Iournal uf Orthopaedic Research 10285-299 Raven Press, Ltd., New York 0 1992 Orthopedic Research Socicty

Tissue Ingrowth into Titanium and Hydroxyapatite-Coated Implants During Stable and Unstable Mechanical Conditions *TKjeld Sgballe, *tEbbe S. Hansen, $Helle B.-Rasmussen, $Peter H. Jgrgensen, and Cody Bunger *Biomerhanics Laborutory, Orthopedic Hospital, University Hospital of Aarhus; and ?Institute o j Experimental Clinical Research, $Institute of Pathology, A a r h s AmtsJygehus; and §Department of Connective Tissue Biology, Institute of Anatomy, University of Aarhus, Aarhus, Denmark

Summary: Lack of initial mechanical stability of cementless prostheses may be responsible for fibrous tissue fixation of prosthetic components to bone. TO study the influence of micromovements on bony ingrowth into titanium alloy (Ti) and hydroxyapatite (HA)-coated implants, a loaded unstable device producing movements of 500 pm during each gait cycle was developed. Mechanically stable implants served as controls. The implants were inserted into the weight-bearing regions of all four femoral condyles in each of seven mature dogs. Histological analysis after 4 weeks of implantation showed a fibrous tissue membrane surrounding both Ti and HA-coated implants subjected to micromovements, whereas variable amounts of bony ingrowth were obtained in mechanically stable implants. The pushout test showed that the shear strength of unstable Ti and HA implants was significantly reduced as compared with the corresponding mechanically stable implants (p < 0.01). However, shear strength values of unstable HA-coated implants were significantly greater than those of unstable ‘l’i implants (p < 0.01) and comparable to those of stable Ti implants. The greatest shear strength was obtained with stable HA-coated implants, which was threefold stronger as compared with the stable Ti implants (p < 0.001). Quantitative determination of bony ingrowth agreed with the mechanical test except for the stronger anchorage of unstable HA implants as compared with unstable Ti implants, where no difference in bony ingrowth was found. Unstable HA-coated implants were surrounded by a fibrous membrane containing islands of fibrocartilage with higher collagen concentration, whereas fibrous connective tissue with lower collagen concentration was predominant around unstable Ti implants. In conclusion, micromovements between bone and implant inhibited bony ingrowth and led to the development of a fibrous membrane. The presence of iibrocartilage and a higher collagen concentration in the fibrous membrane may be responsible for the increased shear strength of unstable HA implants. Mechanically stable implants with HA coating had the strongest anchorage and the greatest amount of bony ingrowth. Key Words: Bone ingrowth-Fibrous membraneHydroxyapatite-Micromotion-Porous ingrowth-Titanium.

Porous coated prostheses were designed to obtain permanent fixation by ingrowth of bone into the porous implant surface. However, recent histological analyses of cementless tibia1 components (1 1,16) and total hip replacements (10,12) retrieved

Received November 21, 1990, accepted October 1, 1991. Address correspondence and reprint requests to Dr. K. S@balleat Biomechanics Laboratory, Orthopaedic Hospital, Randersvej 1, DK-8200 Aarhus N. Denmark This work was presented at the 36th Annual Meeting, Orthopaedic Research Society, New Orleans, Louisiana, 1990.

285

286

K . SOBALLE ET AL.

from humans indicate that most of the components are primarily fixed to the skeleton by fibrous tissue. Reduced bony ingrowth and weakened mechanical anchorage of the porous coated implant may occur if the bone stock is osteopenic (34), in the absence of initial contact between bone and implant (9,35-37,40), if the pore size of the surface coating is too small (38), or with lack of initial mechanical stability between bone and implant (7,8,14,17,21, 30,40). However, hydroxyapatite (HA) coating has been demonstrated to increase bone ingrowth and anchorage strength of porous coated implants (5,13,15,31,35-37,4 1). Experimental studies of cementless tibial components implanted into cadaver tibias have shown micromovements ranging between 200 and 500 pm at loads in the low physiologic range (6,33,39,47). Studies on femoral components in hip arthroplasty have revealed micromovements of 100-500 pm (46), or even less (50) between bone and implant. In clinical studies, inducible displacement of noncemented tibial trays in total knee replacements measured l year postoperatively by roentgen stereophotogrammetry is in the range of 400-1,300 pm. indicating prosthetic fixation by a fibrous membrane (32). Movements between bone and implant have been shown to prevent bony ingrowth and result in development of a fibrous tissue membrane, particularly if this motion occurs during the healing process after implantation (8,14,21,30,44). However, we found no reports on experimental models of controlled relative movements between bone and implant during each gait cycle. The purpose of the present study was to develop a new dynamic system to investigate the tissue response around porous coated implants subjected to controlled micromovements and to study the effect of HA coating in this unstable mechanical milieu.

titanium (Ti) and HA-coated implants was performed in the weight-bearing region of the femoral condyles. In randomized order, unstable HA implants were allocated to the medial or lateral condyle in the right knee and the stable HA implants were allocated to the corresponding condyle in the left knee. The unstable Ti implants were inserted in the remaining condyle in the left knee and the stable Ti implants were inserted in the corresponding condyle of the right knee. Thus, one unstable and one stable implant was inserted in each knee. The dogs were killed after 4 weeks and the results were evaluated by pushout test, histology, and histomorphometry. Dynamic Device

The principle for producing micromovements is shown in Fig. 1 . The system consisted of an implantable dynamic device manufactured in Ti alloy (Fig. 2), which was inserted into the weight-bearing part of the femoral condyle. When the knee was loaded during gait, load transfer from the tibial part

MATERIALS AND METHODS Experimental Design

Eight mature Labrador dogs with an average age of 28 (range 15-35) months and weighing an average of 31 (range 24-39) kg comprised the material. The dogs were bred for scientific purposes and were handled according to the Danish law on animal experimentation. Preoperative radiography of both knees ensured that all animals were skeletally normal and mature. Implantation of stable and unstable

J Orthop Res, Vol. 10, N o . 2,1992

FIG. 1. The unstable device implanted in the distal femoral condyle is shown at left. Note the protrusion of polyethylene, which transmits the load from the tibial part of the knee. During weight bearing, the implant will be displaced in the axial direction and the spring is tightened. When the leg is unloaded the tightened spring will move the implant back to the initial position. A titanium ring inserted into the subchondral part of the hole serves as a bearing and centralizer for the polyethylene plug. To the right is the stable device, which was completely unloaded (not shown).

