Injury, 6, 277-285
Incorporation and VitaIlium
277
of stainless in bone
steel, titanium
L. Linder and J. Lundskog The Laboratory of Experimental Biology, Department University of Giiteborg, Sweden Summary
Operative trauma, implant instability and corrosion are recognized as aetiological factors in the production of the soft-tissue capsule often present round a metal implant in bone. In the present experimental study, the isolated effect of corrosion was evaluatedunder standardized experimental conditions where operative trauma was minimal and implant movement was negligible. The reaction of bone to stainless steel, titanium and Vitallium was studied by histological and biomechanical methods, and compared to that provoked by copper. No soft-tissue capsule was found round the modern materials, but copper gave rise to a marked tissue reaction. It is concluded that atraumatic operative techniques and implant immobility are more effective for stable incorporation of metallic implants in bone than the nature of the implant itself, provided that one of the modern, low-corrosive metals or alloys is used. MODERN metallic implant materials (AN 316 stainless steel, titanium and Vitalhum) are all considered to be well tolerated by living tissues. Implanted in muscle, a fibrous capsule will always be found to surround the implant, regardless of the material used (Laing et al., 1967). Implanted in bone, a similar connective tissue layer of varying thickness is often found interposed between the bone and the implant surface (Hickman et al., 1958; Hicks, 19.58; Cohen, 1961; Emneus and Stenram, 1965; Homsy et al., 1972). However, a true bone-metal interface has also been reported (Hickman et al., 1958; Hicks, 1958; Wagner, 1963; B&remark et al., 1969; Charnley, 1970; Lundskog, 1972). Since a soft-tissue layer round an implant in bone reduces the stability of the implant with possible implant loosening as a result, it is of great practical importance to elucidate the aetiology of such a layer.
of Anatomy,
Traditionally, the connective tissue reaction has been attributed to implant corrosion, since corrosion products are frequently found in the tissue surrounding the implant (Venable et al., 1937; Wagner, 1963; Emneus and Stenram, 1965; Laing et al., 1967). Sometimes, however,
2.5
2.5
Fig. 1.-Schematic picture of the implants used in the study. All measurements are in mm. The narrow segment is introduced into the hole drilled in the bone. The wider segment stabilizes the implant during the first weeks after implantation, and is used as the connection point in the mechanical test. even macroscopically visible tissue contamination from implants appears to cause little or no reaction (B&remark et al., 1969). Other factors may be responsible for the marked tissue reaction around implants. Lundskog (1972), working with titanium, has shown
Injury: the British Journal of Accident Surgery Vol. ~/NO. 4
278 Tab/e /.-Composition W ATi 24 ATi 45 AISI 316 Vitallium
14-16
of the low-corrosive CO
bal.
The figures were obtained
Fig. 2.-Surface x
materials tested in the study
Fe
0
N
0.05 0.05 bal. 3.0 max.
0.10 0.35 -
0.03 0.07 -
from the manufacturers
c 0.05 0.10 0.05 max. 0.15 max.
Ii 0.012 0.012 -
Pd -
Ti bal. bal. -
Cr 17.4 19-21
and are given in percentages.
topography of implants of 4 materials tested, as seen by scanning electron microscopy 300. A, Stainless steel; B, Titanium ATi 45; C, Titanium ATi 24; D, Vitallium.
at
279
Linder and Lundskog: Stainless Steel, Titanium and Vitallium Tab/e 1. (continued) Ni
MO
Mn
Si
-
-
-
12.5 9-l 1
2.7 -
0.6 1 .O max.
1 .7 2.0 max.
The figures for titanium are maximal values.
that excessive operative trauma at the time of implant insertion causes soft-tissue formation round the implant and Uhthoff (1973) and Cameron et al. (1973) have shown that movement of the implant during the first weeks after implantation is still another cause of soft-tissue formation. It thus appears to be of interest to establish the importance of corrosion as compared to operative trauma and implant movement as an aetiological factor in soft-tissue reaction in bone. In the experimental model used by Lundskog (1972) both operative trauma and implant movement could be controlled to such a degree that no soft-tissue reaction occurred. For this reason it was decided to use this experimental model for comparative studies of the tissue reaction to various implants made of stainless steel, titanium and Vitallium. Copper, a metal well known for causing tissue reaction, was included in the study as a control material. MATERIALS AND METHODS Implants of stainless steel (AISI 316), titanium (ATi 45 and ATi 24, Avesta Jernverks AB, Sweden), wrought Vitallium, and copper were used. The composition of the low-corrosive materials used is given in Table Z, and the shape of the implants is demonstrated in Fig. 1. In earlier investigations the importance of the microstructure of the implant surface has been evaluated (Lundskog, 1972) and for this Tab/e
//.-Distribution
of the animals
Group 1 (VitalliumTitanium ATi 24) Number of animals Observation time
reason care was taken to obtain the same surface finish on all implants. However, because of the different mechanical properties of the implant materials, the requirement of an identical surface finish could not be met (Fig. 2), even if the surface topography appeared the same to the naked eye. All implants except those of Vitallium were machined with steel tools. The Vitallium implants were ground by a grinding plate, type 57A 46 18 BBE. The steel implants were not treated electrolytically. After fabrication, all implants were cleaned in alcohol in ultra sound before being autoclaved. After the experiments all implants except those of copper were submitted to surface analysis, performed with a scanning electron microscope (Cambridge JSM-UB). Micrographs, x 300 and x 3000, were obtained. Twenty-one rabbits of both sexes, 6 months old and weighing approximately 3.5 kg, were used in the investigation. The animals were divided into 4 groups in which the tissue reaction provoked by stainless steel, titanium ATi 45, Vitallium and copper respectively was analysed in relation to that of titanium ATi 24. Titanium ATi 24 was implanted in all animals, because of personal experience with this material, and thus each animal served as its own control. The distribution of the animals and implant materials into groups as well as the observation time for each group are shown in Table ZZ. For implantation, the method described by Lundskog (1972) was used. Under general barbiturate anaesthesia and under strictly aseptic conditions, a skin flap, based anteriorly, was raised over the proximal medial side of the tibia exposing the deep fascia. After careful dissection and diathermy, a flap of the fascia, based posteriorly, was raised and the periosteum exposed. Care was taken to avoid bleeding. With a sharp punch, a small piece of periosteum was removed and a guide hole drilled in the exposed cortical
6 7-7.5
Group 2 Stainless steel AISI 316Titanium ATi 24)
6 months
7 months
Group 3 (Titanium ATi 45Titanium ATi 24)
5 1 O-l 0.5 months
Group 4 (CopperTitanium ATi 24)
4 4 months (1) 3 weeks (3)
The groups are characterized by different test materials, titanium ATi 24 serving as the control material throughout. One implant of the test material and one of the control material was inserted in each animal, each tibia receiving one implant.
