J. BIOMED. MATER. RES.
VOL. 11, PP. 165-178 (1977)
The Restoration of Articular Surfaces Overlying Replamineform Porous Biomaterials* RICHARD T. CHIROFF and RODNEY A. WHITE, Orthopaedic Research Laboratory, State University of New York, Upstate Medical Center, Syracuse, New York, and E. W. WHITE, J. N. WEBER, and D. ROY, Materials Research Laboratory, The Pennsylvania State University, University Park, Pennsylvania
Summary Replamineform porous implants (4 mm X 4 mm diameter) were placed into full-thickness cartilage and bone defects of the weight-bearing surface of the lateral femoral condyles of adult male white rabbits. These were analyzed a t 1 day, 1 week, 6 weeks, 3 months, and 6 months for 1) ingrowth of tissue within the implants and 2) restoration of the articular surface overlying them. Appropriate unfilled, but similar, control defects were also studied. Mineralized bone was seen within the substance of both the TiO, and hydroxyapatite implants at 1 week; this extremely rapid response was present in every specimen studied and was not seen with uA1203 or control animals. With the passage of time, maturation of this bone ingrowth occurred so that by 3 months, they were all incorporated into the surrounding bone. Only the hydroxyapatite implants showed consistent regenerative healing of hyaline articular cartilage from the margins of the defectaswith the passage of time; this occurred whenever the subchondral bone adjac9nt to the defect proliferated in a “creeping” fashion over the articular aspect of the implant, and the undamaged cartilage then followed it. Fibrocartilage, and not hyaline cartilage, formed the articular surface over the Ti02 and aAlZO8 implants and in the controls.
INTRODUCTION Recently, Baker and co-workers’ have shown that articular cartilage can, in fact, be stimulated to regenerate with the aid of electrical stimulation. One of the problems which these investigators encountered was related to the rapidity with which fibrous tissue formed *Presented in part a t the Sixth International Biomaterials Symposium, Clemson University, Clemson, South Carolina, April, 1974. 16;i
@ 1977 by John Wiley & Sons, Inc.
166
CHIROFF ET AL.
in the cartilage defects which were created. This rather dense scar had the ability to prevent the closure of the defects in the articular cartilage by a rather simple mechanical barrier effect in the control animals. The unusual mechanical and physical properties associated with the internal pore structure configuration of skeletal calcium carbonate deposited by echinoderms and corals2 provide ideal qualities for implant purposes because of the remarkable uniformity of pore size and complete interconnection of the pores. Both the pore size and size of interconnection fall well within the ranges found to be suitable for optimum bone tissue ingrowth. Subsequently, techniques were developed4 to replicate the invertebrate skeletal microstructures in a wide range of the most promising materials, including metals, ceramics, and polymers. For hard tissue applications, the coral genus Porites appears t o offer a n ideal microstructural configuration. Its skeleton has pores with diameters in the range 140-160 Fm, and all of the pores are interconnected. The precursor carbonate is soft and easily machined into any conceivable geometry (plugs, screws, plates, pins, or customshaped devices) prior to replication of the microstructure in another material. As described e l ~ e w h e r eit, ~is possible to prepare implants which are wholly porous, or implants which are partially porous and partially solid, e.g., a plug with a solid core surrounded by a n outer porous surface. I n addition, various materials with the porosity of Porztes have been placed in the distal femora and proximal tibiae of adult dogs, and the tissue ingrowth has been described.6 Although these new Replatnineform biomaterials have potential application to a wide range of orthopedic problems including hard tissue replacement and repair, in addition t o the attachment of prosthetic devices, this study was undertaken and designed in order to examine whether specific porous biomaterials could facilitate or enhance reparative regeneration of hyaline articular cartilage.
METHODS AND MATERIALS The three materials studied were alumina (crAlz03), titania (Tion), and hydroxyapatite [C~,O(PO,)~(OH),].The most satisfactory aAl2O3 replicas have been made using Alcoa A-15SG alumina. This starting material has been throughly characterized
RESTORATION I N POROUS BIOMATERIALS
1 67
by standard CESEMI proc~dures.~The median particle size (number count basis) is 0.35 pm, and the maximum and minimum sizes are 10 and 0.1 gm, respectively. The size distribution is significantly bimodal. The particle mass distribution has also been determined; in this case, the median size is 4.2 pm. Alcoa A-15SG is 99.5+y0 A1203 with the remainder composed primarily of SiOz, Na20, and Fez03. Procedures for preparing Replamineform alumina have been published.s Final firing conditions are at 1650°C for 3 hr. The best starting material for preparing the Ti02 (rutile) implant test specimens has been found to be Raker reagent-grade TiOz (anatase). In this material, 50% of thr particles are smaller than 0.1 pm. The material is 99.0% Ti02 with a maximum of 0.023% trace metal content. Fired a t 1425°C for 3-4 hr, this material undergoes a 29% linear shrinkage. Microporosity is less than that of the aA1203,but sintered grain size is larger. Hydroxyapatite test specimens were prepared by hydrothermal conversion of the skeletal CaCOj from the coral Porites as described by Roy and Kurtossy-Linnehan. Complete hydrothermal conversion of the aragonite (CaCoa) to hydroxyapatite was accomplished by a 24 hr treatment in a fluid medium of (NH4)2HP04and HzO maintained a t 3OOOC and 15,000 psi. The apatite formed by this process completcly replaced Porites carbonate as confirmed by both petrographic microscopy and x-ray powder diffraction Electron microprobe analyses show that the composition of the apatite is the same as that obtained for apatites that crystallized a t higher temperatures and are stoichiometric, i.e., C a : P = 10:6. Scanning electron micrographs demonstrate the morphological integrity of the resultant apatite. Each of these materials was prepared in one of two configurations: entirely porous 4 mm diameter x 4 nim long cylinders, or 4 mm x 4 mm cylinders, one of the ends of which could be nonporous to a depth of 1 mm. Prior to implantation, they werr sterilized by exposure to gaseous ethylene oxide a t 130°F for 4 hr. The model used for the study was the creation of full-thickness defects of the articular cartilage and subchondral bone of the weightbearing surfaces of the lateral femoral condyles of male, adult New Zealand white rabbits ranging in size from 4.0 to 5.8 kg. Anesthesia for all surgical procedures was as follows. Chlorpromazine hydrochloride (Thorazine, 25 mg/ml; Smith Kline & French Labs., Phila-
168
CHIROFF ET AL.
