- Email: [email protected]
Synovial macrophage response to aluminium oxide ceramic and cobalt-chrome alloy wear particles in rats
Scone macroph~e response to aluminiumoxideceramicand cobalt-chromealloywear particles In rats .
D.W.Howieand B. Vernon-ibex Dep8ftment of ~atholo9~ Un~e~i~ of Adelaide, Ade~8ide, South Austriltia, Australia (Received 28 May f 98 7; accepted 30 September 198 7)
The tissue response to intra-articular injection into rat knees of similar concentrations of sterile cobaltchrome (Co+) alloy end aiuminium oxide (AI,O,) ceramic prosthesis wear particles was assessed at 1,4 and 13 wk. A semiquantitative method of histological assessment showed a co~elation be~een the numbers of particles and macrophage response in the tissues, and a difference in the response to different concentrations of Co-Cr particles. At ? wk, the macrophage response to Co-Cr particles was significantly greater @ < 0.01) than that to A&O, particles but no difference was detected at 4 and 13 wk. The difference may be due to necrosis of macrophages induced by Co-Cr particles. Keywords: Bi~ompafibi/i~,
wear particles, alumina ceramic, cobalt-cbmme
Aluminium oxide (AI,O,) ceramic total hip prostheses have been developed as the ceramic has a number of claimed advantages which include: better biocompatibility than stainless steel and cobalt-chrome (Co-Cr) alloys’-4; a very low coefficient of friction5’6; a low wear rate7+; less wear particle production; and a lower toxicity of ceramic wear particles to tissues compared with particles of other materials”, ’ ’ . The ceramic prostheses currently used may have either a ceramic on ceramic, or ceramic on ultra high molecular weight polyethylene articulation, usually have a metal femoral stem, and may be implanted with or without the use of acrylic cement. The clinical results of Al,Os ceramic prostheses, in the short term, have been reported to be as good as the results using prostheses made of conventional materials7*“-‘4. However, Trepte et a/.‘s reported a high incidence of femoral loosening of uncemented components of ceramic total hip arthroplasties at a shortterm follow-up. Moreover, O’Leary and Mallory” reported poor results at 2 yr follow-up of 66 Mittelmier ceramic total hip prostheses, with a failure rate requiring revision of 19% and few patients were free of pain. The major cause for failure was component loosening. Whether ceramic prostheses are likely to be superior to prostheses made of other materials, depends on whether ceramics are more biocompatible than metals when these materials are implanted in bone under weight-bearing Correspondence to Dr D.W. Howia, Orthopaedic Surgery Unit, Flinders Medical Centre, Bedford Park, South Australia 5042, Australia. Q 1988 442
Biometerieis
1988, Vol9
September
alloy, synovia
conditions, and whether ceramics cause fewer problems due to friction, wear and wear particle release. Severe wear of ceramic prostheses has been repot-ted following malposition of components’3 and wear rates of ceramic on ceramic prostheses components of up to five times that seen in ideal situations have been reported if there is minor incongruity between articulating components’7. Studies in vitro and in vivo suggest that Al,Os ceramic particles have little effect on cells and tissues”, 18*19. Problems in these studies were that particles were injected in saline which will not prevent clumping of particles, no method to exclude infection was mentioned and no quantitative assessment of the tissue response was made. Some of the tissue responses described in these studies might indicate mild to moderate toxicity but are difficult to interpret as no controls were included. This study was undertaken to compare the tissue response to intra-articular injection of Al,Os particles and Co-Cr particles using a semiquantitative method of assessment.
~ATEKIALS
AND METHODS
Preparation of Co-Cr particles Wear particles of Co-Cr alloy (Vitallium, Howmedica Inc.) were prepared as previously described” with the following modifications. instead of using a milling fluid constituted from 20% fetal calf serum added to RPMI f 640 medium, homologous serum prepared from J.C. Lewis rats and Butterworth Et Co (Publishers) Ltd. 0142-96
12/88/050442-07$03.00
Synovial macrophage response: D. W. Howie and B. Vernon-Roberts
diluted 1 : 50 in sterile normal saline (NaC12 0.9% w/v) was used. Before milling, the Vitallium screws and chamber cups were washed and acid-neutralized according to ASTM standards for biocompatibility testing of metals and sterilized by autoclaving”. The size distribution of particles was determined by differential sedimentation” and particles of 3 pm and less were utilized for injection. The concentration of metal and proportion of Co and Cr in the particle suspension were assessed by atomic absorption spectroscopy (Varian Techtron Model 1200)22 after dissolving particles by heating in concentrated HN03 and then in HCI (Ref. 23). The content of metal was measured against Co and Cr salt standards and against dry weighed samples of Co-Cr alloyz4. The metal particle suspension prepared in the shaker was found to contain 1 .I 1 mg/ml Co and 0.55 mg/ml Cr, i.e. a Co : Cr ratio of 2 : 1. Since Co comprises 60% of the Vitallium alloy, the total metal content of the particle suspension was 1.85 mg/ml. This particle suspension was diluted to 0.74 and 0.30 mg/ml, and these suspensions were termed highdose and low-dose Co-Cr suspensions.
Preparation of A1203 particles AI,O, particles were produced by shaking irregularly shaped pieces of A1203 ceramic (Bioiox, Richards Medical), approximately 2-5 mm in diameter, in a shaker which consisted of two unused Biolox Al,03 ceramic hollow femoral head components of a Mittelmier total hip prosthesis which were clamped together. The particles were milled in a solution of homologous J.C. Lewis rat serum diluted 1 : 50 in normal saline. Before milling, the A1203 shaker and pieces were washed in a warm ultrasound bath in a solution of Sparkleen diluted 1 : 10 in distilled water. Sparkleen (Fisher Co.) is an organic solvent used to wash total joint replacement prostheses. After five rinses in distilled water, the particles were washed in dilute HNO, to remove possible metal contaminants, rinsed five times in distilled water, washed in an ultrasound bath in boiling distilled water, and then autoclaved. These procedures fulfilled the requirement that materials be prepared in a manner similar to their preparation for human implantationz’. Particles above 3 iurn in diameter were separated from the prepared suspension by differential sedimentation2’ and discarded. The weight of theA1203 particles in the remaining suspension was calculated by weighing oven-dried samples of the suspension. The particles were prepared as a suspension in dilute serum to a concentration of 0.74 mg/ ml AI,O,, which was the same concentration of metal as in the high-dose Co-Cr suspension. As the ratio of the specific gravity of A1203 to Co-Cr is approximately 2.5 : 8, the A1,03 suspension was diluted by this ratio to obtain approximately equivalent numbers of ceramic and Co-Cr particles in the respective suspensions. The particles were sterilized by auto~laving before injection.
examination of these knees (Figures 7 and 2) demonstrated macrophage proliferation and phagocytosis of alumina wear particles.
Techniques for injection and killing of animals Sixty male J.C. Lewis rats 8-l 6 wks old and weighing 180-300 g, were injected with the high-dose Co-Cr particle suspension in one knee and the A120, particle suspension in the other knee. Another 60 similar rats were injected with low-dose Co-Cr particle suspension in one knee and dilute serum control solution in the other knee. The rats were randomly allocated to groups so that there were 20 highdose metal and ceramic injected rats and 20 low-dose metal and saline injected rats in each group. The rats in these groups were killed at I,4 and 13 wks following injection. Injection and killing were performed under anaesthetic using 3% halothane in 50% N,O in oxygen. Before injection and killing, the legs were shaved and prepared with 1% alcoholic chlorhexidine. Injection, using a separate 0.5 ml syringe (B-D Plastipak 0.5 ml insulin syringe) with 27.5 gauge needle for each knee was performed through the patella tendon while the knee was held at 30 degrees of flexion. Distension of the
Figure7 Photomi~mgraph of the synovium of a rat knee 1 wk follow;ng intra-a~;cutar injection of unsized A/,0, ceramic wear particles. There has been marked accumulation of macrophages associated with the presence of wear particles. (Haematoxylin and eosin, X SO].