Injury:
280
the British Journal
of Accident
Surgery
Vol. G/No. 4
in lateral projections taken after sacrifice. A, Stainless steel implant. Arrows indicate the region where control measurements demonstrated a ‘ neck ’ of the imolant. B. Titanium ATi 24 imolant. Both periosteal and end&teal bone proliferation is C, VitaIlium implant; D, Copper implant (4 months).
Fig. 3.-Radiographs
obvious
but the implant
was completely
loose.
bone. Round this hole, a circular area of periosteum, of 4.0 mm diameter, was removed by another punch with a central guide peg. Finally, with a flat drill, guided by a central peg as well, a circular cortical defect, of 3.6 mm diameter, was made centrally in the periosteal defect. The drill used in this procedure was made of the same material as the implant later to be inserted in the hole. The bone was continuously irrigated with saline during drilling, which was done at 2000 rpm. One implant was then inserted in the cortical defect in each tibia. The implant was secured by a suture through its horizontal hole. The wound was closed by interrupted silk sutures, sprayed with a plastic wound dressing (Nobecutane) and covered with surgical tape. One week and two days before sacrifice, each animal received oxytetracycline (Terramycin) intraperitoneally in a dose of 10 mg per kg. Before sacrifice, the animals, under urethane anaesthesia, were submitted to microangiography. Five hundred ml of a 1 : 1 mixture of Pelican ink (C 1 l/l431 a) in physiological saline were
infused into the abdominal aorta. The animals of the copper-titanium group did not receive tetracycline injections and were not submitted to the microangiographic procedure because the toxic effect of copper had been obvious from general inspection and conventional histology in earlier pilot studies. After sacrifice and microangiography the implantation area was inspected and radiographed. The bone specimens with the implants in situ were then submitted to a mechanical test, described in detail by Lundskog (1972). In this procedure the implants are extracted axially and the extraction force is plotted against the displacement of th.e implant. From each curve thus obtained, the maximal extraction force can be determined. For histological evaluation, each tibia was divided into two longitudinal halves through the middle of the implant site. One of the halves was fixed in 5 per cent formaldehyde, decalcified in 30 per cent formic acid, and embedded in paraffin. Sections 4 I_Iand 100 I_Ithick were cut,
Linder and Lundskog
: Stainless
Steel, Titanium and Vitallium
the former being stained with haematoxylin and eosin for routine histology, the latter cleared by the Spalteholz method for microangiographic studies. The other half of the tibia, after freezedrying, was embedded in methyl methacrylate in the undecalcified state. Sections, 300-500 p in thickness, were obtained by saw-cutting. Tetracycline labelling was studied with the aid of a Leitz Orthoplane microscope in reflected light by the Ploem method. f
Fig. 4.-Diagrammatic
representation of the results The mean value and the extreme values of the maximal extraction forces of the materials in each group are given. Note, however, that the high extraction forces for some stainless steel implants can be explained by the bottle-neck shape of these implants. of the mechanical
test.
RESULTS General inspection All skin incisions healed by first intention and were barely visible. No signs of infection or discoloration of the tissues surrounding the implants were seen. All implants, except those of copper, were firmly incorporated in the bone and tested with a pair of tweezers. After extraction, the endosteal ends of the implants were found to have been invested by a thin bony capsule. No soft tissue was seen to adhere to the implants and the implants had the same surface appearance as when they were inserted. The copper implants were unstable, surrounded by fibrous tissue and necrotic material, some of which usually adhered to the implants. At 16 weeks, periosteal bone formation was evident round the copper implant, which was buried in new bone. Radiography The results of the general inspection were confirmed. The cortical bone had normal thickness
281
round all implants except the copper ones. A thin bony lamella round the part of the implant protruding into the medullary cavity as well as a small bony collar round the implant on the
Fig. S.-Histological section from the bone bordering a titanium ATi 24 implant. The bone has a mature appearance and the vascular channels are directed parallel to the implant surface. At the arrow, a small vessel is seen in the interface zone (H. and E. x 150.)
periosteal side were evident (Fig. 3A, B and C). After 3 weeks, the bone round the copper implants showed signs of periosteal thickening, and at 16 weeks both periosteal and endosteal thickening (Fig. 3D). Mechanical test The mean value of the maximal extraction forces for the steel implants was 7.6 kp, and for the corresponding titanium implants the mean value was 3.6 kp. The titanium ATi 45 implants had a mean of 7.2 kp and their titanium ATi 24 controls 9.7 kp. The figure for Vitallium was 7.2 kp and for the titanium ATi 24 controls of this group 5.8 kp. The mean for copper was 0.7 kp and for the corresponding titanium ATi 24 implants the mean value was 1.8 kp. The results are given diagrammatically in Fig. 4, where the mean values for each group as well as the extreme values can be seen. The discrepancy in extraction forces in the steel-titanium group initiated control measurements of all implants. These showed that 4 out
282
Injury: the British Journal of Accident Surgery Vol. ~/NO. 4
of 6 steel implants were slightly bottle-shaped, as seen in Fig. 3A. The thinner segment corresponded to 25 per cent of the implant length, and the difference in diameters between the ‘ normal ’ and the narrow part was 1 per cent. The defective steel implants had extraction forces of 9.2, 9.7, 9.3 and 10.0 kp, whereas the truly cylindrical steel implants both had extraction forces of 4.0 kp. The difference in the maximal extraction forces of the bottle-shaped and cylindrical steel
Fig. 6.-Histological appearance of the bone beside a copper implant, inserted 4 months earlier. Cancellous bone proliferation in the medullary cavity is present. Close to the implant is connective tissue, infiltrated with an abundance of round cells. The original cortical bone has assumed a cancellous appearance (H. and E. x 9).