TITANIUM A N D HYDROXYAPATITE-COATED IMPLANTS

287

during the following flexion phase from 33 to 40". Trotting would double the total amplitude of movement, which normally was 30". From postmortem radiographic measurements and clinical observation in the present series, it was demonstrated that contact between the dynamic device and the tibia1 part of the knee was maintained within the range of 30-95" of knee flexion with the implantation technique used.

3

2

4

4a

FIG. 2. The unstable device consists of seven components all manufactured from titanium alloy (Ti-6AI-4V) as the porous coated Ti implant. A hollow titanium cylinder (1) has selftapping threads to ensure firm fixation in the bone. A spring (2) is placed inside the cylinder and held in place by an adjustable screw (3) at one end. In the other end a titanium piston (4) can move freely i n the axial direction. When mounted, the platform (4a) on the piston projects exactly 500 p m over the end of the titanium cylinder. When the implant (5) is screwed onto the threads of the piston and an axial load is applied on the polyethylene plug ( 6 ) ,the implant will move until it is stopped by reaching the titanium cylinder and the movement is limited to 500 pm. To prevent rotation of the piston, one end of the spring is fixed to the piston (4) and the other t o the screw (3),which is locked into the titanium cylinder by a small polyethylene plug inserted into the threads of the screw. A hole through the piston and the polyethylene plug connects the compartment in the titanium cylinder with the knee joint. The coating is removed at the distal end of the implant (5a), to prevent bony ingrowth in this area. 1 , Titanium alloy cylinder; 2, spring: 3, screw; 4, piston with a threaded part; 4a, platform at the base of the piston thread; 5, test implant; 5a, base of test implant; 6, polyethylene plug.

of the knee displaced the implant in the axial direction and tightened the spring. When the leg was unloaded, the tightened spring moved the implant back to the initial position. Thus, a controlled movement occurred during each gait cycle. In this study, the movement was predetermined to -500 k 15 pm. The coil spring was 2.3 mm in diameter and 4.5 mm long, turned from Ti alloy wire, and adjusted to a stiffness of -14 Nimm with a preload of 0.5 N, the total displacement force being 10 N. The site of placement of the device in the femoral condyle was chosen from electrogoniometric studies of normal gait in dogs (I), which showed that paw contact occurred at 40" of flexion in the stifle joint. According to these studies, load would occur during the following 7" extension and subsequent flexion of 7". Paw lift occurred again at 40" of flexion. The stifle was then flexed to 60" before an extension in preparation for the next support phase. Thus, during walking phase, the knee joint would be loaded during extension from 40 to 33" as well as

Mechanical Testing Measurements of the dynamic devices were performed before and after implantation using a universal test machine (Instron Ltd., High Wycombe, Bucks HP12 3SY, England). The dynamic device was placed on a metal platform and a rod was placed in the upper holding device for an axial compression test. An initial load of 0.1 N was used to define the contact position. A displacement rate of 2 mmimin was used for all tests, and loaddeformation curves were obtained by an X-Y recorder. The stiffness was obtained from the slope of the straight-line portion of the load-displacement curve. The value was corrected for the testing system compliance. The springs were adjusted until the desired stiffness (14 N/mm) was obtained. A preload of 0.5 N was approached. After 4 weeks of implantation, the loaded devices were removed and tested mechanically using the same procedure as before implantation. Values from mechanical testing of the springs are given in Fig. 3. The mean displacement measured was 490 p m (SEM 20). Only minor variation in stiffness and preload was registered between the devices. The stiffness and preload increased equally and significantly during the 4 weeks of implantation from 14 to 16 Nimm and from 0.5 to 4 N, respectively. Stiffness and preload were not different between devices with HA and Ti implants .

Implants The two types of implant coatings used were plasma sprayed HA and plasma sprayed Ti alloy. The cylindrical Ti and HA-coated plugs were 6.0 mm in diameter with an overall length of 10 mm, fabricated with a precision of k0.05 mm. Ti implants consisted of a solid Ti-6A1-4V alloy core with a coating of Ti-6A1-4V deposited by plasma-spray technique, resulting in a mean pore size of 300 pm.

J Orthop Res, Vol. 10, No. 2 , 1992

288

K . SgBALLE ET AL. BEFORE IMPLANTATION (N =14)

AFTER IMPLANTATION (N = 14)

109-

10-

8-

8-

9-

h

7-

5 6D

:5 -

-I

FIG. 3. Results from mechanical test of the dynamic devices.

4-

3-

21I

The HA-coated implants consisted of analogous Ti porous coated implants on which a 50-75-pm layer of spray-dried synthetic HA (CdP ratio 1.641.70) as determined by Auger electron spectroscopy was deposited by plasma-spraying technique. The substrate Ti for the HA-coated implants was appropriately undersized to provide identical implant diameters after coating. According to the manufacturer, the strength of attachment between the HA and the substrate as determined by American Society for Testing and Materials (ASTM) standard C-633 for cohesive strength of coatings to metal revealed minimum tensile strength of 5,000 lbiin' (34.5 MPa) and a minimum shear strength of 3,000 Ibiin' (20.7 MPa). X-ray diffraction analysis of the ceramic coating compared with ASTM powder diffraction standard 9-432 for HA showed pure HA with no trace of tricalcium phosphate at the level of sensitivity, which is -3%, indicating that there is at least 97% HA. Sterilization by gamma irradiation did not change the purity of HA. Fourier transform infrared analysis showed a crystallinity averaging 75% (crystal sizes detectable in excess of 0.1 kmj. The crystalline composition was HA with no additional phases, and the carbonate content was less than 1%. The surface roughness (Ra) of Ti and HA-coated implants was determined using a roughness meter (Perthen, Hannover, Germany) with a stylus tip radius of 3 pm. Ra (international parameter of roughness) is the arithmetic mean of the departures of the roughness profile from the mean line. Ra was 41 pm for HA implants and 47 pm for Ti implants. The profile depth was also measured and was shown to be 445 Frn for HA and 496 pm for Ti implants.