delphia, Pa., 19101) (1.5 mg/kg) was injected intramuscularly 45 min prior t o surgery; sodium pentobarbital (Nembutal 50 mg/ml; Abbott Labs., North Chicago, Ill. 60064) (10 mg/kg) was injected intravenously, and after 2-3 min the animals were shaved and skin prepared by washing for 5 min with surgical soap and applying an aqueous solution of benzalkonium chloride (Zephiran, 1/750 solution ; Winthrop Labs., New York, N.Y. 10016). Ketamine hydrochloride (Ketaject, 10 mg/ml; Bristol Labs., Syracuse, N.Y. 13201) (10 mg/kg) was injected intravenously, with careful monitoring of the rabbits’ respiratory rate and air exchange as apnea was likely t o occur, and the surgical procedure as described below begun. Additional ketamine hydrochloride (10 mg/kg) was injected intramuscularly a t 5-10 min intervals during the course of the procedures if restlessness occurred. All of the surgical instruments were steamautoclaved, and sterile technique was maintained throughout the procedures. The knee joint of each rabbit was exposed by a longitudinal skin incision, placed anterolaterally, about 1 in. long. Dissection was carried, in layers, through the subcutaneous tissue, investing fascia, extensor retinaculum, joint capsule, and synovium t o expose the lateral femoral condyle. The tendon of the digital extensor muscle was released from its origin on the anterior aspect of the condyle, exposing the articular surface of that structure. Using a scalpel, a 2 mm ellipse of articular cartilage was removed in order to facilitate placement of a 1 mm stainless steel drill bit a t the selected site on the weight-bearing surface of the condyle. Progressive enlargement of the defect to a depth of 4 mm was accomplished by the use of drill bits 1 mm, 2 mm, and 4 mm in diameter. The defect thus created comprised the entire width of the articular cartilage of the condyle in most cases. Implants in the form of 4 mm X 4 mm cylinders were inserted into the defects such that their exposed surfaces lay just below the margins of the articular cartilage; control specimens consisted of unfilled, similar defects. The joints were thoroughly irrigated with sterile saline, and any debris resulting from creation of the defects was removed. All of the implanted materials were lightly press-fitted into the defects and a full range of passive motion of the joint was possible. The synovium was closed with a continuous 4-0 chromatized catgut suture ; the subcutaneous tissue with interrupted sutures of the same material; and the skin with inter-
RESTORATION IN POROUS BIOMATERIALS
169
rupted 3-0 braided silk. The initial six rabbits had their operative wounds dressed with vibesate (Aeroplast dressing; Park, Davis & Go., Detroit, Mich. 48232), spray-on plastic bandage; these wounds healed poorly, however, and superficial infections with Pasteurella multocida occurred. These infections were treated locally by opening the skin incisions, cleansing them thoroughly with surgical soap, and administering 50 mg oral doses of tylosin (Tylocine, 50 mg; Lilly and Co., Indianapolis, Ind. 46206) daily for 5 days. With daily antiseptic dressing changes, it was determined that only one deep infection involving the joint had occurred; this was the right knee of animal number 3, a control specimen. With discontinuation of the dressing described above and with application following surgery of a dry, sterile gauze dressing, held in place by paper tape, all subsequent wounds healed uneventfully. Immediately postoperatively, and with the animals still anesthetized, standard roentgenograms were taken of the distal femora; these were repeated a t weekly intervals until the animals were killed on a preselected schedule, so that representative specimens of each type could bc analyzed a t 1 day, 1 week, 6 weeks, 3 months, and 6 months postoperatively. At the end of each experimental period, the animals were killed by the intravenous injection of pentobarbital sodium (Somlethol, 6 g/ml; J. A. Webster, Inc., No. Billerica, Mass. 01862) (6 g/2.25 kg wt). The distal femora were immediately removed from the animals and the soft tissue dissected from them. Color and black-and-white photographs of each specimen were taken, and the gross appearance of each healing defect correlated with the histology seen. Previously describedL0techniques of hard tissue preparation for microradiography and fully mineralized histology were used on each specimen. The microradiographs were made from 100 pm thick sections; these were then ground, by hand, to thickness of 35-70 pm and then stained with Paragon (Paragon Multiple Stain; C. & C. Paragon Co., Inc., Bronx, N.Y.) and Safranin-0 (Safranin-0 Stain; Fisher Scientific Co., Fair Lawn, N.J. 07430) for histologic evaluation. Adjacent sections from each methacrylate-embedded specimen were examined with the scanning electron microscope and the electron microprobe. These latter analyses were for the elemental constituents of calcium, phosphorous, and sulfur in all specimens, and aluminum and titanium where applicable. The use of the electron
170
CHIROFF ET AL.
microprobe for sulfur as a means of characterizing hyaline cartilage, fibrocartilage, and bone has recently been described ;I1 this was correlated, as much as possible, with the histology seen.
