Bone-particle-impregnated bone cement: An in vivo weight-bearing study K.R. Dai Department of Orthopaedics, Ninth People's Hospital, Shanghai Second Medical University, Shanghai 200011, Peopk's Republic of China
Y.K.Liu,* J. B. Park, and C. R. Clarkt Departments of Biomedical Engineering and Orthopaedics,' University of lowa, lowa City, lowa 52242
K. Nishiyama Department of Orthopaedics, Keio University, Tokyo, Japan Z. K.Zheng
Department of Orthopaedics, Ninth People's Hospital, Shanghai Second Medical University, Shanghai 200011, People's Republic of China To evaluate an experimental inorganicbone-particle-impregnated bone cement, canine hip prostheses were implanted in dogs using a regular bone cement on one side and the experimental bone cement on the other. In a preliminary feasibility study, bone ingrowth into the resorbed bone-particle spaces was established 3 months after implantation in three dogs. In a more detailed study, twenty-eight (28) dogs were divided in four groups to delineate the effects of time on the phenomena of bony ingrowth. One month after implantation, active bone ingrowth into the bone cement was obvious. By 3 months postimplantation, the ingrowth appeared to have traversed the thickness of the bone-particle-impregnated cement.
By the fifth month, most of the interconnected inorganic bone particles were replaced by new bone. At the end of a year, the ingrown bone was mature and negligible new bone activity was present. Biomechanical pushout tests closely corroborated the histologic observations. The maximum shear strength of the cement/bone interface of the experimental side reached 3.6 times that of the control side at 5 months postimplantation. No further improvements were seen at 12 months postimplantation. A viable bone/cement interface may result in a better orthopedic implant fixation system by combining the advantages of both cement for immediate rigidity and biological ingrowth for longterm stability.
INTRODUCTION
Total (hip) joint replacement is a major success in orthopedic surgery. By and large, two methods are employed for the fixation of load-bearing im*To whom correspondence should be addressed. Part of this paper was presented at 35th Annual Meeting of the Orthopaedic Research Society, Las Vegas, NV, February 6-9, 1989. Journal of Biomedical Materials Research, Vol. 25, 141-156 (1991) CCC 0021-9304/91/020141-16$04.00 0 1991 John Wiley & Sons, Inc.
DAI ET AL.
142
plants to the musculoskeletal system: (a) use of grouting materials, e.g., polymethylmethacrylate (PMMA), as bone cement between bone and prosthesis; (b) direct apposition of tissue onto porous or nonporous implant surfaces. Both fixation techniques have advantages and disadvantages. The main advantages of cement fixation are: fast fixation time and surgically forgiving. The chief disadvantages of PMMA are: complications arising from cement polymerization and distribution, and the inherent material mismatch among implant, cement, and bone. The main merit of tissue ingrowth fixation is its viability. Its demerits are (a) the initial fixation time is long compared to cement fixation, (b) it is surgically not forgiving, and (c) an even greater mismatch of material properties with bone as compared to PMMA. Our group has undertaken a systematic study to explore the potentialities of mixing resorbable inorganic bone particles of optimal size (150-300 pm) with currently used bone cement to provide good initial fixation while allowing tissue ingrowth into the cement to maintain long-term stability. We have shown previously’” that incorporation of 30% (by weight) inorganic bone particles improved the fatigue life of bone cement by an order of magnitude without unduly lowering its tensile strength. Recently, we have shown in an in vivo rabbit study that when plugs of as-received PMMA (control) and inorganic-bone-particle-impregnated PMMA (experimental) were implanted into the femora, bony ingrowth was obvious on the experimental plugs. However, because the preparation was a non-weight-bearing one, there was biomechanical evidence of resorption when no physiologically significant stresses were present directly in the ingrowth t i s ~ u eThe . ~ interfacial shear stresses in push-out tests reached a maximum at 7 weeks and started to fall significantly at 9 weeks. The present study is a logical extension to evaluate the use of inorganicbone-particle-impregnated PMMA in an in vivo weight-bearing preparation. The premise of the protocol is that (a) the inorganic-bone-particle-impregnated PMMA plays the beneficial roles of a grout, and (b) once the bone particles were resorbed in vivo and bony ingrowth into the voids takes place we gain all the merits of biological fixation. The study was divided into two phases. Phase I was to test the feasibility of bone ingrowth into inorganic-bone-particle-impregnated PMMA in a weightbearing preparation, and, if affirmative, to undertake Phase 11- a kinetic study of the ingrowth phenomenon as a function of time. PHASE I: MATERIALS A N D METHOD
