- Email: [email protected]
The effect of titanium plasma-sprayed implants on trabecular bone healing in the goat
Biomaterials 19 (1998) 1093 — 1099
The effect of titanium plasma-sprayed implants on trabecular bone healing in the goat S. Vercaigne!,", J.G.C. Wolke!, I. Naert", J.A. Jansen!,* ! Unit Oral Function/Department of Biomaterials, University of Nijmegen, P.O. Box 9101, 6500 Nijmegen, The Netherlands " BIOMAT Research Group, Department of Prosthetic Dentistry, University of Leuven, School of Dentistry, Oral Pathology and Maxillofacial Surgery, Kapucijnenvoer 7, 3000 Leuven, Belgium Received 25 July 1997; accepted 25 January 1998
Abstract The bone response to different titanium plasma-sprayed implants was evaluated in a goat model. Therefore, beam-shaped implants were installed into the trabecular femoral condyles of 10 goats. These implants were provided with three different titanium plasma-sprayed coatings (Ti2, Ti3 and Ti4) with a Ra of 16.5, 21.4 and 37.9 lm, respectively. An Al O grit-blasted implant (Ti-un) 2 3 with a Ra of 4.7 lm was used as control. After an implantation period of 3 months, the implants were evaluated histologically and histomorphometrically. Only one implant (Ti3) was not recovered after the evaluation period. Light microscopy showed a limited amount of bone for the various implants. Most of the implants showed a different degree of fibrous tissue alternating with direct bone contact. Complete fibrous encapsulation of the implants was observed in some of the sections. No signs of delamination of the plasma-sprayed coating was visible. No significant difference in bone contact were measured between the different types of implants (P'0.05). Histomorphometrical analysis revealed significantly higher bone mass close to the implant (0—500 lm) for the Ti3, Ti4 and Ti-un implants placed in the medial femoral condyle and the Ti4 implants placed in the lateral condyle. At distance (500—1500 lm), no difference in bone mass measurements between the different implants was observed (P'0.05). ( 1998 Published by Elsevier Science Ltd. All rights reserved Keywords: Implants; Titanium; Plasma-sprayed coatings; Surface topography
1. Introduction The clinical success of skeletal implants is among others related to the biological fixation of the implant by bone apposition. Several implant as well as non-implant related factors can influence this process. Important implant-related factors to achieve an adequate bone— implant interface are: implant design, implant material and surface quality. The surface quality of an implant will depend on the chemical, physical, mechanical and topographical properties of the surface. The surface topography is related to the degree of roughness of the surface and the orientation of the surface irregularities. Relevant non-implant parameters are: surgical technique, patient health condition, implant loading conditions and implant location (e.g. cortical versus trabecular bone) [1].
* Corresponding author. Tel.: #31-243614920/#31-243614006; fax: #31-243541971; e-mail: [email protected].
Various reports suggest that surface roughness plays a significant role in the final bone integration of an implant. In comparative studies [2—7] between smooth and roughened surfaces, poor experimental results were achieved for the smooth surfaced implants. No anchorage and stability of the ingrowing bone could be established, resulting in fibrous encapsulation of the smooth implants. In contrast, roughened implants showed more bone apposition and a better mechanical interlocking. Besides this mechanical influence, surface roughness is also suggested to increase directly the amount of bone formation along the interface due to surface enlargement [3]. Additionally, local bone conditions can attribute to the etiology of implant loss. In view of this, the overall bone density and thickness of the cortical layer are considered to be relevant parameters [8]. For example, the low density and limited corticalization of trabecular bone are assumed to result in a less primary stability of the implant during the healing period. Less primary stability
0142-9612/98/$19.00 ( 1998 Published by Elsevier Science Ltd. All rights reserved. PII S 0 1 4 2 - 9 6 1 2 ( 9 8 ) 0 0 0 3 9 - 8
1094
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
results in continuous micromovement of the implant during the healing period. This causes soft tissue encapsulation and finally results in implant failure [9—13]. To reduce micromovement and to increase primary bone apposition, the use of roughened implants is recommended [14, 15]. Considering the above mentioned, the aim of the study was to test the hypothesis that surface microroughness as created by titanium plasma-spraying enhances the biological performance of implants inserted in trabecular bone.
2. Materials and methods 2.1. Implant materials For the experiment, 80 beam-shaped TiAl V implants 6 4 were used. The implants measured 8]10]4 mm and were grit-blasted resulting in a Ra of 4.7-lm. They were cleaned ultrasonically in propanol and dried at 50°C. Subsequently, these implants were left uncoated (Ti-un) or a commercially titanium (cp) coating was applied using a plasma-spray process. Table 1 shows the various coatings that were deposited on the substrates. To create various roughnesses, powders with different particle size were used. To compensate for the thickness of the coatings, the implants that were used for titanium plasma-spraying varied within a range of 3.4—3.8 mm. By this, the final diameter of all implants was 4 mm$200 lm. The surface topography of the various implants was characterized by surface profilometry (Tally-Hobson) and scanning electron microscopy (SEM) (Table 1 and Fig. 1). The chemical composition and purity of the coatings was characterized by electron dispersive spectroscopy (EDS) and X-ray diffraction (XRD). The X-ray diffraction and EDS pattern revealed Al O on the 2 3 surface of the grit-blasted implants. Before surgery, the implants were cleaned ultrasonically in 100% ethanol to remove any loose particles and dried at 50°C. Afterwards the implants were autoclaved. 2.2. Experimental design and surgery Ten healthy adult, female Saanen goats, weighing about 60 kg were used in this experiment. The selected animals were tested to ensure that they were Caprine Arthritis Encephalitis free. The animals were housed in a stable. The installation was performed under general anesthesia. The anesthesia was induced by an intravenous injection of pentobarbital and maintained by ethrane 2—3% through a constant volume ventilator, administered through an endo-tracheal tube. The goats were connected to a heart monitor. To reduce the risk of peri-operative infection, the goats were treated according to the following doses of antibiotics: during the opera-
Table 1 Specifications of the four surfaces of the used implants Coating type
Mean powder particle size (lm)
Ra-value (lm)
Coating thickness (lm)
Ti2 Ti3 Ti4 Ti-un
22 75 150
16.2 21.4 37.9 4.7
200—300 300—400 400—600
