- Email: [email protected]
Infection adjacent to titanium and bone cement implants: an experimental study in rabbits
Eiomoferiols 16 (1995) 1273-1277 0 1995 Elsevier Science Limited
Printed in Great Britain. All rights reserved 0142-9612/95/$10.00
Infection adjacent to titanium and bone cement implants: an experimental study in rabbits Lennart Sanzh and Lars Linder Department
of Orthopedics,
University Hospital MAS, S-214 01 Malmd, Sweden
A pure titanium cylinder or a piece of bone cement was implanted in each upper tibia1 metaphysis of 20 rabbits. After 4 months radiograms were taken and 104, 106, or 10’ Staphylococcus aufeus were injected into each leg through central holes in the implants in three groups of animals. Four weeks later new radiograms, bacteriological and histological biopsies were obtained. Three animals died before the end of the experiment. In animals which received lo6 or 10’s. aureus radiographic signs of infection were found in 1l/22 legs with both titanium and bone cement implants. No radiolutient zones ‘developed around the implants. Bacteriological cultures from bone close to the implants were negative in all legs with titanium implants and positive in four legs with cement implants. Seven {cultures were negative in spite of radiographic changes. It is concluded that after a proper time for ,wound healing the bone around unloaded implants of both titanium and bone cement is fairly resistant ‘to infection. In some cases, healing of an infection in the surrounding bone seems possible. ,Yeywords:
Infection,
implant,
bone cement, titanium
IReceived 21 December 1994; accepted 16 March 1995
weeks thereafter3-5. The influence of timing of the bacterial challenge was prominent in the study in rats by Elson et al.“, where 70% of the implants of plain bone cement became infected if challenged on the day of surgery versus 18% if the injection was delayed for
Haematogenous infection is a disheartening complication of an otherwise successful joint replacement. In septic arthritis, controversy exists whether stable implants can be left in place after mechanical cleansing of the joint. It is also not known whether the mode of @plant anchorage (cement versus no cement) has an influence on the further development of the infection. There are several e.xperimental models of implantrelated osteomyelitis, most of which imply bacterial inoculation at the time of implant insertion. All experiments demonstrate an increased susceptibility to infection in the presence of a foreign body. Thus, while inoculation of 2 x 10’ Staphylococcus aureus throtigh drill holes did not cau,se osteomyelitis in rabbit tibiae’, the introduction of an implant reduced the necessary number of bacteria to 106. In dogs, the number of S. aureus necessary to produce infection in half of the legs (Q,J were of the order of 10” around metallic implants and lo* for implants of bone cement cured in vivo as compared to 10s for controls without implants’. Intravenous injection of bacteria to produce a remote implant infection has ‘been used by two groups. Using doses of 10*-lO’” S. aureus or Propionibacterium acnes, Blomgren has regularly been able to produce remote implant infections in rabbits, both when the injections were given on the day of surgery and 4-6 Correspondence
6 weeks.
Both the nature of the implant and the time of bacterial challenge thus seem to influence the development of infection, but so far no study has evaluated the sequelae of late inoculation (local or haematogenous) around different implant materials, such as metals and bone cement. Such a comparison may be especially pertinent in the light of the demonstration that pure titanium in an in vitro system appeared to have bactericidal properties7. This may partly explain the clinical observation that osseointegrated dental implants are hardly ever lost because of infection, despite being placed in communication with the oral cavity*. The aim of the present investigation was to compare the infection susceptibility of the bone surrounding implants of bone cement and pure titanium which had been allowed to heal into the bone before bacterial inoculation. Preliminary attempts to create an infection by intravenous injection of bacteria resulted in profound morbidity among the animals. In order to reduce the number of animals lost because of death or systematic illness, we developed a model in which the bacteria were injected locally with minimal tissue trauma.
to Dr L. !