Bone particles were prepared using the same technique described earlier’ except that dead canine bones were used. A femoral prosthesis with a constant diameter (7 mm) round stem4r5was employed for this phase of the study. The surgical procedure was similar to that used previously: Using an aseptic technique, a lateral approach to the hip joint was made. After osteotomy of the femoral neck, the joint capsule was incised, the ligament teres was cut, and the femoral head was removed from the acetabulum. The intramedullary
BONE-PARTICLE-IMPREGNATED CEMENT
143
cavity was drilled and reamed to 10-mm diameter to a 10-cm length. The cavity was rinsed with saline and suctioned several times, then a vent tube was inserted to allow the escape of trapped blood and air as the bone cement was forced in. The vent tube was removed and the hip prosthesis was placed. Once the cement hardened, the joint was reduced. Standard, commercially available, (control)bone cement was used in one leg and the 30%bone-particleimpregnated-bone cement was used in the contralateral leg. Both hip implantations were done at the same time. The dogs were transported to an outdoor kennel after recovery. A total of three dogs were used in Phase I of the study. After 3 months, the dogs were sacrificed and the femora with implants were retrieved and radiographed. One at a time, the femora were mounted on a Bronwill Scientific sectioning saw and cut with a diamond blade into approximately two millimeter thick sections for mechanical testing. Three histological samples 5 mm thick were obtained at three levels as shown in Figure 1. Push-out mechanical tests were performed on a hydraulically controlled machine (MTS, Minneapolis, MN) using 2.5 mm/min cross-head speed at room temperature. By keeping the specimens moist with Ringer solution, efforts were made not to dry the specimens. The size of the plunger and the support were varied with the size of the bonekement interface along the longitudinal axis of the bone. A special jig, consisting of 8 squared inserts around a hollow brass cylinder as shown in Figure 2, was used to vary the size of the support. Also, the size of the circular plunger was varied to make the closest possible clearance (0.5 mm) between support and plunger. The shape of the interface between bone and bone cement was mostly circular due to pre-drilling of the intramedullary canal. Where the diameter of the bone was considerably larger than the drilled hole, the interface (bonekement) was sometimes irregular in which case only an average diameter was used to calculate the shear area of interface. The load-displacement curve was recorded on an X-Y plotter from which the maximum interfacial shear strength was calculated.
A
HISTOLOGY SPECIMENS
1:
TEST
T w
Figure 1. Sectioning scheme of the retrieved femur with implant for mechanical and histological preparation.
DAI ET AL.
144
1 ’ Adjustable square bars
Figure 2. The special jig designed for mechanical push-out tests using a materials testing machine. The actuator of the testing machine is fitted with variously sized cylindrical plungers to accomplish the specimen push-out.
The histology samples were fixed immediately in a 10% buffered formalin for more than one week following the procedure of Kenner et aL6 The samples were then put into successive ethanol solutions (80,85,90,95, and 100%) for at least 24 h each. The samples were next desiccated in a vacuum for 24 h and cast in polyester resin. After hardening for 2 days, they were cut with a diamond wafer saw into 0.3-mm-thick sections, then ground to 50-100 pm thin by using 320-, 400-, and 600-grid sandpapers on a rotating polishing wheel. The samples were stained with H & E for histologic evaluation. PHASE I: RESULTS A N D DISCUSSION
The mechanical push-out test results of the phase I feasibility study are summarized in Table I and plotted in Figure 3. It is obvious that there is a large variation in the interfacial shear strength due to the anatomic location of the test specimens as well as animal to animal variations. Nevertheless, paired t-statistics showed no significant difference between the experimental and control side for dog number 2. However, the other two dogs showed varying degrees of freedom which had a p value of less than 0.025. The results tend to indicate that the bone particle impregnated bone cement had a higher interfacial shear strength between bone and cement after 3 months of implantation. This is corroborated by the histological results. No new bone formation was evident in Figure 4A around the cement but new bone formation was demonstrated adjacent to the bone-cement interface in the experimental specimens (Fig. 4B).In addition, Figure 4 shows that remnants of dead, inorganic bone which were not resorbed are present simultaneously with new bone cells. Other investigators have tried similar ideas to the present one, i.e., the inclusion of resorbable particles into bone cement of polymethylmethacrylate However, to our knowledge, this is the first time that bony tissue ingrowth into bone-particle-impregnated bone cement was shown in a loadbearing prosthesis preparation.
