ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, May 1998, p. 1266–1268 0066-4804/98/$04.0010 Copyright © 1998, American Society for Microbiology
Vol. 42, No. 5
In Vitro Antimicrobial Effects of Various Combinations of Penicillin and Clindamycin against Four Strains of Streptococcus pyogenes DENNIS L. STEVENS,* KARL J. MADARAS-KELLY,
AND
DAVID M. RICHARDS
Infectious Diseases Section, VA Medical Center, Boise, and Idaho State University College of Pharmacy, Pocatello, Idaho, and University of Washington School of Medicine, Seattle, Washington Received 28 July 1997/Returned for modification 14 December 1997/Accepted 4 March 1998
Previous studies using mouse models of Streptococcus pyogenes necrotizing fasciitis demonstrated that clindamycin had greater efficacy than penicillin. Frequently both agents are used concurrently in the treatment of severe S. pyogenes infections. This study investigated interactions between penicillin and clindamycin. E-test and broth microdilution assays suggested additivity or indifference, while timed-killing assays demonstrated concentration-dependent variable effects. Timed-kill studies utilizing clinical concentrations suggest that there is no antagonism with the combination of drugs but that the combination does not have a bactericidal advantage over either penicillin or clindamycin alone. from 0 to 0.03 mg/ml for penicillin and from 0 to 0.48 mg/ml for clindamycin. Log-phase bacteria were adjusted to 5 3 105 CFU per ml, and the plates were incubated for 24 h at 37°C in 5% CO2 prior to CFU determination. The determinations of MICs by the E-test method were performed according to the manufacturer’s recommendations (18). The concentration ranges were 0.002 to 32.0 mg/ml for penicillin and 0.016 to 256.0 mg/ml for clindamycin (E-test strips; AB Biodisk, Solna, Sweden). Bacteria were spectrophotometrically adjusted to 108 CFU/ml and spread evenly on SBA. E-test strips were placed on dry plates and incubated for 18 h at 37°C in 5% CO2. Broth microdilution assays were performed at concentrations ranging from 1/32 to 2 times the MIC for penicillin and from 1/128 to 8 times the MIC for clindamycin. The dilutions were made in 96-well plates (Corning Glass Works, Corning, N.Y.) in a checkerboard fashion, and the inoculum was prepared as described above. The plates were incubated for 24 h at 37°C in 5% CO2. To evaluate interactions between antibiotics, we calculated the fractional inhibitory concentration (FIC); the following formulas were used to calculate the FIC index: FICA 5 (MICA in combination)/(MICA alone), FICB 5 (MICB in combination)/(MICB alone), and the FIC index 5 FICA 1 FICB, where FICA (FICB) and MICA (MICB) are the FIC and MIC for antibiotic A (B), respectively (6). FIC indices were used to characterize antibiotic interactions as follows: synergy, FIC index # 0.5; additivity, 0.5 , FIC index , 1; indifference, 1 , FIC index # 4; antagonism, FIC index . 4 (7, 18). The dynamics of bacterial killing by penicillin and clindamycin were tested against ATCC 12384 by using each antibiotic alone and in combination, at concentrations equal to 1/2, 1, 2, 4, 8, 16, and 100 times the MIC. The timed-kill studies were performed with a final inoculum of 106 CFU/ml in 10 ml, and the tubes were continuously agitated and incubated for 24 h at 37°C in 5% CO2. Duplicate samples were removed at 0, 2, 4, 10, 18, and 24 h, and dilutions were plated onto SBA for CFU determination (17). Synergy was defined as a .2-log10-unit decrease in CFU per milliliter at 24 h due to the combination compared to that due to the more-active single agent, additivity or indifference was defined as a ,1-log10-unit change in