Preliminary study A preliminary study of the effects of A1203 particles was performed by injecting a sterile sample of the particle suspension before grading of particle size and separation of the smaller particles to be used in the definite study. Five rats received intra-articular injection of particles in both knees and, after killing them at 1 wk, the tissues were processed in the same manner as in the definitive study. Histological
Figure 2 Photomi~rograph of the synovium of a rat knee 1 wk following intra-a&colar injection of unsized particles of A1,03 The wear particles are located predominantly within macrophages. Note the absence of lymphocytes and necrosis confined to occasional macrophages. (Haemetoxylin and eosin, X400).
Biomaterials
1988, t/of 9 September
443
Synovial macrophage response: 0. W. Howie and 8. Vernon-Roberts
suprapatella pouch was taken to indicate a successful injection. The amount injected was 0.01 ml per 30 g body weight, as this was found in pilot studies to be the amount necessary to distend the knees of rats of various sizes. Animals recovered within 1 h and did not limp. They were housed in metal cages and fed a routine cubed diet. At the time of death, the animals were inspected externally for swelling of the knee or tumours. Under anaesthesia, using aseptic techniques including sterile instruments for each knee, an arthrotomy was performed and a synovial biopsy was taken. The biopsy was placed in Stuarts transport media for microbiological culture. After killing, the whole leg was resected and fixed in 10% formalsaline. The knee was held flat on a glass slide to allow consistent sectioning of the knee for histopathology examination. A post-mortem examination was performed on each animal.
Tissue processing for histopathology After fixation in 10% buffered formal-saline for 3 d, the knees were decalcified using a commercial decalcifying solution (Decal, Omega Chemical Corporation, NY, USA). The extent of decalcification was controlled by daily radiographs (Kodak Min-R film, Hewlett Series 43805N Faxitron Cabinet). After neutralization in 5% silver sodium sulphate, the central portion of the knee was removed by sag&al section either side of the patella. A specimen 4 mm thick was obtained, which was then dehydrated through graded alcohol, cleared in chloroform and embedded in paraffin wax. After trimming, beginning at the side of the arthrotomy, a 6 pm section from the middle of the knee was cut using a hand-operated LKB 2259 Multirange Microtome. Sections were stained with haematoxylin and eosin.
Microbiology
techniques to exclude infection
Aerobic and anaerobic cultures were performed on all rat synovial biopsies and injected solutions of metal, ceramic and dilute serum before and after injection. The biopsies were removed from Stuart’s media and, using a stomacher, were homogenized in 0.5 ml of glucose-cooked meat medium. In an anaerobic chamber, the homogenates and the samples of injected materials were inoculated onto a supplemented blood agar (BAYH) plate, an anaerobic blood culture broth and a glucose-cooked meat broth, and these were incubated anaerobically. The sample was then inoculated on to a chocolate agar plate and incubated in COz. Plate cultures were continued for 4 d and broth cultures for 10 d. A subculture was taken if any broth became turbid. Organisms were identified according to standard laboratory techniques25,26.
Histopathological
Biomaterials
Statistical
methods
The correlation between the particle scores and the macrophage scores for each suspension and the significance of the correlation were assessed by Kendall’s Rank Correlation method modified for tied rank’*. The significance of any differences between the particle and macrophage scores of each suspension, was calculated by the Wilcoxon Rank Sum Test for unpaired data’* . The macrophage score was plotted against the particle score for the suspensions, and the significance of differences between correlation coefficient was assessed by covariant analysis”. Because a number of tests for significance of differences and correlations were performed, thereby increasing the likelihood of finding differences, a significance level of p < 0.01 was accepted.
RESULTS Histology At 1 wk following injection, particles of Co-Cr (Figure 3) and A1203 (Figure 4) were both extracellular and intracellular within macrophages. At 4 and 13 wk, the majority of particles were intracellular. The number of particles seen in all injected knees appeared to decrease between 1 and 4 wk. and then remained constant to 13 wk. The predominant cellular response to both materials was a macrophage infiltrate which appeared greater at 1 wk. had decreased at 4 wk. and remained unchanged at 13wk. There were differences in the response to the Co-Cr and A1203 particles at 1 wk. In the Co-Cr injected knees, prominent zones of necrosis of macrophages were seen at the sites of high particle accumulation (Figure 3). By contrast, in the knees injected with A1203 particles in the preliminary study where the particle size was not controlled (Figures 1 and 2) and in the definitive study of particles of 3 pm and smaller (Figure 4). very few macrophages showed evidence of necrosis and zones of necrosis were absent.
scoring
A semiquantitative scoring method was used based on the presence or absence of a particular feature in each successive high-power field of synovium and subsynovium of the section examined microscopically. The cell and particle scores were based on a modification of that used by Mirra et aI.” to assess the tissue response to wear particles in humans. A positive macrophage score was more than 30 macrophages per high-power field (X400, field area 4.7 mm’) and a positive particle score was more than 20 particles. Quantitation was not attempted for necrosis, fibrosis, presence of other cell types and gradings of macrophages or particles scores, since pilot studies had shown this to be
444
unreliable. The macrophage and particle scores were derived as a percentage of the number of high-powered fields of synovium in which these features were present in each section of the rat knee relative to the total number of fields in the section as a whole.
1988, Vol 9 September
Photomicrograph of the synovium of a rat knee I wk following Figure 3 intra-articular injection of Co-0 particles of 3 pm and smaller. There has been extensive necrosis of macrophages which have accumulated in association with wear particles. Lymphocytes have accumulated in the subsynovium. (Haematoxylin and eosin, X 160).
Synovial macrophage response: D. W. Howie and B. Vernon-Roberts
system detects differences in the macrophage response to different numbers of particles of the same material. The low-dose Co-Cr macrophage scores were not significantly different from the AI,O, macrophage scores at 1, 4 and 13 wk following injection.
Particle
Photomicrograph of the subsynovium of a rat knee 1 wk Figure 4 following injection ofA1203patticlas of 3 pm andsmaller. There is a localized accumulation of macrophages in response to particles which are intracellular and extracellular. Necrosis is absent. (Haematoxylin and eosin, X 160).
and macrophage
correlations
There was a high correlation between the particle score and macrophage score following injection of high-dose Co-Cr suspension (r = 0.92,p < 0.001 ), low-dose Co-Cr suspension (r = 0.97, p Q 0.001) and A1203 suspension (1.= 0.83, p < 0.00 1). To determine the macrophage response to the two different concentrations of particles of Co-Cr, the macrophage scores were plotted against the particle scores at each time interval (Figure 5). The lines of best fit have similar slopes, but the low-dose suspension consistently scores a lower particle and macrophage score, confirming the differences described previously. One Week
Moreover, while a lymphocytic infiltrate was prominent 1 wk after the injection of Co-Cr particles (Figure 3), lymphocytes were not a feature of knees injected with AI,O, at any stage (Figures 1 and 4). At 4 and 13 wk, necrosis was confined to the reaction to Co-Cr particles, and lymphocytes were not observed in the reaction to either particle. There was a difference in the response to high-dose and low-dose Co-Cr particle injections: more particles and macrophages were seen in the knees injected with highdose Co-Cr particles and necrosis was more common. The knees injected with saline control solution showed slightly increased cellularity of the subsynovium at 1 wk and were normal thereafter. No tumours were found in any animals.