Fig. 7.-Detail of Fig. 6 at higher magnification. Necrotic tissue is seen in the interface zone. Arrow indicates areas with necrotic bone (H. and E. x 82).
implants thus appears to be the result of the geometrical difference between the implants-an assumption which seems reasonable on the basis of mechanical calculation.
close to the implant surface. Numerous vascular channels were observed at the interface between implant and bone. The perivascular connective tissue did not spread over the surface of the bone to form a capsule, but was strictly confined to the vessel. The bone marrow had a normal appearance. The copper implants, on the other hand, produced a distinctly different picture, regardless of the observation time (Figs. 6 and 7). The cortex was thickened by new bone formation both on the periosteal and endosteal side and had a cancellous appearance. No callus had been laid down close to the implant surface. Empty osteocyte lacunae were commonly seen in this region. Connective tissue with round-cell infiltration was found to surround the implants and round-cell infiltration was also seen in the bone marrow.
Histological analysis The bone round the implants of stainless steel, titanium ATi 45 and 24, and Vitallium, regardless of the observation time, showed the same histological picture (Fig. 5). Lamellar bone was seen bordering the implant, the lamellae being mostly arranged parallel to the implant surface. Further away from the implant there was evidence of osteons running in a spiral or circular fashion round the implant site. Usually, no borderline could be detected between the original cortical bone and the newly formed bone. No continuous connective-tissue layer was seen, either along the sides of the implant, or along the bony floor, deep to the implant. Such a bony floor was seen in most sections and had the same appearance for all materials, although it was most prominent in the titanium ATi 455ATi 24 group. In the bone apparently normal osteocytes were seen, even
Microangiography Due to incomplete filling of the vessels, some specimens could not be examined. However, from the remaining angiograms it was evident that the cortical vessels had regenerated and formed
Linder and Lundskog : Stainless Steel, Titanium and Vitailium
283
surface topography. There was no evidence of corrosion on any of the implants. The steel implants generally had a smoother surface than their titanium controls and the same was true for the titanium ATi 45 implants. The Vitallium and the corresponding titanium implants were of the same roughness (Fig. 2). The surface analysis, however, also revealed differences between implants of the same material, especially implants of titanium ATi 45 and ATi 24.
Fig. 8.-Microangiogram from the bone bordering a steel implant, demonstrating the vascular arrangement in the endosteal bony capsule. (x 68.)
a network round the intracortical part of the implant. The bony capsule surrounding the endosteal aspect of the implant appeared to be supplied both by intracortical and endosteal vessels (Fig. 8). In no case were tom vessels seen in the interface zone, which might have indicated the presence of connective tissue, torn out in the mechanical test. Nor could tortuous vessels, suggestive of granulation-tissue formation, be seen. The vascular architecture appeared to be the same for steel, titanium ATi 45 and 24, and Vitallium. Fluorescence microscopy Tetracycline labelling was seen in all specimens and no difference in distribution or quantity for the different implant materials was detected. The labelling was usually seen round the larger vascular channels, in some instances those directly bordering the implant suface. Often, a small lamella of bone between the implant and the vessel was labelled, suggesting that new bone had been laid down round a vascular channel formerly in contact with the implant. Labelling was also seen at some distance from the implant, e.g. at the anterior tibia1 margin. Scanning electron microscopy All implants except those of copper were examined for signs of corrosion and differences in
DISCUSSION The present study shows that implants of AISI 316 stainless steel, pure titanium (ATi 45 and ATi 24) and wrought Vitallium can be incorporated in live rabbit bone, provided they are inserted with minimal trauma and immobilized during the healing period. This is in contrast to implants made of copper, which, with the same technique, was not accepted by the bone. For all implants, except those of copper, stable incorporation was accomplished in 3 weeks, after which the stability of the implants was consolidated, as shown by the higher extraction forces required in the groups observed for a longer time. This finding agrees with the results obtained by Cameron et al. (1972) and Lundskog (1972). Within each group, however, relatively large variations in maximal extraction forces were found for implants of the same material, and there were also variations between the groups. From earlier investigations (Lundskog, 1972) it is known that variations in the microstructure of the implant surface have a marked influence on the maximal extraction forces. Thus it appears likely that the differences in surface topography, combined with the biological variation of the animal material, explain the variation in the extraction forces. However, the titanium ATi 24 standard was always stable in the bone, and, if the groups are considered separately, there was no evidence that any one of the test materials was less stable than titanium ATi 24. In view of the fact that the implants were neither porous nor threaded, remarkably high extraction forces were recorded. This suggests close apposition of mineralized ground substance to the implants, and in fact no connectivetissue capsule was found round any of the implants except those of copper. Thus, since no adverse tissue reaction to any of the modern implant materials was seen, at least on light microscopy, it is concluded that bone, under conditions, treats these implant optimal materials in the same way.
284
Lundskog (1972), Cameron et al. (1973) and Uhthoff (1973) have pointed out the importance of implant immobility for the successful incorporation of an implant in bone. According to these authors, mechanical loading without implant movement is compatible with incorporation, whereas loading with implant movement gives rise to a soft-tissue capsule round the implant, and the risk of the implant loosening. As the experimental model used in this study prevents implant movement, the results support the theory presented. However, it has also been shown (Lundskog, 1972) that, even in the absence of implant movement, soft tissue may form if thermal injury to the bone is inflicted at the operation. Using the same model as in the present study, Lundskog was able to show that temperatures exceeding 75°C for 30 seconds always inh.ibited implant incorporation, despite the fact that the implants were immobile. Temperatures of this order may very well be reached in clinical practice, for example during drilling, if special precautions are not taken (Lundskog and Nilsson, 1969; Matthews and Hirsch, 1972). One of these is saline irrigation, which lowers the temperature to acceptable values (Lundskog and Nilsson, 1969). Subliminal thermal trauma at the operation and immobility of the implant thus appear to be prerequisites for implant incorporation in bone, while the implant material itself seems to be of less importance provided it is one of the modern, low-corrosive metals or alloys. Copper produced marked bone reaction but this material is known to be very toxic (Venable et al., 1937; Nicole, 1947), possibly via corrosion products (Pappas and Cohen, 1968), and the observed reaction was expected. This proves that chemical trauma may cause tissue reaction, even if the implant is inserted under optimal conditions. As it has been shown that all metallic implants corrode in living tissues (Laing et al., 1967) even though the defects in the metal may be invisible (Hicks and Cater, 1962), it appears likely that corrosion of low intensity was also present with The the tested modern implant materials. corrosion rate, however, was obviously insufficient to cause tissue reaction by itself. A doseresponse relationship appears to exist for this type of trauma, and the present study indicates that the tissue reaction threshold was never reached except in the copper-titanium group. The results strongly support the suggestion by is the true Hicks (1958) that ‘ zero reaction response to the chemically inert foreign body and that the fibrotic reaction is the response to
Injury
: the
British Journal
of Accident
Surgery
Vol. ~/NO. 4
varying degrees of microcorrosion.’ Moreover, the study shows that the mere physical presence of the implants did not cause adverse tissue This finding is interesting, because reaction. th.ere have been reports stating that bone always treats metallic implants as ‘ foreign bodies ’ by connective-tissue encapsulation (Collins, 1954). The modern implantation techniques used in clinical practice usually rely on composite devices which may increase the corrosion rate (Bechtol et al., 1959; Scales et al., 1959; Emneus and Stenram, 1965 ; Colangelo and Greene, 1969). Another factor of importance is mechanical stress which may result in stress corrosion or fretting corrosion. Whether an increased corrosion rate caused by any of these factors can induce tissue reaction is still open to debate, and it seems that this problem can be solved only by further standardized experiments. Even though the conclusions of the present investigation may be somewhat limited by the relatively short observation time, the study indicates that gentle handling of the tissues at the operation and mechanical stability of the implant during the period of bone healing are more important for stable incorporation in bone than is the implant material, as long as it is one of the modern, low-corrosive metals or alloys. Thus, it appears likely that many of the tissue reactions seen round metallic implants in bone may have been falsely attributed to corrosion alone.