J Orthop Res, Vol. 10, No. 2 , 1992

I

I

Polyethylene Plug

The polyethylene plugs were manufactured from ultra-high-molecular-weight polyethylene (UHMWPE) (medical grade), (Chirulen, Ruhrchemie AG, Oberhausen. Germany). The plugs were 25 pm less than the inner diameter of the centralizer and were domed at one end to contact the tibia1 part of the knee. All implants were sterilized by gamma irradiation (35 kGy of CO-60for 14 h, Ris@National Laboratory, Roskilde, Denmark). Surgical Procedures

Anesthesia was induced by intravenous Brietal (methohexital) 5 mgikg after premedication with 1.5 ml 0.1% Combelen [ l0-(3-dimethylaminopropyl)propionyl phenothazine] . After orotracheal intubation, anesthesia was maintained by halothane anesthesia through a constant-volume ventilator. A separate medial and lateral arthrotomy was performed through a common longitudinal prepatellar skin incision. By maximal flexion of the knee, the weight-bearing portion of both femoral condyles was exposed. The implantation site was selected in the central portion of the femoral condyles, which contacts the meniscus and tibia during the stance phase and the walking phase (1). Initially a guide wire was inserted into a 2.1-mm drill hole (Fig. 4A). Drilling was guided by fluoroscopy. The best anteroposterior (AP) visualization of the condyles was in the supine dog with 90" flexion of the hip. The guide wire was placed centrally in the condyle as seen in this projection. In the lateral plane, the tip of the guide wire was directed toward

TITANIUM AND H YDROX YAPATITE-COATED IMPLANTS

289

FIG. 4. Schematic drawing illustrates the sequences in the operative procedure.

the proximal limit of facies patellaris, which is easily identified by fluoroscopy. A cannulated trapdrill (Fig. 4A) expanded the hole to 6-mm width at the 10-mm deep part and 7.5 mm at the 20-mm superficial part, allowing firm fixation of the Ti cylinder in the deep part of the hole and permitting a 0.75-mm gap surrounding the test implant. A handdrilling technique was used to avoid thermal trauma to the bone. Using a specially manufactured tapping instrument, the 5-mm subchondral part of the drill hole was tapped for treads (Fig. 4B) for later emplacement of the centralizer. After insertion of the unstable device (Fig. 4C), a centralizer (Fig. 4D) for the polyethylene was inserted subchondrally to ensure parallel axis with the piston. Then the test implant was screwed onto the thread of the piston (Fig. 4E) and tightened against the platform at the base of the piston thread. Finally, the UHMWPE polyethylene plug, fitting into the centralizer, was screwed in place so that the dome end projected slightly above the femoral articular cartilage (Fig. 4F). The centralizer served as a bearing for the polyethylene plug and restricted movement of the implant to axial translation. Extension of the knee confirmed contact between the meniscus and the polyethylene plug, and axial motion of the system

could be visualized by axial load application on the tibia. The amount of projection of polyethylene yas adjustable to allow exact positioning of the system in the joint contact area. Thus, when assembled and implanted into the weight-bearing part of the femoral condyle, controlled movements would occur during each gait cycle. The stable implants were implanted using an identical procedure (Fig. 1) but without loading the implant to ensure mechanical stability. Soft tissues were closed in a routine manner and a prophylactic antibiotic drug (Anhypen, GistBrocades, Delft, Holland) was administered for 3 days starting 1 h preoperatively. The knees were radiographed on anteroposterior and lateral views immediately postoperatively. The inclination of the implants in relation to the long axis of the femur was measured using a protector. The dogs were allowed free activity and unrestricted weight-bearing after surgery. The animals stayed in individual cages that measured 1.5 X 2.5 m with outdoor training 3 h a day (1.5 x 3.5 m). Gait performance was registered regularly. After 4 weeks the dogs were killed using methohexital and saturated KC1. Ten and 3 days before termination, the dogs were tetracycline double-

J Orthop Res, Vol. 10, No. 2,1992

K . SQBALLE ET AL.

290

labeled by i.v. administration of chlortetracycline (Dumocyclin; Dumex, Copenhagen, Denmark) 15 mg/kg for bone histomorphometry . Immediately postmortem, the knees were opened under sterile conditions, the implants were exposed, and cultures were taken from the implantation site and synovial fluid. Preparation The distal femora were excised, radiographed in the AP and lateral planes, and stored at -20°C for 1 week. Three standardized sections orthogonal to the long axis of the implant were performed on a water-cooled diamond band saw (Exakt-Cutting Grinding System, Exakt Apparatebau, 2000 Norderstedt, Germany). The first cut was 3 mm below the distal end of the implant. The first 3-mm section was stored at - 20°C and used €or mechanical testing, the second 3-mm section served for histomorphometric evaluation on ground stained specimens, and the third 100-pm-thick section was for U V fluorescence microscopy. Specimens for fluorescence microscopy were preserved in 70% alcohol and specimens for grinding were preserved in buffered formalin. Mechanical Testing The pushout tests were performed with a universal test machine (Instron Ltd.) (34-36). The specimens were placed on a metal platform with a central circular opening supporting the bone to urithin SO0 pm of the interface. A metal rod was placed in the upper holding device for axial pushout test of the implant from the surrounding tissues. A displacement rate of 5 mmimin was used for all tests and load-deformation curves were obtained by an X-Y recorder. A contact position at 2 N was used to define the contact position for destructive pushout test. Ultimate shear strength, apparent shear stiffness, and energy absorption were estimated from the load-displacement curves as previously described (34,35). Histological Evaluation Fibrous Membrane

The membranes around the unstable implants were removed in toto from the surrounding bone under a dissecting stereomicroscope. One part of the membrane was prepared for histology and stained with toluidine blue at pH 5 for

J Orthop

Res. Vol. 10, NO.2,1992

qualitative analysis by light microscopy, and the other part was used for determination of hydroxyproline concentration. Fibrocartilaginous tissue in the membrane was quantified using point-counting technique, and an estimate of the percentage of the membrane containing fibrocartilage (chondrocytes) was obtained; fibrous tissue without chondrocytes accounted for the remaining tissue in the membrane. Synovium

Specimens of synovium from the knee joint were taken and stained with Giemsa, hematoxylin and eosin, or Van Cieson for histological evaluation of inflammatory changes. Membrane Thickness