1273
Biomaterials 1995, Vol. 16 No. 16
1274
Infection
MATERIALS
adjacent
to titanium
and bone cement
implants:
L. Sanz&~
and L. Linder
AND METHODS
Surgery Twenty adult 1.7-3.6 kg New Zealand white rabbits of both sexes were operated on under general anaesthesia. Antibiotic prophlaxis with 125 mg of cefuroxime was given prior to surgery. Under aseptic conditions, a cylindrical implant of pure titanium and a piece of bone cement dough of similar size (Palaces, E. Merck, Darmstadt, Germany) were inserted into a 3.6 mm drill hole in the upper tibia1 metaphysis of either leg. The surgical technique has previously been described in detailg. The titanium implants had a central 1 mm hole and in the bone cement implants a 1.2 mm central hole was created with a needle during polymerization. After 12 weeks, radiograms of the legs were taken and the rabbits were divided into three groups: I-III. A in suspension of 5 x 103, 5 x lo5 or 5 x lo7 bacteria 0.1 ml saline from a clinical isolate of S. epidermidis derived from a total joint infection were injected transcutaneously through the hole in the implant into the marrow cavity. The bacterial suspensions were prepared by inoculating the bacteria in tryptic soya broth overnight at 37°C. Bacteria were washed, centrifuged and immersed in isotonic saline. The number of organisms per ml was checked by a serial dilution plate counting technique. The animals were regularly checked for signs of infection and three weeks later new radiograms were made. Since radiograms were unchanged and no clinical signs of infection were found, the animals were considered uninfected. In a new series 9 x 103, 9 x lo5 or 9 x lo7 S. aureus (phage type 85195 from an infected total hip joint) in 0.1 ml saline were injected in the same way through the hole in the implant. Thus, rabbits in group I were inoculated with the lower number of bacteria and those in group III with the higher dose. After another four weeks new radiograms were taken, the animals were killed, and the implants with surrounding bone were removed aseptically using a small saw. The bone surrounding one half of the implant was broken off for culture and the bone surrounding the other half of the implant was fixed and embedded for histology. From the bone marrow, specimens for bacteriological culture were collected. Analysis (1) The severity of radiographic changes such as osteolysis, bone sclerosis and new bone formation were graded from 0 to + + + (Figure Za-c). (2)The biopsies intended for histology were immediately fixed in 10% buffered formaldehyde at 4°C for several days, followed by dehydration in a graded series of ethanol, clearing in xylene, and embedding in methyl methacrylate. Sections were cut at 5 pm on a Jung hard tissue microtome and stained with haematoxylin-eosin and Goldner’s stain. Before fixation the titanium implants were carefully removed from the bone, but the biopsies with the cement plugs were immersed with the implant in place. The cement was later dissolved by the xylene., Biomaterials
1995, Vol. 16
No.16
Figure 1 Examples of radiographic changes 4 weeks after injection of Staphykxoccusaureus into the marrow cavity of rabbit tibia. a, Osteolysis and new bone formation graded + in the posterior cortex of a tibia with a titanium implant. b, Sclerosis and new bone formation graded ++ in both tibiae. c, Pronounced sclerosis and new bone formation graded + + + in a tibia with a bone cement implant.
Infection adjacent to titanium Table 1 Radiographic in rabbit tibia Group
Number of animals
and bone cement
changes, bacteriological
implants:
L. Sanzen
1275
Linder
and L.
cultures and weight change (%) four weeks after injection of Staphylococcus
Weight change
Type of implant
Radiograms
aureus
Positive culture
0
+
++
+++
Bone
Marrow
I
6
+1
Titanium Bone cement
6 6
0 0
0 0
0 0
0 0
0 0
II
6
-3
Titanium Bone cement
3 3
0 2
2 1
1 0
0 1
1 1’
Ill
5
-9
Titanium Bone cement
2 3
1 1
2 0
0 1
0 3
1 1*
*Leg with positive cultures from both bone and bone marrow.
(3)
Tissue biopsies were sonicated and incubated for 48-72 h in tryptic soy broth and subcultured on blood agar plates. Colonies were identified by morphology, phage typing and disc antibiograms.
implants were culture positive in the marrow. No culture showed growth of the injected strain of S. epidermidis. The coincidence of radiographic changes and a positive bacteriology in group II and III is shown in Figure 2. In the low-dose group (I) all cultures were negative.
RESULTS Table 2
At 12 weeks, before bacterial injection, radiograms showed no abnormalities except in one case in group II, where a moderate sclerosis was found in the tibia with the titanium implant. Three weeks following injection of S. ep.idermidis all radiograms were unchanged and the animals showed no local signs of infection such as limp, swelling, tenderness or increased heat. Also, the animals were generally at ease, and had gained weight. During the following four weeks, after the S. aureus injection, the animals became ill and lost weight in the two groups (II and III) with the largest numbers of inoculated bacteria (Table 1). Three animals died during the investigation period, one in group :I1and two in group III. Individual data for all 20 rabbits are found in Table 2.
Radiography No changes occurred following injection of S. epidermidis. Four weeks after injection of S. aureus some radiograms showed different grades of osteolysis, sclerosis and new bone formation in the upper half of the tibia (Figure 1a-c). No radiolucent zones close to the implants or bone sequestrae were seen. Changes were found in 6111 tibiae with titanium implants and 5111 with bone cement implants in the two groups (II and III) with the largest inoculates, and none in the group (I) with the smallest inoculates (Table 2).