BONE-PARTICLE-IMPREGNATED CEMENT
145
TABLE I Phase I Results of Mechanical Push-Out Tests; the Maximum Interfacial Shear Strength between Bone and Bone Cement Experimental Side (MPa) Dog 1 Dog 2
Sec. No. 1 2 3 4 5 6
1.42 1.41 2.07 1.22 0.96 0.97
Average
2.06 0.92 1.86 0.45 0.45 0.38
1.64 1.00 1.48 1.18 0.67 0.59
Dog 3
Average
1.29 0.80 0.92 0.19 0.13 0.42
1.29 1.06 0.83 0.41 0.23 0.32
1.42 0.66 0.51 1.87 0.59 0.41
Control Side (MPa) Dog 1 Dog 2
Sec. No.
Dog 3
1.05 0.28 0.19 0.55 0.46 0.39
1.52 2.09 1.39 0.50 0.09 0.16
Paired t test t-value Degree of freedom p value
2.93 5 C0.025
2.54 5 ~0.05
0.12 5 NS**
3.36 5 c0.01
*Paired t-tests were made along the section between experimental and control side. **NS:not significant.
il
U
Cnntrol Experimental
-
250
1.5
- 200 MAXIMUM INTERFACIAL SHEAR STRENGTH
I 4-d -
150
I00
0.5
SO
0
1
2
3
4
Proximal
5
6
n-
Distal
SAMPLE POSITION
Figure 3. Plot of the maximum interfacial shear strength versus sample position. Phase I of study.
DAI ET AL.
146 PHASE 11: MATERIALS A N D METHODS
After affirmation of bony ingrowth into inorganic-bone-particleimpregnated PMMA 3 months postimplantation in three animals, we proceeded to the kinetic study of Phase 11. Twenty-eight (28) mongrel dogs, with weight ranging from 12.0 to 19.5 kg, were randomly assigned to each of four groups. The prostheses, shown in Figure 5, were designed based on a geometric study of 50 grown mongrel dogs. The design of the canine formal prosthesis was modified in order to facilitate the precoating procedure and to enhance the torsional strength between the stem and the precoated cement. The stem of the 316 L stainless-steel prosthesis was straight with a V-shaped cross-section of 5.5 mm diameter and a 135" stem to neck angle. Three sizes of femoral heads were manufactured with 13-, 16-, and 19-mm diameters in order to fit different canine acetabula. All of the prostheses were precoafed with PMMA (Simplex-P) under pressure in accordance with the procedure outlined in Barb et al.4and Park et aL5After precoating, the stems were cylindrical with a coating thickness of at least 1 mm. The surgical techniques of prosthesis placement were similar to Phase I. After completion of surgery of both hindlimbs, the dogs were suspended in the cage by a cotton girdle around their abdomens. This technique of suspension totally prevented the possible pressure necrosis and slight superficial infection which might have occurred during Phase I of our study. After 3 days, the dogs were released from their suspensions and roamed freely in their cages. Both hindlimbs appear to be similarly loaded as no discernable limp was observed in any of the dogs. No force plate analyses were done to verify quantitatively the above observation. The experimental animals were divided into four groups, each consisting of 7 dogs. One, three, five and twelve months after the hemiarthroplasty, the appropriate group was sacrificed. Tetracycline HC1 was injected i.m. at a dose of 30 mg/kg 48 h before sacrifice for fluorescence labeling. Both femora were excised with the implants after radiographically ascertaining that the artificial femoral heads were in their normal positions. The femur with implant was cut into 10 5-mm sections with a diamond saw. The fourth and seventh sections were used for histology and the remainder used for mechanical push-out tests in a similar fashion as described in Phase I and Figure 2. PHASE 11: RESULTS
Some of the samples taken at the first, third, fifth, and twelfth month postsurgery were stained with H & E for histologic assessment. No obvious tissue reactions were found on either hindlimb. A fibrous tissue layer was found in a few slices of the control side. These layers were thin ( 4 0 pm) and were often broken into several parts. No such fibrous layer could be found on the experimental side, i.e., the side with 30% inorganic-bone-particle impregnation of the bone cement. In lieu of a fibrous tissue layer, new bone ingrowth into the experimental cement increased with time. The new bone formation was always at or near the bonekement interface, thus implying that the new
BONE-PARTICLE-IMPREGNATED CEMENT
Figure 4. H & E histological sections, (A) control side, (B) experimental side ( ~ 1 0 0 )Note . bony tissue in the experimental cement. P = prosthesis, C = cement, B = bone. Phase I of study.