Strains of Streptococcus pyogenes remain exquisitely sensitive to penicillin in vitro, yet clinical failures of penicillin have been reported (1, 4, 6, 8, 10, 16). In experimental models of severe S. pyogenes infection, penicillin is less effective than erythromycin or clindamycin despite penicillin’s superior antimicrobial activity (16). Because resistance to erythromycin in Japan and Scandinavia has been frequently reported and because some of these strains are also resistant to clindamycin, the choice of antibiotics has become a more difficult task (13, 15). One approach to treat severe invasive S. pyogenes infection has been to utilize a combination of penicillin and clindamycin. The rationale is that penicillin provides coverage against 100% of S. pyogenes strains and that clindamycin has demonstrated greater efficacy in experimental models of necrotizing fasciitis (16). Despite this, there are no data to support the use of this combination. In fact, the use of a beta-lactam antibiotic together with a protein synthesis inhibitor may result in antagonism both in vitro (9) and in vivo (12). Therefore, the present study was undertaken to investigate the in vitro antimicrobial effect of combinations of penicillin and clindamycin against S. pyogenes by using E tests, the broth microdilution method, and timed bacterial kill curves. Penicillin G (benzylpenicillin; Sigma, St. Louis, Mo.) and clindamycin HCl (Cleocin; Upjohn, Kalamazoo, Mich.) were tested against a standard strain of S. pyogenes (ATCC 12384; American Type Culture Collection, Rockville, Md.) and three clinical strains (DLS 88003, DLS 88008, and DLS 96004) isolated from patients with streptococcal toxic shock syndrome. The organisms were cultured in Todd-Hewitt Broth (Difco Laboratories, Detroit, Mich.) and grown on tryptic soy agar with 5% sheep blood (SBA) plates (PML Microbiologicals, Tualatin, Oreg.). For broth microdilution, the MICs and minimum bactericidal concentrations (MBCs) were determined and defined according to the National Committee for Clinical Laboratory Standards guidelines (14). Appropriate concentrations of both antibiotics were diluted twofold with concentrations ranging * Corresponding author. Mailing address: Infectious Diseases Section (Building 45), Veterans Affairs Medical Center, 500 West Fort St., Boise, ID 83702. Phone: (208) 422-1599. Fax: (208) 422-1365. E-mail:
[email protected]. 1266
VOL. 42, 1998
NOTES
1267
TABLE 1. Effects of antibiotics upon the timed killing of the ATCC 12384 strain of S. pyogenes Multiple of MIC
FIG. 1. Kill rate constant versus the multiple of the MIC for the ATCC 12384 strain. The graph shows the concentration-dependent variation of kill rate for clindamycin (E), penicillin ({), and the combination (‚) in relation to that for the control (h) on the basis of the timed-killing data. Kill rate constants were determined by a linear fit of the 0- and 24-h time points.
CFU per milliliter, and antagonism was defined as a .2-log10 unit increase in CFU per milliliter (7, 18). Synergy testing by the E-test was performed by placing E strips on the blood agar plates in a cross formation, with a 90° angle at the intersection of the respective MICs (18). The plates were incubated and recorded as described above, and the nature of the interaction was determined by the FIC index (18). The penicillin mean MIC and MBC as determined by broth microdilution were both 0.015 mg/ml, and the E-test method resulted in an MIC of 0.012 mg/ml. The clindamycin mean MIC and MBC as determined by broth microdilution were 0.06 and 0.22 mg/ml, respectively, and the E test indicated an MIC of 0.16 mg/ml. The killing of S. pyogenes at 24 h by combinations of penicillin and clindamycin was variable depending upon the multiple of MIC being studied (Table 1). At one-half the MIC the combination exhibited synergy, as evidenced by a 2.3-log10-unit reduction in CFU per milliliter compared to penicillin alone.
1/2 1 2 4 8 16 100
Difference in killinga (log10 CFU/ml) for combination at 24 h vs: Combination at 0 h
Penicillin at 24 hb
Clindamycin at 24 h
21.96 21.30 21.10 21.10 20.84 20.53 21.02
22.29 (S) 10.07 (I) 12.89 (An) 12.17 (An) 11.64 (I) 11.60 (I) 11.28 (I)
23.16 10.14 20.48 10.17 10.61 11.07 10.39
a Cultures were incubated with or without antibiotics for 24 h at which time duplicate samples were assayed for CFU for each multiple of MIC. Results are the mean CFU from four separate experiments. Negative values indicate that the combination at 24 h resulted in greater killing. b Letters in parentheses characterize the effect of the combination versus that of penicillin alone. S, synergy; An, antagonism; I, indifference.
