it is assumed that the effect of these measures must have been similar in both study periods. This study, together with similar earlier investigations, indicates that the use of as the primary aminoglycoside in the empirical treatment of infection in cancer patients does not lead to widespread increases in resistance, and is associated with a decrease in resistance to and. This information and the fact that in-vitro resistance amongst Gram-negative bacteria is less common to than to other aminoglycosides, suggests that would be an appropriate choice as the primary aminoglycoside for use in empirical therapy at centres where resistance is a problem. Acknowledgement This study was completed with research support from Bristol-Myers Squibb Company US Pharmaceutical Group. References Betts, R. F., Valenti, W. M., Chapman, S. W., Chonmaitree, T., Mowrer, G., Pincus, P. etal. (1984). Five-year surveillance of aminoglycoside usage in a university hospital. Annals of Internal Medicine 1, 219-22. Cunha, B. A. (1987). The prevention of aminoglycoside resistance by the use of as the primary hospital aminoglycoside. Advances in Therapy 4, 33-9. Gerding, D. N., & Larson, T. A. (1985). Aminoglycoside resistance in Gram-negative bacilli during increased use. American Journal of Medicine 79, Suppl. 1A, 1-7.
Susceptibility and aminoglycoside usage 7 Gerding, D. N., Buxton, A. E., Hughes, R. A., Cleary, P. P., Arbaczawski, J. & Stamm, W. E. (1979). Nosocomial multiply resistant Klebsiella pneumoniae: epidemiology of an outbreak of apparent index case origin. Antimicrobial Agents and Chemotherapy 15, 608-15. Levine, J. F., Maslow, M. J., Leibowitz, R. E., Pollock, A. A., Hanna, B. A., Scaeffer, S. et al. (1985). Amikacin-resistant Gram-negative bacilli: correlation of occurrence with use. Journal of Infectious Diseases 151, 295-3. Moody, M. M., de Jongh, C. A., Schimpff, S. C. & Tillman, G. L. (1982). Long-term use; effects on aminoglycoside susceptibility patterns of Gram-negative bacilli. Journal of the American Medical Association 248, 99-202. Ogle, J. W., Reller, L. B. & Vasil, M. L. (1988). Development of resistance in Pseudomonas aeruginosa to imipenem, norfloxacin, and ciprofloxacin during therapy: proof provided by typing with a DNA probe. Journal of Infectious Diseases 158, 537-41. Wielunsky, E., Drucker, M., Cohen, T. & Reisner, S. H. (1983). Replacement of gentamicjn by as a means of decreasing resistant Gram-negative rods in a neonatal intensive care unit. Journal of Medical Science (Israel) 19, 16-08. Young, L. S., Martin, W. J., Meyer, R. D., Weinstein, R. J. & Anderson, E. T. (1977). Gram-negative rod bacteremia: microbiologic, immunologic and therapeutic considerations. Annals of Internal Medicine 86, 456-71.