147
DAI ET AL.
148
t 1 @ 5,5-
A
-
-A
N Figure 5. Design variables of the canine femoral prosthesis used in this study.
bone was associated with resorption of the inorganic bone particles. For those sections which were not stained, the bone and bone cement interface was searched for fluorescence. When a field of fluorescence was found, a photograph was taken. The same field was photographed again under light microscopy. The two photographs were compared. If bone structure was found at the same sites where fluorescence was exhibited, we took it as evidence of new bone ingrowth. After one month of implantation, inorganic bone-particles at the periphery of the bone cement were replaced by new ingrown bone as shown in Figure 6. Typically, Figure 6A illustrates fluorescence and 6B the corresponding bony ingrowth seen under light microscopy. Figures 7A and 7B show the fluorescence and the corresponding new bone ingrowth 3 months postimplantation. Note that bone had advanced practically through the whole thickness of the bone cement and had almost reached the surface of the precoated prosthesis. Most of the fields examined still showed marked fluorescence, indicating that new bone is still being laid down. In the 5month samples, new bone ingrowth appeared to be nearly completed as shown in Figures 8A and 8B. The bone architecture appears more mature. In the 12-
BONE-PARTICLE-IMPREGNATED CEMENT
149
(b) Figure 6. One month postimplantation. (A) Fluorescence micrograph; (B) light micrograph of the same location. Legends in micrograph: c = cortical bone, b = bone particles and bone cement and p = precoated prosthesis. Note new bone ingrowth at bonekement interface (40 x 4).
month specimens shown in Figure 9A and 9B, the fluorescence intensity was markedly reduced with neither new evidence of active bone ingrowth nor architectural changes. The results of the mechanical push-out tests are tabulated in Table I1 and analyzed statistically. The mean and standard deviation of the maximum in-
DAI ET AL.
150
(b) Figure 7. Three months postimplantation. (A) Fluorescence micrograph, (B) light micrograph of the same location. Legends in micrograph: c = cortical bone, b = bone particles and bone cement and p = precoated prosthesis. Note bone ingrowth has traversed through practically the whole thickness of the inorganic-bone-particle-impregnated bone cement (100 x 4).
terfacial shear strength at the bonekement interface is graphically displayed as Figure 10. DISCUSSION
More than 30 years have elapsed since Sir John Charnley advocated the use of polymethylmethacrylate bone cement as a grout for total hip arthroplasty.
BONE-PARTICLE-IMPREGNATED CEMENT
151
Figure 8. Five months postimplantation. (A) Fluorescence micrograph; (B) light micrograph of the same location. Legends on micrograph: c = cortical bone, b = bone particles and bone cement and p = precoated prosthesis. Note bone ingrowth as indicated by fluorescence is reduced and the bone architecture appears more mature (100 x 4).
Late loosening and proximal femoral resorption are complications of cemented total joint replacement. Long-term clinical observations and postmortem retrievals of joints have contributed to a better understandin$ of late loosening in cemented joint arthr~plasty.'l-'~ An analysis of previously functioning cemented total joints, with no radiologic evidence of loosening, retrieved at autopsy, have shown the following: (a) Presence of many fatigue
152
DAI ET AL.
Figure 9. One year after implantation. (A) Fluorescence micrograph; (B) light micrograph of the same location. Legends on micrograph: c = cortical bone, b = bone particles and bone cement and p = precoated prosthesis. Note the negligible fluorescence activity and the organized structure of mature bone (100 x 4).