Bacteriology In six legs of four animals (one in group II and three in group III) S. aureus was isolated, and all strains were of the same phage type as that injected (Table 3). Cultures were positivle from bone close to the implant in 4111 legs with cement implants and in Oh1 with titanium implants (,P = 0.09, Fischer’s exact test). Bacteria were also found in the marrow of two of these legs with cement implants. Two legs with titanium
A
Individual data for 20 rabbits’ B
1
0
2
0
3
5
4
-2
5
2
6
0
7
0
6
-8
9
-11
10
5
11 12
-9
13 14
5 -14
15 16
-6
17 18
-15
19
-3
20
-7
C
D
E
F
G
H
1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2
1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3
0 0 0
0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0
1 1 1 1 1 2 3 3 1 2 1 ? ? 2 ? 1 1 2 1 3
0
0 0 0 0 0 0 0
0 0 1 2 0 2 2 0 1
0
0 0 0 0 0 0 0
0 0 0 1 0 0
I
1 3 0 0 0 2 0
0 0 0 0 0 1
0 0 0 0 1 0
1 2 1 ? 1 3
2
1 0 0
0 1
0 0
1 1 1
0 3 2 1 1 0
0 1 0 0 0 0
0 1 0 0 0 0
2 3 ? 2 ? 1
*A is rabbit number. B is weight change (%). C is implant: 1, titanium; 2. bone cement. D is bacterial dose: 1, 104; 2, lo?; 3, 10’. E is radiographic changes: 1, +; 2, ++; 3, + + +. The bacteriological cultures were as follows: F, bone. G, marrow: 0, negative; 1, S. aureus. tl. bone-implant contact: 1. extensive; 2. patchy; 3. none; ?, not classified. I, complication: 1. died; 2, bone sclerosis before bacteriological challenge.
Biomaterials 1995, Vol. 16 No. 16
Infection adjacent to titanium
1276 Positive
roentgenography
Positive
bacteriology
Figure 2 Distribution of positive radiographic and bacteriological findings in 11 rabbit tibia with titanium and bone cement implants inoculated with 9 x 105-9 x lo7 Staphylococcus aureus (groups II and III).
and bone cement implants: L. Sanz&
Thus, there was no correlation tion and positive cultures.
and L. Linder
between osseointegra-
Interface membrane Where a membrane was seen around the titanium implants, it was never inflamed. In the case of bone cement, 7110 interface membranes showed a patchy infiltration of lymphocytic cells or plasma cells (Figure 3). Multinucleated giant cells were always seen around the endosteal aspect of the cement implants, regardless of the presence or absence of a continuous membrane, but never around the titanium implants. Marrow There was no correlation between histology and radiographic changes, except for one case classified radiographically as + + +. In this case with a bone cement implant there was proliferation of bone in the marrow cavity and all marrow species were occupied by loose fibrous tissue with an increased number of plasma cells (Figure 4). Less pronounced fibrosis and/ or bony proliferation was seen in two titanium cases and five cement- cases. There was no correlation between these findings and the bacteriological or the radiographic results.
DISCUSSION
Figure 3 Detail of the membrane adjacent to the endosteal aspect of a cement implant in group II. Bacterial cultures were negative and radiograms showed no reaction. Multinucleated giant cells are seen bordering the cement surface. There is a focal accumulation of inflammatory ceils within the membrane. This amount of plasma cells represents the largest accumulation of inflammatory cells in any of the sections in the entire study (Goldner stain, x200).
This study was based on the clinical situation usually present in hematogenous infections, i.e. the mature interface of a stable implant. In a set of preliminary experiments we tried to infect the implants by intravenous injections of bacteria. In the case of S. epidermidis the animals and implants were unaffected, but in the case of S. aureus more than half of the animals died due to multiple septic emboli after a period of severe illness. For ethical reasons intravenous injections were abandoned. In order to -reduce animal morbidity and mortality while still atraumatically bringing bacteria in close
Histology Six legs (four titanium and two cement implants) did not include a sufficient continuous length of the interface to allow classification of the interface structure. All of these legs were culture negative. Of the remaining 13 titanium implants, 11 were osseointegrated (i.e. the extent of the bone-implant contact was considered normal for the experimental model and implant materialg* lo), one was incompletely invested in bone, and one [group I] invested in a reaction-free fibrous membrane. The corresponding figures for the cement implants were five, six, and. four (one in groups I and II, two in group III). Osseointegration-bacterial growth In the two titanium cases, which were culture positive in the bone marrow, osseointegration was present. In the four cement cases with positive cultures from bone, one was osseointegrated, one had a limited area of direct bone-cement contact, whereas two had a fibrous tissue interface with a patchy infiltration of plasma cells but no lymphocytes or granulocytes. Biomaterials
1995, Vol. 16 No. 16
Figure 4 The bone adjacent to a cement implant in group III, where bacterial cultures were positive and where radiograms were classified as + + +. There is an intimate contact between the cement and the bone, in some places without intervening cells. There is a marked proliferation of bone and all marrow spaces are fibrotic (Goldner stain, x75).