REFERENCES
BECHTOL, C. O., FERGUSON, A. B. jun. and LAING, I’. G. (1959), Metals and Engineering in Bone and Joint Surgery (1st ed.) Baltimore, Williams &
Wilkins Company. p. 40. BR~~NEMARK, P.-l., BREINE, U., ADELL, R., HANSSON, B. O., LINDSTRBM,J. and OHLSSON,& (1969), ‘ Intraosseous anchorage of dental prostheses. I. Experimental studies, Stand. J. Plast. Reconstr. Surg.,3,8 1. CAMERON, H., MACNAB, 1. and PILLIAR, R. (1972), ‘ Porous surfaced Vitallium staples ‘, S. Afv. J. Surg., 10, 63. CAMERON, H., PILLIAR, R. and MACNAB, 1. (1973), ‘ The effect of movement on the bonding of porous metal to bone ‘, J. Biomed. Mater. Res., 7, 301. CHARNLEY, J. (1970), ‘ The fixation of prostheses in living bone ‘, In: Modern Trends in Biomechanics (ed. Simpson. D. C.), London, Butterworths. ‘ Tissue reactions to metals-the COHEN, 5.-(19&l), influence of surface finish ‘, J. Bone Jt Surg., 43A, 687. COLANGELO,V. J. and GREENE, N. D. (1969), ‘ Corrosion and fracture of type 316 SMO orthopedic implants ‘, J. Biomed. Mater. Res., 3, 247.
Linder and Lundskog
: Stainless
Steel, Titanium and Vitallium
D. H. (1954), ’ Tissue changes in human femurs containing plastic appliances ‘, J. Bone Jt Surg., 36B, 458. EMNBUS,H. and STENRAM,U. (1965), ‘ Metal implants in the human body. A histopathological study ‘,
COLLINS,
Acta Orthop.
Stand.,
36,
115.
HICKMAN,J., CLARKE,E. G. C. and JENNINGS,A. R. (1958), ‘ Structural changes in bone associated with metallic implants ‘, J. Bone Jt Surg., 40B, 799. HICKS, J. H. (1958), ‘ Pathological effects from surgical metal ‘, In : Modern Trends in Surgical Materials (ed. Gillis, L.), London, Butterworths. HICKS, J. H. and CATER, W. H. (1962), ‘ Minor reactions, due to modern metal ‘, J. Bone Jt Surg. 44B, 122. HOMSY, C. A., STANLEY,R. F., ANDERSON,M. S. and KING, J. W. (1972), ‘ Reduction of tissue and bone adhesion to cobalt alloy fixation appliances ‘, J. Biomed.
Mater.
Res., 6, 451.
LAING, P. G., FERGUSON,A. B. and HODGE, E. S. (1967), ‘ Tissue reaction in rabbit muscle exposed to metallic implants ‘, Ibid., 1, 135. LUNDSKOG,J. (1972), ‘ Heat and bone tissue ‘, Stand. J. Plast. Reconstr.
Surg.,
Suppl.
9.
LUNDSKOG,J. and NILSSON, K. (1969), Unpublished data. MATTHEWS,L. S. and HIRSCH, C. (1972), ‘ TemperaRequesrs G(iteborg,
for reprints Sweden.
should
be nddressed
to:-Lars
Linder,
MD,
285
tures measured in human cortical bone when drilling ‘, J. Bone Jt Surg., 54A, 297. NICOLE, R. ‘ Metallschadigung bei (1947), Osteosynthesen. Experimentelle und klinische Untersuchungen tiber Wesen und Bedeutung der Metallose und der Korrosion ‘, He/v. C&r. Acta, Suppl. III. PAPPAS, A. M. and COHEN, J. (1968), ‘ Toxicity of metal particles in tissue culture. Part I, A new assay method using cell counts in the phase of replication ‘, J. Bone Jt Surg., 5OA, 535. SCALES,J. T., WINTER, G. D. and SHIRLEY,H. T. of orthopaedic implants. (1959), ’ Corrosion Screws, plates and femoral nail-plates ‘, Ibid., 41B, 810.