cracks in the bone cement. (b) The retrieved bone cement exhibited extremely low fatigue life. (c) The micromotions between the prosthesis and the femoral cortex, measured as late as 10 years after surgery, were of the same order of magnitude as have been previously reported immediately after insertion of the cemented femoral components. (d) Continued marked reductions of the
BONE-PARTICLE-IMPREGNATEDCEMENT
153
TABLE I1 Summary of Mechanical Push-Out Tests Measuring the Maximum Interfacial Shear Strength between Bone and Bone Cement
Postimplantation Period
Experimental Group (MPa)
1 month 3 months
4.21 2 0.94 (56) 6.73 2 0.42 (56) 7.42 2 0.46 (56)
Control Group (MPa)
Student t-test p value ~
5 months 12 months
2.98 2.19 2.04 2.00
7.35 f 0.50 (56)
(56) & 0.14 (56) 2 0.13 (56) ? 0.20 (56) 2 0.19
Y.K.Liu,* J. B. Park, and C. R. Clarkt Departments of Biomedical Engineering and Orthopaedics,' University of lowa, lowa City, lowa 52242
K. Nishiyama Department of Orthopaedics, Keio University, Tokyo, Japan Z. K.Zheng
Department of Orthopaedics, Ninth People's Hospital, Shanghai Second Medical University, Shanghai 200011, People's Republic of China To evaluate an experimental inorganicbone-particle-impregnated bone cement, canine hip prostheses were implanted in dogs using a regular bone cement on one side and the experimental bone cement on the other. In a preliminary feasibility study, bone ingrowth into the resorbed bone-particle spaces was established 3 months after implantation in three dogs. In a more detailed study, twenty-eight (28) dogs were divided in four groups to delineate the effects of time on the phenomena of bony ingrowth. One month after implantation, active bone ingrowth into the bone cement was obvious. By 3 months postimplantation, the ingrowth appeared to have traversed the thickness of the bone-particle-impregnated cement.
By the fifth month, most of the interconnected inorganic bone particles were replaced by new bone. At the end of a year, the ingrown bone was mature and negligible new bone activity was present. Biomechanical pushout tests closely corroborated the histologic observations. The maximum shear strength of the cement/bone interface of the experimental side reached 3.6 times that of the control side at 5 months postimplantation. No further improvements were seen at 12 months postimplantation. A viable bone/cement interface may result in a better orthopedic implant fixation system by combining the advantages of both cement for immediate rigidity and biological ingrowth for longterm stability.
INTRODUCTION
Total (hip) joint replacement is a major success in orthopedic surgery. By and large, two methods are employed for the fixation of load-bearing im*To whom correspondence should be addressed. Part of this paper was presented at 35th Annual Meeting of the Orthopaedic Research Society, Las Vegas, NV, February 6-9, 1989. Journal of Biomedical Materials Research, Vol. 25, 141-156 (1991) CCC 0021-9304/91/020141-16$04.00 0 1991 John Wiley & Sons, Inc.
DAI ET AL.
142
plants to the musculoskeletal system: (a) use of grouting materials, e.g., polymethylmethacrylate (PMMA), as bone cement between bone and prosthesis; (b) direct apposition of tissue onto porous or nonporous implant surfaces. Both fixation techniques have advantages and disadvantages. The main advantages of cement fixation are: fast fixation time and surgically forgiving. The chief disadvantages of PMMA are: complications arising from cement polymerization and distribution, and the inherent material mismatch among implant, cement, and bone. The main merit of tissue ingrowth fixation is its viability. Its demerits are (a) the initial fixation time is long compared to cement fixation, (b) it is surgically not forgiving, and (c) an even greater mismatch of material properties with bone as compared to PMMA. Our group has undertaken a systematic study to explore the potentialities of mixing resorbable inorganic bone particles of optimal size (150-300 pm) with currently used bone cement to provide good initial fixation while allowing tissue ingrowth into the cement to maintain long-term stability. We have shown previously’” that incorporation of 30% (by weight) inorganic bone particles improved the fatigue life of bone cement by an order of magnitude without unduly lowering its tensile strength. Recently, we have shown in an in vivo rabbit study that when plugs of as-received PMMA (control) and inorganic-bone-particle-impregnated PMMA (experimental) were implanted into the femora, bony ingrowth was obvious on the experimental plugs. However, because the preparation was a non-weight-bearing one, there was biomechanical evidence of resorption when no physiologically significant stresses were present directly in the ingrowth t i s ~ u eThe . ~ interfacial shear stresses in push-out tests reached a maximum at 7 weeks and started to fall significantly at 9 weeks. The present study is a logical extension to evaluate the use of inorganicbone-particle-impregnated PMMA in an in vivo weight-bearing preparation. The premise of the protocol is that (a) the inorganic-bone-particle-impregnated PMMA plays the beneficial roles of a grout, and (b) once the bone particles were resorbed in vivo and bony ingrowth into the voids takes place we gain all the merits of biological fixation. The study was divided into two phases. Phase I was to test the feasibility of bone ingrowth into inorganic-bone-particle-impregnated PMMA in a weightbearing preparation, and, if affirmative, to undertake Phase 11- a kinetic study of the ingrowth phenomenon as a function of time. PHASE I: MATERIALS A N D METHOD
