ABSTRACT
Rifampin monotherapy was compared to the combination of linezolid or vancomycin with rifampin in an experimental rat model of methicillin-resistant Staphylococcus aureus (MRSA) chronic foreign body osteomyelitis. MRSA was inoculated into the proximal tibia, and a titanium wire was implanted. Four weeks after infection, rats were treated intraperitoneally for 21 days with rifampin alone (n = 16), linezolid plus rifampin (n = 14), or vancomycin plus rifampin (n = 13). Thirteen animals received no treatment. At completion of treatment, qualitative cultures of the wire and quantitative cultures of the bone (reported as median values) were performed. Quantitative cultures from the control, rifampin monotherapy, linezolid-plus-rifampin, and vancomycin-plus-rifampin groups revealed 4.54, 0.71, 0.10, and 0.50 log10 CFU/gram of bone, respectively. The bacterial load was significantly reduced in all treatment groups compared to that in the control group. Rifampin resistance was detected in isolates from 10, 2, and 1 animal in the rifampin, linezolid-plus-rifampin, and vancomycin-plus-rifampin groups, respectively. Cultures of the removed wire revealed bacterial growth in 1 and 2 animals in the rifampin and linezolid-plus-rifampin groups, respectively, with no growth in the vancomycin-plus-rifampin group and growth from all wires in the untreated group. In conclusion, we demonstrated that combination treatment with linezolid plus rifampin or vancomycin plus rifampin is effective in an animal model of MRSA foreign body osteomyelitis in the context of retention of the infected foreign body.
The pathogenesis of chronic orthopedic prosthetic infections is related to the presence of organisms growing in biofilm. Within biofilms, there is an altered microenvironment which may compromise antimicrobial action independent of traditional mechanisms of antimicrobial resistance. Proposed mechanisms include decreased antimicrobial activity against stationary-growth-phase bacteria, modification of antimicrobial agents by inactivating enzymes, decreased penetration, and the presence of persister cells (7).
Management of orthopedic device infections typically involves antimicrobial treatment in combination with a surgical approach. In cases of hardware retention, Staphylococcus aureus infections are often treated with the combination of a nonrifamycin antistaphylococcal agent plus rifampin (8, 29). Indeed, rifampin exhibits bactericidal activity against organisms growing in biofilm, but when administered alone, it is associated with rapid emergence of resistance. S. aureus isolates growing in biofilm demonstrate poor antimicrobial susceptibility compared to the planktonic forms of the same isolates (23, 26). Compounding the issue, methicillin-resistant S. aureus (MRSA) strains with reduced susceptibility to vancomycin have been identified (1). Although vancomycin is considered the drug of choice for invasive infections due to MRSA, alternative agents, such as linezolid, can be used in cases of adverse events or nonsusceptibility to vancomycin. In addition, linezolid has excellent oral bioavailability and thus may be conveniently administered. However, clinical data supporting the use of linezolid for orthopedic prosthetic infections are limited to small nonrandomized studies (3, 19, 21, 22, 28).
In a preliminary experimental rat study of methicillin-susceptible S. aureus (MSSA) foreign body infection, we demonstrated that neither vancomycin nor linezolid alone resulted in a significant decrease in the numbers of bacteria in the bone surrounding the implant or on the implant itself (11). No resistance to either antimicrobial was noted at the end of treatment. Herein, we used the same rat model of experimental foreign body infection but evaluated rifampin monotherapy and rifampin in combination with either linezolid or vancomycin against MRSA.
(This work was presented in part at the 50th Interscience Conference on Antimicrobial Agents and Chemotherapy, Boston, MA, 12 to 15 September 2010.)
MATERIALS AND METHODS
Microorganism.MRSA IDRL 6169, originally recovered from a patient with a prosthetic joint infection, was studied.
Antimicrobial agents.Vancomycin was obtained from Hospira Inc., Lake Forest, IL. Linezolid was obtained from Pfizer Inc., New York, NY. Rifampin was obtained from Akorn Strides, Lake Forest, IL. The MIC of each antimicrobial agent was determined by broth microdilution using an inoculum of 5 × 105 CFU/ml according to Clinical and Laboratory Standards Institute (CLSI) guidelines (4). Similarly, the minimum bactericidal concentration (MBC) was determined according to CLSI guidelines (5).