Particle
”
I
0
Macrophage
1
I
I
20
30
40
50
%
Four Weeks 60
g
40
P g
30
9 y;ofl 1, 0
,
,
,
,
,
10
20
30
40
50
PartIcks
%
Thirteen Weeks 60
scores
The high-dose Co-Cr macrophage scores were significantly different from the A1203 particle scores at 1 (p < 0.001). 4 (p Q 0.001) and 13 wk (p < 0.001) following injection and suggest a difference in the macrophage response following injection of these particles of different materials. The high-dose Co-Cr macrophage scores were significantly different from the low-dose Co-Cr macrophage scores at 1 (p < O.OOl), 4 (p < 0.001) and 13 wk (p < 0.00 1) following injection, and suggest that the scoring
I
10
PartIcks
scores
The high-dose Co-Cr particle scores were significantly different from the low-dose scores at 1 (p Q 0.001 ), 4 (p < 0.001) and 13 wk (p < 0.001) following injection, demonstrating that the scoring method used detects differences in the number of particles in knees injected with different concentrations of particles of the same material. There was no significant difference between the highdose Co-Cr particle scores and the A1203 particle scores at 1, 4 and 13 wk. There was a significant difference between the low-dose Co-Cr particle scores and the A1203 particle scores at 1 (p < 0.001). 4 (p Q 0.001) and 13 wk (p g 0.001). These results suggest that the numbers of particles of A1203 in the knees were similar to the numbers following injection of high-dose Co-Cr suspension, but were different following the injection of low-dose Co-Cr suspension.
-
i
:$&.g/ t 0
IO
20
30
I
40
50
PartIcks%
The macrophage response at 1,4 and 13 wk following injection Figure 5 of high-dose l and low-dose 0 Co-&particle suspensions. The lines of best fit have similar shapes, indicating that the macrophage response is related to the number of particles, but is not as severe following injection of smaller numbers of particles of the same material.
Biomaterials
1988, Vol 9 September
445
Symviat macrophage response: D. W. Howie and 6. Vernon-Roberts
The macrophage and particle scores for high-dose and low-dose Co-Cr suspensions were combined and compared with the scores for A&OS suspension (Figure 6). One week falling injection, the lines of best fit between the two materials were significantly different (P < 0.01) and suggest that the macrophage response to Co-Cr particles was approximately twice that to Al20s particles. At 4 and 13 wk, there were no signifi~nt differences between the slope of the lines. These results confirm a difference in the response to Co-Cr particles compared with Al206 particles which had been suggested by the finding of significant differences between the macrophage scores for high-dose Co-Cr and Al2Os. Although significant differences between macrophage scores were also detected at 4 and 13 wk, while no differences were detected for particle scores, the method of covariant analysis did not confirm these differences.
Microbiology The cultures from the synovial biopsies of the rat kn88S and the human tissues were negative, except for one rat knee One Week 60-
I 20
t 30
Particles %
Four Weeks
30
40
50
Particles 96
Thkt8en Weeks 60
I
1
20 30 Particles %
,
1
40
50
The macrophage response at 1,4 and 13 wk following injection Figure 6 of Co-0 @ andAi20s A particle suspensions. The slopes of the lines of best fit are significantly different at 1 wk (p < 0.01). but nor at 4 and 13 wk following injection, indicating a more severe macrophage response to similar numbers of Co-0 pefiicles compamd to A120s particles.
446
Biomaterials
1988, Vat 9 September
biopsy which grew Staphylococcus epidermidis in one broth. This growth was assumed to be a contaminant, as the plate cultures and other broth cultures were negative and the histopathol~y of this knee showed no evidence of acute inflammation.
DISCUSSION Intra-articular injection of Co-Cr alloy and Al20s ceramic particles in the same size range as particles seen in the tissues around human prostheses”, ’ ‘* 13,3’s31 induced a macrophage response in the rat knees which correlated with the amount of particulate deposited in the synovium at the three killing times. The macrophage response to Co-Cr particles differed from that to Al206 particles. One week following injection, the response to Co-Cr particles was greater than that to AI20, particles but there was no difference at 4 and 13 wk. These results suggest Co-Cr particles induce a greater inflammatory response shortly after injection. Both these materials are relatively inert in solid form and Richardson et a1.32 found no difference in the response to their intramuscular implantation. /n viffo comparisons have not been made between these two materials in particle form but there is some suggestion that Co-Cr particles are more toxic. Co-Cr particles in vitro interfere with the phagocytic activity of macrophages and are toxic to macrophages and fibroblasts”* 33. Cobalt has been implicated as the most toxic of the constituent metals of this alloy and this toxicity has been related to its greater solubility34. A1203 particles have been reported to mildly inhibit fibroblast growth in vitro, causing slightly less inhibition than stainless steel particles”. In vivo studies suggest that ceramic particles are relatively inert. Griss et al.” and Harms and Maus18’” concluded that the macrophage response to subcutaneous, intramuscular and intra-a~icular injection of particles of a similar size to those used in the current study was not of concern. These conclusions were based on subjective assessment of the tissue response and did not include a control material nor adequate techniques to exclude infection. Some worrying features were reported in these studies. A neutrophil response which peaked at 1 wk following subcutaneous injection’* and a persisting lymphocyte response ” following intramuscular injection may be interpreted as evidence of mild to moderate toxicity35P3”. These features were not observed in our study. Uchida3’ observed that the intra-articular injection of Al206 ceramic produced a more severe tissue response than polymer and metal particles, including Co-Cr alloy. The results of the current study suggest a difference in the degree of the initial macrophage response to the two types of particle, but no difference in the degree of the persisting macrophage response.‘rhus, it is possible that the continuing macrophage response depends on the mechanical presence of the particles rather than any continuing toxic properties which may or may not persist. The scoring system used in this study bears further discussion. Despite attempts to inject particles of a similar size and concentration, and in approximately similar amounts, it is possible that there were large variations in the number of particles exposed to the synovia of different knees. For this reason, a scoring system was used which related the macrophage response to the number of particles in the synovium. In addition, the whole of the knee was examined to take into account the possible variations in the number of particles
Synovial macrophage response: D. W. Howie and B. Vernon-Roberts
and the macrophage response at different sites. If the required minimum number of macrophages or particles was present in any given high-power field, then a positive score was recorded. This system has limitations, as there may be far larger numbers of either macrophages or particles present in two different areas but the score will be the same once the threshold number is reached to achieve a positive score. To assess the value of this scoring system in detecting differences in particle numbers and macrophages, two different concentrations of Co-Cr particles were used. The scoring system seems justified as the macrophage response correlated with the particle score in all groups and, more importantly, the scoring system detected a difference in the particle and macrophage score following injection of highdose and low-dose Co-Cr suspensions. The macrophage response alone was used to assess the tissue response to particles, as it could be reliably scored. Other features, such as the extent of lymphocyte and neutrophil infiltrate and the degree of necrosis and fibrosis have been used by others to assess the biocompatibility of particles following intra-articular injection. Some of these features were present in the knees in this study but could not be reproducibly scored. In conclusion, the initial macrophage response to CoCr particles was different from that to Al,O, particles. After this initial difference, the macrophage response was related to the number of particles of each material but not to the type of material. It is not known whether the difference in the macrophage response to the different wear particles is important in determining the deleterious effects of wear particles around human prostheses. A material which produces wear particles that provoke less macrophage response in the surrounding tissues would seem desirable but a difference in the macrophage response between the two materials in this study was only detectable at 1 wk following injection. The adverse effect of macrophages in the periprosthetic tissues and their possible role in bone resorption and prosthetic loosening takes many years to become evident. As particles are continuously released into the periprosthetic tissues and possible toxic effects of the particles may have a cumulative effect, differences in the effect of particles seen at 1 wk may be important. Also, in this study, the persisting macrophage response at 4 wk and 3 months was related to the number of particles in the synovium. It is likely, therefore that the number of particles that persist in the tissues is also an important determinant of long-term adverse effects of prostheses wear particles.
ACKNOWLEDGEMENTS The authors gratefully acknowledge the assistance of Dr R. Garrett and Miss S. Hay (Department of Pathology, University of Adelaide); the staff of the Animal Operating Theatre, Medical Illustration and the Bone Laboratory, Institute of Medical and Veterinary Science; the staff of the Photography Unitsof the Department of Pathology, University of Adelaide, and of Flinders Medical Centre and Mrs R. Kenney in preparation of the manuscript. This work was supported in part by grants from the Daws Research Fellowship, Royal Adelaide Hospital, the Australian Orthopaedic Association, the Royal Australasian College of Surgeons, and the National Health and Medical Research Foundation.