Bone particles were prepared using the same technique described earlier’ except that dead canine bones were used. A femoral prosthesis with a constant diameter (7 mm) round stem4r5was employed for this phase of the study. The surgical procedure was similar to that used previously: Using an aseptic technique, a lateral approach to the hip joint was made. After osteotomy of the femoral neck, the joint capsule was incised, the ligament teres was cut, and the femoral head was removed from the acetabulum. The intramedullary
BONE-PARTICLE-IMPREGNATED CEMENT
143
cavity was drilled and reamed to 10-mm diameter to a 10-cm length. The cavity was rinsed with saline and suctioned several times, then a vent tube was inserted to allow the escape of trapped blood and air as the bone cement was forced in. The vent tube was removed and the hip prosthesis was placed. Once the cement hardened, the joint was reduced. Standard, commercially available, (control)bone cement was used in one leg and the 30%bone-particleimpregnated-bone cement was used in the contralateral leg. Both hip implantations were done at the same time. The dogs were transported to an outdoor kennel after recovery. A total of three dogs were used in Phase I of the study. After 3 months, the dogs were sacrificed and the femora with implants were retrieved and radiographed. One at a time, the femora were mounted on a Bronwill Scientific sectioning saw and cut with a diamond blade into approximately two millimeter thick sections for mechanical testing. Three histological samples 5 mm thick were obtained at three levels as shown in Figure 1. Push-out mechanical tests were performed on a hydraulically controlled machine (MTS, Minneapolis, MN) using 2.5 mm/min cross-head speed at room temperature. By keeping the specimens moist with Ringer solution, efforts were made not to dry the specimens. The size of the plunger and the support were varied with the size of the bonekement interface along the longitudinal axis of the bone. A special jig, consisting of 8 squared inserts around a hollow brass cylinder as shown in Figure 2, was used to vary the size of the support. Also, the size of the circular plunger was varied to make the closest possible clearance (0.5 mm) between support and plunger. The shape of the interface between bone and bone cement was mostly circular due to pre-drilling of the intramedullary canal. Where the diameter of the bone was considerably larger than the drilled hole, the interface (bonekement) was sometimes irregular in which case only an average diameter was used to calculate the shear area of interface. The load-displacement curve was recorded on an X-Y plotter from which the maximum interfacial shear strength was calculated.
A
HISTOLOGY SPECIMENS
1:
TEST
T w
Figure 1. Sectioning scheme of the retrieved femur with implant for mechanical and histological preparation.
DAI ET AL.
144
1 ’ Adjustable square bars
Figure 2. The special jig designed for mechanical push-out tests using a materials testing machine. The actuator of the testing machine is fitted with variously sized cylindrical plungers to accomplish the specimen push-out.
The histology samples were fixed immediately in a 10% buffered formalin for more than one week following the procedure of Kenner et aL6 The samples were then put into successive ethanol solutions (80,85,90,95, and 100%) for at least 24 h each. The samples were next desiccated in a vacuum for 24 h and cast in polyester resin. After hardening for 2 days, they were cut with a diamond wafer saw into 0.3-mm-thick sections, then ground to 50-100 pm thin by using 320-, 400-, and 600-grid sandpapers on a rotating polishing wheel. The samples were stained with H & E for histologic evaluation. PHASE I: RESULTS A N D DISCUSSION
The mechanical push-out test results of the phase I feasibility study are summarized in Table I and plotted in Figure 3. It is obvious that there is a large variation in the interfacial shear strength due to the anatomic location of the test specimens as well as animal to animal variations. Nevertheless, paired t-statistics showed no significant difference between the experimental and control side for dog number 2. However, the other two dogs showed varying degrees of freedom which had a p value of less than 0.025. The results tend to indicate that the bone particle impregnated bone cement had a higher interfacial shear strength between bone and cement after 3 months of implantation. This is corroborated by the histological results. No new bone formation was evident in Figure 4A around the cement but new bone formation was demonstrated adjacent to the bone-cement interface in the experimental specimens (Fig. 4B).In addition, Figure 4 shows that remnants of dead, inorganic bone which were not resorbed are present simultaneously with new bone cells. Other investigators have tried similar ideas to the present one, i.e., the inclusion of resorbable particles into bone cement of polymethylmethacrylate However, to our knowledge, this is the first time that bony tissue ingrowth into bone-particle-impregnated bone cement was shown in a loadbearing prosthesis preparation.