Pharmacokinetic studies.The pharmacokinetic profile in rats of rifampin, linezolid, and vancomycin administered intraperitoneally was determined by measuring serum concentrations at 0.5, 1, 2, 4, 6, 8, and 12 h after administration (6). The antimicrobials were administered to 15 healthy male Wistar rats for each antimicrobial at the doses listed in Table 1. Serum drug concentrations were measured at steady state (after the third dose). Blood was collected by cardiac puncture or tail vein bleed at each time point, and serum was separated by centrifugation. Serum was placed on paper disks and allowed to diffuse into the bioassay plates (Mueller-Hinton agar with Bacillus subtilis as the indicator organism) for 30 min and then incubated at 30°C overnight. Zone sizes were measured using calipers, and concentrations were calculated by linear regression from a five-point standard curve. Rat serum was used as a diluent for control disks. Pharmacokinetic parameters were calculated using the PK Solutions 2.0 software program (Summit Research Services, Montrose, CO).
Pharmacokinetic studies of rifampin, linezolid, and vancomycina
Experimental rat model.The experimental model described was developed and performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of the Mayo Clinic. Experimental chronic foreign body osteomyelitis was established in male Wistar rats using a modification of Zak's model of experimental osteomyelitis (18). In the modified model, we did not use sodium morrhuate, a sclerosing agent which induces bone damage, and we implanted a titanium wire which acts as a foreign body (Fig. 1). The left legs of the rats were shaved with an electric clipper, and the skin was disinfected with povidone-iodine, USP, 10% topical solution (prep solution; Novaplus, Novation, Inc., Irving, TX). Surgical anesthesia was induced with ketamine (60 mg/kg of body weight) (Ketasat, Fort Dodge, IA), xylazine (6 mg/kg) (Vettek; Phoenix Scientific Inc., St. Joseph, MO), and acepromazine (1 mg/kg) (Boehringer Ingelheim, St. Joseph, MO) administered intramuscularly. The proximal third of the left tibia was surgically exposed, and a 1.5-mm hole was drilled into the medullary cavity. Fifty microliter of 106 CFU suspension of the MRSA isolate was injected into the bone. Subsequently, a 10-mm by 1-mm titanium wire (Zimmer, Warsaw, IN) was implanted into the bone. The hole was covered with dental gypsum. The skin was closed with sutures, and the wound was sprayed with antiseptic.
Illustration of the Wistar rat model of chronic foreign body osteomyelitis. (Copyright, Mayo Foundation for Medical Education and Research; reproduced with permission.)
Four weeks after establishing infection, treatment was initiated. Sixty-one animals (weighing 215 to 475 g) were arbitrarily assigned to one of four study arms: no treatment (n = 13), rifampin alone (n = 16), linezolid plus rifampin (n = 16), or vancomycin plus rifampin (n = 16). Rifampin was administered at 25 mg/kg intraperitoneally twice daily, linezolid at 35 mg/kg intraperitoneally twice daily, and vancomycin at 50 mg/kg intraperitoneally twice daily. Treatment was administered for 21 days. Animals not surviving 21 days of antimicrobial treatment were removed from the experiment and not further studied. There were two deaths in the linezolid-plus-rifampin group and three deaths in the vancomycin-plus-rifampin group. There were no deaths in the other groups.
Twelve hours after completion of antimicrobial therapy, the rats were sacrificed with 100 mg/kg of pentobarbital and the left tibia was aseptically removed. The tibia was frozen to −70°C. Bone within 5 mm of the implanted titanium wire from each animal was weighed and pulverized for quantitative bacterial culture. The pulverized bone was suspended in 2 ml of Trypticase soy broth (TSB), vortexed for 30 s, sonicated at 40 kHz for 5 min, serially diluted, and plated on sheep blood agar plates. Quantitative culture results for bone were obtained after 48 h of incubation and expressed as log10 CFU per gram of tissue. For qualitative cultures, 8 ml of sterile TSB was added to the remaining 2 ml broth containing pulverized bone and incubated for 48 h. For broth with evidence of growth, subcultures were obtained to confirm the presence of MRSA. The wire was aseptically removed, placed in 2 ml of TSB, vortexed, sonicated, and qualitatively cultured as described above.
Emergence of resistance studies.For MRSA recovered from bone after treatment, the MIC of linezolid, rifampin, or vancomycin was determined for three colonies for each culture.