tion, Albipen' (MycoPharma, De Bilt, The Netherlands) 15%, 3 ml per 50 kg s.c.; one day and three days after the operation, Albipen' LA, 7.5 ml per 50 kg s.c. To place the implants, the animals were immobilized on their back, the hindlimbs shaved, washed and disinfected with povidine-iodine. A longitudinal incision was made on the medial and lateral surface of the left and the right femoral condyles. After exposure of the condyles, pilot holes were drilled. The holes were gradually widened until the final size to harbour the implant. The bone preparation was performed using a very gentle technique and profuse external cooling. After drilling, the implants were placed as tightly as possible (Fig. 2). A primary stability for all implants was reached each time. Following the installment of the implants, the soft tissues were closed in separate layers using resorbable sutures. In this way, each femur received 2 implants (one in the medial and one in the lateral condyle), resulting in 4 implants in each goat. A total of 40 implants was placed: 10 Ti2, 10 Ti3, 10 Ti4 and 10 Ti-un. The implants were inserted following a balanced split plot design. Half of the goats received in vivo fluorochromes (Table 2). These markers were administered at timed intervals. The implants were left in place for 3 months after implantation. 2.3. Processing of the implants At the end of the predetermined period, the animals were killed by an overdose of Narcoved' (Apharmo, Arnheim, The Netherlands). Afterwards, the femoral condyles together with the implants were excised immediately. 2.4. Histological procedures The excess tissue of the excised femoral condyles was removed. The specimens were fixed in 4% buffered formalin solution at pH 7.2—7.4, for one week. Following dehydration by series of ethanol, the tissue blocks were embedded in methylmethacrylate. After polymerization, nondecalcified thin (10 lm) sections were prepared perpendicular to the long axis of the implant, using a diamond-blade microtome technique [16]. The sections were stained with basic fuchsin and methylene blue
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
1095
Fig. 1. Scanning electron micrograph showing the surface topography of the various implants. (A) Titanium plasma-sprayed implant Ti2 (Ra"16.5 lm); (B) Titanium plasma-sprayed implant Ti3 (Ra" 21.4 lm; (C) Titanium plasma-sprayed implant Ti4 (Ra"37.9 lm); (D) grit-blasted implant (Ra"4.7 lm).
for light microscopical evaluation. The histological specimens that were retrieved from the animals which received fluorochromes, were prepared in the same way as described above. After sectioning, reflectant fluorescence microscopy was used for the evaluation. 2.5. Histological and histomorphometrical evaluation
Fig. 2. Photograph of an implant placed into the femoral condyle of a goat.
A light microscope (Leica BV, Rijswijk, The Netherlands) was used for the histological evaluation. In addition, an image analysis technique (Technical Command Language' developed by TNO, Delft, The Netherlands) was used for histomorphometrical evaluation. The following quantitative parameters were assessed: (A) Percentage of bone contact at the interface. The amount of interfacial bone contact was defined as the
1096
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
Table 2 Schedule indicating the time intervals at which five of the ten goats received subcutaneous fluorochrome markers to indicate the remodeling activity after implant installation Weeks before sacrifice
Fluorochrome
Colour
Dose (mg/kg)
11 9 6 1
Tetracycline Calceine Alizerin-complexon Tetracycline
Yellow Green Red Yellow
25 25 25 25
2.6. Statistical analysis All measurements were statistically evaluated using a two-way analysis of variance (ANOVA) and a multiple comparison procedure (Tukey). In addition a simple linear regression test was used to reveal the existence of a correlation between roughness and the percentage of bone contact.
3. Results During the experimental period, two goats had to be sacrificed due to a broken leg. One additional goat was operated, so that the final number of experimental animals was nine. The rest of the animals remained in good health. At the end of the healing period, one implant (Ti3) was lost. No clinical signs of inflammation or adverse tissue reaction were observed around the remaining implants. 3.1. Histology
Fig. 3. Schematic drawing of the histomorphometrical measuring areas: (A) % bone contact; (B) amount of bone (lm2).
percentage of implant length at which there is direct bone-to-implant contact without intervening soft tissue layers. The amount of bone was measured over a distance of 4 mm on the right and the left side of the implant respectively, starting from the caudal part (Fig. 3). (B) ¹he bone mass around the implants. The bone mass in rectangular regions along the implant surface was determined. Three regions of interest (ROI), i.e. ROI 500, ROI 1000 and ROI 1500 were marked (Fig. 3). All ROIs had a width of 0.5 mm and a length of 2.8 mm, resulting in a total area of 1.4 mm2. ROI 500 was defined as a rectangular area in direct contact with the implant. Area ROI 1000 was placed at 500—1000 lm distance from the implant surface and area ROI 1500 at the 1000—1500 lm distance. The bone mass was quantified in lm2]103. All quantitative measurements were performed for three at randomly chosen sections of each implant. The presented data are the average of three measurements. In addition, the results presented are divided in implants inserted into the medial and the lateral femoral condyles respectively.
Histologic assessment revealed that for all prepared specimens the most coronal part of the implants was mostly surrounded with fibrous tissue (Fig. 4). The remaining (apical) part of the implant showed a limited amount of bone. Most of the implants revealed varying areas of direct bone contact and/or fibrous tissue. In some of the sections, we even observed that only fibrous tissue was interposed between the bone and the implant. When newly formed bone was present in the interface, this bone had a trabecular structure. Further, we noticed ingrowth of bone into the porosity of the Ti4 coating (Fig. 5). The histological evaluation did not demonstrate a difference in bone response between the various implants (Fig. 6). In addition, the titanium plasma-sprayed coatings appeared to be stable. No signs of delamination or loosened particles were observed. The fluorochrome labeling demonstrated an active bone deposition and remodeling process. Apparently, bone formation after implant insertion started from the margin of the drill towards the implant. No obvious differences in bone activity around the various implants were found. 3.2. Histomorphometrical evaluation 3.2.1. Bone contact The results of the bone contact measurements for the various implants and implantation sites are listed in Tables 3 and 4. Statistical analysis showed no difference in percentage of bone contact (P'0.05) for the various implants. In addition, no significant difference (P'0.05) in bone contact between implants inserted into the
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
Fig. 4. Light micrograph of a Ti4-coated implant. The most coronal part of all the implants is surrounded with fibrous tissue. Original magnification]11.2. Bar"890 lm.