BONE-PARTICLE-IMPREGNATED CEMENT
145
TABLE I Phase I Results of Mechanical Push-Out Tests; the Maximum Interfacial Shear Strength between Bone and Bone Cement Experimental Side (MPa) Dog 1 Dog 2
Sec. No. 1 2 3 4 5 6
1.42 1.41 2.07 1.22 0.96 0.97
Average
2.06 0.92 1.86 0.45 0.45 0.38
1.64 1.00 1.48 1.18 0.67 0.59
Dog 3
Average
1.29 0.80 0.92 0.19 0.13 0.42
1.29 1.06 0.83 0.41 0.23 0.32
1.42 0.66 0.51 1.87 0.59 0.41
Control Side (MPa) Dog 1 Dog 2
Sec. No.
Dog 3
1.05 0.28 0.19 0.55 0.46 0.39
1.52 2.09 1.39 0.50 0.09 0.16
Paired t test t-value Degree of freedom p value
2.93 5 C0.025
2.54 5 ~0.05
0.12 5 NS**
3.36 5 c0.01
*Paired t-tests were made along the section between experimental and control side. **NS:not significant.
il
U
Cnntrol Experimental
-
250
1.5
- 200 MAXIMUM INTERFACIAL SHEAR STRENGTH
I 4-d -
150
I00
0.5
SO
0
1
2
3
4
Proximal
5
6
n-
Distal
SAMPLE POSITION
Figure 3. Plot of the maximum interfacial shear strength versus sample position. Phase I of study.
DAI ET AL.
146 PHASE 11: MATERIALS A N D METHODS
After affirmation of bony ingrowth into inorganic-bone-particleimpregnated PMMA 3 months postimplantation in three animals, we proceeded to the kinetic study of Phase 11. Twenty-eight (28) mongrel dogs, with weight ranging from 12.0 to 19.5 kg, were randomly assigned to each of four groups. The prostheses, shown in Figure 5, were designed based on a geometric study of 50 grown mongrel dogs. The design of the canine formal prosthesis was modified in order to facilitate the precoating procedure and to enhance the torsional strength between the stem and the precoated cement. The stem of the 316 L stainless-steel prosthesis was straight with a V-shaped cross-section of 5.5 mm diameter and a 135" stem to neck angle. Three sizes of femoral heads were manufactured with 13-, 16-, and 19-mm diameters in order to fit different canine acetabula. All of the prostheses were precoafed with PMMA (Simplex-P) under pressure in accordance with the procedure outlined in Barb et al.4and Park et aL5After precoating, the stems were cylindrical with a coating thickness of at least 1 mm. The surgical techniques of prosthesis placement were similar to Phase I. After completion of surgery of both hindlimbs, the dogs were suspended in the cage by a cotton girdle around their abdomens. This technique of suspension totally prevented the possible pressure necrosis and slight superficial infection which might have occurred during Phase I of our study. After 3 days, the dogs were released from their suspensions and roamed freely in their cages. Both hindlimbs appear to be similarly loaded as no discernable limp was observed in any of the dogs. No force plate analyses were done to verify quantitatively the above observation. The experimental animals were divided into four groups, each consisting of 7 dogs. One, three, five and twelve months after the hemiarthroplasty, the appropriate group was sacrificed. Tetracycline HC1 was injected i.m. at a dose of 30 mg/kg 48 h before sacrifice for fluorescence labeling. Both femora were excised with the implants after radiographically ascertaining that the artificial femoral heads were in their normal positions. The femur with implant was cut into 10 5-mm sections with a diamond saw. The fourth and seventh sections were used for histology and the remainder used for mechanical push-out tests in a similar fashion as described in Phase I and Figure 2. PHASE 11: RESULTS
Some of the samples taken at the first, third, fifth, and twelfth month postsurgery were stained with H & E for histologic assessment. No obvious tissue reactions were found on either hindlimb. A fibrous tissue layer was found in a few slices of the control side. These layers were thin ( 4 0 pm) and were often broken into several parts. No such fibrous layer could be found on the experimental side, i.e., the side with 30% inorganic-bone-particle impregnation of the bone cement. In lieu of a fibrous tissue layer, new bone ingrowth into the experimental cement increased with time. The new bone formation was always at or near the bonekement interface, thus implying that the new
BONE-PARTICLE-IMPREGNATED CEMENT
Figure 4. H & E histological sections, (A) control side, (B) experimental side ( ~ 1 0 0 )Note . bony tissue in the experimental cement. P = prosthesis, C = cement, B = bone. Phase I of study.