Infection adiacent to titanium
and bone cement implants:
L. Sanzk
contact with the implants, the transcutaneous technique was developed. The aim was to evaluate whether one of the implani. materials was more prone to infection than the other. Our doses were chosen from the literature on inoculation in fresh wounds; the lowest dose was expected to infect only cement implants, while the highest was expected to infect all implants. While we do not know whether an injection of bacteria reproduces all the local events during septicemia, there is no doubt that the bacteria are deposited very close to the implant surface, as the injection needle passes through the implant. The correlation between bacterial dose and weight loss after injection of S. aureus indicates that the rabbits did in fact become infected. Although our strains of S. epidermidis and S. aureus were from human implant-related infections, they obviously had different virulence. This is in accordance with earlier studieszs6. In the study in dogs by Petty et al.‘, the IDS,, values for these bacteria differed up to 500 times for infections adjacent to metal implants. However, for cement implants cured . . JR VJVO IDso, 2 x 104, was the same for both S. aureus Consequently, at least some and S. epidermidis. infections ought to have occurred in our groups II and III. Our inability to produce infections with S. epidermidis can eitlher be explained by too low a virulence of the bacterial strain or by a protective effect of a mature bone-implant interface as compared to a fresh operative wound. In the case of S. aureus, infection did occur, but again less easily than expected. In fresh wounds with bone cement implants cured in viva, lo4 S. aureus infected all rats6 and ID,~ for dogs’ was 2 x 104. If metallic implants are used 10” bacteria give a 50% incidence in dogs” and 2 x lo6 bacteria a 100% incidence of persistent infections in rabbits*. In our preliminary experiments with intravenous injection in rabbits severe multiorgan infections developed, proving the virulence of the bacterial strain. In the main experiment no infections were produced with lo4 bacteria, either for cement or titanium. We did not observe more than 50% incidence of radiographic changes when doses of 106-lOa S. aureus were used, and in these legs there were only 4122 with persistent infection, indicating a healing process. This is in contrast to osteomyelitis caused by direct inoculation in fresh wounds, where bacteria could be cultured after several months in 44/ 57 rabbits’ and in 20120 dogsI’. Cultures from bone in two legs close to cement implants were positive without any radiographic changes observed. This was also found by Blomgren et a1.5 in infections caused by Propionibacterium acmes. Comparing the infection rates around the cement and titanium implants, radiographic changes were present in equal numbers, but positive cultures were present in only two legs with titanium implants against four with bone cement implants. Moreover, it is noteworthy that we were unable to culture any bacteria from the bone biopsies close to the titanium implants, whereas
and L. Linder
1277
four bone biopsies close to cement grew S. aureus. We conclude that the infection susceptibility is reduced by a mature interface compared to a fresh wound, but, even so, there still seems to be a difference between a metallic implant and bone cement. Whether our inability to cause a persistent osteomyelitis close to the titanium implants depends on its suggested bactericidal properties7 cannot be answered by this study but certainly merits further investigation.
ACKNOWLEDGEMENTS This study was financially supported by the Riksfijreningen mot reumatism. The authors wish to thank Dr Mats Walder and BjGrn Nilsson at the Department of Clinical Morphology for valuable assistance and Dr Per Gunnar Persson, Department of Pathology, Kristianstad Hospital for help with the evaluation of the histological sections.
REFERENCES 1
2
3
4
5
6
7
8
9 10 11
Andriole VT, Nagel DA, Southwick WO. A paradigm for human chronic osteomyelitis. J Bone Joint Surg (Am) 1973; 55: 1511-1515. Petty W, Spaniewer S, Shuster JJ, Silverthorne C. The influence of skeletal implants on incidence of infection. Experiments in a canine model. J Bone Joint Surg (Am) 1985; 67: 1236-1244. Blomgren G, Lindgren U. The susceptibility of total joint replacement to hematogenous infection in the early postoperative period: an experimental study in the rabbit. Clin Orthop 1980; 151: 308-312. Blomgren G, Lindgren U. Late hematogenous infection in total joint replacement: studies of gentamicin and bone cement in the rabbit. Clin Orthop 1981; 155: 244248. Blomgren G, Lundquist H, Nord C-E, Lindgren U. Late anaerobic haematogenous infection of experimental total joint replacement. A study in the rabbit using Propionibacterium acnes. J Bone Joint Surg (Br) 1981; 63: 614-618. Elson RA, Jephcott AE, McGechie DB, Verettas Q Bacterial infection and acrylic cement in the rat. J Bone Joint Surg (Br) 1977; 59: 452457. Tengvall P, Hijrnsten EG, Elwing H, LundstrGm I. Bactericidal properties of a titanium-peroxy gel obtained from metallic titanium and hydrogen peroxide. J Biomed Mater Res 1990; 24: 319-330. Brinemark P-I, Hansson BO, Adell R, Breine U, Lindstram J,’ Hall& 0, ijhman A. Osseointegrated implants in the treatment of the edentulous jaw. Stand JPIastic Reconstruct Surg 1977; supplement 16. Linder L. Osseointegration of metallic implants I. Acta Orthop Stand 1989; 60: 129-134. Linder L. Reaction of bone to the acute chemical trauma of bone cement. J Bone Joint Surg [Am) 1977; 59: 82-87. Fitzgerald RH. Experimental osteomyelitis: description of a canine model and the role of depot administration of antibiotics in the prevention and treatment of sepsis. J Bone Joint Surg (Am) 1983; 65: 371-380.