RESULTS As has been stated, one control joint was the site of a deep infection. Of the 48 rahbits operated upon, only one was not moving about in its cage by the evening of surgery and normally active by the next day. That exception was a rabbit in which a major portion of the patellar tendon, controlling knee extension, was sectioned a t the time of surgery; this animal’s gait remained impaired and it was sacrificed 1 week following surgery. None of the implants placed was visualized on the initial or followup in vivo x-rays; Figure 1 is a postmortem x-ray of a 1week hydroxy- ‘I apatite specimen. When the joints were opened and the specimens excised, no implant was found to be dislodged from the defect into which it was placed.
Fig. 1. Postmorteni x-riidiograph o f the left femoral condyles of a 1 week, entirely porous hydroxyapatite specimen. This implant may he slightly more depressed in the created defect t.han is desirable.
RESTOR.ATION I N POROUS BIOMATERIALS
171
Gross Appearance At 1 day, all pores were filled with clotted blood, and no tissue ingrowth or overgrowth was present. A t 1 week, essentially all of the defects were still significantly depressed at the margins. The control animals were obviously proliferating tissue from the depths of the defects. This tissue was yellowish-white in color and did not have the “glistening” appearance of the intact articular cartilage at, the margins of the defect. With the porous hydroxyapatite and alumina implants in place, a majority of the cellular proliferation appeared t o arise from the margins of the defect, i.e., the articular cartilage itself. This “creeping” proliferation of tissue as well as the sheen on its surface, not unlike that of the surrounding hyaline cartilage, suggested that attempts a t rrpair of the defect with that tissue were present. I n all longrr-trrm specimens, including controls, the created defects wcrc progressively lcss depressed with the passage of timr, and the absolute thickness of the tissues overgrowing the implants was increased (Figs. 2b and 2c).
Microradiography At 1 week post-insertion, all of the TiOz and hydroxyapatite implants had mineralized tissue present within the substance of their porosities (Fig. 2a); the control defects and aA1203 specimens did not show these changes. By 6 weeks, and thereafter for all specimens, each of the implanted materials had virtually total ingrowth with normal-appearing, mineralized bone (Figs. 2b and 2c).
Histology None of the specimens showed fibrous or other encapsulating responses suggestive of rejection. As regards the tissue growing into the body of the implants, t h r fully mineralized histology confirmed the microradiographic findings ; rrmarkablc, indeed, was the rapid and complete fashion in which the pores of the TiOz and hydroxyapatite became ingrown with bone (Fig. 3a) ; both the control and alumina sites were still filled with clotted blood a t 1 week postoperatively. Analysis of the tissues overlying the implants delineated quite diff erent responses for the hydroxyapatite and ceramic materials.
172
CHIItOFF E T AL.
Fig. 2. Representative microradiographs. (a) One week, entirely porous titania implant. Well within the substance of the Ti02 there is mineralization characteristic of new-formed bone (60X). (b) Six week, entirely porous, hydroxyapatite implant. Proliferation of the subchondral bone from the margins of the created defects has occurred; histology of this specimen is shown in Fig. 3b (40 X). (c) Six month, “capped” hydroxyapatite implant. The radiodense implant is virtually covered by a “plate” of new bone (60X).
I n the case of the former, newly formed bone and, accompanying it, hyaline cartilage came to reconstitute the articular surfaces (Figs. 3b and 3c). With both the alumina and titania, it was fibrocartilage which covered the implants (Fig. 3d). With the passage of time, the absolute thickness of all overgrowing tissues increased with constant filling-in of the depressions overlying the implants by newformed tissue. The fact that no difference in the patterns was seen between the entirely porous or “capped” implants was significant. The control defects were completely filled with bone by 6 months post-creation; their articular surfaces at that, and each earlier, test period were restored with fihrocartilage.
RESTORATION I N POROUS BIOMATERIALS
Fig. 2.
(continued)
173
174
CHIROFF E T AL.
Fig. 3. Fully mineralized histology. (a) One week, entirely porous, titania implant. Normal-appearing, cellular, and mineralized new bone is seen within the pores (Paragon, l 2 0 x ) . (b) Six week, entirely porous hydroxyapatite implant. This is the exact section from which the microradiograph of Fig. 2b was made; the proliferating bone is almost “osteophytic” as it extends towards the center of articular aspect of the implant (Paragon, 40X). (c) Six month, “capped” hydroxyapatite implant. There is total cellular and morphologic continuity between the intact articular cartilage a t the margin of the created defect and that tissue now forming the bone plate shown in Fig. 3b (Paragon, 1 2 0 ~ ) . (d) Three month, entirely porous alumina implant. No marginal proliferation of bone or cartilage is seen; the tissue overlying the implant has the appearance of fibrocartilage (Paragon, 120X).