UHTHOFF, H. K. (1973), ‘ Mechanical factors influencing the holding power of screws in compact bone ‘, Ibid., 55B, 633. VENABLE,C. S., STUCK, W. G. and BEACH,A. (1937), ‘ The effects on bone of the presence of metals; based upon electrolysis. An experimental study ‘, Ann. Surg., 105, 917. WAGNER, H. (1963), ‘Die Einbettung von Metallschrauben im Knochen und die Heilungsvorgange des Knochengewebes unter dem Einfluss der stabilen Osteosynthese ‘, Langenbecks Arch. Klin. Chir., 305, 28. Department
of Anatomy,
University
of
Giiteborg
Fack,
S-400
33
Incorporation and VitaIlium
277
of stainless in bone
steel, titanium
L. Linder and J. Lundskog The Laboratory of Experimental Biology, Department University of Giiteborg, Sweden Summary
Operative trauma, implant instability and corrosion are recognized as aetiological factors in the production of the soft-tissue capsule often present round a metal implant in bone. In the present experimental study, the isolated effect of corrosion was evaluatedunder standardized experimental conditions where operative trauma was minimal and implant movement was negligible. The reaction of bone to stainless steel, titanium and Vitallium was studied by histological and biomechanical methods, and compared to that provoked by copper. No soft-tissue capsule was found round the modern materials, but copper gave rise to a marked tissue reaction. It is concluded that atraumatic operative techniques and implant immobility are more effective for stable incorporation of metallic implants in bone than the nature of the implant itself, provided that one of the modern, low-corrosive metals or alloys is used. MODERN metallic implant materials (AN 316 stainless steel, titanium and Vitalhum) are all considered to be well tolerated by living tissues. Implanted in muscle, a fibrous capsule will always be found to surround the implant, regardless of the material used (Laing et al., 1967). Implanted in bone, a similar connective tissue layer of varying thickness is often found interposed between the bone and the implant surface (Hickman et al., 1958; Hicks, 19.58; Cohen, 1961; Emneus and Stenram, 1965; Homsy et al., 1972). However, a true bone-metal interface has also been reported (Hickman et al., 1958; Hicks, 1958; Wagner, 1963; B&remark et al., 1969; Charnley, 1970; Lundskog, 1972). Since a soft-tissue layer round an implant in bone reduces the stability of the implant with possible implant loosening as a result, it is of great practical importance to elucidate the aetiology of such a layer.
of Anatomy,
Traditionally, the connective tissue reaction has been attributed to implant corrosion, since corrosion products are frequently found in the tissue surrounding the implant (Venable et al., 1937; Wagner, 1963; Emneus and Stenram, 1965; Laing et al., 1967). Sometimes, however,
2.5
2.5
Fig. 1.-Schematic picture of the implants used in the study. All measurements are in mm. The narrow segment is introduced into the hole drilled in the bone. The wider segment stabilizes the implant during the first weeks after implantation, and is used as the connection point in the mechanical test. even macroscopically visible tissue contamination from implants appears to cause little or no reaction (B&remark et al., 1969). Other factors may be responsible for the marked tissue reaction around implants. Lundskog (1972), working with titanium, has shown
Injury: the British Journal of Accident Surgery Vol. ~/NO. 4
278 Tab/e /.-Composition W ATi 24 ATi 45 AISI 316 Vitallium
14-16
of the low-corrosive CO
bal.
The figures were obtained
Fig. 2.-Surface x
materials tested in the study
Fe
0
N
0.05 0.05 bal. 3.0 max.
0.10 0.35 -
0.03 0.07 -
from the manufacturers
c 0.05 0.10 0.05 max. 0.15 max.
Ii 0.012 0.012 -
Pd -
Ti bal. bal. -
Cr 17.4 19-21
and are given in percentages.
topography of implants of 4 materials tested, as seen by scanning electron microscopy 300. A, Stainless steel; B, Titanium ATi 45; C, Titanium ATi 24; D, Vitallium.
at
279
Linder and Lundskog: Stainless Steel, Titanium and Vitallium Tab/e 1. (continued) Ni
MO
Mn
Si
-
-
-
12.5 9-l 1
2.7 -
0.6 1 .O max.
1 .7 2.0 max.
The figures for titanium are maximal values.
that excessive operative trauma at the time of implant insertion causes soft-tissue formation round the implant and Uhthoff (1973) and Cameron et al. (1973) have shown that movement of the implant during the first weeks after implantation is still another cause of soft-tissue formation. It thus appears to be of interest to establish the importance of corrosion as compared to operative trauma and implant movement as an aetiological factor in soft-tissue reaction in bone. In the experimental model used by Lundskog (1972) both operative trauma and implant movement could be controlled to such a degree that no soft-tissue reaction occurred. For this reason it was decided to use this experimental model for comparative studies of the tissue reaction to various implants made of stainless steel, titanium and Vitallium. Copper, a metal well known for causing tissue reaction, was included in the study as a control material. MATERIALS AND METHODS Implants of stainless steel (AISI 316), titanium (ATi 45 and ATi 24, Avesta Jernverks AB, Sweden), wrought Vitallium, and copper were used. The composition of the low-corrosive materials used is given in Table Z, and the shape of the implants is demonstrated in Fig. 1. In earlier investigations the importance of the microstructure of the implant surface has been evaluated (Lundskog, 1972) and for this Tab/e
//.-Distribution
of the animals
Group 1 (VitalliumTitanium ATi 24) Number of animals Observation time
reason care was taken to obtain the same surface finish on all implants. However, because of the different mechanical properties of the implant materials, the requirement of an identical surface finish could not be met (Fig. 2), even if the surface topography appeared the same to the naked eye. All implants except those of Vitallium were machined with steel tools. The Vitallium implants were ground by a grinding plate, type 57A 46 18 BBE. The steel implants were not treated electrolytically. After fabrication, all implants were cleaned in alcohol in ultra sound before being autoclaved. After the experiments all implants except those of copper were submitted to surface analysis, performed with a scanning electron microscope (Cambridge JSM-UB). Micrographs, x 300 and x 3000, were obtained. Twenty-one rabbits of both sexes, 6 months old and weighing approximately 3.5 kg, were used in the investigation. The animals were divided into 4 groups in which the tissue reaction provoked by stainless steel, titanium ATi 45, Vitallium and copper respectively was analysed in relation to that of titanium ATi 24. Titanium ATi 24 was implanted in all animals, because of personal experience with this material, and thus each animal served as its own control. The distribution of the animals and implant materials into groups as well as the observation time for each group are shown in Table ZZ. For implantation, the method described by Lundskog (1972) was used. Under general barbiturate anaesthesia and under strictly aseptic conditions, a skin flap, based anteriorly, was raised over the proximal medial side of the tibia exposing the deep fascia. After careful dissection and diathermy, a flap of the fascia, based posteriorly, was raised and the periosteum exposed. Care was taken to avoid bleeding. With a sharp punch, a small piece of periosteum was removed and a guide hole drilled in the exposed cortical
6 7-7.5
Group 2 Stainless steel AISI 316Titanium ATi 24)
6 months
7 months
Group 3 (Titanium ATi 45Titanium ATi 24)
5 1 O-l 0.5 months
Group 4 (CopperTitanium ATi 24)
4 4 months (1) 3 weeks (3)
The groups are characterized by different test materials, titanium ATi 24 serving as the control material throughout. One implant of the test material and one of the control material was inserted in each animal, each tibia receiving one implant.