Biomaterials 1995. Vol. 16 No. 16
Printed in Great Britain. All rights reserved 0142-9612/95/$10.00
Infection adjacent to titanium and bone cement implants: an experimental study in rabbits Lennart Sanzh and Lars Linder Department
of Orthopedics,
University Hospital MAS, S-214 01 Malmd, Sweden
A pure titanium cylinder or a piece of bone cement was implanted in each upper tibia1 metaphysis of 20 rabbits. After 4 months radiograms were taken and 104, 106, or 10’ Staphylococcus aufeus were injected into each leg through central holes in the implants in three groups of animals. Four weeks later new radiograms, bacteriological and histological biopsies were obtained. Three animals died before the end of the experiment. In animals which received lo6 or 10’s. aureus radiographic signs of infection were found in 1l/22 legs with both titanium and bone cement implants. No radiolutient zones ‘developed around the implants. Bacteriological cultures from bone close to the implants were negative in all legs with titanium implants and positive in four legs with cement implants. Seven {cultures were negative in spite of radiographic changes. It is concluded that after a proper time for ,wound healing the bone around unloaded implants of both titanium and bone cement is fairly resistant ‘to infection. In some cases, healing of an infection in the surrounding bone seems possible. ,Yeywords:
Infection,
implant,
bone cement, titanium
IReceived 21 December 1994; accepted 16 March 1995
weeks thereafter3-5. The influence of timing of the bacterial challenge was prominent in the study in rats by Elson et al.“, where 70% of the implants of plain bone cement became infected if challenged on the day of surgery versus 18% if the injection was delayed for
Haematogenous infection is a disheartening complication of an otherwise successful joint replacement. In septic arthritis, controversy exists whether stable implants can be left in place after mechanical cleansing of the joint. It is also not known whether the mode of @plant anchorage (cement versus no cement) has an influence on the further development of the infection. There are several e.xperimental models of implantrelated osteomyelitis, most of which imply bacterial inoculation at the time of implant insertion. All experiments demonstrate an increased susceptibility to infection in the presence of a foreign body. Thus, while inoculation of 2 x 10’ Staphylococcus aureus throtigh drill holes did not cau,se osteomyelitis in rabbit tibiae’, the introduction of an implant reduced the necessary number of bacteria to 106. In dogs, the number of S. aureus necessary to produce infection in half of the legs (Q,J were of the order of 10” around metallic implants and lo* for implants of bone cement cured in vivo as compared to 10s for controls without implants’. Intravenous injection of bacteria to produce a remote implant infection has ‘been used by two groups. Using doses of 10*-lO’” S. aureus or Propionibacterium acnes, Blomgren has regularly been able to produce remote implant infections in rabbits, both when the injections were given on the day of surgery and 4-6 Correspondence
6 weeks.
Both the nature of the implant and the time of bacterial challenge thus seem to influence the development of infection, but so far no study has evaluated the sequelae of late inoculation (local or haematogenous) around different implant materials, such as metals and bone cement. Such a comparison may be especially pertinent in the light of the demonstration that pure titanium in an in vitro system appeared to have bactericidal properties7. This may partly explain the clinical observation that osseointegrated dental implants are hardly ever lost because of infection, despite being placed in communication with the oral cavity*. The aim of the present investigation was to compare the infection susceptibility of the bone surrounding implants of bone cement and pure titanium which had been allowed to heal into the bone before bacterial inoculation. Preliminary attempts to create an infection by intravenous injection of bacteria resulted in profound morbidity among the animals. In order to reduce the number of animals lost because of death or systematic illness, we developed a model in which the bacteria were injected locally with minimal tissue trauma.
to Dr L. !
