Vol. 61, No. 12INFEcrION AND IMMUNITY, Dec. 1993, p. 5225-52300019-9567/93/125225-06$02.00/0Copyright ©) 1993, American Society for Microbiology

Chronic Staphylococcal Osteomyelitis: a New ExperimentalRat Model

N. SPAGNOLO,1* F. GRECO,2 A. ROSSI,1 L. CIOLLI,3 A. TETI,3 AND P. POSTERARO'Department ofMicrobiology, 1 and Department of Orthopedics, 3 School ofMedicine, Catholic University of

the Sacred Heart, Rome, and Department of Orthopedics, School ofMedicine,University ofAncona, Ancona,2 Italy

Received 24 June 1993/Returned for modification 13 August 1993/Accepted 23 September 1993

A rat model of chronic staphylococcal osteomyelitis was developed. Fibrin glue (5 Il) and Staphylococcusaureus (2 x 106 CFU/5 vIA) were inoculated into the proximal metaphysis of the tibia. The rats were killed atintervals of between 1 and 6 months, and the tibias were removed. Induced lesions were evaluated byradiographic, macroscopic, and histological examinations and bacterial counts. Roentgenograms revealedosteomyelitis in more than 90%o of the tibias. Gross bone pathology revealed skeletal deformation, new boneformation, abscesses, and draining skin fistulas in more than 80%o of cases. Histological examination revealedosteomyelitis in more than 90%o of cases, and bacterial counts were positive in 86% of cases. Only fibrin glue(5 ,lI) was inoculated into controls. Controls showed no osteomyelitic lesions, and counts were negative in sevenof eight control tibias. The main feature of this model is the use of fibrin glue instead of the sclerosing agentsand foreign bodies used in other models. The model reproduces lesions similar to those ofhuman posttraumaticosteomyelitis and can be reliably used in pathophysiological and therapeutic studies.

Osteomyelitis is an infective process in bone and bonemarrow. Despite progress in the field of diagnosis andtherapy of infectious diseases, some aspects of the patho-genesis, early diagnosis, and therapy of osteomyelitis arestill unclear because of the multiple variables involved. It istherefore necessary to follow a methodological path which isboth clinical and experimental. Therefore, animal models ofosteomyelitis able to reproduce the clinical and gross patho-logical phases of the disease must be developed.Animal models of acute and chronic osteomyelitis have

been developed with rats (24, 25, 31), rabbits (1, 17-19, 22,26, 27, 30), dogs (11, 13), guinea pigs (23), and chickens (12).In these models, osteomyelitic lesions are induced by use ofsclerosing agents and foreign bodies with bacterial strains.Sclerosing agents probably cause alterations of the marrowcirculation (10), and intramedullary foreign bodies couldconstitute an inert substratum to which Staphylococcusaureus adheres tenaciously and produces the extracellularhexopolysaccharide matrix called the glycocalyx (18, 19).The aim of our study was to develop a rat model of chronicstaphylococcal osteomyelitis with pathophysiological, clini-cal, and gross pathological characteristics similar to those ofthe human disease by using fibrin glue instead of sclerosingagents or foreign bodies.

MATERIALS AND METHODS

Animals. Seventy-six male outbred Wistar rats (250 to 350g) were used. Seven of these died of anesthetic complica-tions. The rats were individually caged, fed a standard pelletdiet, and supplied with water ad libitum. They were dividedinto two groups, A and B, consisting of 61 and 8 animals,respectively. Group A was treated with fibrin glue and S.aureus, and group B (control) was treated with fibrin gluealone.The two groups were each divided into four subgroups as

* Corresponding author.

a function of the time interval between surgery and death(Table 1). The animals were killed by CO2 asphyxiation.

Bacterial strain and preparation of inocula. The S. aureusstrain used in our model was isolated from a clinical speci-men of a patient with osteomyelitis. This strain was identi-fied by biochemical tests and MIC assays against severaldrugs by use of the Sceptor system (Becton Dickinson). Nostrains other than this one were tested in our model.An S. aureus colony was inoculated on nutrient agar

(Difco) slants and incubated at 37°C for 24 h. Subsequently,a subculture was made in nutrient broth (Difco) for 18 h at37°C. This culture was then diluted to obtain 2 x 106 CFU/5>l.