The bactericidal activity of the combination at the MIC was nearly identical to that of each antibiotic alone. At two and four times the MIC, 2.89- and 2.2-log10-unit increases, respectively, in CFU per milliliter compared to penicillin alone indicated antagonism. All concentrations greater than four times the MIC indicated indifference. Figure 1 summarizes the change in the rate of bacterial killing in relation to the multiple of the MIC studied. Synergy testing using the E-test and broth microdilution methods against three clinical strains and one ATCC strain demonstrated indifferent or additive effects. Mean FIC values were 0.746, 1.346, 0.938, and 0.908 (broth microdilution) and 1.183, 0.959, 0.773, and 1.158 (E test) for isolates ATCC 12384, DLS 88003, DLS 88008, and DLS 96004, respectively. Our results, like that of White et al., showed a good correlation between the E-test method and the traditional broth microdilution method (Mann-Whitney rank sum test; P 5 0.088) (18).
FIG. 2. A timed-killing curve at four times the MIC demonstrates that the curve for the combination of drugs (‚) follows the curve for clindamycin alone (E). This pattern is consistent at all concentrations greater than one-half the MIC. The combination still results in a .2-log10-unit advantage over the unchecked growth of the control (h). The inset displays earlier time points which produce a pattern consistent with the overall curve. {, penicillin.
1268
NOTES
ANTIMICROB. AGENTS CHEMOTHER.
Each timed-kill study evaluated the bactericidal effect of a specific MIC multiple after 24 h. At one-half the MIC there was synergistic interaction between the antibiotics in combination. The combination of antibiotics at the MIC inhibited growth by almost 1 log10 unit compared to the initial inoculum. Specifically, sub-MIC levels of penicillin and clindamycin did not prevent bacterial growth; however, the combination of these antibiotics at such concentrations exhibited enhanced killing compared to either antibiotic alone. In contrast, at two and four times the MIC the combination indicated antagonism, which may be explained in part by the enhanced killing of S. pyogenes by penicillin alone. The enhanced killing effect of beta-lactam antimicrobials at two to four times the MIC has been previously described (5). In addition, the dynamics of bacterial killing of the combination closely follows the kill curve for clindamycin at concentrations higher than the MIC, suggesting that penicillin’s contribution to bacterial killing is being inhibited (Fig. 2). This phenomenon could be explained by the observation that clindamycin affects the expression of penicillin-binding proteins in S. pyogenes (19). Clinically relevant concentrations of penicillin or clindamycin used in treating severe S. pyogenes infections are much higher than those used in this synergy study. For example, typical dosages that might be used in the treatment of severe S. pyogenes infections would include 4 million U of crystalline penicillin G administered every 4 h (24 million U per day) or 900 mg of clindamycin phosphate administered every 8 h. Anticipated peak and trough serum concentrations associated with those dosage regimens would approximate 40 to 80 and 1 to 2 mg/ml, respectively, for penicillin and 10 to 12 and 1 to 1.5 mg/ml, respectively, for clindamycin (3, 11). In the present study, the closest approximation of clinical dosage regimens was 100 times the MIC for both agents, which corresponded to peak concentrations of approximately 1.1 and 13 mg/ml for penicillin and clindamycin, respectively. Clinically relevant concentrations of the antibiotic combination did not exhibit antagonistic effects, nor did they have a bactericidal advantage over penicillin or clindamycin alone. Because there is no corroborative in vivo data for S. pyogenes and this combination (2), additional studies using an animal model will be necessary to further investigate this issue. This study was supported by a grant from the Department of Veterans Affairs Merit Review Program to D.L.S. REFERENCES 1. Adams, E. M., S. Gudmundsson, D. E. Yocum, R. C. Haselby, W. A. Craig, and W. R. Sundstrom. 1985. Streptococcal myositis. Arch. Intern. Med. 145:1020–1023.
