Efficacy of Treatment of Staphylococcal Osteomyelitis

Transcription

1 ANTIMICROBLAL AGENTS AND CHEMOTHERAPY, Dec. 1992, p /92/ $02.00/0 Copyright 1992, American Society for Microbiology Vol. 36, No. 12 Relationship between Antibiotic Concentration in Bone and Efficacy of Treatment of Staphylococcal Osteomyelitis in Rats: Azithromycin Compared with Clindamycin and Rifampin TERRY O'REILLY,l SAMUEL KUNZ,' ERIC SANDE,2 OTO ZAK,1 MERLE A. SANDE,2 AND MARTIN G. TAUBER2* Pharma Research, Ciba-Geigy Ltd., Basel, Switzerland,' and Infectious Diseases Laboratories, San Francisco General Hospital, and Department ofmedicine, University of California, San Francisco, California Received 21 May 1992/Accepted 21 September 1992 We examined the effect of azithromycin (CP-62,993), a new oral macrolide-like antibiotic, alone and in combination with rifampin, as treatment for experimental staphylococcal osteomyelitis. Clindamycin was used as a comparison drug. Rats (n = 10 to 15 per group) were infected by direct instillation of Staphylococcus aureus into the tibial medullary cavity. After 10 days, 21-day treatments with azithromycin (50 mg/kg of body weight, once daily, by the oral route), rifampin (20 mgtkg, once daily, subcutaneously), or clindamycin (90 mg/kg, three times daily, by the oral route) were started. The drugs were used singly or in combination (azithromycin plus rifampin or cindamycin plus rifampin). Peak azithromycin concentrations in bone were >30 times higher than levels in serum, but the drug had little effect on final bacterial titers ( log1o CFU/g of bone; for controls, 6.54 t 0.28 log1o CFU/g). CUindamycin was more active than azithromycin (3.26 ± 2.14 log1o CFU/g of bone; 20%Yo of sterilized bones), but rifampin was the most active single drug ( loglo CFU/g; 53% of sterilized bones). Therapy with rifampin or clindamycin alone was associated with the emergence of resistance. Rifampin plus azithromycin (0.51 ± 1.08 log1o CFU/g of bone; 80%o of sterilized bones) and rifampin plus clindamycin (0.87 ± 1.34 log1o CFU/g of bone; 66% of sterilized bones) were the most active regimens. Thus, azithromycin is ineffective as a single drug for the treatment of experimental staphylococcal osteomyelitis, despite high levels in bone that markedly exceded the MIC, but it may be an attractive partner drug for rifampin. Osteomyelitis is a difficult infection to treat. The majority of cases are caused by staphylococci, gram-negative bacteria, and anaerobes (12). If not treated adequately in the acute phase, there is a high risk that the disease will progress to a chronic infection, possibly becoming refractory to further therapeutic intervention. The treatment of acute-phase osteomyelitis relies on prolonged therapy with antibiotics. Since the duration of treatment is at least 4 to 6 weeks, effective oral antibiotic therapy would represent a significant advantage over standard parenteral regimens. However, most orally active antibiotics achieve concentrations in bone only slightly higher than their MICs (12). Rifampin is highly active against staphylococci, it can be administered orally once daily, and it achieves high levels in tissue, including high intracellular levels in macrophages (4). The drug is used to treat staphylococcal osteomyelitis, but when used as a single agent, emergence of resistance is frequently observed (13). In contrast, when rifampin is used in combination with other antibiotics, the in vivo efficacy of the combination appears to be increased and emergence of resistance can be prevented. The combination of rifampin and ciprofloxacin is used to treat some staphylococcal infections (3), but isolates resistant to ciprofloxacin are emerging (1). Therefore, it is important for investigators to evaluate potential new partner antibiotics for rifampin. Azithromycin (CP-62,993) is a novel orally active azalide * Corresponding author compound displaying a long half-life in serum and tissues, pharmacokinetic properties that make it an attractive drug for use in the treatment of acute-phase osteomyelitis caused by a susceptible organism (5, 6, 8, 18). Furthermore, concentrations of this antibiotic in tissues, including bone (5) and macrophages (9), are higher than its simultaneous concentrations in serum; such properties probably account for the observation that the in vivo efficacy of this drug is superior to what would be expected on the basis of its levels in serum and its in vitro activity (6, 7, 14, 16). Azithromycin is