Statistical methods.Descriptive statistics for MRSA log10 CFU per gram of bone were summarized as medians and interquartile ranges. For statistical purposes, we considered an absence of any growth to be 0.1 log10 CFU/gram of bone and growth in qualitative broth culture (but not in quantitative cultures) to be 0.5 log10 CFU/gram of bone. Differences in median log10 CFU of MRSA per gram of bone among the four groups were compared using the Kruskal-Wallis test. Further, pairwise comparisons of median log10 numbers of CFU of MRSA per gram of bone between the four groups (i.e., control and three treatments) were made using the Wilcoxon rank sum test. Based on Bonferroni's correction for adjustment due to multiple comparisons, P values less than 0.008 were considered statistically significant to account for the six pairwise comparisons. For comparison of emergence of rifampin resistance among the treatment groups, we used Fisher's exact test. P values less than 0.025 were considered statistically significant in order to adjust for the multiple comparisons based on Bonferroni's correction. All tests were two sided. Analysis was performed using SAS software (version 9; SAS Institute, Inc., Cary, NC).
RESULTS
The MICs of the MRSA isolate for each antimicrobial were as follows: oxacillin, >128 μg/ml; vancomycin, 1 μg/ml; rifampin, <0.015 μg/ml; and linezolid, 0.25 μg/ml. The MBCs of the isolate were as follows: vancomycin, 1 μg/ml; rifampin, 0.125 μg/ml; and linezolid, 128 μg/ml. The pharmacokinetics of rifampin, linezolid, and vancomycin are shown in Table 1. The data for linezolid (11) and vancomycin (24) were obtained from previous studies. The area under the concentration-time curve from 0 to 24 h (AUC0-24) was 332 μg·h/ml for rifampin, 278 μg·h/ml for linezolid, and 264 μg·h/ml for vancomycin.
Experimental rat model.Results of quantitative cultures of bone after treatment are summarized in Fig. 2. MRSA was cultured from all harvested bones in the control group at a median of 4.54 (range, 3.46 to 5.80) log10 CFU/gram of bone. After treatment, the median count of MRSA was 0.71 (range, 0.10 to 4.51) log10 CFU/gram of bone in the rifampin group. The median counts were 0.10 (range, 0.10 to 3.11) log10 CFU/gram of bone in the linezolid-plus-rifampin group and 0.50 (range, 0.10 to 2.39) log10 CFU/gram of bone in the vancomycin-plus-rifampin group. Differences in colony counts compared to the control group were significant in all treatment groups (P < 0.0001). There was no significant difference between the rifampin group and the linezolid-plus-rifampin group (P = 0.11), between the rifampin group and the vancomycin-plus-rifampin group (P = 0.29), or between the linezolid-plus-rifampin group and the vancomycin-plus-rifampin group (P = 0.62).
Results of quantitative bone cultures. Median value, interquartile range, and individual results are shown for each study group. The differences between the control and each treatment group were statistically significant (P < 0.0001). Cultures that demonstrated resistance to rifampin after treatment are shown in red.
Cultures of the titanium wires were also obtained. MRSA was cultured from titanium wires in all control animals. There was no growth in 15/16 animals (94%) in the rifampin group, in 12/14 animals (86%) in the linezolid-plus-rifampin group, and in 13/13 animals in the vancomycin-plus-rifampin group.
Emergence of resistance (rifampin MIC after treatment, >1 μg/ml) was detected in isolates from 10 animals (63%) in the rifampin group, 2 animals (14%) in the linezolid-plus-rifampin group, and 1 animal (8%) in the vancomycin-plus-rifampin group. The differences between the monotherapy and combination treatment groups were statistically significant (P = 0.010 between the rifampin monotherapy and linezolid-plus-rifampin groups, and P = 0.006 between the rifampin monotherapy and vancomycin-plus-rifampin groups). The findings remained significant in spite of adjusting for multiple comparisons. There was no emergence of resistance to linezolid or vancomycin in the respective treatment groups.
DISCUSSION
In the present study, we demonstrated that combination treatment with linezolid plus rifampin or vancomycin plus rifampin is more effective than no treatment in an animal model of MRSA foreign body osteomyelitis in the context of retention of the infected foreign body. We studied an MRSA isolate recovered from a patient with prosthetic joint infection. We treated the animals for 21 days; previous studies performed in our laboratory had demonstrated that this is the optimal duration of therapy of experimental osteomyelitis not associated with a foreign body (20). Other investigators have used the same duration of parenteral antimicrobial treatment in a rat model of foreign body-associated osteomyelitis involving the tibia (12, 15).