The membrane thickness was determined on unstained and undecalcified 100-pm-thick sections as the distance between implant surface and tetracycline-labeled bone. It was measured by UV fluorescence microscopy at x 100 magnification with a Zeiss eyepiece micrometer placed perpendicularly to the implant surface (24) at 10 randomly selected locations around the implant. The scale was calibrated at x 100 magnification using a stage micrometer (Zeiss) graduated in 0.01-mm units. Bony Ingrowth

Bone ingrowth, defined as bone tissue in contact with the implant surface, was measured on stained ground specimens of SO-pm thickness. During dehydration, the specimens were stained with 0.4% basic fuchsin. A jar containing the specimen and liquid methylmethacrylate was put under vacuum for -24 h to remove entrapped air. After polymerization at 450-nm wavelength (Kulzer-Exact-Lightpolymerization unit, Norderstedt, Germany), the block was mounted on acrylic slices and ground to -50 pm using a microgrinding system (EXAKTMicro Grinding System, EXAKT Apparaturbau, Norderstedt, Germany) and counterstained with 2% light green for 6 min to allow quantitative histological evaluation of fibrous tissue distribution and bone to implant apposition using transmitted light microscopy at x 100 magnification. The procedure was performed blindly and in random order using a linear intercept technique with a Zeiss integrating plate with 10 gridlines mounted in the eyepiece of the microscope. The number of intersections with bone or fibrous connective tissue in direct contact with the implant surface was counted. Measure-

TITANIUM AND H YDROXYAPATITE-COATED IMPLANTS

ments were made on successive adjacent fields along the entire implant circumference. Approximately 300 test points were evaluated for each irnplant and an estimate of the percentage of the implant surface in contact with either bone or fibrous tissue was obtained, with bone marrow and empty spaces accounting for the remaining coverage of the implant. Gap Healing

In a well-defined area away from the implant surface (-40-220 pm), bone volume (BV) was quantified for evaluation of the amount of bone filling the initial gap. The BV was estimated blindly by pointcounting technique using the central part of the integrating plate with 25 points at x 160 magnification. Measurements were made on successive adjacent fields (-800 points) along the entire implant circumference on ground stained specimens of 50-pm thickness. Determination of Membrane Hydroxyproline Concentration The isolated membranes were freeze dried for 72 h and the dry weight was determined after defatting. The hydroxyproline concentration was determined by a colorimetric method a.m. Woessner (49) using 4-diaminobenzaldehyde after hydrolysis with HCI . The collagen concentration was calculated by multiplication of hydroxyproline concentration by 7.46 (28). Bone tissue from the femoral condyles was also analyzed for hydroxyproline concentration. Statistics From all the parameters, the mean values and SEM were calculated. Before statistical testing, differences between groups were found to be approximately normally distributed by probit analysis. Comparison of the means was performed by twotailed paired and unpaired t test. Analyses of variance for repeated measures were performed on pushout data to determine differences in presence/ absence of HA coating and stability.

291

dogs walked with a limp for 6-8 days, recovered shortly thereafter, and remained healthy and active throughout the study. At autopsy, all polyethylene plugs were in situ. Increased synovial fluid was seen in all knees. The menisci had only minor lesions at the central part. The articular tibia1 cartilage beneath the menisci was fibrillated but grossly intact. The bacterial cultures from the knees were negative. Inclination of the implants in relation to the long axis of the femur varied between 55 and 65" and no difference was found between Ti and HAcoated implants. Mechanical Analysis All unstable specimens with fibrous membrane failed between the surrounding tissue and implant. Most specimens with bony ingrowth (stable implants) failed at the bone-implant interface, but a few stable Ti and HA implants had small amounts of bone on the surface after the pushout test. No failures were observed at the HA-Ti alloy substrate interface. In the course of pushout tests of two unstable HA-coated implants, the load-deformation curves had two peaks with increasing loads. In these cases the first (and lowest) peak value was used. In one case the XY writer failed to draw the curve from the pushout test, but the ultimate shear strength was read directly from the testing machine. The ultimate shear strength (Fig. 5) of unstable Ti and HA implants was significantly reduced as compared with the corresponding stable implants. However, shear strength values of unstable HA-coated implants were fivefold increased compared with unstable Ti implants and obtained the same shear strength as stable Ti implants. The greatest shear strength was obtained by stable HA-coated implants, which was 250% stronger compared with the stable Ti implants (p < 0.01). Shear stiffness and energy absorption largely parallelled those for shear strength (Table 1). The analysis of variance showed that the differences but not the ratio between the fixation of HA and Ti-coated implants depended on the stability conditions, e.g., the strongest effect of HA coating was obtained during stable mechanical conditions (F = 13.2, p < 0.01) (Table 1). Histology

RESULTS One dog was excluded because of continuous limp and diminished weight bearing. The rest of the

Qualitative Description All unstable implants were surrounded by a fibrous membrane (Fig. 6), whereas all stable

J Orthop Res, Vol. 10. No. 2 , I992

292

K . SGRALLE ET AL. HA 7-

2.2-

2.0h

5.

1.8-

'

5 1.6B

1.4-

c

u)

z

1.2-

1

3

1.0

2

0.8

(0

Ti

FIG. 5. Results from the mechanical (pushout) test. The arrows indicate unstable implants, and the two other bars represent stable implants. HA, hydroxyapatite; Ti, titanium.

implants had variable amounts of bone contact (Fig. 7). The membranes surrounding the unstable Ticoated implants consisted of fibrous connective tissue (Fig. 8) with an irregular collagen fiber orientation. Small areas of fibrocartilage were seen, especially at the membrane-bone interface. Histiocytes, lymphocytes, or macrophages were present at the interface between Ti and membrane. Titanium debris or polyethylene debris, which may be seen as bright white under polarized light, was not observed. Few blood vessels were seen in the membranes. The membranes surrounding the HA-coated implants consisted predominantly of large islands of fibrocartilage with uni- and bicellular groups of chondrocytes in lacunae irregularly distributed between the heavy collagenous fiber bundles (Fig. 9). An amorphous purple-staining cartilaginous ground substance (matrix) was present. Blood vessels were seen, but no histiocytes, lymphocytes, or macrophages were seen. Around the periphery of the fibrous membrane, the newly formed trabeculae formed a plate of condensed bone parallel to the outer surface of the im-