1097
Fig. 6. Histological appearance of Ti3 (A) and Ti-un (B). The trabecular structure of the bone around the implants can be recognized. Fibrous tissue alternates with bony attachment. Original magnification]70. Bar"140 lm.
Table 3 Percentages of the measured bone—implant contact over 4 mm on the left and the right side of the implant. Three histological sections per implant were measured Material
Mean % bone contact$ standard deviation
Ti2
9.9$11.9 (n"9) 11.3$13.1 (n"8) 14.3$14.1 (n"9) 14.9$16.9 (n"9)
Ti3 Ti4 Ti-un Fig. 5. Light micrograph of a Ti4 coating showing ingrowth of newly formed bone into the undercuts of the coating. Original magnification]140. Bar"70 lm.
medial and those inserted into the lateral condyles were found. Finally, simple linear regression showed that no relationship existed between roughness and percentage of bone contact (r"0.01).
3.2.2. Bone mass measurements Table 5 shows the results of the bone mass measurements in the different regions of interest. Statistical analysis showed no difference between the different types of implants. On the other hand, a significant difference existed between the different regions of the Ti3, Ti4 and Ti-un implants placed in the medial condyle and the Ti4
1098
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
implants placed in the lateral condyle. ROI 500 of these implants always revealed a higher bone amount than ROI 1000 and ROI 1500 (P(0.05).
4. Discussion and conclusions The results of the current study are in clear contrast with the findings of a lot of other investigators [2—7], who showed a positive effect of surface microroughness on the integration of implants. Several explanations can be given for this discrepancy. First, there is the used surgical procedure. The beamshaped design of our implants did not allow to make standarized drill holes. For example, to create the bone cavity we had to use a conically shaped drill. This resulted always in a gap that was slightly wider at the top than at the bottom of the cavity. Increased gap size and reduced fit can have a negative influence on the integration of an implant [17, 18]. Second, the surface enlargement, as created by very rough coatings, can result in an increased ion release, which can have a negative effect on bone formation [19—22]. We suppose that the high degree of roughness of our grit-blasted and plasma-spray-coated implants in combination with ion release resulted in the observed low percentage of bone contact. Additional evaluations are recommended to confirm this theory. Further, our XRD and EDS measurements revealed the presence of Al O on the surface of our grit-blasted 2 3 implants. In agreement with Gross et al. [23], we suppose that the Al O , as remained from the blasting procedure, 2 3 Table 4 Means and standard deviations for the various implants for the 2 implantation sites. The number of the measured histological sections is shown in parantheses
Medial Lateral
Ti2
Ti3
Ti4
Ti-un
12.5$12 (n"5) 6.8$12.7 (n"4)
20.2$12.8 (n"4) 2.4$4.9 (n"4)
13.9$8.6 (n"4) 14.6$18.9 (n"5)
11.2$13.4 (n"4) 18$20.3 (n"5)
have a negative effect on the bone response. They reported that the release of Al O from an implant sur2 3 face can inhibit the final mineralization of bone. More recently, also Johansson et al. [24] found a deteriorating influence of aluminium on bone formation. On the other hand, Wennerberg et al. [5] could never demonstrate any effect of Al O blasting procedure on the bone response. 2 3 Although, their auger electron spectroscopy (AES) evaluations revealed a higher Al O concentration on 2 3 the surface of the Al O blasted implants. Probably, this 2 3 discrepancy in results is due to a still unknown threshold value for ion release. Further in our histomorphometrical evaluation, we did discern between implants inserted into the medial and into the lateral femoral condyle. This was done, because goats have X-shaped legs. As a consequence, differences in loading condition exist within the femoral condyle which can play an important role in the bone response between implants placed in the medial or lateral condyle [25]. This is confirmed by the studies of Caulier et al. [26] and Corten et al. [27]. In our study such differences were not found. The reason has to be the variations in the implant positioning. The implants could not always be oriented in a similar inclination with respect to the longitudional axis of the femur. This results in different stress transfer patterns for the implants, which also explains the variance in bone contact measurements. For future studies the use of cylindrical implants, which allow a more standardized positioning, is recommended. Although, we have to notice that the results were not completely consistent, our histomorphometrical data suggest that the medially placed implants showed more bone mass close to the implant than the laterally placed implants. We suppose that this is due to a reduced load transmission [28]. The continuous micromovement of a less incorporated implant results in a persistent stimulation of the surrounding bone. As compensation for the low percentage of bone contact, the bone will respond with an increased bone turnover, resulting in an increased bone mass. This theory does not completely corroborate with the generally accepted phenomenon of stress shielding, where bone atrophy occurs at places where too high forces act on an implant or inadequate
Table 5 Measured amount of bone in lm2]103. Three histological sections per implant were measured Coating
Ti2 Ti3 Ti4 Ti-un
Medial
Lateral
ROI 500
ROI 1000
ROI 1500
ROI 500
ROI 1000
ROI 1500
350.0$119.4 415.0$298.7 332.5$70.4 312.5$165.6
182.0$73.3 270.0$115.2 230.0$74.8 152.5$42.7
130.0$52.9 165.0$78.5 127.5$49.9 75$50.7
177.0$149.8 130.0$143.6 400.0$187.7 224.0$249.5
165.0$34.2 145.0$68.5 172$31.9 122.0$76.6
77.5$23.6 85.0$55.6 98.0$14.8 82.0$30.3
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
stress transfer and distribution is present. On the other hand, another important factor in stress shielding is time. Our implantation time was relatively short. We assume that when the implants are left in place for longer periods, also reduction in bone mass will occur at the bone— implant interface. The laterally placed implants, except Ti4, showed no significant difference. The forces acting on the medially placed implants, resulting in a higher mass close to the implants, were not present on the laterally placed implants. This confirms the hypothesis that the biological response of those implants is the result of the mechanical influence. In conclusion, no difference in bone behaviour was observed in our study design between grit-blasted and various titanium plasma-sprayed implants. Therefore, besides mechanical interlocking, the biological efficacy of very rough surfaces remains unconclusive. Acknowledgements The authors thank J.P.C.M. van der Waerden and A.F.M. Leijdekkers-Govers for their participation and the analytical support, H.G. Schaeken for the management of the XRD and J.E.G. Hulshoff for the surgical assistance during the animal experiments.