REFERENCES 1
2
3
4
5
6
7
8
9
10 11
12 13
14 15
16
17
18
19
20
21
22
23
Griss, P., von Andrian-Werburg, H., Krempien. B. and Heimke. G.. Biological activity and histocompatibilityof denseA120s/Mg0ceramic Implants in rats, J. Biomed. Mater. Res 1973, 7.453-462 Griss, P.. Greenspan, DC., Heimke, G., Krempien, B., Bucherger, R., Hensch, L.L. and Jentschura, G., Evaluation of a bioglasscoated AI,Os total hip prosthesis in sheep, .I. Biomed. Mater. Res. 1976, 10, 511-518 Heimke. G., Griss, P., Jentschura, G. and Werner, E., Direct anchorage of A1203-ceramic hip components: three years of clinical experience and results of further animal studies, J. Biomed. Mater. Res. 1979, 13. 57-65 Busing, C.M., d’Hoedt, B., Schulte, W. and Heimke, G., Morphological demonstration of direct deposition of boneon human aluminiumoxide ceramic dental implants, Biomarerials 1983, 4(2), 125-l 27 Dorre. E. and Dawihl, W., Ceramic hip endoprostheses, in Mechanical Properties of Biomarerials (Eds G.W. Hastings and D.F. Williams), John Wiley Et Sons, 1980, pp 113-l 27 Dowson, D. and Linnett, I.W., A study of the wear of ultra molecular weight polyethylene against a high alumina ceramic, in Mechanical Properties of Biomaterials (Eds G.W. Hastings and D.F. Williams), John Wiley 8 Sons, 1980 3-25 Boutin, P.. Arthroplasttc totale de la hanche par prothese en alumine frittee: etude experimentale et premieres applications cliniques, Rev. Chir. Orthop. 1972,58, 229-246 Semlitsch. M.. Lehmann, M., Weber, H., Oderre. E. and Willert, H.G., New prospects for a prolonged functional life-span of artificial hip joints by using the material combination polyethylene/aluminium oxide ceramic/metal, J. Biomed. Mater. Res. 1977. 11, 537-552 Boutin. P. and Blanquaert, D., A study of the mechanical properties of alumina-on-alumina total hip prostheses, Rev. Chir. Orfhop. 198 1,67, 279-287 Harms, J. and Mausle, E.,Tissue reaction toceramic implant material, J. Eiomed. Marer. Res. 1979. 13. 67-87 Stock, D., Diezemann, ED. and Gottstein, J., Results of endoprosthetic hip joint replacement with the aluminium ceramic-metal composite prosthesis ‘Lindenhof’, Arch. Orthop. Trauma Surg. 1980, 97, 712 Zweymuller, K., Bone and joint replacements with bioceramic endoprostheses, Fortschr. Med. 1979, 97(45), 2107-211 1 Griss. P. and Heimke, G., Five years experience with ceramic-metalcomposite hip endoprostheses. I. Clinical evaluation, Acra Orthop. Trauma Sorg. 198 1, 98, 157-l 64 Rampoldi, A., Mittelmeier’s ceramic hip prosthesis, &a/. J. Orfhop. Traumatol. 1984, 10, 305-3 1 1 Trepte, L.T., Gauer, E.F. and Gartner, B.M., Results with ceramic-onceramic sliding endoprosthesis completely or partially fixed without cement, 2 Olthop. 1985, 123, 239-244 O’Leary, J.F.M. and Mallory, T.H., Mittelmier ceramic total hip anhroplasty: two year follow-up, Anna/ Meeting of rhe American Academy of Orthopaedic Surgeons, New Orleans, Feb. 1986 Hinterberger, J. Ungethum, M. and Plitz, W.,Tribological propertiesof aluminium oxide ceramics, in Mechanical Properries of Biomarenkls (Eds G.W. Hastings and D.F. Williams), John Wiley Et Sons, 1980, pp 73-82 Griss. P.. Krempien, B., von Andrian-Werburg, H.F., Heimke, G., Fleiner, R. and Diehm, T., Experimental analysis of ceramic tissue interactions: a morphologic fluoroescenteoptic and radiographic study on dense alumina oxide ceramic in various animals, .I. Biomed. Mater. Res. 1974, 8. 39-48 Plenk, H., Evaluation of the effect of ceramic and different metallic Implant materials on the growth rate of human fibroblast cultures, in Evaluation ofBiomateria/s (Eds G.D. Winter, J.L. Leray and K. de Groot), John Wtley Et Sons, 1980, pp 399-403 Garrett, R.. Wilksch, J. and Vernon-Roberts, B.. Effects of cobaltchrome alloy wear particles on the morphology, viability and phagocytic activrty of murine macrophages in virro, Ausr. J. Exp. Biol. Med. Sci. 1983,61,355-369 ASTM. F36 l-80: Standard procedure for assessment of compatibility of metallic materials for surgical implants with respect to effects of materials on tissues. F86-76: Standard recommended practice for surface preparation and marking of metallic surgical implants. F60378: Standard specification for high-purity dense aluminium oxide for surgrcal implant application. Annual Book of ASTM Standards, Part 46. American Society for Testing and Materials, Philadelphia, 1980 Gatehouse. B.M. and Willis, J.B., Performance of a simple atomic absorption spectrophotometer, Spectrochimica Acra 196 1, 17, 710-718 McPherson, E.L.. Price, J.W. and Scaite. P.H.. Application of atomic
Biomarerials
1988, Vol9 September
447
Synovial macrophage response: 0. W. Howie and 8 Vernon-Roberts
24
25 26 27
26 29
30 31
448
absorption spsctroscopy to the determination of cobalt in steel, alloy steel and nickel, Nature 1963, 399, 371-372 ASTM, D3558-77: Standard test methods for cobalt in water. D1687-77: Standard test methods for chromium in water. Annual Book ofASTM Standards, Part 3 1, American Society for Testing and Materials, Philadelphia, 1977 Cowan, S.T., Manual for the Identihcation of Medical Bacteria (Ed. S.T. Cowan), 2nd Edn,Cambridge University Press, Cambridge, 1975 Holdeman, L.V., Cato, E.P. and Moore, W.E.C., Anaerobic Laboratory Manual, 4th Edn, Virginia Polytechnic Institute, USA, 1977 Mirra. J.M., Amstutz, H.C., Matos, M. and Gold, R., The pathology of the joint tissues and its clinical relevance in prosthetic failure, C/in. Orfhop. 1976,117,221-240 Swinscow, T.D.V., Statistics a? Sguare One, 3rd Edn, British Medical Association, London, UK, 1978 Snedecor, G.W. and Cochrane, W.G., Method of co-variant analysis, in StatisticalMethods, 6th Edn, Iowa State University press, 1967, pp 41 Q-436 Winter, G.D., Tissue reactions to metallic wear and corrosion products in human patients, J. Biomed. Mater. Res. 1974,8, 1 l-26 Vernon-Roberts, 8. and Freeman, M.A.R., The tissue response to total
Biomatenals
1988, Vol9
September
32
33
34
35
36
37
joint replacement prostheses, in The Scientific Basis of Joint Replacement (Eds S.A.U. Swanson and M.A.R. Freeman), Pitman Medical Publishing Co. Ltd., Tunbridge Wells, UK, 1977, pp 86-l 29 Richardson, W.C., Klawitter, J.J., Sauer, 8.W.. Pruitt, J.R. and Hulbert, SF., Soft tissue response to four dense ceramic materials and two clinically used biomaterials,./. Biomed. Mater. Res. 1975,9,73-80 Rae, T., A study on the effects of particulate metals of orthopaedic interest on murine macrophages in vitro, J. Bone Joint Surg. 1975, 570.444-450 Rae, T., The toxicity of metals used in orthopaedic prostheses: an experimental study using cultured human synovial fibroblasts, ./. Bone Joint Surg. 198 1,638,435-440 ANSI/ADA, Recommended standard practices for biological evaluation of dental materials, American National Standards/American Dental Association Document 1979.41 FDI, Recommended standard practices for biological evaluation of dental materials, Federation Oentaire International Document 1979. 198 (International Dental Federation, London) Uchida, S., A histopathologic comparison of the tissue reaction to prosthetic materials in the knee joint of rats, Onhopaedics 1985, 8, 1276-l 280
D.W.Howieand B. Vernon-ibex Dep8ftment of ~atholo9~ Un~e~i~ of Adelaide, Ade~8ide, South Austriltia, Australia (Received 28 May f 98 7; accepted 30 September 198 7)
The tissue response to intra-articular injection into rat knees of similar concentrations of sterile cobaltchrome (Co+) alloy end aiuminium oxide (AI,O,) ceramic prosthesis wear particles was assessed at 1,4 and 13 wk. A semiquantitative method of histological assessment showed a co~elation be~een the numbers of particles and macrophage response in the tissues, and a difference in the response to different concentrations of Co-Cr particles. At ? wk, the macrophage response to Co-Cr particles was significantly greater @ < 0.01) than that to A&O, particles but no difference was detected at 4 and 13 wk. The difference may be due to necrosis of macrophages induced by Co-Cr particles. Keywords: Bi~ompafibi/i~,