Five deaths occurred, all in animals receiving combination therapy. Combination treatment involved two intraperitoneal injections, whereas rifampin was administered as a single injection each time. In addition, the volume of the injected rifampin was lower than that of the other drugs. Of note, in previous experiments with monotherapy we did not observe any rat losses. Hence we suspect that the deaths were related to puncture of a major vessel or a visceral organ and were not due to an effect of the infection or specific drug toxicity. Indeed, one of the deaths occurred immediately after the injection. The animal losses did not affect the statistical power of our study to detect differences among the groups.
In a previous study, we demonstrated that linezolid or vancomycin alone did not eradicate infection in a rat model of MSSA foreign body osteomyelitis (11). More specifically, the median counts were 3.87 log10 CFU/gram of bone for vancomycin alone, 4.98 for linezolid, and 5.45 for the untreated control group. MSSA was recovered from wire cultures in 12/15 animals (80%) in the vancomycin group, 14/15 animals (93%) in the linezolid group, and 10/12 animals (83%) in the control group. Differences among the groups for either the bone or the wire cultures were not statistically significant after adjusting for multiple comparisons. In the present study, we have demonstrated in vivo activity of rifampin administered alone or in combination with linezolid or vancomycin against MRSA. Rifampin monotherapy was associated with a significantly higher emergence of resistance than combination treatment. Nonetheless, rifampin resistance was detected even with combination treatment. Emergence of resistance should be considered in cases of clinical treatment failure, even with the use of rifampin combination therapy. Even in the context of resistant isolates, low colony counts were measured by quantitative cultures. There was no emergence of resistance for either linezolid or vancomycin. Notably, though, both antimicrobials were given in combination with rifampin and not as monotherapy. Still, in our prior study (11) where these agents were given alone, we did not note emergence of resistance.
To our knowledge, this is the first article to report the outcome of combination treatment with linezolid plus rifampin in an animal model of foreign body infection with bone involvement. Our findings are in accordance with results reported by other investigators in different models employing guinea pigs or rats (2, 16). In these models, multiperforated polytetrafluoroethylene (Teflon) tissue cages were subcutaneously implanted in the flanks of the study animals. There was no involvement of bone with infection. The efficacy of combination treatment with rifampin and other antistaphylococcal agents, including teicoplanin (27) and daptomycin (13), has been shown in other animal experiments where, again, tissue cages were employed. The activity of combination treatment with rifampin plus linezolid was also shown in a rat model of vascular graft infection (25).
We harvested the bone for culture 12 h after treatment cessation. Other investigators have harvested bone at later time points (9, 14, 17). Antibiotic carryover may explain culture negativity, particularly for the wires, a potential limitation of our study. Notably, in a tissue cage-associated MRSA infection model, Baldoni et al. showed 50 to 60% cure rates against adherent bacteria 5 days after cessation of combination treatment with linezolid and rifampin (2).
The pharmacokinetic studies showed that the linezolid, vancomycin, and rifampin levels obtained a half-hour after injection were above the peak levels that are considered therapeutic in humans. The half-life of linezolid and vancomycin was shorter than that in humans, and vancomycin trough levels were undetectable. Nevertheless, culture results suggest activity of the study drugs at the concentrations administered. Rifampin levels were detectable at trough. Regarding drug-drug interactions, in healthy male volunteers, coadministration of linezolid with rifampin resulted in a decrease in the levels of linezolid (10). The clinical relevance of this observation is unknown. The medication is not an inducer of the cytochrome P-450 enzymes, and the mechanism of interaction has not been elucidated. There is no known interaction between vancomycin and rifampin.
In conclusion, we demonstrated that the combination of linezolid plus rifampin or vancomycin plus rifampin is efficacious in a rat model of MRSA chronic osteomyelitis. Combination treatment was associated with lower rates of emergence of rifampin resistance. Resistance to either linezolid or vancomycin was not demonstrated. The results provide data that may be useful to clinicians treating orthopedic prosthetic or other biofilm-associated infections.
ACKNOWLEDGMENTS
We thank Pfizer, Inc., for supporting the studies performed herein.
FOOTNOTES
- Received 1 June 2010.
- Returned for modification 15 November 2010.
- Accepted 18 December 2010.
- Accepted manuscript posted online 28 December 2010.
- Copyright © 2011, American Society for Microbiology