Injury:
280
the British Journal
of Accident
Surgery
Vol. G/No. 4
in lateral projections taken after sacrifice. A, Stainless steel implant. Arrows indicate the region where control measurements demonstrated a ‘ neck ’ of the imolant. B. Titanium ATi 24 imolant. Both periosteal and end&teal bone proliferation is C, VitaIlium implant; D, Copper implant (4 months).
Fig. 3.-Radiographs
obvious
but the implant
was completely
loose.
bone. Round this hole, a circular area of periosteum, of 4.0 mm diameter, was removed by another punch with a central guide peg. Finally, with a flat drill, guided by a central peg as well, a circular cortical defect, of 3.6 mm diameter, was made centrally in the periosteal defect. The drill used in this procedure was made of the same material as the implant later to be inserted in the hole. The bone was continuously irrigated with saline during drilling, which was done at 2000 rpm. One implant was then inserted in the cortical defect in each tibia. The implant was secured by a suture through its horizontal hole. The wound was closed by interrupted silk sutures, sprayed with a plastic wound dressing (Nobecutane) and covered with surgical tape. One week and two days before sacrifice, each animal received oxytetracycline (Terramycin) intraperitoneally in a dose of 10 mg per kg. Before sacrifice, the animals, under urethane anaesthesia, were submitted to microangiography. Five hundred ml of a 1 : 1 mixture of Pelican ink (C 1 l/l431 a) in physiological saline were
infused into the abdominal aorta. The animals of the copper-titanium group did not receive tetracycline injections and were not submitted to the microangiographic procedure because the toxic effect of copper had been obvious from general inspection and conventional histology in earlier pilot studies. After sacrifice and microangiography the implantation area was inspected and radiographed. The bone specimens with the implants in situ were then submitted to a mechanical test, described in detail by Lundskog (1972). In this procedure the implants are extracted axially and the extraction force is plotted against the displacement of th.e implant. From each curve thus obtained, the maximal extraction force can be determined. For histological evaluation, each tibia was divided into two longitudinal halves through the middle of the implant site. One of the halves was fixed in 5 per cent formaldehyde, decalcified in 30 per cent formic acid, and embedded in paraffin. Sections 4 I_Iand 100 I_Ithick were cut,
Linder and Lundskog
: Stainless
Steel, Titanium and Vitallium
the former being stained with haematoxylin and eosin for routine histology, the latter cleared by the Spalteholz method for microangiographic studies. The other half of the tibia, after freezedrying, was embedded in methyl methacrylate in the undecalcified state. Sections, 300-500 p in thickness, were obtained by saw-cutting. Tetracycline labelling was studied with the aid of a Leitz Orthoplane microscope in reflected light by the Ploem method. f
Fig. 4.-Diagrammatic
representation of the results The mean value and the extreme values of the maximal extraction forces of the materials in each group are given. Note, however, that the high extraction forces for some stainless steel implants can be explained by the bottle-neck shape of these implants. of the mechanical
test.
RESULTS General inspection All skin incisions healed by first intention and were barely visible. No signs of infection or discoloration of the tissues surrounding the implants were seen. All implants, except those of copper, were firmly incorporated in the bone and tested with a pair of tweezers. After extraction, the endosteal ends of the implants were found to have been invested by a thin bony capsule. No soft tissue was seen to adhere to the implants and the implants had the same surface appearance as when they were inserted. The copper implants were unstable, surrounded by fibrous tissue and necrotic material, some of which usually adhered to the implants. At 16 weeks, periosteal bone formation was evident round the copper implant, which was buried in new bone. Radiography The results of the general inspection were confirmed. The cortical bone had normal thickness
281
round all implants except the copper ones. A thin bony lamella round the part of the implant protruding into the medullary cavity as well as a small bony collar round the implant on the
Fig. S.-Histological section from the bone bordering a titanium ATi 24 implant. The bone has a mature appearance and the vascular channels are directed parallel to the implant surface. At the arrow, a small vessel is seen in the interface zone (H. and E. x 150.)
periosteal side were evident (Fig. 3A, B and C). After 3 weeks, the bone round the copper implants showed signs of periosteal thickening, and at 16 weeks both periosteal and endosteal thickening (Fig. 3D). Mechanical test The mean value of the maximal extraction forces for the steel implants was 7.6 kp, and for the corresponding titanium implants the mean value was 3.6 kp. The titanium ATi 45 implants had a mean of 7.2 kp and their titanium ATi 24 controls 9.7 kp. The figure for Vitallium was 7.2 kp and for the titanium ATi 24 controls of this group 5.8 kp. The mean for copper was 0.7 kp and for the corresponding titanium ATi 24 implants the mean value was 1.8 kp. The results are given diagrammatically in Fig. 4, where the mean values for each group as well as the extreme values can be seen. The discrepancy in extraction forces in the steel-titanium group initiated control measurements of all implants. These showed that 4 out
282
Injury: the British Journal of Accident Surgery Vol. ~/NO. 4
of 6 steel implants were slightly bottle-shaped, as seen in Fig. 3A. The thinner segment corresponded to 25 per cent of the implant length, and the difference in diameters between the ‘ normal ’ and the narrow part was 1 per cent. The defective steel implants had extraction forces of 9.2, 9.7, 9.3 and 10.0 kp, whereas the truly cylindrical steel implants both had extraction forces of 4.0 kp. The difference in the maximal extraction forces of the bottle-shaped and cylindrical steel
Fig. 6.-Histological appearance of the bone beside a copper implant, inserted 4 months earlier. Cancellous bone proliferation in the medullary cavity is present. Close to the implant is connective tissue, infiltrated with an abundance of round cells. The original cortical bone has assumed a cancellous appearance (H. and E. x 9).