Fibrin glue. Reconstituted lyophilized human fibrin (Tis-sucol 0.5 kit; Immuno, Vienna, Austria) was used. Fivemicroliters of fibrin glue was inoculated into each tibia.

Experimental protocol. The rats were anesthetized in-traperitoneally with 2 mg of diazepam (Valium; Roche,Milan, Italy) and intramuscularly with 2.5 mg of ketamine(Ketalar; Parke-Davis, Milan, Italy) per 100 g of bodyweight. Both hind legs were shaved and disinfected withpolyvinylpyrrolidone-iodine (Betadine; Chinoin, Milan, Ita-ly). The anterior tibial metaphysis of each leg was surgicallyexposed, and a hole was drilled through the cortex by use ofa high-speed drill with a 0.6-mm-diameter bit.Group A was inoculated with fibrin glue (5 pl) and S.

aureus (2 x 106 CFU/5 pl), and group B was inoculated withfibrin glue alone (5 pl). The cortical bore was then closedwith bone wax to avoid leakage of the inoculum, and themuscle fascias and skin were sutured.

Assessment. After the animals had been killed, the lesionswere assessed in a blinded manner on the basis of roentgen-ograms, gross bone pathology, histological examination, andmicrobiological analysis. Two reviewers helped the authorswith the radiographic assessment, gross bone pathology, andhistological examination. Gross bone pathology was deter-mined when the tibias were removed. Radiographic assess-ment and histological examination were performed at theend of the experiment.

5225

on January 26, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

5226 SPAGNOLO ET AL.

TABLE 1. Observation time and distribution of subgroups

Subgroup Days No. of rats

Al 30 9A2 60 10A3 90 12A4 180 30

B1 30 2B2 60 2B3 90 2B4 180 2

Both tibias were removed. One, chosen at random, wasused for microbiological analysis, and the other was used forradiographic and histological investigations.

Gross bone pathology was determined in both tibiasaccording to the grading of Rissing et al. (25): 0, absence ofabscesses, sequestra, active bone formation, and erythema;1, minimal erythema, without abscesses or new bone forma-tion; 2, erythema, with widening of the head and shaft andnew bone formation; 3, abscesses, with new bone formation,sinus tract drainage, or grossly purulent exudate; 4, severebone resorption, abscesses, and diaphyseal or total tibialinvolvement.For microbiological analysis, muscle and connective tis-

sue were removed from the tibia, which was then crushedwith bone rongeurs and pulverized with a mortar and pestleunder sterile conditions. The powder was suspended in 5 mlof saline, and serial dilutions of up to 106 were prepared. Onemilliliter of each dilution was added to 9 ml of nutrient agar,placed in an incubator at 37°C, and cultured in triplicate inpetri dishes. All colonies were counted after 24 and 48 h ofincubation, and the mean number of microorganisms per

milliliter of solution was calculated.Standard-view roentgenograms of the tibia were taken.

For each tibia, periosteal reaction, bone deformation, wid-ening of the diaphysis, osteolysis, and osteosclerosis weredetermined. The frequency of each of these alterations wasrecorded.For histological analysis, the samples were fixed in neutral

buffered formalin solution, decalcified in 4 N formic acid-sodium citrate (10%), embedded in paraffin, and cut into6-,um longitudinal sections. The sections were stained withhematoxylin and eosin and Masson's stain. The followingalterations, characteristic of osteomyelitis, were assessed:bone marrow inflammation, cortical alterations, subperios-teal, medullary, and intracortical abscesses, and sequestra.