wear particles, alumina ceramic, cobalt-cbmme
Aluminium oxide (AI,O,) ceramic total hip prostheses have been developed as the ceramic has a number of claimed advantages which include: better biocompatibility than stainless steel and cobalt-chrome (Co-Cr) alloys’-4; a very low coefficient of friction5’6; a low wear rate7+; less wear particle production; and a lower toxicity of ceramic wear particles to tissues compared with particles of other materials”, ’ ’ . The ceramic prostheses currently used may have either a ceramic on ceramic, or ceramic on ultra high molecular weight polyethylene articulation, usually have a metal femoral stem, and may be implanted with or without the use of acrylic cement. The clinical results of Al,Os ceramic prostheses, in the short term, have been reported to be as good as the results using prostheses made of conventional materials7*“-‘4. However, Trepte et a/.‘s reported a high incidence of femoral loosening of uncemented components of ceramic total hip arthroplasties at a shortterm follow-up. Moreover, O’Leary and Mallory” reported poor results at 2 yr follow-up of 66 Mittelmier ceramic total hip prostheses, with a failure rate requiring revision of 19% and few patients were free of pain. The major cause for failure was component loosening. Whether ceramic prostheses are likely to be superior to prostheses made of other materials, depends on whether ceramics are more biocompatible than metals when these materials are implanted in bone under weight-bearing Correspondence to Dr D.W. Howia, Orthopaedic Surgery Unit, Flinders Medical Centre, Bedford Park, South Australia 5042, Australia. Q 1988 442
Biometerieis
1988, Vol9
September
alloy, synovia
conditions, and whether ceramics cause fewer problems due to friction, wear and wear particle release. Severe wear of ceramic prostheses has been repot-ted following malposition of components’3 and wear rates of ceramic on ceramic prostheses components of up to five times that seen in ideal situations have been reported if there is minor incongruity between articulating components’7. Studies in vitro and in vivo suggest that Al,Os ceramic particles have little effect on cells and tissues”, 18*19. Problems in these studies were that particles were injected in saline which will not prevent clumping of particles, no method to exclude infection was mentioned and no quantitative assessment of the tissue response was made. Some of the tissue responses described in these studies might indicate mild to moderate toxicity but are difficult to interpret as no controls were included. This study was undertaken to compare the tissue response to intra-articular injection of Al,Os particles and Co-Cr particles using a semiquantitative method of assessment.
~ATEKIALS
AND METHODS
Preparation of Co-Cr particles Wear particles of Co-Cr alloy (Vitallium, Howmedica Inc.) were prepared as previously described” with the following modifications. instead of using a milling fluid constituted from 20% fetal calf serum added to RPMI f 640 medium, homologous serum prepared from J.C. Lewis rats and Butterworth Et Co (Publishers) Ltd. 0142-96
12/88/050442-07$03.00
Synovial macrophage response: D. W. Howie and B. Vernon-Roberts
diluted 1 : 50 in sterile normal saline (NaC12 0.9% w/v) was used. Before milling, the Vitallium screws and chamber cups were washed and acid-neutralized according to ASTM standards for biocompatibility testing of metals and sterilized by autoclaving”. The size distribution of particles was determined by differential sedimentation” and particles of 3 pm and less were utilized for injection. The concentration of metal and proportion of Co and Cr in the particle suspension were assessed by atomic absorption spectroscopy (Varian Techtron Model 1200)22 after dissolving particles by heating in concentrated HN03 and then in HCI (Ref. 23). The content of metal was measured against Co and Cr salt standards and against dry weighed samples of Co-Cr alloyz4. The metal particle suspension prepared in the shaker was found to contain 1 .I 1 mg/ml Co and 0.55 mg/ml Cr, i.e. a Co : Cr ratio of 2 : 1. Since Co comprises 60% of the Vitallium alloy, the total metal content of the particle suspension was 1.85 mg/ml. This particle suspension was diluted to 0.74 and 0.30 mg/ml, and these suspensions were termed highdose and low-dose Co-Cr suspensions.
Preparation of A1203 particles AI,O, particles were produced by shaking irregularly shaped pieces of A1203 ceramic (Bioiox, Richards Medical), approximately 2-5 mm in diameter, in a shaker which consisted of two unused Biolox Al,03 ceramic hollow femoral head components of a Mittelmier total hip prosthesis which were clamped together. The particles were milled in a solution of homologous J.C. Lewis rat serum diluted 1 : 50 in normal saline. Before milling, the A1203 shaker and pieces were washed in a warm ultrasound bath in a solution of Sparkleen diluted 1 : 10 in distilled water. Sparkleen (Fisher Co.) is an organic solvent used to wash total joint replacement prostheses. After five rinses in distilled water, the particles were washed in dilute HNO, to remove possible metal contaminants, rinsed five times in distilled water, washed in an ultrasound bath in boiling distilled water, and then autoclaved. These procedures fulfilled the requirement that materials be prepared in a manner similar to their preparation for human implantationz’. Particles above 3 iurn in diameter were separated from the prepared suspension by differential sedimentation2’ and discarded. The weight of theA1203 particles in the remaining suspension was calculated by weighing oven-dried samples of the suspension. The particles were prepared as a suspension in dilute serum to a concentration of 0.74 mg/ ml AI,O,, which was the same concentration of metal as in the high-dose Co-Cr suspension. As the ratio of the specific gravity of A1203 to Co-Cr is approximately 2.5 : 8, the A1,03 suspension was diluted by this ratio to obtain approximately equivalent numbers of ceramic and Co-Cr particles in the respective suspensions. The particles were sterilized by auto~laving before injection.
examination of these knees (Figures 7 and 2) demonstrated macrophage proliferation and phagocytosis of alumina wear particles.
Techniques for injection and killing of animals Sixty male J.C. Lewis rats 8-l 6 wks old and weighing 180-300 g, were injected with the high-dose Co-Cr particle suspension in one knee and the A120, particle suspension in the other knee. Another 60 similar rats were injected with low-dose Co-Cr particle suspension in one knee and dilute serum control solution in the other knee. The rats were randomly allocated to groups so that there were 20 highdose metal and ceramic injected rats and 20 low-dose metal and saline injected rats in each group. The rats in these groups were killed at I,4 and 13 wks following injection. Injection and killing were performed under anaesthetic using 3% halothane in 50% N,O in oxygen. Before injection and killing, the legs were shaved and prepared with 1% alcoholic chlorhexidine. Injection, using a separate 0.5 ml syringe (B-D Plastipak 0.5 ml insulin syringe) with 27.5 gauge needle for each knee was performed through the patella tendon while the knee was held at 30 degrees of flexion. Distension of the
Figure7 Photomi~mgraph of the synovium of a rat knee 1 wk follow;ng intra-a~;cutar injection of unsized A/,0, ceramic wear particles. There has been marked accumulation of macrophages associated with the presence of wear particles. (Haematoxylin and eosin, X SO].