[9] [10]
[11]
[12]
[13]
[14]
[15]
[16]
[17] [18]
References [1] Albrektsson T, Bras nemark P-I, Hansson H-A, Lindstro¨m J. Osseointegrated titanium implants. Requirements for ensuring a long-lasting, direct bone-to-implant anchorage in man. Acta Orthop Scand 1981;52:155. [2] Buser D, Schenk RK. Influence of surface characteristics on bone integration of titanium implants. A histomorphometric study in miniature pigs. J Biomed Mat Res 1991;25:889—902. [3] Gotfredsen K, Nimb L, Hjo¨rting-Hansen E, Jensen JS, Holme´n, A. Histomorphometric and removal torque analysis for TiO 2 blasted titanium implants. Clin Oral Impl Res 1992;3:77—84. [4] Wennerberg A, Albrektsson T, Lausmaa J. Torque and histomorphometric evaluation of c.p. titanium screws blasted with 25- and 75-lm-sized particles Al O . J Biomed Mat Res 1996;30:251—60. 2 3 [5] Wennerberg A, Albrektsson T, Johansson C, Andersson B. Experimental study of turned and grit-blasted screw-shaped implants with special emphasis on effects of blasting material and surface topography. Biomaterials 1996;17:15—22. [6] Wennerberg A, Albrektsson T, Andersson B, Krol JJ. A histomorphometric and removal torque of screw-shaped titanium implants with three different surface topographies. Clin Oral Impl Res 1995;6:24—30. [7] Wennerberg A, Albrektsson T, Andersson B. An animal study of c.p. titanium screws with different surface topographies. J Mater Sci Mater Med 1995;6:302—9. [8] Lekholm U, Zarb GA. Patient selection and preparation. In: Bras nemark P-I, Zarb G, Albrektsson T, editors. Tissue integrated
[19] [20]
[21]
[22] [23]
[24]
[25]
[26]
[27]
[28]
1099
prosthesis: osseointegration in clinical dentistry. Chicago Quintessence Publ. Co. Inc., 1985. Jaffin RA, Berman CL. The excessive loss of Bras nemark fixtures in type IV bone: a 5-year analysis. J Periodontol 1991;62:2—4. Cune MS, de Putter C, Hoogstraten J. Treatment outcome with implant-retained overdentures: Part I—clinical findings and predictability of clinical treatment outcome. J Prosthet Dent 1996;72: 144—51. Cune MS, de Putter C, Hoogstraten J. Treatment outcome with implant-retained overdentures: Part II—patient statisfaction and predictability of subjective treatment outcome. J Prosthet Dent 1996;72:152—8. Hutton JE, Heath MR, Chai JY et al. Factors related to success and failure rates at 3-year follow-up in a multicenter study of overdentures supported by Bras nemark implants. J Oral Maxillofac Implants 1995;10:33—42. Cameron HU, Pilliar RM, Macnab I. The effect of movement on the bonding of porous metal to bone. J Biomed Mat Res 1973;7: 301—11. Maniatopoulos C, Pilliar RM, Smith DC. Threaded versus porous-surfaced designs for implant stabilization in bone-endodontic implant model. J Biomed Mat Res 1986;20:1309—33. Søballe K. Hydroxyapatite ceramic coating for bone implant fixation. Mechanical and histological studies in dogs. Acta Orthoped Scand 1993;64(Suppl):255. Van Der Lubbe HBM, Klein CPAT, de Groot K. A simple method for preparing thin (10 lm) histological sections of undecalcified plastic embedded bone with implants. Stain Technol 1988;63:171—7. Cameron HU, Pilliar RM, Macnab I. The rate of bone ingrowth into porous metal. J Biomed Mat Res 1976;10:295—9. Sandborn PM, Cook SD, Spires WP, Kesters MA. Tissue response to porous- coated implants lacking initial bone apposition. J Arthroplasty 1989;3(4):337—46. Black J. Does corrosion matter? J Bone J Surg 1988;70b:517—20. Thompson GJ, Puleo DA. Effects of sublethal metal ion concentrations on osteogenic cells derived from bone marrow stromal cells. J Appl Biomaterials 1995;6:249—58. Osborn JF, Willick P, Neenen N. The release of titanium into human bone from a titanium implant coated with plasma-sprayed titanium. Adv Biomater 1990;9:75—80. Wennerberg A. On surface roughness and implant incorporation. PhD Thesis, Go¨teborg, 1996. Gross U, Strunz V. The interface of various glasses and glass ceramics with a bony implantation bed. J Biomed Mat Res 1985;19:251—71. Johansson C, Albrektsson T, Thomsen P, Sennerby L, Lodding A, Odelius H. Tissue reaction to titanium-6aluminium-4vanadium alloy. Eur J Exp Musculoskel Res 1992;1:161—9. Heimke G, Griss P, Werner E, Jentschura G. The effects of mechanical factors on biocompatibility tests. J Biomed Engng 1981;3:209—13. Caulier H, van der Waerden JPCM, Paquay YCGJ et al. Effect of calcium phosphate (Ca-P) coatings on trabecular bone response: a histological study. J Biomed Mat Res 1995;29:1061—9. Corten FGA, Caulier H, van der Waerden JPCM, Kalk W, Corstens FHM, Jansen JA. The assessment of bone surrounding implants in goats: ex vivo measurements by dual X-ray absorptiometry. Biomaterials 1997;6:495—501. Williams D. Concise encyclopedia of medical and dental materials. Oxford: Pergamon Press, 2nd ed., 1990:51—9.