J Orthvp Res, Vol. 10, No. 2, 1992

plant around both HA and Ti implants, and many osteoblasts were seen at the border between membrane and surrounding bone, where both bone resorption and formation of woven bone occurred. Around all unstable implants, a high remodeling activity was seen, which increased toward the bone-membrane interface as evidenced by many resorptive foci and much osteoid formation. The high remodeling activity at the periphery of the membrane was further demonstrated by fluorescence microscopy, which showed accumulation of tetracycline-labeled bone at the border of the drill hole. The initial gaps surrounding stable HA-coated implants were filled by woven bone with high remodeling activity. Bone tissue was in intimate contact with the stable HA-coated implants, only occasionally interposed by fibrous tissue. The stable Ti implants displayed several areas without direct bone-implant apposition and many areas with interposition of fibrous tissue. No inflammatory reaction was seen around stable implants. Synovium Histology of stained synovia showed slight synovial reaction evidenced by the presence of plasma cells and lymphocytes. Histomorphometry

Thickness of Fibrous Membrane The mean membrane thickness was 585 pm (SEM 45) around unstable Ti implants and 624 km (SEM 42) around unstable HA-coated implants subjected to micromovements (NS). Bone and Fibrous Tissue Ingrowth All stable implants were surrounded by variable amounts of bone. Under stable conditions, HA coating increased bony ingrowth fivefold compared with Ti (p < 2 X (Fig. IOA). The unstable implants had sporadic presence of bone tissue (p < 0.01), and no difference was found between the two types of coating (Fig. 10A). The unstable implants had more fibrous tissue ingrowth (Fig. 10B) compared with the stable implants (p < 0.0001). Very small amounts of fibrous tissue were observed in contact with stable HA-coated implants, whereas

293

TITANIUM A N D HYDROXYAPATITE-COATED IMPLANTS

TABLE 1. Pushout values for titanium and hydroxyapatite coated implants under vtahle and unstable conditions after 4 weeks of implantation Titanium implants ~~

Ultimate shear strength (MPa) Apparent shear stiffness (MPdmm) Energy to failure (Urn)

Hydroxyapatite implants ~

Stable

Unstable

Stable

Unstable

0.63 (0 16) 3.9 (I 4) 1.7 (0 5 )

0.12 (0.01)"

2.18 (0.24)h 20.3 (3.7Ib 6.2 (0.8)b

0.63 (0. l)",' 3.0 (0.7)"s" 3.2 (1.0)"'

0.29 (0.0s)" 0.66 (0.1)"

Values are mean (SEM) (N = 7). Registration of stiffness and energy absorption failed for one stable titanium implant. LI Unstable implant significantly different from corresponding stable implant (p > 0.01). Stable hydroxyapatite-coated implant significantly different from stable titanium implant (p < 0.05). Unstable hydroxyapatite-coated implant significantly different from unstable titanium-coated implant (p < 0.01).

one-third of Ti implant surfaces were covered by fibrous tissue (Fig. 10B). Gup Healing

HA coating increased bone formation in the gap under stable conditions compared with Ti implants (p < 0.01), whereas practically no bone was found in the gap around the unstable implants (p < O.OOOOl), irrespective of the type of coating (Fig. 11).

Hydroxyproline Concentration Determination of hydroxyproline concentration from one membrane from a Ti implant failed because the sample glass broke during the procedure. The mean hydroxyproline concentration (given in percent of dry weight) in the membrane around Ti implants was 4.9 and 6.7% around HA-coated implants (p < 0.05). The corresponding collagen concentration was 37 and 5796, respectively, which means that the collagen concentration was 55%

FIG. 6. Photomicrograph from an HA-coated implant subjected to micromovements shows fibrous tissue around HA-coated implant (light green, basic fuchsin, grounded section; original magnification x6).

. I Orthop Res, Vol. 10, No. 2 , 1992

294

K . SgBALLE ET AL.

FIG. 7. Photomicrograph from a stable hydroxyapatite-coated implant shows bone ingrowth across the initial gap and bone apposition on the implant (light green, basic fuchsin, grounded section; original magnification x6).

greater in the membranes around HA-coated implants compared with membranes around Ti-coated implants (p < 0.05). The hydroxyproline concentration in trabecular bone from the femoral condyles was 3.41%, corresponding to a collagen concentration of 25.5%. DISCUSSION

The present study demonstrates that controlled movements of implants in the range of 500 pm relative to the surrounding bone inhibit bony ingrowth and result in fibrous tissue membrane formation around implants, irrcspective of the type of surface coating. These findings are in agreement with Pilliar et al. (30), who found a fibrous tissue membrane in a dynamically loaded canine intramedullary model with an implant pore size of 50-400 pm. Cameron (8) demonstrated that gross motion of an unstable canine osteotomy site fixed with a porous coated staple prevented bony ingrowth. Ingrowth of bone was also precluded in a dynamically loaded hinged

J Orthop Res, Vol. 10, No. 2 , 1992

total knee with a pore size of 87-1 10 pm (14), and in a porous coated loaded Ti segmental prosthesis with a pore size of 270 pm, a fibrous layer was commonly demonstrated at the bone-implant interface (21). The membranes described in these reports (14,21,30) consisted of fibrous connective tissue except from one report (8) in which some areas with cartilage were found after an observation period of 16 weeks. The rationale for selecting 500-pm movement in this study was that recent reports have demonstrated relative bone-implant movements of 100600 pm when testing cementless tibia1 trays (5,33,39,47) and hip femoral components (46). In another study, using the same device, we selected 150 pm movements (33a). In the system used, the degree of implant movements was not controlled throughout the observation period, but a mechanical test of the spring in the cylinder was performed before and after implantation. The increased stiffness that developed during the implantation period probably is caused by the presence of fibrous tissue surrounding the springs. The device was designed

TITANIUM A N D H YDROXYAPATITE-COATED IMPLANTS

295

FIG. 8. Photomicrograph of membrane from the unstable titanium-coated implant shows fibrous connective tissue (toluidine blue at pH 5 ; original magnification ~ 2 5 0 ) .