Preliminary study A preliminary study of the effects of A1203 particles was performed by injecting a sterile sample of the particle suspension before grading of particle size and separation of the smaller particles to be used in the definite study. Five rats received intra-articular injection of particles in both knees and, after killing them at 1 wk, the tissues were processed in the same manner as in the definitive study. Histological
Figure 2 Photomi~rograph of the synovium of a rat knee 1 wk following intra-a&colar injection of unsized particles of A1,03 The wear particles are located predominantly within macrophages. Note the absence of lymphocytes and necrosis confined to occasional macrophages. (Haemetoxylin and eosin, X400).
Biomaterials
1988, t/of 9 September
443
Synovial macrophage response: 0. W. Howie and 8. Vernon-Roberts
suprapatella pouch was taken to indicate a successful injection. The amount injected was 0.01 ml per 30 g body weight, as this was found in pilot studies to be the amount necessary to distend the knees of rats of various sizes. Animals recovered within 1 h and did not limp. They were housed in metal cages and fed a routine cubed diet. At the time of death, the animals were inspected externally for swelling of the knee or tumours. Under anaesthesia, using aseptic techniques including sterile instruments for each knee, an arthrotomy was performed and a synovial biopsy was taken. The biopsy was placed in Stuarts transport media for microbiological culture. After killing, the whole leg was resected and fixed in 10% formalsaline. The knee was held flat on a glass slide to allow consistent sectioning of the knee for histopathology examination. A post-mortem examination was performed on each animal.
Tissue processing for histopathology After fixation in 10% buffered formal-saline for 3 d, the knees were decalcified using a commercial decalcifying solution (Decal, Omega Chemical Corporation, NY, USA). The extent of decalcification was controlled by daily radiographs (Kodak Min-R film, Hewlett Series 43805N Faxitron Cabinet). After neutralization in 5% silver sodium sulphate, the central portion of the knee was removed by sag&al section either side of the patella. A specimen 4 mm thick was obtained, which was then dehydrated through graded alcohol, cleared in chloroform and embedded in paraffin wax. After trimming, beginning at the side of the arthrotomy, a 6 pm section from the middle of the knee was cut using a hand-operated LKB 2259 Multirange Microtome. Sections were stained with haematoxylin and eosin.
Microbiology
techniques to exclude infection
Aerobic and anaerobic cultures were performed on all rat synovial biopsies and injected solutions of metal, ceramic and dilute serum before and after injection. The biopsies were removed from Stuart’s media and, using a stomacher, were homogenized in 0.5 ml of glucose-cooked meat medium. In an anaerobic chamber, the homogenates and the samples of injected materials were inoculated onto a supplemented blood agar (BAYH) plate, an anaerobic blood culture broth and a glucose-cooked meat broth, and these were incubated anaerobically. The sample was then inoculated on to a chocolate agar plate and incubated in COz. Plate cultures were continued for 4 d and broth cultures for 10 d. A subculture was taken if any broth became turbid. Organisms were identified according to standard laboratory techniques25,26.
Histopathological
Biomaterials
Statistical
methods
The correlation between the particle scores and the macrophage scores for each suspension and the significance of the correlation were assessed by Kendall’s Rank Correlation method modified for tied rank’*. The significance of any differences between the particle and macrophage scores of each suspension, was calculated by the Wilcoxon Rank Sum Test for unpaired data’* . The macrophage score was plotted against the particle score for the suspensions, and the significance of differences between correlation coefficient was assessed by covariant analysis”. Because a number of tests for significance of differences and correlations were performed, thereby increasing the likelihood of finding differences, a significance level of p < 0.01 was accepted.
RESULTS Histology At 1 wk following injection, particles of Co-Cr (Figure 3) and A1203 (Figure 4) were both extracellular and intracellular within macrophages. At 4 and 13 wk, the majority of particles were intracellular. The number of particles seen in all injected knees appeared to decrease between 1 and 4 wk. and then remained constant to 13 wk. The predominant cellular response to both materials was a macrophage infiltrate which appeared greater at 1 wk. had decreased at 4 wk. and remained unchanged at 13wk. There were differences in the response to the Co-Cr and A1203 particles at 1 wk. In the Co-Cr injected knees, prominent zones of necrosis of macrophages were seen at the sites of high particle accumulation (Figure 3). By contrast, in the knees injected with A1203 particles in the preliminary study where the particle size was not controlled (Figures 1 and 2) and in the definitive study of particles of 3 pm and smaller (Figure 4). very few macrophages showed evidence of necrosis and zones of necrosis were absent.
scoring
A semiquantitative scoring method was used based on the presence or absence of a particular feature in each successive high-power field of synovium and subsynovium of the section examined microscopically. The cell and particle scores were based on a modification of that used by Mirra et aI.” to assess the tissue response to wear particles in humans. A positive macrophage score was more than 30 macrophages per high-power field (X400, field area 4.7 mm’) and a positive particle score was more than 20 particles. Quantitation was not attempted for necrosis, fibrosis, presence of other cell types and gradings of macrophages or particles scores, since pilot studies had shown this to be
444
unreliable. The macrophage and particle scores were derived as a percentage of the number of high-powered fields of synovium in which these features were present in each section of the rat knee relative to the total number of fields in the section as a whole.
1988, Vol 9 September
Photomicrograph of the synovium of a rat knee I wk following Figure 3 intra-articular injection of Co-0 particles of 3 pm and smaller. There has been extensive necrosis of macrophages which have accumulated in association with wear particles. Lymphocytes have accumulated in the subsynovium. (Haematoxylin and eosin, X 160).
Synovial macrophage response: D. W. Howie and B. Vernon-Roberts
system detects differences in the macrophage response to different numbers of particles of the same material. The low-dose Co-Cr macrophage scores were not significantly different from the AI,O, macrophage scores at 1, 4 and 13 wk following injection.
Particle
Photomicrograph of the subsynovium of a rat knee 1 wk Figure 4 following injection ofA1203patticlas of 3 pm andsmaller. There is a localized accumulation of macrophages in response to particles which are intracellular and extracellular. Necrosis is absent. (Haematoxylin and eosin, X 160).
and macrophage
correlations
There was a high correlation between the particle score and macrophage score following injection of high-dose Co-Cr suspension (r = 0.92,p < 0.001 ), low-dose Co-Cr suspension (r = 0.97, p Q 0.001) and A1203 suspension (1.= 0.83, p < 0.00 1). To determine the macrophage response to the two different concentrations of particles of Co-Cr, the macrophage scores were plotted against the particle scores at each time interval (Figure 5). The lines of best fit have similar slopes, but the low-dose suspension consistently scores a lower particle and macrophage score, confirming the differences described previously. One Week
Moreover, while a lymphocytic infiltrate was prominent 1 wk after the injection of Co-Cr particles (Figure 3), lymphocytes were not a feature of knees injected with AI,O, at any stage (Figures 1 and 4). At 4 and 13 wk, necrosis was confined to the reaction to Co-Cr particles, and lymphocytes were not observed in the reaction to either particle. There was a difference in the response to high-dose and low-dose Co-Cr particle injections: more particles and macrophages were seen in the knees injected with highdose Co-Cr particles and necrosis was more common. The knees injected with saline control solution showed slightly increased cellularity of the subsynovium at 1 wk and were normal thereafter. No tumours were found in any animals.
Particle
”
I
0
Macrophage
1
I
I
20
30
40
50
%
Four Weeks 60
g
40
P g
30
9 y;ofl 1, 0
,
,
,
,
,
10
20
30
40
50
PartIcks
%
Thirteen Weeks 60
scores
The high-dose Co-Cr macrophage scores were significantly different from the A1203 particle scores at 1 (p < 0.001). 4 (p Q 0.001) and 13 wk (p < 0.001) following injection and suggest a difference in the macrophage response following injection of these particles of different materials. The high-dose Co-Cr macrophage scores were significantly different from the low-dose Co-Cr macrophage scores at 1 (p < O.OOl), 4 (p < 0.001) and 13 wk (p < 0.00 1) following injection, and suggest that the scoring
I
10
PartIcks
scores
The high-dose Co-Cr particle scores were significantly different from the low-dose scores at 1 (p Q 0.001 ), 4 (p < 0.001) and 13 wk (p < 0.001) following injection, demonstrating that the scoring method used detects differences in the number of particles in knees injected with different concentrations of particles of the same material. There was no significant difference between the highdose Co-Cr particle scores and the A1203 particle scores at 1, 4 and 13 wk. There was a significant difference between the low-dose Co-Cr particle scores and the A1203 particle scores at 1 (p < 0.001). 4 (p Q 0.001) and 13 wk (p g 0.001). These results suggest that the numbers of particles of A1203 in the knees were similar to the numbers following injection of high-dose Co-Cr suspension, but were different following the injection of low-dose Co-Cr suspension.