Scanning Electron Microscopy and Electron Microprobe Analysis These studies confirmed the findings already described as far as both the intra-implant and articular surface responses. When the electron microprobe was used to determine the sulfation of that tissue which had come to lie b c t w c n thc implants and the joint cavity, it was found that a grcatcr weight-percent sulfur was present in the hydroxyapatite specimens than with the other two, or the controls.
RESTORATION I N POROUS BIOMATERIALS
Fig. 3.
(continued)
175
176
CHIROFF ET AL.
Fig. 3.
(continued)
DISCUSSION Regeneration of lost or damaged parts in lower animal forms is well established. With evolutionary ascendancy, however, it has been held that higher forms, particularly mammals, have apparently lost that regenerative capacity. Beckerl2 and Smith13have shown that the capacity for true regeneration of portions amputated and then subjected to appropriate electrical stimulation remains. It is not fanciful or whimsical to note that, in the human, the only tissue which truly heals itself in a regenerative fashion is bone; that is, for complete and satisfactory resumption of pre-injury status, bone tissue cannot simply heal itself with a fibrous scar-it must be bone. Since bone is in intimate contact with yet another of the body’s hard tissues, cartilage, there may be some carry-over of the principles involved in bone regeneration in healing which could also apply to articular cartilage. Recent 1\-ork1J4has shown that defects in hyalinc cartilage can heal themselves by the regeneration of an exactly similar and living tissue. When placed in either cntirely porous or surface-impermeable configurations immediately beneath the damaged articular cartilage of
RESTORATION I N POROUS BIOMATERIALS
I77
a weight-bearing joint in a rabbit, the hydroxyapatite implants enhanced and facilitated regenerative healing of the marginal hyaline cartilage and subchondral bone to restore the articular surface. In the cases of the TiOz and (rA1203 implants, as well as the control animals, the articular surface was reconstituted by fibrocartilage. If a mammalian species such as the rabbit has the capacity to regenerate its hyaline articular cartilage across a 4 mm defect, it is certainly not possible to say that our hopes for the restoration of function to diseased or damaged joints will necessarily follow. However, the marriage of appropriate surgical techniques, specific materials, and, possibly, porous implant devices could lead to improved treatment of many individuals who are, and will be, disabled by one form or another of comprised function of their diarthrodial joints. Furthermore, as optimal materials and configurations emerge, it should be possible to translate the knowledge gained towards usefulness in treating some of the wide variety of naturally occurring arthritic and arthrotic states in the canine, bovine, and equine species.'"'' The authors wish to gratefully acknowledge the assistance of Frank P. Romano, I11 and Leo Tarhay in the completion of this research. This work was supported by a National Science Foundation Grant No. GH-38944.
References 1 . B. Baker, R. 0. Becker, and J. Spadaro, Clin. Orthop., 102, 251 (1974). 2. J. N. Weber, R. Greer, B. Voight, E. W. White, and H. Roy, J . Ultrastruct. Res., 26, 355 (1969). 3. S. F. Hulbert, F. A. Young, R.. S. Mathews, J. J. Klawitter, C. D. Talbert, and F. H. Stelling, J . Riomed. Muter. Res., 4, 433 (1970). 4. It. A. White, J . N. Weber, and E. W. White, Science, 176, 922 (1972). 3 . E. W. White, J. N. Weber, I). Roy, E. L. Owen, 11. T . Chiroff, and H . A. White, J . Riomed. Muter. Res., 9, 23, (1975). 6. It. T. Chiroff, E. W. White, J. Weber, and I>. Roy, J . Biomed. Muter. Res., 9, 29 (1975). 7. J. Lebiedzik, K. G. Burke, S. Troutman, G. G. Johnson, Jr., and E. W. White, in Scanning Electron Microscopy/l.W3, IITNI Publications, Chicago, 1973, p. 121. 8. L. Tarhay, W. It. Buessem, arid E. W. White, Ceramic Bull., 52, 394 (1973). 9. D. M. Roy and S. Kurtoasy-Linnehan, Nature, 247, 220 (1974). 10. J. Jowsey, P. J. Kelly, B. L. Kiggs, A. J. Bianco, Jr., 11. A. Scholz, and J . Gershon-Cohen, J . Bone Joint Surg., 47A, 785 (1965). 11. R. T. Chiroff, It. A. White, 13. W. White, and L. Tarhay, Calcif. Tissue Res., 17, 87 (1974).
178 12. 13. 14. 15.
CHIROFF ET AL. R. 0. Becker, Nature, 235, 109 (1972). S. 0. Smith, Anat. Rec., 158, 89 (1967). W. E. Riddle, J . Amer. Vet. Med. Assoc., 157, 1471 (1970). E. P. Leonard, Orthopaedic Surgery ofthe Dog and Cat, Saunders, Philadelphia,
1971. 16. S. E. Olsson, J . Small Animal Prac., 12, 333 (1971). 17. C. W. Schwabe, Veterinary Medicine and Human Health, Williams and Wilkins, Baltimore, 1964.