The effect of titanium plasma-sprayed implants on trabecular bone healing in the goat S. Vercaigne!,", J.G.C. Wolke!, I. Naert", J.A. Jansen!,* ! Unit Oral Function/Department of Biomaterials, University of Nijmegen, P.O. Box 9101, 6500 Nijmegen, The Netherlands " BIOMAT Research Group, Department of Prosthetic Dentistry, University of Leuven, School of Dentistry, Oral Pathology and Maxillofacial Surgery, Kapucijnenvoer 7, 3000 Leuven, Belgium Received 25 July 1997; accepted 25 January 1998
Abstract The bone response to different titanium plasma-sprayed implants was evaluated in a goat model. Therefore, beam-shaped implants were installed into the trabecular femoral condyles of 10 goats. These implants were provided with three different titanium plasma-sprayed coatings (Ti2, Ti3 and Ti4) with a Ra of 16.5, 21.4 and 37.9 lm, respectively. An Al O grit-blasted implant (Ti-un) 2 3 with a Ra of 4.7 lm was used as control. After an implantation period of 3 months, the implants were evaluated histologically and histomorphometrically. Only one implant (Ti3) was not recovered after the evaluation period. Light microscopy showed a limited amount of bone for the various implants. Most of the implants showed a different degree of fibrous tissue alternating with direct bone contact. Complete fibrous encapsulation of the implants was observed in some of the sections. No signs of delamination of the plasma-sprayed coating was visible. No significant difference in bone contact were measured between the different types of implants (P'0.05). Histomorphometrical analysis revealed significantly higher bone mass close to the implant (0—500 lm) for the Ti3, Ti4 and Ti-un implants placed in the medial femoral condyle and the Ti4 implants placed in the lateral condyle. At distance (500—1500 lm), no difference in bone mass measurements between the different implants was observed (P'0.05). ( 1998 Published by Elsevier Science Ltd. All rights reserved Keywords: Implants; Titanium; Plasma-sprayed coatings; Surface topography
1. Introduction The clinical success of skeletal implants is among others related to the biological fixation of the implant by bone apposition. Several implant as well as non-implant related factors can influence this process. Important implant-related factors to achieve an adequate bone— implant interface are: implant design, implant material and surface quality. The surface quality of an implant will depend on the chemical, physical, mechanical and topographical properties of the surface. The surface topography is related to the degree of roughness of the surface and the orientation of the surface irregularities. Relevant non-implant parameters are: surgical technique, patient health condition, implant loading conditions and implant location (e.g. cortical versus trabecular bone) [1].
* Corresponding author. Tel.: #31-243614920/#31-243614006; fax: #31-243541971; e-mail: [email protected].
Various reports suggest that surface roughness plays a significant role in the final bone integration of an implant. In comparative studies [2—7] between smooth and roughened surfaces, poor experimental results were achieved for the smooth surfaced implants. No anchorage and stability of the ingrowing bone could be established, resulting in fibrous encapsulation of the smooth implants. In contrast, roughened implants showed more bone apposition and a better mechanical interlocking. Besides this mechanical influence, surface roughness is also suggested to increase directly the amount of bone formation along the interface due to surface enlargement [3]. Additionally, local bone conditions can attribute to the etiology of implant loss. In view of this, the overall bone density and thickness of the cortical layer are considered to be relevant parameters [8]. For example, the low density and limited corticalization of trabecular bone are assumed to result in a less primary stability of the implant during the healing period. Less primary stability
0142-9612/98/$19.00 ( 1998 Published by Elsevier Science Ltd. All rights reserved. PII S 0 1 4 2 - 9 6 1 2 ( 9 8 ) 0 0 0 3 9 - 8
1094
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
results in continuous micromovement of the implant during the healing period. This causes soft tissue encapsulation and finally results in implant failure [9—13]. To reduce micromovement and to increase primary bone apposition, the use of roughened implants is recommended [14, 15]. Considering the above mentioned, the aim of the study was to test the hypothesis that surface microroughness as created by titanium plasma-spraying enhances the biological performance of implants inserted in trabecular bone.
2. Materials and methods 2.1. Implant materials For the experiment, 80 beam-shaped TiAl V implants 6 4 were used. The implants measured 8]10]4 mm and were grit-blasted resulting in a Ra of 4.7-lm. They were cleaned ultrasonically in propanol and dried at 50°C. Subsequently, these implants were left uncoated (Ti-un) or a commercially titanium (cp) coating was applied using a plasma-spray process. Table 1 shows the various coatings that were deposited on the substrates. To create various roughnesses, powders with different particle size were used. To compensate for the thickness of the coatings, the implants that were used for titanium plasma-spraying varied within a range of 3.4—3.8 mm. By this, the final diameter of all implants was 4 mm$200 lm. The surface topography of the various implants was characterized by surface profilometry (Tally-Hobson) and scanning electron microscopy (SEM) (Table 1 and Fig. 1). The chemical composition and purity of the coatings was characterized by electron dispersive spectroscopy (EDS) and X-ray diffraction (XRD). The X-ray diffraction and EDS pattern revealed Al O on the 2 3 surface of the grit-blasted implants. Before surgery, the implants were cleaned ultrasonically in 100% ethanol to remove any loose particles and dried at 50°C. Afterwards the implants were autoclaved. 2.2. Experimental design and surgery Ten healthy adult, female Saanen goats, weighing about 60 kg were used in this experiment. The selected animals were tested to ensure that they were Caprine Arthritis Encephalitis free. The animals were housed in a stable. The installation was performed under general anesthesia. The anesthesia was induced by an intravenous injection of pentobarbital and maintained by ethrane 2—3% through a constant volume ventilator, administered through an endo-tracheal tube. The goats were connected to a heart monitor. To reduce the risk of peri-operative infection, the goats were treated according to the following doses of antibiotics: during the opera-
Table 1 Specifications of the four surfaces of the used implants Coating type
Mean powder particle size (lm)
Ra-value (lm)
Coating thickness (lm)
Ti2 Ti3 Ti4 Ti-un
22 75 150
16.2 21.4 37.9 4.7
200—300 300—400 400—600