-
i
:$&.g/ t 0
IO
20
30
I
40
50
PartIcks%
The macrophage response at 1,4 and 13 wk following injection Figure 5 of high-dose l and low-dose 0 Co-&particle suspensions. The lines of best fit have similar shapes, indicating that the macrophage response is related to the number of particles, but is not as severe following injection of smaller numbers of particles of the same material.
Biomaterials
1988, Vol 9 September
445
Symviat macrophage response: D. W. Howie and 6. Vernon-Roberts
The macrophage and particle scores for high-dose and low-dose Co-Cr suspensions were combined and compared with the scores for A&OS suspension (Figure 6). One week falling injection, the lines of best fit between the two materials were significantly different (P < 0.01) and suggest that the macrophage response to Co-Cr particles was approximately twice that to Al20s particles. At 4 and 13 wk, there were no signifi~nt differences between the slope of the lines. These results confirm a difference in the response to Co-Cr particles compared with Al206 particles which had been suggested by the finding of significant differences between the macrophage scores for high-dose Co-Cr and Al2Os. Although significant differences between macrophage scores were also detected at 4 and 13 wk, while no differences were detected for particle scores, the method of covariant analysis did not confirm these differences.
Microbiology The cultures from the synovial biopsies of the rat kn88S and the human tissues were negative, except for one rat knee One Week 60-
I 20
t 30
Particles %
Four Weeks
30
40
50
Particles 96
Thkt8en Weeks 60
I
1
20 30 Particles %
,
1
40
50
The macrophage response at 1,4 and 13 wk following injection Figure 6 of Co-0 @ andAi20s A particle suspensions. The slopes of the lines of best fit are significantly different at 1 wk (p < 0.01). but nor at 4 and 13 wk following injection, indicating a more severe macrophage response to similar numbers of Co-0 pefiicles compamd to A120s particles.
446
Biomaterials
1988, Vat 9 September
biopsy which grew Staphylococcus epidermidis in one broth. This growth was assumed to be a contaminant, as the plate cultures and other broth cultures were negative and the histopathol~y of this knee showed no evidence of acute inflammation.
DISCUSSION Intra-articular injection of Co-Cr alloy and Al20s ceramic particles in the same size range as particles seen in the tissues around human prostheses”, ’ ‘* 13,3’s31 induced a macrophage response in the rat knees which correlated with the amount of particulate deposited in the synovium at the three killing times. The macrophage response to Co-Cr particles differed from that to Al206 particles. One week following injection, the response to Co-Cr particles was greater than that to AI20, particles but there was no difference at 4 and 13 wk. These results suggest Co-Cr particles induce a greater inflammatory response shortly after injection. Both these materials are relatively inert in solid form and Richardson et a1.32 found no difference in the response to their intramuscular implantation. /n viffo comparisons have not been made between these two materials in particle form but there is some suggestion that Co-Cr particles are more toxic. Co-Cr particles in vitro interfere with the phagocytic activity of macrophages and are toxic to macrophages and fibroblasts”* 33. Cobalt has been implicated as the most toxic of the constituent metals of this alloy and this toxicity has been related to its greater solubility34. A1203 particles have been reported to mildly inhibit fibroblast growth in vitro, causing slightly less inhibition than stainless steel particles”. In vivo studies suggest that ceramic particles are relatively inert. Griss et al.” and Harms and Maus18’” concluded that the macrophage response to subcutaneous, intramuscular and intra-a~icular injection of particles of a similar size to those used in the current study was not of concern. These conclusions were based on subjective assessment of the tissue response and did not include a control material nor adequate techniques to exclude infection. Some worrying features were reported in these studies. A neutrophil response which peaked at 1 wk following subcutaneous injection’* and a persisting lymphocyte response ” following intramuscular injection may be interpreted as evidence of mild to moderate toxicity35P3”. These features were not observed in our study. Uchida3’ observed that the intra-articular injection of Al206 ceramic produced a more severe tissue response than polymer and metal particles, including Co-Cr alloy. The results of the current study suggest a difference in the degree of the initial macrophage response to the two types of particle, but no difference in the degree of the persisting macrophage response.‘rhus, it is possible that the continuing macrophage response depends on the mechanical presence of the particles rather than any continuing toxic properties which may or may not persist. The scoring system used in this study bears further discussion. Despite attempts to inject particles of a similar size and concentration, and in approximately similar amounts, it is possible that there were large variations in the number of particles exposed to the synovia of different knees. For this reason, a scoring system was used which related the macrophage response to the number of particles in the synovium. In addition, the whole of the knee was examined to take into account the possible variations in the number of particles
Synovial macrophage response: D. W. Howie and B. Vernon-Roberts
and the macrophage response at different sites. If the required minimum number of macrophages or particles was present in any given high-power field, then a positive score was recorded. This system has limitations, as there may be far larger numbers of either macrophages or particles present in two different areas but the score will be the same once the threshold number is reached to achieve a positive score. To assess the value of this scoring system in detecting differences in particle numbers and macrophages, two different concentrations of Co-Cr particles were used. The scoring system seems justified as the macrophage response correlated with the particle score in all groups and, more importantly, the scoring system detected a difference in the particle and macrophage score following injection of highdose and low-dose Co-Cr suspensions. The macrophage response alone was used to assess the tissue response to particles, as it could be reliably scored. Other features, such as the extent of lymphocyte and neutrophil infiltrate and the degree of necrosis and fibrosis have been used by others to assess the biocompatibility of particles following intra-articular injection. Some of these features were present in the knees in this study but could not be reproducibly scored. In conclusion, the initial macrophage response to CoCr particles was different from that to Al,O, particles. After this initial difference, the macrophage response was related to the number of particles of each material but not to the type of material. It is not known whether the difference in the macrophage response to the different wear particles is important in determining the deleterious effects of wear particles around human prostheses. A material which produces wear particles that provoke less macrophage response in the surrounding tissues would seem desirable but a difference in the macrophage response between the two materials in this study was only detectable at 1 wk following injection. The adverse effect of macrophages in the periprosthetic tissues and their possible role in bone resorption and prosthetic loosening takes many years to become evident. As particles are continuously released into the periprosthetic tissues and possible toxic effects of the particles may have a cumulative effect, differences in the effect of particles seen at 1 wk may be important. Also, in this study, the persisting macrophage response at 4 wk and 3 months was related to the number of particles in the synovium. It is likely, therefore that the number of particles that persist in the tissues is also an important determinant of long-term adverse effects of prostheses wear particles.
ACKNOWLEDGEMENTS The authors gratefully acknowledge the assistance of Dr R. Garrett and Miss S. Hay (Department of Pathology, University of Adelaide); the staff of the Animal Operating Theatre, Medical Illustration and the Bone Laboratory, Institute of Medical and Veterinary Science; the staff of the Photography Unitsof the Department of Pathology, University of Adelaide, and of Flinders Medical Centre and Mrs R. Kenney in preparation of the manuscript. This work was supported in part by grants from the Daws Research Fellowship, Royal Adelaide Hospital, the Australian Orthopaedic Association, the Royal Australasian College of Surgeons, and the National Health and Medical Research Foundation.