active against most gram-positive organisms, including Staphylococcus aureus (10, 20). In the present study, we used azithromycin in an animal model of acute osteomyelitis to address two objectives: first, to examine the efficacy of azithromycin as a single drug in relationship to its achievable concentrations in bone and its in vitro activity and, second, to evaluate the potential of azithromycin as a partner drug for rifampin in the treatment of staphylococcal osteomyelitis. Clindamycin was used as a comparison antibiotic in the studies. MATERIALS AND METHODS Antibiotics. Azithromycin was obtained from Pfizer Central Research, Groton, Conn. Clindamycin (Dalacin C) was obtained from Upjohn, Purrs, Belgium. Rifampin (Rimactan) was obtained from CIBA-GEIGY Limited, Basel, Switzerland. Bacterial strain. S. aureus 1098 was used in the studies. It

2 2694 O'REILLY ET AL. was stored frozen (-125 C) in brain heart infusion broth supplemented with 10% (vol/vol) fetal calf serum at a density of 4 x 1010 CFU/ml and was used withoutsubculturing. Rats were inoculated with 2 x 109 CFU per tibia. MIC and MBC determinations. The in vitro antibiotic susceptibilities (MICs and MBCs) of S. aureus 1098 were determined by first determining the MIC by a broth tube macrodilution method with an inoculum of 6.0 log1o CFU/ml and then determining the MBC by subculturing each antibiotic dilution onto agar plates; brain heart infusion (BBL Microbiology Systems, Cockeysville, Md.) medium was used. The MIC was defined as the lowest antibiotic concentration that inhibited visible growth, and the MBC was defined as the lowest concentration of antibiotic that reduced the inoculum by 99.9%. Rat model of acute staphylococcal osteomyelitis. The experimental protocols were approved by the Ethical Committee of the Kantonales Veterinaramt of Basel Stadt and were based on previous descriptions (17, 22). One hundred five adult male Madorin rats (weight, approximately 200 g each) were used in the experiments. The rats were first sedated subcutaneously (s.c.) with 0.5 mg of fluanisone plus 0.01 mg of fentanyl citrate (Hypnorm; Janssen) and were then anesthetized with an intramuscular injection of 0.3 mg of midazolam-0.67 mg of fluanisone mg of fentanyl citrate. The left tibia was surgically exposed, and a 1-mm hole was bored with a dental drill into the medullary cavity of the proximal tibia. Bones were infected by first injecting 0.05 ml of 5% sodium morrhuate (Torigan Laboratories, Queens Village, N.Y.); this was followed by an injection of 0.05 ml of the bacterial inoculum. The hole was plugged with dental gypsum (Contura, Zurich, Switzerland), and the wound was closed. The animals were then returned to individual cages and treatment was started 10 days later. In a subset of animals, both tibiae were infected in order to assess the reproducibility of infection in the model. In the untreated animals, bacterial titers in the two tibiae were virtually identical ( versus log1o CFU/g of bone); therefore, subsequent results are for the left tibia only. Bone bacteriology. Four days after the termination of therapy, rats were sacrificed and the infected tibiae were removed, dissected free of adhering soft tissue, weighed, and then frozen in liquid nitrogen. The bones were then powdered by using individual metal ball mills (Retsch, Haan, Federal Republic of Germany). The powder was then mixed with 2 to 3 ml of 0.9% NaCl, and the resultant homogenate was serially diluted and plated onto brain heart infusion agar for determination of CFU. No loss of bacterial viability occurred during this procedure. Homogenates were also plated onto antibiotic-containing plates (10,g/ml for azithromycin or clindamycin, 100 p.g/ml for rifampin) in order to detect antibiotic-resistant subpopulations. Additionally, 0.5 ml of the homogenates was placed in 10 ml of brain heart infusion broth, and the mixture was cultured for 48 h in order to test for sterility of the bone. Results were expressed as loglo CFU per gram of bone. Determination of antibiotic concentrations in serum and bone. The pharmacokinetics of the antibiotics used in the present study were determined in groups of uninfected animals administered a single dose of antibiotic and then sacrificed sequentially over a 24-h time period. Serum was prepared from blood taken from the vena cava, and bone homogenates were prepared as described above. Antibiotic concentrations were determined by agar diffusion bioassays with Micrococcus