Received November 3, 1975 Revised March 8, 1976
VOL. 11, PP. 165-178 (1977)
The Restoration of Articular Surfaces Overlying Replamineform Porous Biomaterials* RICHARD T. CHIROFF and RODNEY A. WHITE, Orthopaedic Research Laboratory, State University of New York, Upstate Medical Center, Syracuse, New York, and E. W. WHITE, J. N. WEBER, and D. ROY, Materials Research Laboratory, The Pennsylvania State University, University Park, Pennsylvania
Summary Replamineform porous implants (4 mm X 4 mm diameter) were placed into full-thickness cartilage and bone defects of the weight-bearing surface of the lateral femoral condyles of adult male white rabbits. These were analyzed a t 1 day, 1 week, 6 weeks, 3 months, and 6 months for 1) ingrowth of tissue within the implants and 2) restoration of the articular surface overlying them. Appropriate unfilled, but similar, control defects were also studied. Mineralized bone was seen within the substance of both the TiO, and hydroxyapatite implants at 1 week; this extremely rapid response was present in every specimen studied and was not seen with uA1203 or control animals. With the passage of time, maturation of this bone ingrowth occurred so that by 3 months, they were all incorporated into the surrounding bone. Only the hydroxyapatite implants showed consistent regenerative healing of hyaline articular cartilage from the margins of the defectaswith the passage of time; this occurred whenever the subchondral bone adjac9nt to the defect proliferated in a “creeping” fashion over the articular aspect of the implant, and the undamaged cartilage then followed it. Fibrocartilage, and not hyaline cartilage, formed the articular surface over the Ti02 and aAlZO8 implants and in the controls.
INTRODUCTION Recently, Baker and co-workers’ have shown that articular cartilage can, in fact, be stimulated to regenerate with the aid of electrical stimulation. One of the problems which these investigators encountered was related to the rapidity with which fibrous tissue formed *Presented in part a t the Sixth International Biomaterials Symposium, Clemson University, Clemson, South Carolina, April, 1974. 16;i
@ 1977 by John Wiley & Sons, Inc.
166
CHIROFF ET AL.
in the cartilage defects which were created. This rather dense scar had the ability to prevent the closure of the defects in the articular cartilage by a rather simple mechanical barrier effect in the control animals. The unusual mechanical and physical properties associated with the internal pore structure configuration of skeletal calcium carbonate deposited by echinoderms and corals2 provide ideal qualities for implant purposes because of the remarkable uniformity of pore size and complete interconnection of the pores. Both the pore size and size of interconnection fall well within the ranges found to be suitable for optimum bone tissue ingrowth. Subsequently, techniques were developed4 to replicate the invertebrate skeletal microstructures in a wide range of the most promising materials, including metals, ceramics, and polymers. For hard tissue applications, the coral genus Porites appears t o offer a n ideal microstructural configuration. Its skeleton has pores with diameters in the range 140-160 Fm, and all of the pores are interconnected. The precursor carbonate is soft and easily machined into any conceivable geometry (plugs, screws, plates, pins, or customshaped devices) prior to replication of the microstructure in another material. As described e l ~ e w h e r eit, ~is possible to prepare implants which are wholly porous, or implants which are partially porous and partially solid, e.g., a plug with a solid core surrounded by a n outer porous surface. I n addition, various materials with the porosity of Porztes have been placed in the distal femora and proximal tibiae of adult dogs, and the tissue ingrowth has been described.6 Although these new Replatnineform biomaterials have potential application to a wide range of orthopedic problems including hard tissue replacement and repair, in addition t o the attachment of prosthetic devices, this study was undertaken and designed in order to examine whether specific porous biomaterials could facilitate or enhance reparative regeneration of hyaline articular cartilage.
METHODS AND MATERIALS The three materials studied were alumina (crAlz03), titania (Tion), and hydroxyapatite [C~,O(PO,)~(OH),].The most satisfactory aAl2O3 replicas have been made using Alcoa A-15SG alumina. This starting material has been throughly characterized
RESTORATION I N POROUS BIOMATERIALS
1 67
by standard CESEMI proc~dures.~The median particle size (number count basis) is 0.35 pm, and the maximum and minimum sizes are 10 and 0.1 gm, respectively. The size distribution is significantly bimodal. The particle mass distribution has also been determined; in this case, the median size is 4.2 pm. Alcoa A-15SG is 99.5+y0 A1203 with the remainder composed primarily of SiOz, Na20, and Fez03. Procedures for preparing Replamineform alumina have been published.s Final firing conditions are at 1650°C for 3 hr. The best starting material for preparing the Ti02 (rutile) implant test specimens has been found to be Raker reagent-grade TiOz (anatase). In this material, 50% of thr particles are smaller than 0.1 pm. The material is 99.0% Ti02 with a maximum of 0.023% trace metal content. Fired a t 1425°C for 3-4 hr, this material undergoes a 29% linear shrinkage. Microporosity is less than that of the aA1203,but sintered grain size is larger. Hydroxyapatite test specimens were prepared by hydrothermal conversion of the skeletal CaCOj from the coral Porites as described by Roy and Kurtossy-Linnehan. Complete hydrothermal conversion of the aragonite (CaCoa) to hydroxyapatite was accomplished by a 24 hr treatment in a fluid medium of (NH4)2HP04and HzO maintained a t 3OOOC and 15,000 psi. The apatite formed by this process completcly replaced Porites carbonate as confirmed by both petrographic microscopy and x-ray powder diffraction Electron microprobe analyses show that the composition of the apatite is the same as that obtained for apatites that crystallized a t higher temperatures and are stoichiometric, i.e., C a : P = 10:6. Scanning electron micrographs demonstrate the morphological integrity of the resultant apatite. Each of these materials was prepared in one of two configurations: entirely porous 4 mm diameter x 4 nim long cylinders, or 4 mm x 4 mm cylinders, one of the ends of which could be nonporous to a depth of 1 mm. Prior to implantation, they werr sterilized by exposure to gaseous ethylene oxide a t 130°F for 4 hr. The model used for the study was the creation of full-thickness defects of the articular cartilage and subchondral bone of the weightbearing surfaces of the lateral femoral condyles of male, adult New Zealand white rabbits ranging in size from 4.0 to 5.8 kg. Anesthesia for all surgical procedures was as follows. Chlorpromazine hydrochloride (Thorazine, 25 mg/ml; Smith Kline & French Labs., Phila-
168
CHIROFF ET AL.