tion, Albipen' (MycoPharma, De Bilt, The Netherlands) 15%, 3 ml per 50 kg s.c.; one day and three days after the operation, Albipen' LA, 7.5 ml per 50 kg s.c. To place the implants, the animals were immobilized on their back, the hindlimbs shaved, washed and disinfected with povidine-iodine. A longitudinal incision was made on the medial and lateral surface of the left and the right femoral condyles. After exposure of the condyles, pilot holes were drilled. The holes were gradually widened until the final size to harbour the implant. The bone preparation was performed using a very gentle technique and profuse external cooling. After drilling, the implants were placed as tightly as possible (Fig. 2). A primary stability for all implants was reached each time. Following the installment of the implants, the soft tissues were closed in separate layers using resorbable sutures. In this way, each femur received 2 implants (one in the medial and one in the lateral condyle), resulting in 4 implants in each goat. A total of 40 implants was placed: 10 Ti2, 10 Ti3, 10 Ti4 and 10 Ti-un. The implants were inserted following a balanced split plot design. Half of the goats received in vivo fluorochromes (Table 2). These markers were administered at timed intervals. The implants were left in place for 3 months after implantation. 2.3. Processing of the implants At the end of the predetermined period, the animals were killed by an overdose of Narcoved' (Apharmo, Arnheim, The Netherlands). Afterwards, the femoral condyles together with the implants were excised immediately. 2.4. Histological procedures The excess tissue of the excised femoral condyles was removed. The specimens were fixed in 4% buffered formalin solution at pH 7.2—7.4, for one week. Following dehydration by series of ethanol, the tissue blocks were embedded in methylmethacrylate. After polymerization, nondecalcified thin (10 lm) sections were prepared perpendicular to the long axis of the implant, using a diamond-blade microtome technique [16]. The sections were stained with basic fuchsin and methylene blue
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
1095
Fig. 1. Scanning electron micrograph showing the surface topography of the various implants. (A) Titanium plasma-sprayed implant Ti2 (Ra"16.5 lm); (B) Titanium plasma-sprayed implant Ti3 (Ra" 21.4 lm; (C) Titanium plasma-sprayed implant Ti4 (Ra"37.9 lm); (D) grit-blasted implant (Ra"4.7 lm).
for light microscopical evaluation. The histological specimens that were retrieved from the animals which received fluorochromes, were prepared in the same way as described above. After sectioning, reflectant fluorescence microscopy was used for the evaluation. 2.5. Histological and histomorphometrical evaluation
Fig. 2. Photograph of an implant placed into the femoral condyle of a goat.
A light microscope (Leica BV, Rijswijk, The Netherlands) was used for the histological evaluation. In addition, an image analysis technique (Technical Command Language' developed by TNO, Delft, The Netherlands) was used for histomorphometrical evaluation. The following quantitative parameters were assessed: (A) Percentage of bone contact at the interface. The amount of interfacial bone contact was defined as the
1096
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
Table 2 Schedule indicating the time intervals at which five of the ten goats received subcutaneous fluorochrome markers to indicate the remodeling activity after implant installation Weeks before sacrifice
Fluorochrome
Colour
Dose (mg/kg)
11 9 6 1
Tetracycline Calceine Alizerin-complexon Tetracycline
Yellow Green Red Yellow
25 25 25 25
2.6. Statistical analysis All measurements were statistically evaluated using a two-way analysis of variance (ANOVA) and a multiple comparison procedure (Tukey). In addition a simple linear regression test was used to reveal the existence of a correlation between roughness and the percentage of bone contact.
3. Results During the experimental period, two goats had to be sacrificed due to a broken leg. One additional goat was operated, so that the final number of experimental animals was nine. The rest of the animals remained in good health. At the end of the healing period, one implant (Ti3) was lost. No clinical signs of inflammation or adverse tissue reaction were observed around the remaining implants. 3.1. Histology
Fig. 3. Schematic drawing of the histomorphometrical measuring areas: (A) % bone contact; (B) amount of bone (lm2).
percentage of implant length at which there is direct bone-to-implant contact without intervening soft tissue layers. The amount of bone was measured over a distance of 4 mm on the right and the left side of the implant respectively, starting from the caudal part (Fig. 3). (B) ¹he bone mass around the implants. The bone mass in rectangular regions along the implant surface was determined. Three regions of interest (ROI), i.e. ROI 500, ROI 1000 and ROI 1500 were marked (Fig. 3). All ROIs had a width of 0.5 mm and a length of 2.8 mm, resulting in a total area of 1.4 mm2. ROI 500 was defined as a rectangular area in direct contact with the implant. Area ROI 1000 was placed at 500—1000 lm distance from the implant surface and area ROI 1500 at the 1000—1500 lm distance. The bone mass was quantified in lm2]103. All quantitative measurements were performed for three at randomly chosen sections of each implant. The presented data are the average of three measurements. In addition, the results presented are divided in implants inserted into the medial and the lateral femoral condyles respectively.
Histologic assessment revealed that for all prepared specimens the most coronal part of the implants was mostly surrounded with fibrous tissue (Fig. 4). The remaining (apical) part of the implant showed a limited amount of bone. Most of the implants revealed varying areas of direct bone contact and/or fibrous tissue. In some of the sections, we even observed that only fibrous tissue was interposed between the bone and the implant. When newly formed bone was present in the interface, this bone had a trabecular structure. Further, we noticed ingrowth of bone into the porosity of the Ti4 coating (Fig. 5). The histological evaluation did not demonstrate a difference in bone response between the various implants (Fig. 6). In addition, the titanium plasma-sprayed coatings appeared to be stable. No signs of delamination or loosened particles were observed. The fluorochrome labeling demonstrated an active bone deposition and remodeling process. Apparently, bone formation after implant insertion started from the margin of the drill towards the implant. No obvious differences in bone activity around the various implants were found. 3.2. Histomorphometrical evaluation 3.2.1. Bone contact The results of the bone contact measurements for the various implants and implantation sites are listed in Tables 3 and 4. Statistical analysis showed no difference in percentage of bone contact (P'0.05) for the various implants. In addition, no significant difference (P'0.05) in bone contact between implants inserted into the
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
Fig. 4. Light micrograph of a Ti4-coated implant. The most coronal part of all the implants is surrounded with fibrous tissue. Original magnification]11.2. Bar"890 lm.