1273
Biomaterials 1995, Vol. 16 No. 16
1274
Infection
MATERIALS
adjacent
to titanium
and bone cement
implants:
L. Sanz&~
and L. Linder
AND METHODS
Surgery Twenty adult 1.7-3.6 kg New Zealand white rabbits of both sexes were operated on under general anaesthesia. Antibiotic prophlaxis with 125 mg of cefuroxime was given prior to surgery. Under aseptic conditions, a cylindrical implant of pure titanium and a piece of bone cement dough of similar size (Palaces, E. Merck, Darmstadt, Germany) were inserted into a 3.6 mm drill hole in the upper tibia1 metaphysis of either leg. The surgical technique has previously been described in detailg. The titanium implants had a central 1 mm hole and in the bone cement implants a 1.2 mm central hole was created with a needle during polymerization. After 12 weeks, radiograms of the legs were taken and the rabbits were divided into three groups: I-III. A in suspension of 5 x 103, 5 x lo5 or 5 x lo7 bacteria 0.1 ml saline from a clinical isolate of S. epidermidis derived from a total joint infection were injected transcutaneously through the hole in the implant into the marrow cavity. The bacterial suspensions were prepared by inoculating the bacteria in tryptic soya broth overnight at 37°C. Bacteria were washed, centrifuged and immersed in isotonic saline. The number of organisms per ml was checked by a serial dilution plate counting technique. The animals were regularly checked for signs of infection and three weeks later new radiograms were made. Since radiograms were unchanged and no clinical signs of infection were found, the animals were considered uninfected. In a new series 9 x 103, 9 x lo5 or 9 x lo7 S. aureus (phage type 85195 from an infected total hip joint) in 0.1 ml saline were injected in the same way through the hole in the implant. Thus, rabbits in group I were inoculated with the lower number of bacteria and those in group III with the higher dose. After another four weeks new radiograms were taken, the animals were killed, and the implants with surrounding bone were removed aseptically using a small saw. The bone surrounding one half of the implant was broken off for culture and the bone surrounding the other half of the implant was fixed and embedded for histology. From the bone marrow, specimens for bacteriological culture were collected. Analysis (1) The severity of radiographic changes such as osteolysis, bone sclerosis and new bone formation were graded from 0 to + + + (Figure Za-c). (2)The biopsies intended for histology were immediately fixed in 10% buffered formaldehyde at 4°C for several days, followed by dehydration in a graded series of ethanol, clearing in xylene, and embedding in methyl methacrylate. Sections were cut at 5 pm on a Jung hard tissue microtome and stained with haematoxylin-eosin and Goldner’s stain. Before fixation the titanium implants were carefully removed from the bone, but the biopsies with the cement plugs were immersed with the implant in place. The cement was later dissolved by the xylene., Biomaterials
1995, Vol. 16
No.16
Figure 1 Examples of radiographic changes 4 weeks after injection of Staphykxoccusaureus into the marrow cavity of rabbit tibia. a, Osteolysis and new bone formation graded + in the posterior cortex of a tibia with a titanium implant. b, Sclerosis and new bone formation graded ++ in both tibiae. c, Pronounced sclerosis and new bone formation graded + + + in a tibia with a bone cement implant.
Infection adjacent to titanium Table 1 Radiographic in rabbit tibia Group
Number of animals
and bone cement
changes, bacteriological
implants:
L. Sanzen
1275
Linder
and L.
cultures and weight change (%) four weeks after injection of Staphylococcus
Weight change
Type of implant
Radiograms
aureus
Positive culture
0
+
++
+++
Bone
Marrow
I
6
+1
Titanium Bone cement
6 6
0 0
0 0
0 0
0 0
0 0
II
6
-3
Titanium Bone cement
3 3
0 2
2 1
1 0
0 1
1 1’
Ill
5
-9
Titanium Bone cement
2 3
1 1
2 0
0 1
0 3
1 1*
*Leg with positive cultures from both bone and bone marrow.
(3)
Tissue biopsies were sonicated and incubated for 48-72 h in tryptic soy broth and subcultured on blood agar plates. Colonies were identified by morphology, phage typing and disc antibiograms.
implants were culture positive in the marrow. No culture showed growth of the injected strain of S. epidermidis. The coincidence of radiographic changes and a positive bacteriology in group II and III is shown in Figure 2. In the low-dose group (I) all cultures were negative.
RESULTS Table 2
At 12 weeks, before bacterial injection, radiograms showed no abnormalities except in one case in group II, where a moderate sclerosis was found in the tibia with the titanium implant. Three weeks following injection of S. ep.idermidis all radiograms were unchanged and the animals showed no local signs of infection such as limp, swelling, tenderness or increased heat. Also, the animals were generally at ease, and had gained weight. During the following four weeks, after the S. aureus injection, the animals became ill and lost weight in the two groups (II and III) with the largest numbers of inoculated bacteria (Table 1). Three animals died during the investigation period, one in group :I1and two in group III. Individual data for all 20 rabbits are found in Table 2.
Radiography No changes occurred following injection of S. epidermidis. Four weeks after injection of S. aureus some radiograms showed different grades of osteolysis, sclerosis and new bone formation in the upper half of the tibia (Figure 1a-c). No radiolucent zones close to the implants or bone sequestrae were seen. Changes were found in 6111 tibiae with titanium implants and 5111 with bone cement implants in the two groups (II and III) with the largest inoculates, and none in the group (I) with the smallest inoculates (Table 2).