RESULTS

Radiographic examination. Radiographic findings for the61 tibias from group A are shown in Table 2. At least two

radiographic alterations were observed in 90.2% and at leastthree were observed in 70.5% of the tibias from group A.Periosteal reactions were observed in 43 of the 61 tibias fromgroupA (70.5%), bone deformation was observed in 31 tibias(50.8%), widening of the shaft was observed in 52 tibias(85.2%), osteolysis was observed in 60 tibias (98.4%), andosteosclerosis was observed in 49 tibias (80.3%).One month after surgery, the radiographic images showed

osteolytic areas at the site of operation, initial periostealreactions and, less frequently, abscesses in the soft tissuescommunicating with the bone marrow cavity (Fig. 1A).Moreover, a marked alteration of the bone, with widening ofthe shaft, was observed.Two months after surgery, the osteolytic areas were more

extensive and displayed a thick sclerotic border (Fig. 1B).Three and 6 months after the operation, new bone formationwas accentuated (Fig. 1C and D). At 6 months, there was

sometimes radiographic evidence of sequestra (Fig. 1D).For six of the eight tibias from group B, roentgenograms

showed small osteolytic areas surrounded by a thin scleroticborder at the site of operation. In no case was there a

periosteal reaction, bone deformation, or widening of theshaft.

Gross bone pathology. Gross bone pathology was studiedfor the 122 tibias (61 rats) from group A and the 16 tibias (8rats) from group B. The results for group A are shown inTable 3.

In group A, 80.4% of tibias had a score of > 1 and showedsigns of osteomyelitis, whereas only 19.6% had a score of 0or 1, without signs of osteomyelitis. The mean score as-

signed to group A tibias was 2.2.A score of 1 was assigned to 4 group B tibias (25%), and a

score of 0 was assigned to the remaining 12 (75%). The meanscore for controls was 0.25.For group A, macroscopic examination during the first 3

months after surgery revealed large abscesses either con-

fined to the subperiosteal site or, in cases in which there wasa periosteal rupture, invading the soft tissues (Fig. 2A).These abscesses communicated with the bone and some-

times drained externally by way of bright red fistulas. Theskin was atrophic and adhered to the underlying layers.

Six months after surgery, gross bone pathology was char-acterized by marked metadiaphyseal alterations, with ery-thema, new bone formation, and small intracortical ab-scesses. In some cases, the bone collapsed and causedaccentuated axial deviation of the skeletal segment.

Histological examination. Histological examination was

performed on the 61 tibias from group A and the 8 tibias fromgroup B. Osteomyelitis was observed in 55 of the 61 tibiasfrom group A (90.2%). In particular, osteomyelitis was

observed in 8 of the 9 tibias from subgroup Al (88.9%), 9 ofthe 10 tibias from subgroup A2 (90.0%), 10 of the 12 tibiasfrom subgroup A3 (83.3%), and 28 of the 30 tibias from

TABLE 2. Radiographic assessment

No. of tibias with the following alteration (% of total)Subgroup (totalno. of tibias) Periosteal Bone Widening of Osteolysis Osteosclerosis

reaction deformation the shaft

Al (9) 8 (88.9) 7 (77.7) 4 (44.4) 9 (100) 5 (55.6)A2 (10) 8 (80.0) 7 (70.0) 9 (90.0) 10 (100) 6 (60.0)A3 (12) 7 (58.3) 8 (66.7) 9 (75.0) 12 (100) 9 (75.0)A4 (30) 20 (66.7) 9 (30.0) 20 (66.7) 29 (96.7) 28 (93.3)

INFECr. IMMUN.

on January 26, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

CHRONIC OSTEOMYELITIS MODEL 5227

FIG. 1. (A) Roentgenogram of a tibia from subgroup Al. Note the large abscess communicating with the bone and the extensive osteolyticarea surrounded by a sclerotic border. (B) Roentgenogram of a tibia from subgroup A2. Note the periosteal reaction and the abscesscommunicating with the bone. (C) Roentgenogram of a tibia from subgroup A3. Note the alteration of the anteromedial cortex, the extensiveperiosteal reaction, and the osteosclerosis. (D) Roentgenogram of a tibia from subgroup A4. Note the deformation of the medial profile of thetibia, the periosteal reaction, and the widening of the shaft. A large sequestrum can be seen within the osteolytic area surrounded by asclerotic border.

subgroup A4 (93.3%). Samples from group B did not showhistological signs of osteomyelitis.