luteus ATCC 9341 as the indicator strain, ANTIMICROB. AGENTS CHEMOTHER. whereas the rifampin assay used Antibiotic Medium No. 1 (BBL Microbiology Systems), the assays for clindamycin awlazithromycian used Antibiotic Medium No. 1 supplemented with 5 ml of 1 N NaOH per liter. In addition, five infected rats were treated with each of the study drugs according to the regimens outlined below in order to determine the trough antibiotic concentrations in bone at the end of 21 days of therapy. Animals given azithromycin and rifampin were sacrificed 24 h after the last dose, while the interval was 8 h for animals given clindamycin. When calculating antibiotic concentrations, the value corresponding to the lower limit of detectability was used when drug was not detectable. Experimental treatment. Beginning 10 days postinfection, animals were treated for 21 days. Azithromycin was given orally (p.o.) once a day at 50 mg/kg of body weight, clindamycin (90 mg/kg) was administered p.o. every 8 h, and rifampin was administered s.c. once a day at a dose of 20 mg/kg. Statitical analysis. Data are expressed as means ± standard deviations and as the percentage of sterile bones. When determining the statistical differences between drug titers in bone, animals with sterile bones were assigned values of 0 CFU/g of bone. Comparisons between treatment groups were performed by analysis of variance (Fisher significant difference method for pairwise comparisons). Determination of differences in the treatments with respect to ratios of bones sterilized was performed by logistic regression analysis, RESULTS MIC and MBC determinations. The MICs and MBCs of the study drugs for S. aureus 1098 were as follows: azithromycin, 1.0 and 1.0,ug/ll; rifampin 0.06 and 0.06 p,g/ml; and clindamycin 0.12 and 0.25 ±g/ml, respectively. Pharmacokinetics in serum and bone. The pharmacokinetics of rifampin in the model of osteomyelitis after a s.c. dose of 20 mg/kg have been determined previously (2). In that study, peak levels in serum were 12.6 pg/ml, with a half-life (tl,) of 5.5 h, and peak levels in bone were 19.1,g/g, with a tla similar to that in serum. In the present study, administration of azithromycin (50 mg/kg, p.o.) to uninfected animals resulted in a mean peak concentration in serum of h) and a mean peak concentration in bone of ug/g (tj/2, 20.7,ug/g (tj12, >24 h) (Table 1). As a reference tissue, liver accumulated azithromycin to a peak level of 42.8 ± 20.7 p,g/g, which was similar to the level of accumulation reported previously (6, 18). Clindamycin (90 mg/kg, p.o.) application to uninfected rats yielded peak concentrations of 2.98 pg/ml in serum and 3.22 pg/ml in bone, and t. values were 4.2 and 7.2 h for serum and bone, respectively (Table 1). At the end of 21 days of therapy of infected animals, trough antibiotic levels of clindamycin and rifampin in bone were below the level of detectability (-1 pg/g; Table 1). In contrast, azithromycin accumulated in bone during therapy and resulted in trough concentrations of 92.7 pg/g of bone 24 h after the last treatment; this was more than four times greater than the peak concentration after the initial dose (P < 0.001; Table 1). Furthermore, in infected rats azithromycin was present in bone at a concentration of 46.6 ± 5.7 p,g/g 4 days after the last dose. Efficacy of single-drug therapy. Despite high concentrations in bone relative to in vitro MIC-MBC (Table 1), azithromycin was minimally effective. After 21 days of therapy, bacterial titers in bone had declined by approxi-

3 VOL. 36, 1992 EXPERIMENTAL S. AUREUS OSTEOMYELITIS 2695 TABLE 1. Pharmacokinetic data for azithromycin, rifampin, and clindamycin in experimental staphylococcal osteomyelitis in rats Uninfected rats Infected rats Dr[ g (dose Site Posttherapy trough TrouglVMlC Drug/k](doste)St Peak concn tl h[en) Peak/MIC concn (isg/nml or Tratoug I (pg/ml or mg/g) t(h[en) ratio mg/g)rai Azithromycin (50, p.o.) Serum 0.63 ± <1 NDa Bone 20.7 ± 15.6 > ± Clindamycin (90, p.o.) Serum 2.98 ± ND Bone 3.22 ± <1 <8 Rifampin (20, s.c.)r Serum ND Bone <1 <15 a ND, not determined. b Data for rifampin were from a previous report (2). mately 1 log1o CFU/g compared with the titers in controls, and none of the animals had sterile bones (Table 2). Quantitation of bacterial titers was apparently not hampered by the high concentrations of azithromycin in bone (carryover effect), since bacterial counts on the plates diluted out in the expected 10-fold steps all the way to the