Bacteriology In six legs of four animals (one in group II and three in group III) S. aureus was isolated, and all strains were of the same phage type as that injected (Table 3). Cultures were positivle from bone close to the implant in 4111 legs with cement implants and in Oh1 with titanium implants (,P = 0.09, Fischer’s exact test). Bacteria were also found in the marrow of two of these legs with cement implants. Two legs with titanium
A
Individual data for 20 rabbits’ B
1
0
2
0
3
5
4
-2
5
2
6
0
7
0
6
-8
9
-11
10
5
11 12
-9
13 14
5 -14
15 16
-6
17 18
-15
19
-3
20
-7
C
D
E
F
G
H
1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2
1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3
0 0 0
0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0
1 1 1 1 1 2 3 3 1 2 1 ? ? 2 ? 1 1 2 1 3
0
0 0 0 0 0 0 0
0 0 1 2 0 2 2 0 1
0
0 0 0 0 0 0 0
0 0 0 1 0 0
I
1 3 0 0 0 2 0
0 0 0 0 0 1
0 0 0 0 1 0
1 2 1 ? 1 3
2
1 0 0
0 1
0 0
1 1 1
0 3 2 1 1 0
0 1 0 0 0 0
0 1 0 0 0 0
2 3 ? 2 ? 1
*A is rabbit number. B is weight change (%). C is implant: 1, titanium; 2. bone cement. D is bacterial dose: 1, 104; 2, lo?; 3, 10’. E is radiographic changes: 1, +; 2, ++; 3, + + +. The bacteriological cultures were as follows: F, bone. G, marrow: 0, negative; 1, S. aureus. tl. bone-implant contact: 1. extensive; 2. patchy; 3. none; ?, not classified. I, complication: 1. died; 2, bone sclerosis before bacteriological challenge.
Biomaterials 1995, Vol. 16 No. 16
Infection adjacent to titanium
1276 Positive
roentgenography
Positive
bacteriology
Figure 2 Distribution of positive radiographic and bacteriological findings in 11 rabbit tibia with titanium and bone cement implants inoculated with 9 x 105-9 x lo7 Staphylococcus aureus (groups II and III).
and bone cement implants: L. Sanz&
Thus, there was no correlation tion and positive cultures.
and L. Linder
between osseointegra-
Interface membrane Where a membrane was seen around the titanium implants, it was never inflamed. In the case of bone cement, 7110 interface membranes showed a patchy infiltration of lymphocytic cells or plasma cells (Figure 3). Multinucleated giant cells were always seen around the endosteal aspect of the cement implants, regardless of the presence or absence of a continuous membrane, but never around the titanium implants. Marrow There was no correlation between histology and radiographic changes, except for one case classified radiographically as + + +. In this case with a bone cement implant there was proliferation of bone in the marrow cavity and all marrow species were occupied by loose fibrous tissue with an increased number of plasma cells (Figure 4). Less pronounced fibrosis and/ or bony proliferation was seen in two titanium cases and five cement- cases. There was no correlation between these findings and the bacteriological or the radiographic results.
DISCUSSION
Figure 3 Detail of the membrane adjacent to the endosteal aspect of a cement implant in group II. Bacterial cultures were negative and radiograms showed no reaction. Multinucleated giant cells are seen bordering the cement surface. There is a focal accumulation of inflammatory ceils within the membrane. This amount of plasma cells represents the largest accumulation of inflammatory cells in any of the sections in the entire study (Goldner stain, x200).
This study was based on the clinical situation usually present in hematogenous infections, i.e. the mature interface of a stable implant. In a set of preliminary experiments we tried to infect the implants by intravenous injections of bacteria. In the case of S. epidermidis the animals and implants were unaffected, but in the case of S. aureus more than half of the animals died due to multiple septic emboli after a period of severe illness. For ethical reasons intravenous injections were abandoned. In order to -reduce animal morbidity and mortality while still atraumatically bringing bacteria in close
Histology Six legs (four titanium and two cement implants) did not include a sufficient continuous length of the interface to allow classification of the interface structure. All of these legs were culture negative. Of the remaining 13 titanium implants, 11 were osseointegrated (i.e. the extent of the bone-implant contact was considered normal for the experimental model and implant materialg* lo), one was incompletely invested in bone, and one [group I] invested in a reaction-free fibrous membrane. The corresponding figures for the cement implants were five, six, and. four (one in groups I and II, two in group III). Osseointegration-bacterial growth In the two titanium cases, which were culture positive in the bone marrow, osseointegration was present. In the four cement cases with positive cultures from bone, one was osseointegrated, one had a limited area of direct bone-cement contact, whereas two had a fibrous tissue interface with a patchy infiltration of plasma cells but no lymphocytes or granulocytes. Biomaterials
1995, Vol. 16 No. 16
Figure 4 The bone adjacent to a cement implant in group III, where bacterial cultures were positive and where radiograms were classified as + + +. There is an intimate contact between the cement and the bone, in some places without intervening cells. There is a marked proliferation of bone and all marrow spaces are fibrotic (Goldner stain, x75).