Fig. 7.-Detail of Fig. 6 at higher magnification. Necrotic tissue is seen in the interface zone. Arrow indicates areas with necrotic bone (H. and E. x 82).
implants thus appears to be the result of the geometrical difference between the implants-an assumption which seems reasonable on the basis of mechanical calculation.
close to the implant surface. Numerous vascular channels were observed at the interface between implant and bone. The perivascular connective tissue did not spread over the surface of the bone to form a capsule, but was strictly confined to the vessel. The bone marrow had a normal appearance. The copper implants, on the other hand, produced a distinctly different picture, regardless of the observation time (Figs. 6 and 7). The cortex was thickened by new bone formation both on the periosteal and endosteal side and had a cancellous appearance. No callus had been laid down close to the implant surface. Empty osteocyte lacunae were commonly seen in this region. Connective tissue with round-cell infiltration was found to surround the implants and round-cell infiltration was also seen in the bone marrow.
Histological analysis The bone round the implants of stainless steel, titanium ATi 45 and 24, and Vitallium, regardless of the observation time, showed the same histological picture (Fig. 5). Lamellar bone was seen bordering the implant, the lamellae being mostly arranged parallel to the implant surface. Further away from the implant there was evidence of osteons running in a spiral or circular fashion round the implant site. Usually, no borderline could be detected between the original cortical bone and the newly formed bone. No continuous connective-tissue layer was seen, either along the sides of the implant, or along the bony floor, deep to the implant. Such a bony floor was seen in most sections and had the same appearance for all materials, although it was most prominent in the titanium ATi 455ATi 24 group. In the bone apparently normal osteocytes were seen, even
Microangiography Due to incomplete filling of the vessels, some specimens could not be examined. However, from the remaining angiograms it was evident that the cortical vessels had regenerated and formed
Linder and Lundskog : Stainless Steel, Titanium and Vitailium
283
surface topography. There was no evidence of corrosion on any of the implants. The steel implants generally had a smoother surface than their titanium controls and the same was true for the titanium ATi 45 implants. The Vitallium and the corresponding titanium implants were of the same roughness (Fig. 2). The surface analysis, however, also revealed differences between implants of the same material, especially implants of titanium ATi 45 and ATi 24.
Fig. 8.-Microangiogram from the bone bordering a steel implant, demonstrating the vascular arrangement in the endosteal bony capsule. (x 68.)
a network round the intracortical part of the implant. The bony capsule surrounding the endosteal aspect of the implant appeared to be supplied both by intracortical and endosteal vessels (Fig. 8). In no case were tom vessels seen in the interface zone, which might have indicated the presence of connective tissue, torn out in the mechanical test. Nor could tortuous vessels, suggestive of granulation-tissue formation, be seen. The vascular architecture appeared to be the same for steel, titanium ATi 45 and 24, and Vitallium. Fluorescence microscopy Tetracycline labelling was seen in all specimens and no difference in distribution or quantity for the different implant materials was detected. The labelling was usually seen round the larger vascular channels, in some instances those directly bordering the implant suface. Often, a small lamella of bone between the implant and the vessel was labelled, suggesting that new bone had been laid down round a vascular channel formerly in contact with the implant. Labelling was also seen at some distance from the implant, e.g. at the anterior tibia1 margin. Scanning electron microscopy All implants except those of copper were examined for signs of corrosion and differences in
DISCUSSION The present study shows that implants of AISI 316 stainless steel, pure titanium (ATi 45 and ATi 24) and wrought Vitallium can be incorporated in live rabbit bone, provided they are inserted with minimal trauma and immobilized during the healing period. This is in contrast to implants made of copper, which, with the same technique, was not accepted by the bone. For all implants, except those of copper, stable incorporation was accomplished in 3 weeks, after which the stability of the implants was consolidated, as shown by the higher extraction forces required in the groups observed for a longer time. This finding agrees with the results obtained by Cameron et al. (1972) and Lundskog (1972). Within each group, however, relatively large variations in maximal extraction forces were found for implants of the same material, and there were also variations between the groups. From earlier investigations (Lundskog, 1972) it is known that variations in the microstructure of the implant surface have a marked influence on the maximal extraction forces. Thus it appears likely that the differences in surface topography, combined with the biological variation of the animal material, explain the variation in the extraction forces. However, the titanium ATi 24 standard was always stable in the bone, and, if the groups are considered separately, there was no evidence that any one of the test materials was less stable than titanium ATi 24. In view of the fact that the implants were neither porous nor threaded, remarkably high extraction forces were recorded. This suggests close apposition of mineralized ground substance to the implants, and in fact no connectivetissue capsule was found round any of the implants except those of copper. Thus, since no adverse tissue reaction to any of the modern implant materials was seen, at least on light microscopy, it is concluded that bone, under conditions, treats these implant optimal materials in the same way.
284
Lundskog (1972), Cameron et al. (1973) and Uhthoff (1973) have pointed out the importance of implant immobility for the successful incorporation of an implant in bone. According to these authors, mechanical loading without implant movement is compatible with incorporation, whereas loading with implant movement gives rise to a soft-tissue capsule round the implant, and the risk of the implant loosening. As the experimental model used in this study prevents implant movement, the results support the theory presented. However, it has also been shown (Lundskog, 1972) that, even in the absence of implant movement, soft tissue may form if thermal injury to the bone is inflicted at the operation. Using the same model as in the present study, Lundskog was able to show that temperatures exceeding 75°C for 30 seconds always inh.ibited implant incorporation, despite the fact that the implants were immobile. Temperatures of this order may very well be reached in clinical practice, for example during drilling, if special precautions are not taken (Lundskog and Nilsson, 1969; Matthews and Hirsch, 1972). One of these is saline irrigation, which lowers the temperature to acceptable values (Lundskog and Nilsson, 1969). Subliminal thermal trauma at the operation and immobility of the implant thus appear to be prerequisites for implant incorporation in bone, while the implant material itself seems to be of less importance provided it is one of the modern, low-corrosive metals or alloys. Copper produced marked bone reaction but this material is known to be very toxic (Venable et al., 1937; Nicole, 1947), possibly via corrosion products (Pappas and Cohen, 1968), and the observed reaction was expected. This proves that chemical trauma may cause tissue reaction, even if the implant is inserted under optimal conditions. As it has been shown that all metallic implants corrode in living tissues (Laing et al., 1967) even though the defects in the metal may be invisible (Hicks and Cater, 1962), it appears likely that corrosion of low intensity was also present with The the tested modern implant materials. corrosion rate, however, was obviously insufficient to cause tissue reaction by itself. A doseresponse relationship appears to exist for this type of trauma, and the present study indicates that the tissue reaction threshold was never reached except in the copper-titanium group. The results strongly support the suggestion by is the true Hicks (1958) that ‘ zero reaction response to the chemically inert foreign body and that the fibrotic reaction is the response to
Injury
: the
British Journal
of Accident
Surgery
Vol. ~/NO. 4
varying degrees of microcorrosion.’ Moreover, the study shows that the mere physical presence of the implants did not cause adverse tissue This finding is interesting, because reaction. th.ere have been reports stating that bone always treats metallic implants as ‘ foreign bodies ’ by connective-tissue encapsulation (Collins, 1954). The modern implantation techniques used in clinical practice usually rely on composite devices which may increase the corrosion rate (Bechtol et al., 1959; Scales et al., 1959; Emneus and Stenram, 1965 ; Colangelo and Greene, 1969). Another factor of importance is mechanical stress which may result in stress corrosion or fretting corrosion. Whether an increased corrosion rate caused by any of these factors can induce tissue reaction is still open to debate, and it seems that this problem can be solved only by further standardized experiments. Even though the conclusions of the present investigation may be somewhat limited by the relatively short observation time, the study indicates that gentle handling of the tissues at the operation and mechanical stability of the implant during the period of bone healing are more important for stable incorporation in bone than is the implant material, as long as it is one of the modern, low-corrosive metals or alloys. Thus, it appears likely that many of the tissue reactions seen round metallic implants in bone may have been falsely attributed to corrosion alone.