147
DAI ET AL.
148
t 1 @ 5,5-
A
-
-A
N Figure 5. Design variables of the canine femoral prosthesis used in this study.
bone was associated with resorption of the inorganic bone particles. For those sections which were not stained, the bone and bone cement interface was searched for fluorescence. When a field of fluorescence was found, a photograph was taken. The same field was photographed again under light microscopy. The two photographs were compared. If bone structure was found at the same sites where fluorescence was exhibited, we took it as evidence of new bone ingrowth. After one month of implantation, inorganic bone-particles at the periphery of the bone cement were replaced by new ingrown bone as shown in Figure 6. Typically, Figure 6A illustrates fluorescence and 6B the corresponding bony ingrowth seen under light microscopy. Figures 7A and 7B show the fluorescence and the corresponding new bone ingrowth 3 months postimplantation. Note that bone had advanced practically through the whole thickness of the bone cement and had almost reached the surface of the precoated prosthesis. Most of the fields examined still showed marked fluorescence, indicating that new bone is still being laid down. In the 5month samples, new bone ingrowth appeared to be nearly completed as shown in Figures 8A and 8B. The bone architecture appears more mature. In the 12-
BONE-PARTICLE-IMPREGNATED CEMENT
149
(b) Figure 6. One month postimplantation. (A) Fluorescence micrograph; (B) light micrograph of the same location. Legends in micrograph: c = cortical bone, b = bone particles and bone cement and p = precoated prosthesis. Note new bone ingrowth at bonekement interface (40 x 4).
month specimens shown in Figure 9A and 9B, the fluorescence intensity was markedly reduced with neither new evidence of active bone ingrowth nor architectural changes. The results of the mechanical push-out tests are tabulated in Table I1 and analyzed statistically. The mean and standard deviation of the maximum in-
DAI ET AL.
150
(b) Figure 7. Three months postimplantation. (A) Fluorescence micrograph, (B) light micrograph of the same location. Legends in micrograph: c = cortical bone, b = bone particles and bone cement and p = precoated prosthesis. Note bone ingrowth has traversed through practically the whole thickness of the inorganic-bone-particle-impregnated bone cement (100 x 4).
terfacial shear strength at the bonekement interface is graphically displayed as Figure 10. DISCUSSION
More than 30 years have elapsed since Sir John Charnley advocated the use of polymethylmethacrylate bone cement as a grout for total hip arthroplasty.
BONE-PARTICLE-IMPREGNATED CEMENT
151
Figure 8. Five months postimplantation. (A) Fluorescence micrograph; (B) light micrograph of the same location. Legends on micrograph: c = cortical bone, b = bone particles and bone cement and p = precoated prosthesis. Note bone ingrowth as indicated by fluorescence is reduced and the bone architecture appears more mature (100 x 4).
Late loosening and proximal femoral resorption are complications of cemented total joint replacement. Long-term clinical observations and postmortem retrievals of joints have contributed to a better understandin$ of late loosening in cemented joint arthr~plasty.'l-'~ An analysis of previously functioning cemented total joints, with no radiologic evidence of loosening, retrieved at autopsy, have shown the following: (a) Presence of many fatigue
152
DAI ET AL.
Figure 9. One year after implantation. (A) Fluorescence micrograph; (B) light micrograph of the same location. Legends on micrograph: c = cortical bone, b = bone particles and bone cement and p = precoated prosthesis. Note the negligible fluorescence activity and the organized structure of mature bone (100 x 4).
cracks in the bone cement. (b) The retrieved bone cement exhibited extremely low fatigue life. (c) The micromotions between the prosthesis and the femoral cortex, measured as late as 10 years after surgery, were of the same order of magnitude as have been previously reported immediately after insertion of the cemented femoral components. (d) Continued marked reductions of the
BONE-PARTICLE-IMPREGNATEDCEMENT
153
TABLE I1 Summary of Mechanical Push-Out Tests Measuring the Maximum Interfacial Shear Strength between Bone and Bone Cement
Postimplantation Period
Experimental Group (MPa)
1 month 3 months
4.21 2 0.94 (56) 6.73 2 0.42 (56) 7.42 2 0.46 (56)
Control Group (MPa)
Student t-test p value ~
5 months 12 months
2.98 2.19 2.04 2.00
7.35 f 0.50 (56)
(56) & 0.14 (56) 2 0.13 (56) ? 0.20 (56) 2 0.19