Infection adiacent to titanium
and bone cement implants:
L. Sanzk
contact with the implants, the transcutaneous technique was developed. The aim was to evaluate whether one of the implani. materials was more prone to infection than the other. Our doses were chosen from the literature on inoculation in fresh wounds; the lowest dose was expected to infect only cement implants, while the highest was expected to infect all implants. While we do not know whether an injection of bacteria reproduces all the local events during septicemia, there is no doubt that the bacteria are deposited very close to the implant surface, as the injection needle passes through the implant. The correlation between bacterial dose and weight loss after injection of S. aureus indicates that the rabbits did in fact become infected. Although our strains of S. epidermidis and S. aureus were from human implant-related infections, they obviously had different virulence. This is in accordance with earlier studieszs6. In the study in dogs by Petty et al.‘, the IDS,, values for these bacteria differed up to 500 times for infections adjacent to metal implants. However, for cement implants cured . . JR VJVO IDso, 2 x 104, was the same for both S. aureus Consequently, at least some and S. epidermidis. infections ought to have occurred in our groups II and III. Our inability to produce infections with S. epidermidis can eitlher be explained by too low a virulence of the bacterial strain or by a protective effect of a mature bone-implant interface as compared to a fresh operative wound. In the case of S. aureus, infection did occur, but again less easily than expected. In fresh wounds with bone cement implants cured in viva, lo4 S. aureus infected all rats6 and ID,~ for dogs’ was 2 x 104. If metallic implants are used 10” bacteria give a 50% incidence in dogs” and 2 x lo6 bacteria a 100% incidence of persistent infections in rabbits*. In our preliminary experiments with intravenous injection in rabbits severe multiorgan infections developed, proving the virulence of the bacterial strain. In the main experiment no infections were produced with lo4 bacteria, either for cement or titanium. We did not observe more than 50% incidence of radiographic changes when doses of 106-lOa S. aureus were used, and in these legs there were only 4122 with persistent infection, indicating a healing process. This is in contrast to osteomyelitis caused by direct inoculation in fresh wounds, where bacteria could be cultured after several months in 44/ 57 rabbits’ and in 20120 dogsI’. Cultures from bone in two legs close to cement implants were positive without any radiographic changes observed. This was also found by Blomgren et a1.5 in infections caused by Propionibacterium acmes. Comparing the infection rates around the cement and titanium implants, radiographic changes were present in equal numbers, but positive cultures were present in only two legs with titanium implants against four with bone cement implants. Moreover, it is noteworthy that we were unable to culture any bacteria from the bone biopsies close to the titanium implants, whereas
and L. Linder
1277
four bone biopsies close to cement grew S. aureus. We conclude that the infection susceptibility is reduced by a mature interface compared to a fresh wound, but, even so, there still seems to be a difference between a metallic implant and bone cement. Whether our inability to cause a persistent osteomyelitis close to the titanium implants depends on its suggested bactericidal properties7 cannot be answered by this study but certainly merits further investigation.
ACKNOWLEDGEMENTS This study was financially supported by the Riksfijreningen mot reumatism. The authors wish to thank Dr Mats Walder and BjGrn Nilsson at the Department of Clinical Morphology for valuable assistance and Dr Per Gunnar Persson, Department of Pathology, Kristianstad Hospital for help with the evaluation of the histological sections.
REFERENCES 1
2
3
4
5
6
7
8
9 10 11
Andriole VT, Nagel DA, Southwick WO. A paradigm for human chronic osteomyelitis. J Bone Joint Surg (Am) 1973; 55: 1511-1515. Petty W, Spaniewer S, Shuster JJ, Silverthorne C. The influence of skeletal implants on incidence of infection. Experiments in a canine model. J Bone Joint Surg (Am) 1985; 67: 1236-1244. Blomgren G, Lindgren U. The susceptibility of total joint replacement to hematogenous infection in the early postoperative period: an experimental study in the rabbit. Clin Orthop 1980; 151: 308-312. Blomgren G, Lindgren U. Late hematogenous infection in total joint replacement: studies of gentamicin and bone cement in the rabbit. Clin Orthop 1981; 155: 244248. Blomgren G, Lundquist H, Nord C-E, Lindgren U. Late anaerobic haematogenous infection of experimental total joint replacement. A study in the rabbit using Propionibacterium acnes. J Bone Joint Surg (Br) 1981; 63: 614-618. Elson RA, Jephcott AE, McGechie DB, Verettas Q Bacterial infection and acrylic cement in the rat. J Bone Joint Surg (Br) 1977; 59: 452457. Tengvall P, Hijrnsten EG, Elwing H, LundstrGm I. Bactericidal properties of a titanium-peroxy gel obtained from metallic titanium and hydrogen peroxide. J Biomed Mater Res 1990; 24: 319-330. Brinemark P-I, Hansson BO, Adell R, Breine U, Lindstram J,’ Hall& 0, ijhman A. Osseointegrated implants in the treatment of the edentulous jaw. Stand JPIastic Reconstruct Surg 1977; supplement 16. Linder L. Osseointegration of metallic implants I. Acta Orthop Stand 1989; 60: 129-134. Linder L. Reaction of bone to the acute chemical trauma of bone cement. J Bone Joint Surg [Am) 1977; 59: 82-87. Fitzgerald RH. Experimental osteomyelitis: description of a canine model and the role of depot administration of antibiotics in the prevention and treatment of sepsis. J Bone Joint Surg (Am) 1983; 65: 371-380.
Biomaterials 1995. Vol. 16 No. 16