REFERENCES 1
2
3
4
5
6
7
8
9
10 11
12 13
14 15
16
17
18
19
20
21
22
23
Griss, P., von Andrian-Werburg, H., Krempien. B. and Heimke. G.. Biological activity and histocompatibilityof denseA120s/Mg0ceramic Implants in rats, J. Biomed. Mater. Res 1973, 7.453-462 Griss, P.. Greenspan, DC., Heimke, G., Krempien, B., Bucherger, R., Hensch, L.L. and Jentschura, G., Evaluation of a bioglasscoated AI,Os total hip prosthesis in sheep, .I. Biomed. Mater. Res. 1976, 10, 511-518 Heimke. G., Griss, P., Jentschura, G. and Werner, E., Direct anchorage of A1203-ceramic hip components: three years of clinical experience and results of further animal studies, J. Biomed. Mater. Res. 1979, 13. 57-65 Busing, C.M., d’Hoedt, B., Schulte, W. and Heimke, G., Morphological demonstration of direct deposition of boneon human aluminiumoxide ceramic dental implants, Biomarerials 1983, 4(2), 125-l 27 Dorre. E. and Dawihl, W., Ceramic hip endoprostheses, in Mechanical Properties of Biomarerials (Eds G.W. Hastings and D.F. Williams), John Wiley Et Sons, 1980, pp 113-l 27 Dowson, D. and Linnett, I.W., A study of the wear of ultra molecular weight polyethylene against a high alumina ceramic, in Mechanical Properties of Biomaterials (Eds G.W. Hastings and D.F. Williams), John Wiley 8 Sons, 1980 3-25 Boutin, P.. Arthroplasttc totale de la hanche par prothese en alumine frittee: etude experimentale et premieres applications cliniques, Rev. Chir. Orthop. 1972,58, 229-246 Semlitsch. M.. Lehmann, M., Weber, H., Oderre. E. and Willert, H.G., New prospects for a prolonged functional life-span of artificial hip joints by using the material combination polyethylene/aluminium oxide ceramic/metal, J. Biomed. Mater. Res. 1977. 11, 537-552 Boutin. P. and Blanquaert, D., A study of the mechanical properties of alumina-on-alumina total hip prostheses, Rev. Chir. Orfhop. 198 1,67, 279-287 Harms, J. and Mausle, E.,Tissue reaction toceramic implant material, J. Eiomed. Marer. Res. 1979. 13. 67-87 Stock, D., Diezemann, ED. and Gottstein, J., Results of endoprosthetic hip joint replacement with the aluminium ceramic-metal composite prosthesis ‘Lindenhof’, Arch. Orthop. Trauma Surg. 1980, 97, 712 Zweymuller, K., Bone and joint replacements with bioceramic endoprostheses, Fortschr. Med. 1979, 97(45), 2107-211 1 Griss. P. and Heimke, G., Five years experience with ceramic-metalcomposite hip endoprostheses. I. Clinical evaluation, Acra Orthop. Trauma Sorg. 198 1, 98, 157-l 64 Rampoldi, A., Mittelmeier’s ceramic hip prosthesis, &a/. J. Orfhop. Traumatol. 1984, 10, 305-3 1 1 Trepte, L.T., Gauer, E.F. and Gartner, B.M., Results with ceramic-onceramic sliding endoprosthesis completely or partially fixed without cement, 2 Olthop. 1985, 123, 239-244 O’Leary, J.F.M. and Mallory, T.H., Mittelmier ceramic total hip anhroplasty: two year follow-up, Anna/ Meeting of rhe American Academy of Orthopaedic Surgeons, New Orleans, Feb. 1986 Hinterberger, J. Ungethum, M. and Plitz, W.,Tribological propertiesof aluminium oxide ceramics, in Mechanical Properries of Biomarenkls (Eds G.W. Hastings and D.F. Williams), John Wiley Et Sons, 1980, pp 73-82 Griss. P.. Krempien, B., von Andrian-Werburg, H.F., Heimke, G., Fleiner, R. and Diehm, T., Experimental analysis of ceramic tissue interactions: a morphologic fluoroescenteoptic and radiographic study on dense alumina oxide ceramic in various animals, .I. Biomed. Mater. Res. 1974, 8. 39-48 Plenk, H., Evaluation of the effect of ceramic and different metallic Implant materials on the growth rate of human fibroblast cultures, in Evaluation ofBiomateria/s (Eds G.D. Winter, J.L. Leray and K. de Groot), John Wtley Et Sons, 1980, pp 399-403 Garrett, R.. Wilksch, J. and Vernon-Roberts, B.. Effects of cobaltchrome alloy wear particles on the morphology, viability and phagocytic activrty of murine macrophages in virro, Ausr. J. Exp. Biol. Med. Sci. 1983,61,355-369 ASTM. F36 l-80: Standard procedure for assessment of compatibility of metallic materials for surgical implants with respect to effects of materials on tissues. F86-76: Standard recommended practice for surface preparation and marking of metallic surgical implants. F60378: Standard specification for high-purity dense aluminium oxide for surgrcal implant application. Annual Book of ASTM Standards, Part 46. American Society for Testing and Materials, Philadelphia, 1980 Gatehouse. B.M. and Willis, J.B., Performance of a simple atomic absorption spectrophotometer, Spectrochimica Acra 196 1, 17, 710-718 McPherson, E.L.. Price, J.W. and Scaite. P.H.. Application of atomic
Biomarerials
1988, Vol9 September
447
Synovial macrophage response: 0. W. Howie and 8 Vernon-Roberts
24
25 26 27
26 29
30 31
448
absorption spsctroscopy to the determination of cobalt in steel, alloy steel and nickel, Nature 1963, 399, 371-372 ASTM, D3558-77: Standard test methods for cobalt in water. D1687-77: Standard test methods for chromium in water. Annual Book ofASTM Standards, Part 3 1, American Society for Testing and Materials, Philadelphia, 1977 Cowan, S.T., Manual for the Identihcation of Medical Bacteria (Ed. S.T. Cowan), 2nd Edn,Cambridge University Press, Cambridge, 1975 Holdeman, L.V., Cato, E.P. and Moore, W.E.C., Anaerobic Laboratory Manual, 4th Edn, Virginia Polytechnic Institute, USA, 1977 Mirra. J.M., Amstutz, H.C., Matos, M. and Gold, R., The pathology of the joint tissues and its clinical relevance in prosthetic failure, C/in. Orfhop. 1976,117,221-240 Swinscow, T.D.V., Statistics a? Sguare One, 3rd Edn, British Medical Association, London, UK, 1978 Snedecor, G.W. and Cochrane, W.G., Method of co-variant analysis, in StatisticalMethods, 6th Edn, Iowa State University press, 1967, pp 41 Q-436 Winter, G.D., Tissue reactions to metallic wear and corrosion products in human patients, J. Biomed. Mater. Res. 1974,8, 1 l-26 Vernon-Roberts, 8. and Freeman, M.A.R., The tissue response to total
Biomatenals
1988, Vol9
September
32
33
34
35
36
37
joint replacement prostheses, in The Scientific Basis of Joint Replacement (Eds S.A.U. Swanson and M.A.R. Freeman), Pitman Medical Publishing Co. Ltd., Tunbridge Wells, UK, 1977, pp 86-l 29 Richardson, W.C., Klawitter, J.J., Sauer, 8.W.. Pruitt, J.R. and Hulbert, SF., Soft tissue response to four dense ceramic materials and two clinically used biomaterials,./. Biomed. Mater. Res. 1975,9,73-80 Rae, T., A study on the effects of particulate metals of orthopaedic interest on murine macrophages in vitro, J. Bone Joint Surg. 1975, 570.444-450 Rae, T., The toxicity of metals used in orthopaedic prostheses: an experimental study using cultured human synovial fibroblasts, ./. Bone Joint Surg. 198 1,638,435-440 ANSI/ADA, Recommended standard practices for biological evaluation of dental materials, American National Standards/American Dental Association Document 1979.41 FDI, Recommended standard practices for biological evaluation of dental materials, Federation Oentaire International Document 1979. 198 (International Dental Federation, London) Uchida, S., A histopathologic comparison of the tissue reaction to prosthetic materials in the knee joint of rats, Onhopaedics 1985, 8, 1276-l 280