2. Allan, J. D., and R. C. Moellering. 1985. Antimicrobial combinations in the therapy of infections due to gram-negative bacilli. Am. J. Med. 78:65–76. 3. American Society of Health-System Pharmacists. 1997. AHFS drug information 97: penicillin. American Society of Health-System Pharmacists, Inc., Bethesda, Md. 4. Brook, I. 1985. Role of beta-lactamase-producing bacteria in the failure of penicillin to eradicate group A streptococci. Pediatr. Infect. Dis. J. 4:491– 495. 5. Craig, W. A., and S. C. Ebent. 1991. Killing and regrowth of bacteria in vitro: a review. Scand. J. Infect. Dis. 74:63–70. 6. Eagle, H. 1952. Experimental approach to the problem of treatment failure with penicillin. I. Group A streptococcal infection in mice. Am. J. Med. 13:389–399. 7. Eliopoulos, G. M., and R. C. Moellering. 1991. Antimicrobial combinations, p. 432–492. In V. Lorian (ed.), Antibiotics in laboratory medicine. The Williams & Wilkins Co., Baltimore, Md. 8. Gatanaduy, A. S., E. L. Kaplan, B. B. Huwe, C. McKay, and L. W. Wannamaker. 1980. Failure of penicillin to eradicate group A streptococci during an outbreak of pharyngitis. Lancet ii:498–502. 9. Jawetz, E., J. B. Gunnison, R. S. Speck, and V. R. Coleman. 1951. Studies on antibiotic synergism and antagonism: the interference of chloramphenicol with the action of penicillin. Arch. Intern. Med. 87:349–359. 10. Kim, K. S., and E. L. Kaplan. 1985. Association of penicillin tolerance with failure to eradicate group A streptococci from patients with pharyngitis. J. Pediatr. 107:681–684. 11. LeFrock, J. L., A. Molavi, and R. A. Prince. 1982. Symposium on antimicrobial therapy: clindamycin. Med. Clin. North Am. 66:103–120. 12. Lepper, M. H., and H. F. Dowling. 1951. Treatment of pneumococcic meningitis with penicillin compared with penicillin plus aureomycin: studies including observations on an apparent antagonism between penicillin and aureomycin. Arch. Intern. Med. 88:489–494. 13. Maruyama, S., H. Yoshioka, K. Fujita, M. Takimoto, and Y. Satake. 1979. Sensitivity of group A streptococci to antibiotics. Am. J. Dis. Child. 133: 1143–1145. 14. National Committee for Clinical Laboratory Standards. 1993. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. M7-A3. (Abstract.) National Committee for Clinical Laboratory Standards, Villanova, Pa. 15. Seppala, H., A. Nissenen, H. Jarvinen, S. Huovinen, T. Henriksson, E. Herva, S. E. Holm, M. Jahkola, M. L. Katila, T. Klaukka, S. Kontiainen, O. Liimatainen, S. Oinonen, L. Passi-Metsomaa, and P. Huovinen. 1992. Resistance to erythromycin in group A streptococci. N. Engl. J. Med. 326:292– 297. (Abstract.) 16. Stevens, D. L., A. E. Bryant-Gibbons, R. Bergstrom, and V. Winn. 1988. The Eagle effect revisited: efficacy of clindamycin, erythromycin, and penicillin in the treatment of streptococcal myositis. J. Infect. Dis. 158:23–28. 17. Stevens, D. L., S. Yan, and A. E. Bryant. 1993. Penicillin binding protein expression at different growth stages determines penicillin efficacy in vitro and in vivo: an explanation for the inoculum effect. J. Infect. Dis. 167:1401– 1405. 18. White, R. L., D. S. Burgess, M. Manduru, and J. A. Bosso. 1996. Comparison of three different in vitro methods of detecting synergy: time-kill, checkerboard, and E test. Antimicrob. Agents Chemother. 40:1914–1918. 19. Yan, S., G. A. Bohach, and D. L. Stevens. 1994. Persistent acylation of high-molecular weight penicillin binding proteins by penicillin induces the postantibiotic effect in Streptococcus pyogenes. J. Infect. Dis. 170:609–614.