delphia, Pa., 19101) (1.5 mg/kg) was injected intramuscularly 45 min prior t o surgery; sodium pentobarbital (Nembutal 50 mg/ml; Abbott Labs., North Chicago, Ill. 60064) (10 mg/kg) was injected intravenously, and after 2-3 min the animals were shaved and skin prepared by washing for 5 min with surgical soap and applying an aqueous solution of benzalkonium chloride (Zephiran, 1/750 solution ; Winthrop Labs., New York, N.Y. 10016). Ketamine hydrochloride (Ketaject, 10 mg/ml; Bristol Labs., Syracuse, N.Y. 13201) (10 mg/kg) was injected intravenously, with careful monitoring of the rabbits’ respiratory rate and air exchange as apnea was likely t o occur, and the surgical procedure as described below begun. Additional ketamine hydrochloride (10 mg/kg) was injected intramuscularly a t 5-10 min intervals during the course of the procedures if restlessness occurred. All of the surgical instruments were steamautoclaved, and sterile technique was maintained throughout the procedures. The knee joint of each rabbit was exposed by a longitudinal skin incision, placed anterolaterally, about 1 in. long. Dissection was carried, in layers, through the subcutaneous tissue, investing fascia, extensor retinaculum, joint capsule, and synovium t o expose the lateral femoral condyle. The tendon of the digital extensor muscle was released from its origin on the anterior aspect of the condyle, exposing the articular surface of that structure. Using a scalpel, a 2 mm ellipse of articular cartilage was removed in order to facilitate placement of a 1 mm stainless steel drill bit a t the selected site on the weight-bearing surface of the condyle. Progressive enlargement of the defect to a depth of 4 mm was accomplished by the use of drill bits 1 mm, 2 mm, and 4 mm in diameter. The defect thus created comprised the entire width of the articular cartilage of the condyle in most cases. Implants in the form of 4 mm X 4 mm cylinders were inserted into the defects such that their exposed surfaces lay just below the margins of the articular cartilage; control specimens consisted of unfilled, similar defects. The joints were thoroughly irrigated with sterile saline, and any debris resulting from creation of the defects was removed. All of the implanted materials were lightly press-fitted into the defects and a full range of passive motion of the joint was possible. The synovium was closed with a continuous 4-0 chromatized catgut suture ; the subcutaneous tissue with interrupted sutures of the same material; and the skin with inter-
RESTORATION IN POROUS BIOMATERIALS
169
rupted 3-0 braided silk. The initial six rabbits had their operative wounds dressed with vibesate (Aeroplast dressing; Park, Davis & Go., Detroit, Mich. 48232), spray-on plastic bandage; these wounds healed poorly, however, and superficial infections with Pasteurella multocida occurred. These infections were treated locally by opening the skin incisions, cleansing them thoroughly with surgical soap, and administering 50 mg oral doses of tylosin (Tylocine, 50 mg; Lilly and Co., Indianapolis, Ind. 46206) daily for 5 days. With daily antiseptic dressing changes, it was determined that only one deep infection involving the joint had occurred; this was the right knee of animal number 3, a control specimen. With discontinuation of the dressing described above and with application following surgery of a dry, sterile gauze dressing, held in place by paper tape, all subsequent wounds healed uneventfully. Immediately postoperatively, and with the animals still anesthetized, standard roentgenograms were taken of the distal femora; these were repeated a t weekly intervals until the animals were killed on a preselected schedule, so that representative specimens of each type could bc analyzed a t 1 day, 1 week, 6 weeks, 3 months, and 6 months postoperatively. At the end of each experimental period, the animals were killed by the intravenous injection of pentobarbital sodium (Somlethol, 6 g/ml; J. A. Webster, Inc., No. Billerica, Mass. 01862) (6 g/2.25 kg wt). The distal femora were immediately removed from the animals and the soft tissue dissected from them. Color and black-and-white photographs of each specimen were taken, and the gross appearance of each healing defect correlated with the histology seen. Previously describedL0techniques of hard tissue preparation for microradiography and fully mineralized histology were used on each specimen. The microradiographs were made from 100 pm thick sections; these were then ground, by hand, to thickness of 35-70 pm and then stained with Paragon (Paragon Multiple Stain; C. & C. Paragon Co., Inc., Bronx, N.Y.) and Safranin-0 (Safranin-0 Stain; Fisher Scientific Co., Fair Lawn, N.J. 07430) for histologic evaluation. Adjacent sections from each methacrylate-embedded specimen were examined with the scanning electron microscope and the electron microprobe. These latter analyses were for the elemental constituents of calcium, phosphorous, and sulfur in all specimens, and aluminum and titanium where applicable. The use of the electron
170
CHIROFF ET AL.
microprobe for sulfur as a means of characterizing hyaline cartilage, fibrocartilage, and bone has recently been described ;I1 this was correlated, as much as possible, with the histology seen.