1097
Fig. 6. Histological appearance of Ti3 (A) and Ti-un (B). The trabecular structure of the bone around the implants can be recognized. Fibrous tissue alternates with bony attachment. Original magnification]70. Bar"140 lm.
Table 3 Percentages of the measured bone—implant contact over 4 mm on the left and the right side of the implant. Three histological sections per implant were measured Material
Mean % bone contact$ standard deviation
Ti2
9.9$11.9 (n"9) 11.3$13.1 (n"8) 14.3$14.1 (n"9) 14.9$16.9 (n"9)
Ti3 Ti4 Ti-un Fig. 5. Light micrograph of a Ti4 coating showing ingrowth of newly formed bone into the undercuts of the coating. Original magnification]140. Bar"70 lm.
medial and those inserted into the lateral condyles were found. Finally, simple linear regression showed that no relationship existed between roughness and percentage of bone contact (r"0.01).
3.2.2. Bone mass measurements Table 5 shows the results of the bone mass measurements in the different regions of interest. Statistical analysis showed no difference between the different types of implants. On the other hand, a significant difference existed between the different regions of the Ti3, Ti4 and Ti-un implants placed in the medial condyle and the Ti4
1098
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
implants placed in the lateral condyle. ROI 500 of these implants always revealed a higher bone amount than ROI 1000 and ROI 1500 (P(0.05).
4. Discussion and conclusions The results of the current study are in clear contrast with the findings of a lot of other investigators [2—7], who showed a positive effect of surface microroughness on the integration of implants. Several explanations can be given for this discrepancy. First, there is the used surgical procedure. The beamshaped design of our implants did not allow to make standarized drill holes. For example, to create the bone cavity we had to use a conically shaped drill. This resulted always in a gap that was slightly wider at the top than at the bottom of the cavity. Increased gap size and reduced fit can have a negative influence on the integration of an implant [17, 18]. Second, the surface enlargement, as created by very rough coatings, can result in an increased ion release, which can have a negative effect on bone formation [19—22]. We suppose that the high degree of roughness of our grit-blasted and plasma-spray-coated implants in combination with ion release resulted in the observed low percentage of bone contact. Additional evaluations are recommended to confirm this theory. Further, our XRD and EDS measurements revealed the presence of Al O on the surface of our grit-blasted 2 3 implants. In agreement with Gross et al. [23], we suppose that the Al O , as remained from the blasting procedure, 2 3 Table 4 Means and standard deviations for the various implants for the 2 implantation sites. The number of the measured histological sections is shown in parantheses
Medial Lateral
Ti2
Ti3
Ti4
Ti-un
12.5$12 (n"5) 6.8$12.7 (n"4)
20.2$12.8 (n"4) 2.4$4.9 (n"4)
13.9$8.6 (n"4) 14.6$18.9 (n"5)
11.2$13.4 (n"4) 18$20.3 (n"5)
have a negative effect on the bone response. They reported that the release of Al O from an implant sur2 3 face can inhibit the final mineralization of bone. More recently, also Johansson et al. [24] found a deteriorating influence of aluminium on bone formation. On the other hand, Wennerberg et al. [5] could never demonstrate any effect of Al O blasting procedure on the bone response. 2 3 Although, their auger electron spectroscopy (AES) evaluations revealed a higher Al O concentration on 2 3 the surface of the Al O blasted implants. Probably, this 2 3 discrepancy in results is due to a still unknown threshold value for ion release. Further in our histomorphometrical evaluation, we did discern between implants inserted into the medial and into the lateral femoral condyle. This was done, because goats have X-shaped legs. As a consequence, differences in loading condition exist within the femoral condyle which can play an important role in the bone response between implants placed in the medial or lateral condyle [25]. This is confirmed by the studies of Caulier et al. [26] and Corten et al. [27]. In our study such differences were not found. The reason has to be the variations in the implant positioning. The implants could not always be oriented in a similar inclination with respect to the longitudional axis of the femur. This results in different stress transfer patterns for the implants, which also explains the variance in bone contact measurements. For future studies the use of cylindrical implants, which allow a more standardized positioning, is recommended. Although, we have to notice that the results were not completely consistent, our histomorphometrical data suggest that the medially placed implants showed more bone mass close to the implant than the laterally placed implants. We suppose that this is due to a reduced load transmission [28]. The continuous micromovement of a less incorporated implant results in a persistent stimulation of the surrounding bone. As compensation for the low percentage of bone contact, the bone will respond with an increased bone turnover, resulting in an increased bone mass. This theory does not completely corroborate with the generally accepted phenomenon of stress shielding, where bone atrophy occurs at places where too high forces act on an implant or inadequate
Table 5 Measured amount of bone in lm2]103. Three histological sections per implant were measured Coating
Ti2 Ti3 Ti4 Ti-un
Medial
Lateral
ROI 500
ROI 1000
ROI 1500
ROI 500
ROI 1000
ROI 1500
350.0$119.4 415.0$298.7 332.5$70.4 312.5$165.6
182.0$73.3 270.0$115.2 230.0$74.8 152.5$42.7
130.0$52.9 165.0$78.5 127.5$49.9 75$50.7
177.0$149.8 130.0$143.6 400.0$187.7 224.0$249.5
165.0$34.2 145.0$68.5 172$31.9 122.0$76.6
77.5$23.6 85.0$55.6 98.0$14.8 82.0$30.3
S. Vercaigne et al. / Biomaterials 19 (1998) 1093—1099
stress transfer and distribution is present. On the other hand, another important factor in stress shielding is time. Our implantation time was relatively short. We assume that when the implants are left in place for longer periods, also reduction in bone mass will occur at the bone— implant interface. The laterally placed implants, except Ti4, showed no significant difference. The forces acting on the medially placed implants, resulting in a higher mass close to the implants, were not present on the laterally placed implants. This confirms the hypothesis that the biological response of those implants is the result of the mechanical influence. In conclusion, no difference in bone behaviour was observed in our study design between grit-blasted and various titanium plasma-sprayed implants. Therefore, besides mechanical interlocking, the biological efficacy of very rough surfaces remains unconclusive. Acknowledgements The authors thank J.P.C.M. van der Waerden and A.F.M. Leijdekkers-Govers for their participation and the analytical support, H.G. Schaeken for the management of the XRD and J.E.G. Hulshoff for the surgical assistance during the animal experiments.