to permit axial movements, but small rotatory and rocking movements of the implants could not be avoided. The maximal movements in these directions was limited due to the design of the device. All components in the spring system were manufactured from the same Ti alloy as the implant to avoid electrical potential and corrosion of the metal, which might affect the surrounding bone. The reaction on the device within the knee joint did not result in persistent intraarticular inflammation, as evidenced by synovial histology. The fact that only slight fibrillation of the tibia1 cartilage could be demonstrated was expected because the polyethylene plug was easily displaced in the proximal direction when the knee was loaded. An interesting observation in the present study was the stronger fibrous anchorage of unstable HAcoated implants compared with unstable Ti implants. This finding may be ascribed to the presence of fibrocartilage and increased collagen concentration in the membrane surrounding HA-coated implants. From these findings it is suggested that in

addition to the well-known osteoconductive effect (13,15,31,35,36,42), HA also seems to have the capacity to induce increased collagen synthesis from the cells during unstable mechanical conditions. In vitro studies have shown morphological changes and increased cell proliferation and DNA synthesis in cultured fibroblastic and human bone cells when HA particles were added to the culture (19). The HA particles were found to be phagocytized by the fibroblastic cells within 24 h, and the changes in fibroblastic behavior were suggested to result from release of Ca2+ions from the phagocytized HA particles because ionic calcium is a mediator and regulator implicated in many cellular functions (19). The increased collagen content found in the membranes around the unstable HA-coated implants in the present series is consistent with the hypothesis that some of the HA surface is dissolved (5,45)and that Ca2' and PO, ions are released and taken up by the fibroblasts, which might result in increased synthesis of collagen. Hence, different metabolic conditions adjacent to the implant may explain the

J

Orthop Res, Vol. 10, NO. 2 , 1992

296

K . SQBALLE ET AL.

FIG. 9. Photomicrograph of membrane from the unstable hydroxyapatite-coated implant shows islands of fibrocartilage with chondrocytes in lacunae (toluidine blue at pH 5; original magnification x250).

development of different types of connective tissues around HA and Ti-coated implants, because the mechanical environment was equal, at least initially. In the course of the implantation period, the development of fibrocartilage and increased collaHA T

45 7

8

40 h

8

v Q,

OI

20 )

zi0

m

E

e

00

252015-

m O

10-

E

Q,

3530-

Q,

--

90 1

gen synthesis in the membrane likely results in increased stability and reduced motion of HA-coated implants as demonstrated in the present study, because fibrocartilage probably has stronger biomechanical properties than fibrous connective tissue

TI

.H.-A

T

T

80 70 60

-

50

FIG. 10. Results from histomorphometry. A: Bony ingrowth. B: Fibrous tissue ingrowth. The arrows indicate unstable implants, and the two other bars represent stable implants. HA, hydroxyapatite; Ti, titanium.

-

0)

3

.-ccncn TI

UI

5-

10

HA

0-

n

I

J Orthop Rcs, V d . 10, N O . 2 , 1992

TITANIUM A N D HYDROXYAPATITE-COATED IMPLANTS

HA

20 15

TI

10

5 HA

0

-

4

e

FIG. 11. Results from histomorphometry of gap healing in the initial gap surrounding the implant. The arrows indicate unstable implants, and the two other bars represent stable implants. HA, hydroxyapatite; Ti, titanium.

does. This increased stability may be favorable for the further development of the membrane. Whether release of Ca2+ ions trigger the development of fibrocartilage is yet to be investigated. The process of bone ingrowth can be compared with fracture healing (40), in which the outcome depends on the initial mechanical stability of the fracture. Rigidly fixed fractures are healing via direct bone formation, whereas unstable mechanical conditions at the fracture site may result in healing via the cartilaginous phase followed by endochondral ossification (2,25). The final outcome of unstable mechanical conditions may be nonunion with presence of a fibrous tissue layer between the fracture ends, which may be analogous to the fibrous membrane developed around the unstable implants in the present study. The generally accepted hypothesis for these events is that differentiation into osteoblasts occurs where there is a high oxygen tension, whereas differentiation into chondrocytes occurs where the oxygen tension is low (4,23). This is in agreement with the fact that the region over the fracture, in which the cartilage is produced, is devoid of capillaries (2,20). Thus, as suggested by Page et al. (2Y), relative movements at the fracture site lead both to the inhibition of angiogenesis and to the subsequent differentiation of cartilage. Whether the fibrocartilage found around the unstable HA-coated implants in the present study will be replaced later by bone by endochondral ossification is yet unknown, but stud-

297

ies on that question are in progress. The finding that the anchorage of unstable HA-coated implants was equal to those of stable Ti implants was surprising, because fibrous tissue ingrowth was expected to have inferior mechanical properties compared with those of bony ingrowth. One explanation might be that the maturation of the new formed woven bone around the stable Ti implants was still incomplete and not stronger than the fibrocartilaginous tissue with a high collagen concentration around the unstable HA-coated implants because the strength of the interface depends on maturation of the mineralized bone matrix of the ingrown trabeculae (3). Our previous studies on unloaded stable implants surrounded by- a gap _ _ have demonstrated increased anchorage and bony ingrowth into HA-coated implants compared with porous coated Ti implants (35-37). This was confirmed by the present study concerning the stable implants that were unloaded to ensure rigid stability. Despite the fact that interface fibrous membranes adjacent to aseptically loosened cemented and noncemented prostheses from humans (10,12,18,26) have been shown to contain numerous macrophages, giant cells, and high levels of prostaglandin E, and collagenase (18), it has been suggested that ingrowth of fibrous tissue could be beneficial for energy absorption by providing better distribution of stresses (22,43,48). In a recent study, Longo et al. (27) demonstrated a stable fibrous tissue interface around press-fit carbon composite femoral stems in dogs, which obtained clinical results comparable to HA-coated stems that were anchored by bone apposition after a l-year observation period. They concluded that bone bonding of the implant is not essential for implant success but instead could lead to adverse bone remodeling. According to our results after 4 weeks of implantation, the fibrous membrane around Ti-coated implants had almost no capacity of fixation. However, the fibrocartilaginous membrane around unstable HA-coated implants was found to be significantly stronger and might be sufficient to dissipate stresses in total joint arthroplasties. The long-term course of fibrous membrane fixation and the reaction of surrounding bone is not known, but the different types of fibrous tissue surrounding unstable Ti and HAcoated implants might represent different phases in endochondral ossification, the fibrocartilaginous tissue around HA implants representing the most developed stage.