lowest dilution. Clindamycin administered singly was more effective than azithromycin in reducing bacterial titers (P < 0.01), yielding a reduction of about 3 log1o CFU/g compared with the titers in controls and sterilizing 20% of the bones (not significant for controls). Rifampin was the most active single drug in reducing the CFU in bone (P < 0.01 versus clindamycin or azithromycin), with final bone bacterial titers being 5 log1o CFU/g lower than titers in controls and resulting in sterilization of 53% of the bones (P < 0.05 versus controls). Development of resistance during single-drug therapy. Single-drug therapy with azithromycin was not associated with the development of azithromycin-resistant organisms, while 2 of 15 rats treated with rifampin alone and 3 of 10 animals treated with clindamycin alone had organisms resistant to the treatment drug at the end of therapy. The percentage of resistant organisms relative to the total number of bacteria recovered varied greatly for both drugs (between 0.01 and 100% of the recovered organisms were resistant). Efficacy of combination therapy. With respect to reductions in CFU per gram of bone, the combination of azithro- TABLE 2. Efficacy of antibacterial therapy for experimental staphylococcal osteomyelitis with azithromycin, cindamycin, or rifampin alone or in combination Bacterial titers Treatment No. of in bone (loglo % Sterile rats CFU/g [mean bone + SD]) Controls ± 0.28a 0 Azithromycin ± 0.46b 0 Clindamycin ± 2.14b,c 20 Rifampin ± 1.92",d 53e Clindamycin-rifampin cf 66e Azithromycin-rifampin ± 1.08d"f 80e a Significantly different from all active treatment groups (P < 0.05). b The three single-drug therapies were significantly (P < 0.05) different from each other. c p < dp < P < 0.05 compared with controls and azithromycin; not significant between the three groups. f Not significant. mycin plus rifampin (final titer, log1o CFU/g of bone) was significantly more effective than either drug alone (P < 0.05; Table 2). This improved efficacy appeared to be real and not a carryover effect of the high concentrations of azithromycin in bone, since all three rats with positive cultures had counts in the 10-1 dilution, and these counts were similar to counts in the same dilution for samples obtained from animals treated with rifampin alone. Clindamycin plus rifampin (final titers, log1o CFU/g) was significantly better than clindamycin alone (P < 0.01), but the effects of the clindamycin-rifampin combination were not statistically different from those of either rifampin alone or rifampin plus azithromycin (Table 2). Although azithromycin plus rifampin yielded the highest proportion of sterile bones (80%), this combination was not statistically better than clindamycin plus rifampin or rifampin alone (Table 2). Furthermore, the addition of clindamycin or azithromycin to rifampin completely prevented the emergence of the antibiotic resistance observed with single-drug therapy. DISCUSSION Azithromycin as single agent was less effective than the two comparison drugs, clindamycin and rifampin, in the model of S. aureus osteomyelitis described here. The dosage used in our studies resulted in concentrations of azithromycin in serum similar to those achieved in human serum after a 500-mg oral dose (0.4,ug/ml) (8). The pharmacokinetic results obtained in the present study are in line with those obtained in previous experimental studies with similar dosage regimens (peak levels in serum, 1.2,ug/ml [6] or 0.29,g/ml [18]). Furthermore, the tj/2 in serum in the present study was similar to those obtained previously in rats (6, 18) and humans (5). As expected on the basis of previous work that showed selective drug uptake into various tissues (5, 6, 8, 18) and further supported by our data confirming accumulation of drug in liver, the relatively low concentration in serum was associated with concentrations in rat bone more than 30 times higher than the peak concentration in serum after a single oral dose. Note, however, that the azithromycin level in rat bone at 24 h ( ,ug/g) was considerably greater than that observed in humans approximately 1 day following a 500-mg p.o. dose (-1 p,g/g) (5). Azithromycin accumulated in infected bone during therapy; the level in bone was about fourfold greater than the peak level after a single dose. This was in contrast to the results obtained with the two comparison drugs; trough concentrations of both comparison drugs were undetectable at the end of therapy.