RESULTS As has been stated, one control joint was the site of a deep infection. Of the 48 rahbits operated upon, only one was not moving about in its cage by the evening of surgery and normally active by the next day. That exception was a rabbit in which a major portion of the patellar tendon, controlling knee extension, was sectioned a t the time of surgery; this animal’s gait remained impaired and it was sacrificed 1 week following surgery. None of the implants placed was visualized on the initial or followup in vivo x-rays; Figure 1 is a postmortem x-ray of a 1week hydroxy- ‘I apatite specimen. When the joints were opened and the specimens excised, no implant was found to be dislodged from the defect into which it was placed.
Fig. 1. Postmorteni x-riidiograph o f the left femoral condyles of a 1 week, entirely porous hydroxyapatite specimen. This implant may he slightly more depressed in the created defect t.han is desirable.
RESTOR.ATION I N POROUS BIOMATERIALS
171
Gross Appearance At 1 day, all pores were filled with clotted blood, and no tissue ingrowth or overgrowth was present. A t 1 week, essentially all of the defects were still significantly depressed at the margins. The control animals were obviously proliferating tissue from the depths of the defects. This tissue was yellowish-white in color and did not have the “glistening” appearance of the intact articular cartilage at, the margins of the defect. With the porous hydroxyapatite and alumina implants in place, a majority of the cellular proliferation appeared t o arise from the margins of the defect, i.e., the articular cartilage itself. This “creeping” proliferation of tissue as well as the sheen on its surface, not unlike that of the surrounding hyaline cartilage, suggested that attempts a t rrpair of the defect with that tissue were present. I n all longrr-trrm specimens, including controls, the created defects wcrc progressively lcss depressed with the passage of timr, and the absolute thickness of the tissues overgrowing the implants was increased (Figs. 2b and 2c).
Microradiography At 1 week post-insertion, all of the TiOz and hydroxyapatite implants had mineralized tissue present within the substance of their porosities (Fig. 2a); the control defects and aA1203 specimens did not show these changes. By 6 weeks, and thereafter for all specimens, each of the implanted materials had virtually total ingrowth with normal-appearing, mineralized bone (Figs. 2b and 2c).
Histology None of the specimens showed fibrous or other encapsulating responses suggestive of rejection. As regards the tissue growing into the body of the implants, t h r fully mineralized histology confirmed the microradiographic findings ; rrmarkablc, indeed, was the rapid and complete fashion in which the pores of the TiOz and hydroxyapatite became ingrown with bone (Fig. 3a) ; both the control and alumina sites were still filled with clotted blood a t 1 week postoperatively. Analysis of the tissues overlying the implants delineated quite diff erent responses for the hydroxyapatite and ceramic materials.
172
CHIItOFF E T AL.
Fig. 2. Representative microradiographs. (a) One week, entirely porous titania implant. Well within the substance of the Ti02 there is mineralization characteristic of new-formed bone (60X). (b) Six week, entirely porous, hydroxyapatite implant. Proliferation of the subchondral bone from the margins of the created defects has occurred; histology of this specimen is shown in Fig. 3b (40 X). (c) Six month, “capped” hydroxyapatite implant. The radiodense implant is virtually covered by a “plate” of new bone (60X).
I n the case of the former, newly formed bone and, accompanying it, hyaline cartilage came to reconstitute the articular surfaces (Figs. 3b and 3c). With both the alumina and titania, it was fibrocartilage which covered the implants (Fig. 3d). With the passage of time, the absolute thickness of all overgrowing tissues increased with constant filling-in of the depressions overlying the implants by newformed tissue. The fact that no difference in the patterns was seen between the entirely porous or “capped” implants was significant. The control defects were completely filled with bone by 6 months post-creation; their articular surfaces at that, and each earlier, test period were restored with fihrocartilage.
RESTORATION I N POROUS BIOMATERIALS
Fig. 2.
(continued)
173
174
CHIROFF E T AL.
Fig. 3. Fully mineralized histology. (a) One week, entirely porous, titania implant. Normal-appearing, cellular, and mineralized new bone is seen within the pores (Paragon, l 2 0 x ) . (b) Six week, entirely porous hydroxyapatite implant. This is the exact section from which the microradiograph of Fig. 2b was made; the proliferating bone is almost “osteophytic” as it extends towards the center of articular aspect of the implant (Paragon, 40X). (c) Six month, “capped” hydroxyapatite implant. There is total cellular and morphologic continuity between the intact articular cartilage a t the margin of the created defect and that tissue now forming the bone plate shown in Fig. 3b (Paragon, 1 2 0 ~ ) . (d) Three month, entirely porous alumina implant. No marginal proliferation of bone or cartilage is seen; the tissue overlying the implant has the appearance of fibrocartilage (Paragon, 120X).