Extraperiosteal, subperiosteal, intracortical, and in-tramedullary abscesses were observed in group A. Threeareas could be clearly distinguished in the abscesses: cen-tral, medial, and peripheral. The central area containedpolymorphonuclear cells, fibrinoid necrosis, and an occa-sional foam cell (Fig. 2B and C and 3A). The medial areashowed mononuclear cells, identified as lymphocytes,monocytes, and macrophages, and polymorphonuclear cells,including eosinophils. The peripheral area was composed offibrous tissue and enclosed the abscesses. No granulomaswere observed.

Sequestra, i.e., irregular fragments of infected necroticbone surrounded by polymorphonuclear cells, were ob-served within the abscesses (Fig. 3B). A progressive in-crease in the frequency of sequestra with increase in the timeinterval between the induction of the lesion and histologicalexamination was observed. In particular, no sequestra wereobserved in subgroup Al, but sequestra were observed in 2of the 10 tibias from subgroup A2, 4 of the 12 tibias fromsubgroup A3, and 12 of the 30 tibias from subgroup A4.

TABLE 3. Gross bone pathology

Group or subgroup No. of tibias with the following scorea (% of total)(total no. of tibias) 0 1 2 3 4

A (122) 8 (6.5) 16 (13.1) 41 (33.7) 57 (46.7) 0Al (18) 4 (22.2) 0 3 (16.7) 11 (61.2) 0A2 (20) 0 2 (10.0) 4 (20.0) 14 (70.0) 0A3 (24) 0 7 (29.2) 10 (41.7) 7 (29.2) 0A4 (60) 4 (6.7) 7 (11.7) 24 (40.0) 25 (41.6) 0

a Grading system of Rissing et al. (25).

Osteoclasts were sometimes seen in areas of resorption onthe surface of the sequestra.Hematopoietic marrow was replaced by polymorphonu-

clear cells, fibroblasts, and macrophages. The experimentallesions were confined to the tibial metaphysis and only rarelyextended distally, without ever completely involving thediaphysis. The epiphysis was always intact because of theprotective effect of the growth cartilage.

In the early stages of infection, the metadiaphyseal cortexwas partially destroyed, creating a passage between the bonemarrow and the abscesses. Six months after the operation,new bone formation was observed and the cortex wasrepaired in a deformed fashion.

Microbiological analysis. Bacterial counts were performedfor the 61 tibias from group A and the 8 tibias from group B.The results for group A are shown in Table 4.One tibia from group B had a positive culture with a

bacterial count of 5 x 105 CFU/ml. The strain isolated wasidentified by use of the Sceptor system as S. epidermidis. Inan MIC assay with the same drugs as those tested for the S.aureus strain used in our model, the isolated strain showed adifferent sensitivity pattern. Thus, the positive culture for atibia from group B was probably due to contamination duringsurgery.

DISCUSSION

This study was aimed at developing a rat model of chronicstaphylococcal osteomyelitis with pathophysiological, clini-cal, radiographic, and histological characteristics similar tothose of human disease. We chose the rat because purchaseand housing costs are low and this animal tolerates surgicaltrauma and long-term high-dose antibiotic therapy. More-over, the rat tibia is sufficiently small to allow pulverizationfor bacterial counts, and inbred strains can be used for

VOL. 61, 1993

on January 26, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

5228 SPAGNOLO ET AL.

'81V

'%~~~~~~~~~~~~~~~~~~zFIG. 2. (A) tibia from subgroup A2. Note the voluminous abscess in the soft tissues which communicated with the bone by way of a fistula.

(B) Longitudinal section of a tibia from subgroup A2 showing the central area of the abscess. Note polymorphonuclear cells and fibrinoidnecrosis. Magnification, x336. (C) Longitudinal section of a tibia from subgroup A2 showing the detail of the central area of the abscess,containing polymorphonuclear cells. Magnification, x840.

experimentally studying the immunological aspects of thepathogenesis of osteomyelitis. The main disadvantage of therat model is the size of the animal, which considerably limitsthe possibility of experimenting with complex surgical tech-niques.A few rat (24, 25, 31), rabbit (1, 17-19, 22, 26, 27, 30), and

canine (11, 13) models of osteomyelitis used sclerosingagents or foreign bodies. Sclerosing agents (e.g., sodiummorrhuate or barium sulfate) probably cause alterations ofthe marrow circulation, according to some authors (10).Intramedullary foreign bodies (e.g., silicone catheters, in-tramedullary nails, and polyvinyl alcohol sponges), alone orin combination with sclerosing agents, could constitute aninert substratum to which S. aureus adheres tenaciously andproduces the extracellular hexopolysaccharide matrix calledthe glycocalyx (18, 19).