4 2696 O'REILLY ET AL. Thus, the pharmacokinetics of azithromycin in the osteomyelitis model described here were as expected, but the drug showed little effect against a susceptible Staphylococcus species in vivo, despite levels in bone that were substantially greater than the MIC for the organism (the ratio between trough concentrations in bone at the end of therapy and the MIC was -90). In contrast, the ratios of concentrations in bone to the MIC were lower for clindamycin and rifampin, yet these compounds demonstrated better in vivo activity compared with that of azithromycin. The reasons for the discrepancy between the expected and observed in vivo efficacy of azithromycin in our model are unclear and are compounded by the observation that azithromycin potentiated the effect of rifampin. This is similar to observations made with vancomycin, which was also minimally effective alone but potentiated the effect of rifampin (2). In congruence with previous results (13), development of resistance did not appear to be a problem with azithromycin, in contrast to the case with both clindamycin and rifampin when they were each used alone. Several explanations could account for the results obtained with azithromycin in the present study. (i) Because the in vitro activity of azithromycin is greatly reduced at low ph (15), the acidification of tissues that commonly occurs during infection, although not specifically known to occur in patients with osteomyelitis, may result in a reduced potency of azithromycin at this target tissue; (ii) the unique pharmacokinetics of azithromycin may sequester the agent from bacteria in the extracellular fluid space of the bone; and (iii) the bacteria may phenotypically demonstrate increased resistance or tolerance in the present model of osteomyelitis, stressing the efficacies of the study drugs. With respect to this last possibility, preliminary experiments indicate that the MIC-to-MBC ratios of azithromycin for anaerobically grown S. aureus (1:32) were higher than those for aerobically grown organisms (1:1), suggesting that the phenotypic tolerance of this strain to azithromycin can be invoked by anaerobiosis. Despite the minimal activity of azithromycin when it was used as a single drug, the combination of azithromycin with rifampin proved to be highly effective in our model. On the other hand, the combination of clindamycin-rifampin was statistically not more effective than rifampin alone, even though clindamycin was more active than azithromycin when they were used as single agents. These results underscore the difficulties in predicting the in vivo efficacies of antibiotics in osteomyelitis on the basis of in vitro studies or the levels of antibiotics in bone (11, 12). Although rifampin was the most active single drug in the present study, there was, as expected, development of resistance associated with this regimen (13); an important objective of the addition of a second antibiotic was the prevention of this phenomenon, which was achieved with either azithromycin or clindamycin İt cannot be overemphasized that studies in animal models are inherently of limited predictive value for the human clinical situation. For example, the disease process induced in our model was only an imperfect imitation of chronic osteomyelitis in humans, which may result in differences in the pathogen-drug-host defense interactions that can affect the results of therapy. Nevertheless, our results indicate that azithromycin may be an attractive partner drug for rifampin in the treatment of staphylococcal osteomyelitis, even though traditional macrolides do not play a role in the treatment of osteomyelitis (12). Azithromycin plus rifampin was at least as efficacious as clindamycin plus rifampin in reducing bacterial titers in bone and sterilizing bone tissue. ANTIMICROB. AGENTs CHEMOTHER. Furthermore, the pharmacokinetic properties of both azithromycin and rifampin allow treatment with single daily oral doses, an attractive dosing schedule for a disease that uniformly requires prolonged therapy. Quinolones, such as ciprofloxacin, can also be used in combination with rifampin in combating staphylococcal osteomyelitis (19, 21), but they are contraindicated in children, an age group in which staphylococcal osteomyelitis