[9] [10]
[11]
[12]
[13]
[14]
[15]
[16]
[17] [18]
References [1] Albrektsson T, Bras nemark P-I, Hansson H-A, Lindstro¨m J. Osseointegrated titanium implants. Requirements for ensuring a long-lasting, direct bone-to-implant anchorage in man. Acta Orthop Scand 1981;52:155. [2] Buser D, Schenk RK. Influence of surface characteristics on bone integration of titanium implants. A histomorphometric study in miniature pigs. J Biomed Mat Res 1991;25:889—902. [3] Gotfredsen K, Nimb L, Hjo¨rting-Hansen E, Jensen JS, Holme´n, A. Histomorphometric and removal torque analysis for TiO 2 blasted titanium implants. Clin Oral Impl Res 1992;3:77—84. [4] Wennerberg A, Albrektsson T, Lausmaa J. Torque and histomorphometric evaluation of c.p. titanium screws blasted with 25- and 75-lm-sized particles Al O . J Biomed Mat Res 1996;30:251—60. 2 3 [5] Wennerberg A, Albrektsson T, Johansson C, Andersson B. Experimental study of turned and grit-blasted screw-shaped implants with special emphasis on effects of blasting material and surface topography. Biomaterials 1996;17:15—22. [6] Wennerberg A, Albrektsson T, Andersson B, Krol JJ. A histomorphometric and removal torque of screw-shaped titanium implants with three different surface topographies. Clin Oral Impl Res 1995;6:24—30. [7] Wennerberg A, Albrektsson T, Andersson B. An animal study of c.p. titanium screws with different surface topographies. J Mater Sci Mater Med 1995;6:302—9. [8] Lekholm U, Zarb GA. Patient selection and preparation. In: Bras nemark P-I, Zarb G, Albrektsson T, editors. Tissue integrated
[19] [20]
[21]
[22] [23]
[24]
[25]
[26]
[27]
[28]
1099
prosthesis: osseointegration in clinical dentistry. Chicago Quintessence Publ. Co. Inc., 1985. Jaffin RA, Berman CL. The excessive loss of Bras nemark fixtures in type IV bone: a 5-year analysis. J Periodontol 1991;62:2—4. Cune MS, de Putter C, Hoogstraten J. Treatment outcome with implant-retained overdentures: Part I—clinical findings and predictability of clinical treatment outcome. J Prosthet Dent 1996;72: 144—51. Cune MS, de Putter C, Hoogstraten J. Treatment outcome with implant-retained overdentures: Part II—patient statisfaction and predictability of subjective treatment outcome. J Prosthet Dent 1996;72:152—8. Hutton JE, Heath MR, Chai JY et al. Factors related to success and failure rates at 3-year follow-up in a multicenter study of overdentures supported by Bras nemark implants. J Oral Maxillofac Implants 1995;10:33—42. Cameron HU, Pilliar RM, Macnab I. The effect of movement on the bonding of porous metal to bone. J Biomed Mat Res 1973;7: 301—11. Maniatopoulos C, Pilliar RM, Smith DC. Threaded versus porous-surfaced designs for implant stabilization in bone-endodontic implant model. J Biomed Mat Res 1986;20:1309—33. Søballe K. Hydroxyapatite ceramic coating for bone implant fixation. Mechanical and histological studies in dogs. Acta Orthoped Scand 1993;64(Suppl):255. Van Der Lubbe HBM, Klein CPAT, de Groot K. A simple method for preparing thin (10 lm) histological sections of undecalcified plastic embedded bone with implants. Stain Technol 1988;63:171—7. Cameron HU, Pilliar RM, Macnab I. The rate of bone ingrowth into porous metal. J Biomed Mat Res 1976;10:295—9. Sandborn PM, Cook SD, Spires WP, Kesters MA. Tissue response to porous- coated implants lacking initial bone apposition. J Arthroplasty 1989;3(4):337—46. Black J. Does corrosion matter? J Bone J Surg 1988;70b:517—20. Thompson GJ, Puleo DA. Effects of sublethal metal ion concentrations on osteogenic cells derived from bone marrow stromal cells. J Appl Biomaterials 1995;6:249—58. Osborn JF, Willick P, Neenen N. The release of titanium into human bone from a titanium implant coated with plasma-sprayed titanium. Adv Biomater 1990;9:75—80. Wennerberg A. On surface roughness and implant incorporation. PhD Thesis, Go¨teborg, 1996. Gross U, Strunz V. The interface of various glasses and glass ceramics with a bony implantation bed. J Biomed Mat Res 1985;19:251—71. Johansson C, Albrektsson T, Thomsen P, Sennerby L, Lodding A, Odelius H. Tissue reaction to titanium-6aluminium-4vanadium alloy. Eur J Exp Musculoskel Res 1992;1:161—9. Heimke G, Griss P, Werner E, Jentschura G. The effects of mechanical factors on biocompatibility tests. J Biomed Engng 1981;3:209—13. Caulier H, van der Waerden JPCM, Paquay YCGJ et al. Effect of calcium phosphate (Ca-P) coatings on trabecular bone response: a histological study. J Biomed Mat Res 1995;29:1061—9. Corten FGA, Caulier H, van der Waerden JPCM, Kalk W, Corstens FHM, Jansen JA. The assessment of bone surrounding implants in goats: ex vivo measurements by dual X-ray absorptiometry. Biomaterials 1997;6:495—501. Williams D. Concise encyclopedia of medical and dental materials. Oxford: Pergamon Press, 2nd ed., 1990:51—9.