In this model, we used fibrin glue instead of sclerosingagents or foreign bodies. Fibrin glue is a biocompatiblematerial which does not cause aseptic bone necrosis anddoes not constitute an inert substratum, because of itsresorption in a few days. S. aureus could bind to thismolecule through a receptor-like structure, so fibrin gluecould constitute an optimal medium for bacterial growth inloco, thus trapping bacteria and preventing sepsis. Theadhesive and hemostatic properties of fibrin glue justify itswide use in surgery (29). Furthermore, some authors haveattributed to fibrin glue the property of stimulating andaccelerating repair osteogenesis via a mechanism of theosteoconductive type (2, 3, 4, 5, 6, 16).

Radiographic, macroscopic, microscopic, and microbio-logical analyses were carried out on the tibias removed afterthe animals had been killed. Since it has been demonstrated(25) that there is no relationship between clinical and blood

chemistry parameters and the progression of experimentalosteomyelitic lesions, no clinical or laboratory assessmentswere done for the live animals.An analysis of the results permits us to suggest that this

model of chronic osteomyelitis could correspond to theclinical and biological aspects of human disease.The kind and the development of radiographic alterations

of the tibias from group A were similar to those observed inhuman disease. Alterations of the bone structure were morefrequent in the early stages of disease, and osteolytic areasof increasing dimensions were observed, depending on theprogression of the lesions. The osteolytic lesions weresurrounded by sclerotic borders, which were thicker insubgroup A4. Sequestra were detected less frequently radio-graphically than by histological examination because of thelimited resolution power of roentgenograms. In fact, inhuman osteomyelitis, radiographic images very often do notreveal sequestra, which are clearly identified by computer-ized tomography.Gross bone pathology examination in the early stages of

disease revealed extensive destructive lesions involving thesoft tissues, with draining skin fistulas. In later stages, newbone formation replaced the destructive process, analogousto that observed in human osteomyelitis.

Histological examination revealed osteomyelitis in morethan 90% of cases, demonstrating the reliability of themodel. Whereas in the first months following induction ofthe lesions large abscesses associated with osseous necrosiswere prevalent, 6 months after surgery, osteogenetic pro-cesses predominated. New bone formation often took placealong abnormal load-bearing lines and resulted in deforma-tion of the bone. Osteomyelitis became chronic because thesequestra could be neither resorbed, in that they were

INFECT. IMMUN.

on January 26, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

CHRONIC OSTEOMYELITIS MODEL 5229

FIG. 3. (A) Longitudinal section of a tibia from subgroup A3 showing the detail of an abscess. Note the exuberant fibrinoid necrosiscontaining phlogistic cells. Magnification, x 1,000. (B) Longitudinal section of a tibia from subgroup A4. Note the voluminous sequestrumsurrounded by polymorphonuclear cells. The infected necrotic bone has large lacunas filled by purulent exudate. Magnification, x200.

isolated from the vital connective tissue, nor expelled.Sequestra were detected in only a few cases, most likelybecause many of them had been missed in the preparation ofhistological sections.The lesions were confined to the tibial metaphysis and

only rarely extended distally, without ever completely in-volving the diaphysis.

Osteoclasts were always observed in fragments of infectednecrotic bone, contradicting the thesis of Collins, accordingto whom osteoclasts are a cell population rarely found in thecourse of osteomyelitis (9). It has been hypothesized on thebasis of in vitro studies that in infective bone disease (andprobably also in physiological conditions) there could bemechanisms of bone resorption mediated by polymorphonu-clear cells and monocytes (14, 15, 20, 21). However, it is

TABLE 4. Microbiological analysis of group A tibias

No. of tibias with thefollowing culture result

Group or subgroup CFU (106/ml) for S. aureus(total no. of tibias) (mean + SD) (% of total)

Positive Negative

A (61) Not determined 52 (85.2) 9 (14.8)Al (9) 2.95 ± 3.16 8 (88.9) 1 (11.1)A2 (10) 1.78 ± 2.90 10 (100) 0A3 (12) 2.63 + 3.15 10 (83.3) 2 (16.7)A4 (30) 0.77 ± 1.34 24 (80.0) 6 (20.0)

known that bone marrow monocytes can differentiate intoosteoclasts. At present there is no clear evidence that cellsother than osteoclasts can resorb either normal or patholog-ical bone.