occurs frequently. It appears that further studies that will examine the value of azithromycin as a partner drug for rifampin in the therapy of staphylococcal osteomyelitis are warranted. ACKNOWLEDGMENTS The technical assistance of M. Hattenberger, J. Vaxelaire, and S. Tobler is gratefully acknowledged. We are grateful to George Preiswerk for assistance with the statistical evaluations. This study was supported by a grant from Pfizer Laboratories, New York, N.Y. REFERENCES 1. Daum, T. E., D. R. Schaberg, M. S. Terpenning, W. S. Sottile, and C. A. Kauffnan Increasing resistance of Staphylococcus aureus to ciprofloxacin. Antimicrob. Agents Chemother. 34: Dworldn, R., G. Modin, S. Kunz, R. Rich, 0. Zak, and M. Sande Comparative efficacies of ciprofloxacin, pefloxacin, and vancomycin in combination with rifampin in a rat model of methicillin-resistant Staphylococcus aureus chronic osteomyelitis. Antimicrob. Agents Chemother. 34: Dworldn, R. J., M. A. Sande, B. L. Lee, and H. F. Chambers Treatment of right-sided Staphylococcus aureus endocarditis in intravenous drug users with ciprofloxacin and rifampicin. Lancet ii: Farr, B., and G. L. Mandell Rifampin, p In G. L. Mandell, R. G. Douglas, and J. E. Bennett (ed.), Antimicrobial therapy. John Wiley & Sons, Inc., New York. 5. Foulds, G., R M. Shepard, and R. B. Johnson The pharmacokinetics of azithromycin in human serum and tissues. J. Antimicrob. Chemother. 25(Suppl. A): Girard, A. E., D. Girard, A. R Engish, T. D. Gootz, C. R Clnochowski, J. A. Faiella, S. L. Haskell, and J. A. Resma Pharmacokinetics and in vivo studies with azithromycin (CP-62,993), a new macrolide with an extended half-life and excellent tissue distribution. Antimicrob. Agents Chemother. 31: Girard, A. E., D. Girard, and J. A. Retsema Correlation of the extravascular pharmacokinetics of azithromycin with in-vivo efficacy in models of localized infection. J. Antimicrob. Chemother. 25(Suppl. A): Lode, H The pharmacokinetics of azithromycin and their clinical significance. Eur. J. Clin. Microbiol. Infect. Dis. 10: McDonald, P. J., and H. Pruul Phagocyte uptake and transport of azithromycin. Eur. J. Cin. Microbiol. Infect. Dis. 10: Neu, H. C Clinical microbiology of azithromycin. Am. J. Med. 91(Suppl. 3A):12S-18S. 11. Norden, C. W Lessons learned from animal models of osteomyelitis. Rev. Infect. Dis. 10: Norden, C. W Osteomyelitis, p In G. L. Mandell, R. G. Douglas and J. E. Bennett (ed.), Principles and practice of infectious diseases, 3rd ed. Churchill Livingstone, New York. 13. Norden, C. W., and M. Shaffer Treatment of experimental chronic osteomyelitis due to Staphylococcus aureus with vancomycin and rifampin. J. Infect. Dis. 147: Pechere, J. C The activity of azithromycin in animal models of infection. Eur. J. Clin. Microbiol. Infect. Dis. 10: Retsema, J. A., A. E. Girrd, L. A. Brenn, C. R. Cimochowsld, and J. A. Faiela Lack of emergence of signifi-

5 VOL. 36, 1992 cant resistance in vitro and in vivo to the new azalide antibiotic azithromycin. Eur. J. Clin. Microbiol. Infect. Dis. 10: Retsema, J. A., A. E. Girard, D. Girard, and W. B. Millisen Relationship of high tissue concentrations of azithromycin to bactericidal activity and efficacy in vivo. J. Antimicrob. Chemother. 25(Suppl. A): Rissing, J. P., T. B. Buxton, R. S. Weinstein, and R. K Shockley Model of experimental chronic osteomyelitis in rats. Infect. Immun. 47: Shepard, R. M., and F. C. Falkner Pharmacokinetics of azithromycin in rats and dogs. J. Antimicrob. Chemother. 25(Suppl. A): Waldvogel, F. A Use of quinolones for the treatment of osteomyelitis and septic arthritis. Rev. Infect. Dis. 11(Suppl. EXPERIMENTAL S. AUREUS OSTEOMYELITIS ): Williams, J. D Spectrum of activity of azithromycin. Eur. J. Clin. Microbiol. Infect. Dis. 10: Wispelvey, B., and W. M. Scheld Ciprofloxacin in the treatment of Staphylococcus aureus osteomyelitis: a review. Diagn. Microbiol. Infect. Dis. 13: Zak, O., F. Zak, R. Rich, W. Tosch, F. Kradolfer, and W. M. Scheld Experimental staphylococcal osteomyelitis in rats: therapy with rifampin and cloxacillin, alone or in combination, p In P. Periti and G. G. Grassi (ed.), Current chemotherapy and immunotherapy. Proceedings of the 12th International Congress of Chemotherapy. American Society for Microbiology, Washington, D.C.