J Orthop Res, V d . 10, N o . 2 , 1992

K . SgBALLE ET A L .

298

CONCLUSIONS

Micromotion between bone and implant inhibited bony ingrowth and resulted in development of a fibrous membrane around both Ti and HA-coated implants. The fibrous membrane consisted predominantly of fibrocartilaginous tissue and had a higher concentration of collagen when the implant was coated with HA. The presence of fibrocartilage and increased collagen concentration may be responsible for the stronger fibrous anchorage of unstable HA-coated implants as compared with unstable Ti implants. The best anchorage and the greatest amount of bony ingrowth was obtained by the mechanically stable implant coated with HA. Acknowledgment: We thank Flemming Melsen, M.D., Ph.D., U n i v e r s i t y I n s t i t u t e of P a t h o l o g y , A a r h u s Amtssygehus, Denmark, for his valuable advice in histomorphometric analysis. Biomet Inc. U.S.A. kindly delivered t h e implants. Financial support w a s provided by: Danish Rheumatism Association, Danish Medical Research Council; Aarhus University Research Foundation, Danish Foundation for the Advancement of Medical Science, Direktflr Madsen og hustru Olga Madsens Fond, Ferd. & Ellen Hindsgauls F o n d ; and Institute of Experimental Clinical Research, University of Aarhus and K o n g Kristian D e n Tiendes Fond.

REFERENCES I Adrian MJ, Roy WE, Karpovic PV: Normal gait of the dog: an electrogoniometric study. Am J Vet Res 27:9&95, 1966 2 Ashhurst DE: The influence on mechanical conditions on the healing of experimental fractures in the rabbit: a microscopical 5tudy. Philos Trans R Soc Land [Bioll 313:271-302, 1986 3 Barth E, Rflnningen H , Solheim LF, Saethren B: Bone ingrowth into weight-bearing porous fiber titanium implants. Mechanical and biomechanical correlations. J Orfhop Res 4:356361, 1986 4. Bassett CAL, Hemnann I: Influence of oxygen concentration and mechanical factors on differentiation of connective tissues in vitro. Nature 190:460-461, 1961 5 . Beight J, Radin S, Cuckler J, Ducheyne P: Effect of solubility of calcium phosphate coatings on mechanical fixation of porous ingrown implants. In: Transactions of the 35th Annual Meeting, Orthopaedic Research Society, 1989, p 334 6. Branson PJ, Steege JW, Wixson RL, Lewis J, Stulberg SD: Rigidity of internal fixation with uncemented tibial knee implants. J Arthroplasty 4:21-26, 1989 7. Brunski JB: The influence of force, motion, and related quantities on the response of bone to implants. In: Noncemented Total Hip Arthroplasty, ed by RH Fitzgerald Jr, New York, Raven Press, 1988, pp 7-21 8. Cameron HU: The effect of movement on the bonding of porous metal on bone. J Biomed Mafer Res 7:301-3 11, 1973 9. Carlsson L, Rostlund T, Albrektsson B, Albrektsson T: Implant fixation improved by close fit. Acta Orthop Scnnd 59:272-275, 1988

J

Orthop Res, Vol. 10, N o . 2 , 1992

10. Collier JP, Mayor ME, Chae JC, Surprenant BS, Surprenant HP, Dauphinais LA: Macroscopic and microscopic evidence of prosthetic fixation with porous-coated materials. Clin Orthop 2351173-180, 1988 11. Cook SD, Barrack RL, Thomas KA, Haddad RJ: Quantitative analysis of tissue growth into human porous total hip components. J Arthroplasty 3:24%262, 1988 12. Cook SD, Thomas KA, Haddad RJ: Histological analysis of retrieved human porous-coated totdl joint components. Clin Orthop 234:9CL101, 1988 13. Cook SD, Thomas KA, Kay JF, Jarcho M: Hydroxyapatite coated titanium for orthopaedic implant application. Clin Orthop 232:225-243, 1988 14. Ducheyne P, De Meester P, Aernoudt E, Martens M; Mulier C: Influence of a functional dynamic loading on bone ingrowth into surface pores of orthopaedic implants. JBiomed Mater Res 11:811638, 1977 15. Ducheyne P, Hench LL, Kagdn A, Marfens M, Bursens A, Mulier C: Effect of hydroxyapatite impregnation on skeletal bonding of porous coated implants. J Biomed Mater Res 14:225-237. 1980 16. Engh GA, Bobyn JD, Petersen TL: Radiographic and histologic study of porous coated tibial component fixation in cementless total knee arthroplasty. Orthopedics I1:725-731, 1988 17. Eschenroeder HC, Jones LC, Hungerford DS: Biological ingrowth into a porous metal surface following established fibrous reaction. Transaciions of the 34th Annual Meeting, Orthopaedic Research Society, 1988, p 333 18. Goldring SR, Schiller AL, Roelke M, Rourke CM, O’Niell DA, Hams WH: The synovial-like membrane at the bone cement interface in loose total hip replacements and its proposed role in bone lysis. J Bone Joint Surg [Am] 65:575-584, 1983 19. Gregoire M, Orly I, Kerebel L-M, Kerebel B: In vitro effects of calcium phosphate biomaterials on fibroblastic cell behavior. Biol Cell 59:255-260, 1987 20. Ham AW: A histological study of the early phases of bone repair. J Bone Joint Surg [Am] 12:827-844, 1930 21. Heck DA, Nakajima 1, Kelly PJ, Chao EY: The effect of load alteration on the biological and biomechanical performance of a titanium fiber-metal segmental prosthesis. / B o n e Joint Surg [Am] 68:118-126, 1986 22. Hori RY, Lewis JL: Mechanicdl properties of the fibrous tissue found at the bone-cement interface following total joint replacement. J Biomed Mater Res 16:911-927, 1982 23. Jargiello DM, Caplan AI: The establishment of vascularderived microenvironments in the developing chick wing. Dev Bio197:364-374, 1983 24. Kimmel DB, Webster SSJ: Measurements of area, perimeter, and distance: details of data collection in bone histomorphometry. In: Bone Histomorphometry: Techniques and Interpretation, ed by RR Recker, Boca Raton, Florida, CRC Press, 1983, pp 89-108 25. Lane JM, Werntz JR: Biology of fracture healing. In: Fracture Healing (Bnstol-MyersiZimmer Orthopaedic Symposium), ed by JM Lane, New York, Edinburgh, London, Melbourne, Churchill Livingstone, 1987, pp 451-59 26. Lennox DW, Schofield BH, McDonald DF, Riley LH: A histologic comparison of aseptic loosening of cemented, press-fit, and biologic ingrowth prostheses. Clin Orfhop 225: 171-191, 1987 27. Longo JA. Magee FP, Mather SE, Yapp RA, Koeneman JB, Weinstein AM: Comparison of HA and non-HA coated carbon composite femoral stems. In: Trnnsactions of the 35th Annual Meeting, Orthopaedic Research Society, 1989, p 384 28. Neuman RE, Logan MA: The determination of collagen and elastin in tissues. J Biol Chem 186549-556, 1950 29. Page M, Hogg J, Ashhurst DE: The effect of mechanical