Microbiological analysis, showing a high percentage ofpositive cultures (86%), also confirms the reliability of themodel.The chronic osteomyelitis described in this study is local-

ized and can be classified as type 3A, according to theclassification of Cierny et al. (7, 8). Given the methods ofinduction of the disease, our model can be consideredposttraumatic chronic osteomyelitis, according to the clas-sification of Waldvogel and coworkers (28). Because of thehigh level of positivity of induced osteomyelitic lesions andthe probable analogy with human disease, we believe thatthis model could be reliably used in pathogenetic and ther-apeutic studies.

REFERENCES1. Andriole, V. T., D. A. Nagel, and W. 0. Southwick 1973. A

paradigm for human chronic osteomyelitis. J. Bone Jt. Surg.Am. Vol. 55:1511-1515.

2. Arbes, H., P. B4sch, F. LiUntner, and M. Salzer. 1981. Firstclinical experience with heterologous cancellous bone graftingcombined with the fibrin adhesive system (F.A.S.). Arch.Orthop. Trauma Surg. 98:183-188.

3. Bfhler, N., P. Bfisch, G. Sandbach, G. Schiag, J. Eschberger,and L. Schmid. 1977. Der Einfluss von homologem Fibrinogenauf die Osteotomieheilung beim Kaninchen. Unfallheilkunde80:501-508.

VOL. 61, 1993

on January 26, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

5230 SPAGNOLO ET AL.

4. Bosch, P. 1981. Die Fibrinspongiosaplastik. ExperimentelleUntersuchungen und klinische Erfahrung. Wien. Klin. Wochen-schr. Suppl. 93:1-26.

5. Bosch, P., F. Braun, and H. P. Spingler. 1977. Die Technik derFibrinspongiosaplastik. Arch. Orthop. Unfall. Chir. 90:63-75.

6. Bosch, P., F. Lintner, and F. Braun. 1979. Die autologe Spon-giosatransplantation unter Anwendung des Fibrinklebesystemsim Tierexperiment. Wien. Klin. Wochenschr. 91:628-633.

7. Cierny, G., and J. T. Mader. 1984. Adult chronic osteomyelitis.Orthopedics 7:1557-1564.

8. Cierny, G., J. T. Mader, and J. J. Penninck 1985. A clinicalstaging system of adult osteomyelitis. Contemp. Orthop. 10:17-37.

9. Collins, D. H. 1966. Infective disease of bone, p. 209-227. In0. G. Dodge (ed.), Pathology of bone. Butterworths, London.

10. Crane, L. R., C. C. Kapdi, J. N. Wolfe, B. K Silberberg, andA. M. Lerner. 1977. Xeroradiographic, bacteriologic and patho-logic studies in experimental staphylococcus osteomyelitis.Proc. Soc. Exp. Biol. Med. 156:303-314.

11. Deysine, M., E. Rosario, and H. D. Isenberg. 1976. Acutehematogenous osteomyelitis: an experimental model. Surgery79:97-99.

12. Emslie, K. R., and S. Nade. 1983. Acute hematogenous staphy-lococcal osteomyelitis. A description of the natural history in anavian model. Am. J. Pathol. 110:333-345.

13. Fitzgerald, R. H., Jr. 1983. Experimental osteomyelitis: descrip-tion of a canine model and the role of depot administration ofantibiotics in the prevention and treatment of sepsis. J. Bone Jt.Surg. Am. Vol. 65:371-380.

14. Gillespie, W. J., and A. R. Allardyce. 1990. Mechanisms of bonedegradation in infection: a review of current hypotheses. Ortho-pedics 13:407-410.