REFERENCES
BECHTOL, C. O., FERGUSON, A. B. jun. and LAING, I’. G. (1959), Metals and Engineering in Bone and Joint Surgery (1st ed.) Baltimore, Williams &
Wilkins Company. p. 40. BR~~NEMARK, P.-l., BREINE, U., ADELL, R., HANSSON, B. O., LINDSTRBM,J. and OHLSSON,& (1969), ‘ Intraosseous anchorage of dental prostheses. I. Experimental studies, Stand. J. Plast. Reconstr. Surg.,3,8 1. CAMERON, H., MACNAB, 1. and PILLIAR, R. (1972), ‘ Porous surfaced Vitallium staples ‘, S. Afv. J. Surg., 10, 63. CAMERON, H., PILLIAR, R. and MACNAB, 1. (1973), ‘ The effect of movement on the bonding of porous metal to bone ‘, J. Biomed. Mater. Res., 7, 301. CHARNLEY, J. (1970), ‘ The fixation of prostheses in living bone ‘, In: Modern Trends in Biomechanics (ed. Simpson. D. C.), London, Butterworths. ‘ Tissue reactions to metals-the COHEN, 5.-(19&l), influence of surface finish ‘, J. Bone Jt Surg., 43A, 687. COLANGELO,V. J. and GREENE, N. D. (1969), ‘ Corrosion and fracture of type 316 SMO orthopedic implants ‘, J. Biomed. Mater. Res., 3, 247.
Linder and Lundskog
: Stainless
Steel, Titanium and Vitallium
D. H. (1954), ’ Tissue changes in human femurs containing plastic appliances ‘, J. Bone Jt Surg., 36B, 458. EMNBUS,H. and STENRAM,U. (1965), ‘ Metal implants in the human body. A histopathological study ‘,
COLLINS,
Acta Orthop.
Stand.,
36,
115.
HICKMAN,J., CLARKE,E. G. C. and JENNINGS,A. R. (1958), ‘ Structural changes in bone associated with metallic implants ‘, J. Bone Jt Surg., 40B, 799. HICKS, J. H. (1958), ‘ Pathological effects from surgical metal ‘, In : Modern Trends in Surgical Materials (ed. Gillis, L.), London, Butterworths. HICKS, J. H. and CATER, W. H. (1962), ‘ Minor reactions, due to modern metal ‘, J. Bone Jt Surg. 44B, 122. HOMSY, C. A., STANLEY,R. F., ANDERSON,M. S. and KING, J. W. (1972), ‘ Reduction of tissue and bone adhesion to cobalt alloy fixation appliances ‘, J. Biomed.
Mater.
Res., 6, 451.
LAING, P. G., FERGUSON,A. B. and HODGE, E. S. (1967), ‘ Tissue reaction in rabbit muscle exposed to metallic implants ‘, Ibid., 1, 135. LUNDSKOG,J. (1972), ‘ Heat and bone tissue ‘, Stand. J. Plast. Reconstr.
Surg.,
Suppl.
9.
LUNDSKOG,J. and NILSSON, K. (1969), Unpublished data. MATTHEWS,L. S. and HIRSCH, C. (1972), ‘ TemperaRequesrs G(iteborg,
for reprints Sweden.
should
be nddressed
to:-Lars
Linder,
MD,
285
tures measured in human cortical bone when drilling ‘, J. Bone Jt Surg., 54A, 297. NICOLE, R. ‘ Metallschadigung bei (1947), Osteosynthesen. Experimentelle und klinische Untersuchungen tiber Wesen und Bedeutung der Metallose und der Korrosion ‘, He/v. C&r. Acta, Suppl. III. PAPPAS, A. M. and COHEN, J. (1968), ‘ Toxicity of metal particles in tissue culture. Part I, A new assay method using cell counts in the phase of replication ‘, J. Bone Jt Surg., 5OA, 535. SCALES,J. T., WINTER, G. D. and SHIRLEY,H. T. of orthopaedic implants. (1959), ’ Corrosion Screws, plates and femoral nail-plates ‘, Ibid., 41B, 810.
UHTHOFF, H. K. (1973), ‘ Mechanical factors influencing the holding power of screws in compact bone ‘, Ibid., 55B, 633. VENABLE,C. S., STUCK, W. G. and BEACH,A. (1937), ‘ The effects on bone of the presence of metals; based upon electrolysis. An experimental study ‘, Ann. Surg., 105, 917. WAGNER, H. (1963), ‘Die Einbettung von Metallschrauben im Knochen und die Heilungsvorgange des Knochengewebes unter dem Einfluss der stabilen Osteosynthese ‘, Langenbecks Arch. Klin. Chir., 305, 28. Department
of Anatomy,
University
of
Giiteborg
Fack,
S-400
33