Scanning Electron Microscopy and Electron Microprobe Analysis These studies confirmed the findings already described as far as both the intra-implant and articular surface responses. When the electron microprobe was used to determine the sulfation of that tissue which had come to lie b c t w c n thc implants and the joint cavity, it was found that a grcatcr weight-percent sulfur was present in the hydroxyapatite specimens than with the other two, or the controls.
RESTORATION I N POROUS BIOMATERIALS
Fig. 3.
(continued)
175
176
CHIROFF ET AL.
Fig. 3.
(continued)
DISCUSSION Regeneration of lost or damaged parts in lower animal forms is well established. With evolutionary ascendancy, however, it has been held that higher forms, particularly mammals, have apparently lost that regenerative capacity. Beckerl2 and Smith13have shown that the capacity for true regeneration of portions amputated and then subjected to appropriate electrical stimulation remains. It is not fanciful or whimsical to note that, in the human, the only tissue which truly heals itself in a regenerative fashion is bone; that is, for complete and satisfactory resumption of pre-injury status, bone tissue cannot simply heal itself with a fibrous scar-it must be bone. Since bone is in intimate contact with yet another of the body’s hard tissues, cartilage, there may be some carry-over of the principles involved in bone regeneration in healing which could also apply to articular cartilage. Recent 1\-ork1J4has shown that defects in hyalinc cartilage can heal themselves by the regeneration of an exactly similar and living tissue. When placed in either cntirely porous or surface-impermeable configurations immediately beneath the damaged articular cartilage of
RESTORATION I N POROUS BIOMATERIALS
I77
a weight-bearing joint in a rabbit, the hydroxyapatite implants enhanced and facilitated regenerative healing of the marginal hyaline cartilage and subchondral bone to restore the articular surface. In the cases of the TiOz and (rA1203 implants, as well as the control animals, the articular surface was reconstituted by fibrocartilage. If a mammalian species such as the rabbit has the capacity to regenerate its hyaline articular cartilage across a 4 mm defect, it is certainly not possible to say that our hopes for the restoration of function to diseased or damaged joints will necessarily follow. However, the marriage of appropriate surgical techniques, specific materials, and, possibly, porous implant devices could lead to improved treatment of many individuals who are, and will be, disabled by one form or another of comprised function of their diarthrodial joints. Furthermore, as optimal materials and configurations emerge, it should be possible to translate the knowledge gained towards usefulness in treating some of the wide variety of naturally occurring arthritic and arthrotic states in the canine, bovine, and equine species.'"'' The authors wish to gratefully acknowledge the assistance of Frank P. Romano, I11 and Leo Tarhay in the completion of this research. This work was supported by a National Science Foundation Grant No. GH-38944.
References 1 . B. Baker, R. 0. Becker, and J. Spadaro, Clin. Orthop., 102, 251 (1974). 2. J. N. Weber, R. Greer, B. Voight, E. W. White, and H. Roy, J . Ultrastruct. Res., 26, 355 (1969). 3. S. F. Hulbert, F. A. Young, R.. S. Mathews, J. J. Klawitter, C. D. Talbert, and F. H. Stelling, J . Riomed. Muter. Res., 4, 433 (1970). 4. It. A. White, J . N. Weber, and E. W. White, Science, 176, 922 (1972). 3 . E. W. White, J. N. Weber, I). Roy, E. L. Owen, 11. T . Chiroff, and H . A. White, J . Riomed. Muter. Res., 9, 23, (1975). 6. It. T. Chiroff, E. W. White, J. Weber, and I>. Roy, J . Biomed. Muter. Res., 9, 29 (1975). 7. J. Lebiedzik, K. G. Burke, S. Troutman, G. G. Johnson, Jr., and E. W. White, in Scanning Electron Microscopy/l.W3, IITNI Publications, Chicago, 1973, p. 121. 8. L. Tarhay, W. It. Buessem, arid E. W. White, Ceramic Bull., 52, 394 (1973). 9. D. M. Roy and S. Kurtoasy-Linnehan, Nature, 247, 220 (1974). 10. J. Jowsey, P. J. Kelly, B. L. Kiggs, A. J. Bianco, Jr., 11. A. Scholz, and J . Gershon-Cohen, J . Bone Joint Surg., 47A, 785 (1965). 11. R. T. Chiroff, It. A. White, 13. W. White, and L. Tarhay, Calcif. Tissue Res., 17, 87 (1974).
178 12. 13. 14. 15.
CHIROFF ET AL. R. 0. Becker, Nature, 235, 109 (1972). S. 0. Smith, Anat. Rec., 158, 89 (1967). W. E. Riddle, J . Amer. Vet. Med. Assoc., 157, 1471 (1970). E. P. Leonard, Orthopaedic Surgery ofthe Dog and Cat, Saunders, Philadelphia,
1971. 16. S. E. Olsson, J . Small Animal Prac., 12, 333 (1971). 17. C. W. Schwabe, Veterinary Medicine and Human Health, Williams and Wilkins, Baltimore, 1964.
Received November 3, 1975 Revised March 8, 1976