15. Gray, T. K., C. D'Amico, R. Kaplan, R. C. Dodd, D. Mesler, andM. S. Cohen. 1986. Mononuclear phagocytes secrete a proteinthat directly resorbs devitalized bone particles. Bone Miner.1:235-245.

16. Greco, F., L. de Palma, N. Specchia, and P. Lisai. 1988.Experimental investigation into reparative osteogenesis withfibrin adhesive. Arch. Orthop. Trauma Surg. 107:97-103.

17. Johansson, A., 0. Svensson, G. Blomgren, G. Eliasson, and C. E.Nord. 1991. Anaerobic osteomyelitis: a new experimental rabbitmodel. Clin. Orthop. 265:297-301.

18. Lambe, D. W., Jr., K. P. Ferguson, K. J. Mayberry-Carson, B.Tober-Meyer, and J. W. Costerton. 1991. Foreign-body-associ-ated experimental osteomyelitis induced with Bacteroides fra-gilis and Staphylococcus epidermidis in rabbits. Clin. Orthop.266:285-294.

19. Mayberry-Carson, K. J., B. Tober-Meyer, J. K. Smith, D. W.Lambe, Jr., and J. W. Costerton. 1984. Bacterial adherence andglycocalyx formation in osteomyelitis experimentally inducedwith Staphylococcus aureus. Infect. Immun. 43:825-833.

20. Minkin, C., R. Posek, and J. Newbrey. 1981. Mononuclearphagocytes and bone resorption. Metab. Bone Dis. Relat. Res.2:363-369.

21. Mundy, G. R., A. J. Altman, M. D. Gondek, and J. G. Bandelin.1977. Direct resorption of bone by human monocytes. Science196:1109-1111.

22. Norden, C. W., and E. Kennedy. 1970. Experimental osteomy-elitis. I. A description of the model. J. Infect. Dis. 122:410-418.

23. Passl, R., C. H. Mtiller, C. C. Zielinskd, and M. M. Eibl. 1984. Amodel of experimental post-traumatic osteomyelitis in guineapigs. J. Trauma 24:323-326.

24. Power, M. E., M. E. Olson, P. A. G. Domingue, and J. W.Costerton. 1990. A rat model of Staphylococcus aureus chronicosteomyelitis that provides a suitable system for studying thehuman infection. J. Med. Microbiol. 33:189-198.

25. Rissing, J. P., T. B. Buxton, R. S. Weinstein, and R. K. Shockley.1985. Model of experimental chronic osteomyelitis in rats.Infect. Immun. 47:581-586.

26. Scheman, L., M. Lanota, and P. Lewin. 1941. The production ofexperimental osteomyelitis: preliminary report. JAMA 117:1525-1529.

27. Van Wingerden, G. I., V. Lolans, and G. G. Jackson. 1974.Experimental pseudomonas osteomyelitis: treatment with siso-micin and carbenicillin: J. Bone Jt. Surg. Am. Vol. 56:1452-1458.

28. Waldvogel, F. A., G. Medoff, and M. N. Swartz. 1970. Osteo-myelitis: a review of clinical features, therapeutic consider-ations and unusual aspects. N. Engl. J. Med. 282:198-206,260-266, 316-322.

29. Weber, S. C., and M. W. Chapman. 1984. Adhesives in ortho-paedic surgery. A review of the literature and in vitro bondingstrengths of bone-bonding agents. Clin. Orthop. 191:249-261.

30. Whalen, J. L., R. H. Fitzgerald, Jr., and R. T. Morrissy. 1988.A histological study of acute hematogenous osteomyelitis fol-lowing physeal injuries in rabbits. J. Bone Jt. Surg. Am. Vol.70:1383-1392.

31. Zak, O., F. Zak, R. Rich, W. Tosch, F. Kradolfer, and W.Scheld. 1982. Experimental staphylococcal osteomyelitis in rats:therapy with rifampin and cloxacillin alone or in combination, p.973-974. In P. Perti and G. Grassi (ed.), Current chemotherapyand immunotherapy. American Society for Microbiology,Washington, D.C.

INFEcr. IMMUN.

on January 26, 2020 by guesthttp://iai.asm

.org/D

ownloaded from