TITANIUM AND HYDROXYAPATITE-COATED IMPLANTS stability on the macromolecules of the connective tissue matrices produced during fracture healing. I , The collagens. Histachem J 18:251-256, 1986 30. Pilliar RM, Cameron HU, Welsh RP, Binnington AG: Radiographical and morphological studies of load-bearing porous surfaced structured implants. Clin Orthop 156:249-257,

40.

41.

1981 31. Rivero DP, Fox J. Skipor AK, Urban RM, Galante JO: Cal-

cium phosphate-coated porous titanium implants for enhanced skeletal fixation. J Biomed Mater Res 22:191-201, 1988 32. Ryd L: Micromotion in knee arthroplasty. A c f a Orthop Scand 57(suppl 220): 1-79, 1986 33. Shimagaki H, Bechtold JE, Sherman R, Gustilo RB: Initial

stability of tibial components in cementless total knee arthroplasty . In: Transactions of the 34th Annual Meeting, Orthopaedic Research Society, 1988, p 477 33a.Syiballe K, B-Rasmussen H, Hansen ES, Biinger C: Hydroxyapatite coating modifies implant membrane formation. Controlled micromotion studied in dogs. Acta Orthop Scand (in press) 34. Seballe K, Hansen ES, B-Kasmussen H, Hjortdal VE, Juhl GI, Pedersen CM, Hvid I, Biinger C : Fixation of titanium and hydroxyapatite coated implants in osteopenia. J Arf hroplasty 6:307-316, 1991 35. Seballe K, Hansen ES, B-Rasmussen H, Hjortdal VE, Juhl GI, Pedersen CM. Hvid I, Biinger C: Gap healing enhanced by hydroxyapatite coating. Clin Orthop 272:300-307, 1991 36. Syiballe K, Hansen ES, B-Rasmussen H, Pedersen CM, Biinger C: Hydroxyapatite coating enhances fixation of porous coated implants: a comparison in dogs between press fit and non-interference fit. Acta Orthop Scand 61:299-309, 1990 37. Syiballe K, Hansen ES, Brockstedt-Rasmussen H, Pedersen

42.

43.

44.

45.

46.

47.

48.

49.

CM, Biinger C: Bone graft incorporation around titanium and hydroxyapatite coated implants in dogs. Clin Orthop 2741282-293, 1992

38. Spector M, Harmon SL, Kreutner A: Characteristics of tissue growth into proplast and porous polyethylene implants in bone. J Biomed Mater Res 13577492, 1979 39. Strickland AB, Chan KH, Andnacchi TP, Miller J: The initial fixation of porous coated tibial components evaluated by the study of rigid body motion under static load. In: Trans-

SO.

299

actions of the 34th Annual Meeting, Orthopaedic Research Society, 1988, p 476 Sumner DR, Galante JO: Bone ingrowth. In: Surgery ofthe Musculoskeletal Syslem, 2nd ed, ed by CMcC Evarts, New York, Churchill-Livingstone, 1990, pp 151-176 Sumner DR, Kienapfel H , Galante JO: Metallic implants. In: Bone grafts and bone substitutes, ed by MB Habal, AH Reddi, Orlando, FL, W.B. Saunders (in press) Thomas KA, Kay JF, Cook SD, Jarcho M: The effect of surface macrotexture and hydroxyapatite coating on the mechanical strengths and histological profiles of titanium implant materials. J Biomed Mater Res 21:1395-1414, 1987 Tibrewal SO, Grant KA, Goodfellow JW: The radiolucent line beneath the tibial components of the Oxford meniscal knee. J Bone Joint Surg [Br] 66523-528, 1984 Uhthoff HK: Mechanical factors influencing the holding power of screws in compact bone. J Bone Joint Surg [Br] 55:633-639, 1973 van Blitterswijk CA, Grote JJ, Kuijpers W, Blok-van Hoek CJG, Daems WT: Bioreactions at the tissue/hydroxyapatite interface. Biomaterials 6:243-251, 1985 Vanderby R, Manley PA, Kohles SS, Belloli DM, McBeath AA: A micromotion comparison of cemented and porous ingrowth total hip replacements in a canine model. In: Transactions of the 35th Annual Meefing, Orthopaedic Research Society, 1989, p 577 Volz KG, Nisbet JK. Lee RW, McMurtry MG: The mechanical stability of various noncemented tibial components. Clin Orthop 226:3842, 1988 Walker PS, Onchi K, Kurosawa H, Rodger RF: Approaches to the interface problem in total joint arthroplasty. Clin Or!hop 182199-108, 1984 Woessner JF: Determination of hydroxyproline in connective tissues. In: The Methodology of Connective Tissue Research, ed by DA Hall, Oxford, Joynsson-Bruvvers, 1976, pp 227-233 Zalenski E, Jasty M, O’Conner DO, Page A, Krushell R, Bragdon C, Russotti G, Harris WH: Micromotion of poroussurfaced, cementless prostheses following 6 months of in vivo bone ingrowth in a canine model. In: Transactions of the 35th Annual Meeting, Orthopaedic Research Society, 1989, p 377

J

Orthop Res, Vol. 10, NO. 2 , 1992