Effect of an in Vivo Environment on the Strength of Bone Cement W. ROSTOKER, Department of Materials Engineering, University of Illinois, Chicago Circle, Chicago, Illinois 60680, P. LEREIM and J. 0. GALANTE, Department of Orthopedic Surgery, Rush-Presbyterian-St. Luke’s Medical Center, Chicago, Illinois 60612 Summary Bar-shaped polymethylmethacrylate test specimens removed from rabbits after implantation for times up to 26 months showed a significant change in fracture stress as determined by three-point bending in the period between 12 and 26 months. There were no adverse findings in the tissue which developed around the bone-cement test bars.
INTRODUCTION The fixation of most human total-joint- prostheses to bone is achieved by the use of polymethylmethacrylate. This material acts as a grout providing a mechanical interlock and a force-transfer function between the device and the host bone. Accordingly, it must sustain the stresses generated by load transfer without cracking. Since the material is not ductile, stresses can only be sustained or relaxed by fracture. If the cement embedment were to crack once, it would crack repeatedly and this would lead to loosening of the device, one of the serious complications of total hip replacement. Several authors have dealt with the problems concerning the mechanical aspects of m e t h y l m e t h a ~ r y l a t e . Others ~ ~ ~ ~ ~have ~ ~ studied the effect of time upon the mechanical properties of polymethylmethacrylate. Jaffe et al.5 (1974) stored bone-cement samples in bovine serum at 37°C up to two years and found deterioration neither in static properties nor in compression-fatigue behavior. Freitag and Cannon3 (1977) found higher fracture toughness when testing was conducteu on cement specimens immersed in bovine serum at 37°C compared with specimens tested in air. Journal of Biomedical Materials Research, Vol. 13, 365-370 (1979) 0021-9~04/79/O01~~-0~65$01.00 Q 1979 John Wiley & Sons, Inc.
366
ROSTOKER, LEREIM, AND GALANTE
Degradation processes are usually environment-related so that it is important to design an in uiuo experiment. Time is also an important factor and since degradation processes are generally not well understood, accelerated tests are usually of only speculative significance. The purpose of this study was to evaluate the bending strength of' polymethylmethacrylate beams implanted in rabbit muscle for periods up to 26 months.
MATERIALS A N D METHODS Seventy-one New Zealand male, white rabbits, 8-9 lb in weight, were used as hosts. Freshly molded bars of Simplex P bone cement* were implanted under sterile surgical conditions in the paravertebral muscles, six implants per animal, except one rabbit that received only five. In the expectation that some animals might die before their planned time of sacrifice, the number of animals was increased according to the length of the experiment. The number of unplanned deaths was not as large as anticipated so that the number of test specimens evaluated did increase with implant period. The specimens of bone cement were of the following dimensions: 11/2 X '/4 X l/4 in. (38.1 X 6.3 X 6.3 mm). They were molded at the time of implantation using a partitioned metal mold. Moderate hand pressures were applied to achieve proper fill. Removal from the mold was done as soon as polymerization was completed and immediate implantation was performed. After sacrifice the bone-cement bars were immersed in 0.9% NaCl solution until mechanical testing which was performed within 24 hr. Mechanical tests to fracture were performed in air in a three-point bending mode. The specimens were supported by 2.3-mm-diam polished steel rods at a distance of 31.7 mm. The mid-span contact line was engaged by a polished steel bar of 4.8 mm diam, to which the weights were attached. After each load increment, a time lapse of 30 sec was allowed before another increment was added. The dead-weight increments were small compared with the ultimate load, i.e., 0.4-2.3 kg in 31.8 kg. The outer fiber stress at fracture was calculated: c = 2Fl/bh2 * Howmedica, packages dated July-August, 1974.
I N VIVO ENVIRONMENT ON BONE CEMENT
:367
where F is the applied force; 1, the distance between supporting beams; 6 , the width; and h , the height of the specimen. 1 = 31.7 mm, b = h = 6.3 mm. A one-way analysis of variance was used to analyze statistically the fracture data. Histological examination was performed on the tissue surrounding the bone-cement bars. The tissue was fixed immediately in 10% buffered formaldehyde for at least 48 hr, then dehydrated through grades of alcohol. The tissue specimens were embedded in paraffin from which 6-pm-thick sections were cut. The plane of section contained the area where the bone-cement bars had been implanted. Six sections were obtained for each of the specimens examined. The sections were stained with hematoxylin eosin.
RESULTS The outer fiber stress at fracture was 7010 psi (48.3 MPa) one day after molding. Six and twelve months after implantation, it rose to 7080 psi (48.8 MPa) and 7370 psi (50.9 MPa), respectively. From 12 to 26 months after implantation, there was a significant drop in stress at fracture to 6760 psi (46.6 MPa) ( P = 0.023) as shown in Table I. A fibrous layer of variable thickness (30-300 pm; average, 100 Fm) developed around the bone-cement bars. Adjacent to the cement a discontinuous seam of cells, commonly macrophages, was seen. Between this layer and the muscle there was a zone of fatty, loose connective tissue of variable thickness. In a few samples, an increased number of cells was found which reflected a low-grade inflammatory reaction. There was no obvious foreign-body reaction, nor was there any sign of tissue necrosis in any of the samples. One-third of the samples showed bone formation in and around the fibrous layer-both woven bone and mature bone. There were no histological differences between samples removed at 6,12, and 24 months after implantation, TABLE I Time Dependence of Fracture Stress of Bone Cement Tested in Flexure Period 1 2 3
4
Time After EnvironNo. of Outer Fiber Stress a t Fracture Molding ment Specimens Mean, psi (MPa) Std. Dev., psi (MPa) 1 day 6months 12 months 26 months
air inuiuo in viuo
in uiuo
8 12 16 35
7010 (48.3) 7080 (48.8) 7370 (50.9) 6760 (46.6)
549 (3.8) 710 (4.9) 829 (5.7) 864 (6.0)
368
ROSTOKER, LEREIM, AND GALANTE
nor did the organs show changes which could be referred to the implants.
DISCUSSION With regard to the loading system, it was found that fracture involved a small delayed-failure process: upon reaching a given load level, fracture was not immediate, but might occur as much as 30 sec later. Thus, any continuous loading procedure did not reflect the realities of the fracture process. Moreover, a loading rate did not easily translate into a strain rate in 3-point bending; the bending of a low-modulus material brought into question corrections for large deflections;ll the short beam geometry could also require corrections for surface stress. With all of these considerations the simple deadweight loading system described above was used. The breaking load converted to uncorrected surface tensile stress provided a measure which could serve in the present experiment to give statistical indication of a trend in change of strength. It was not intended that these fracture stresses be used for any design calculation. There are rather few comparable measurements in the literature. The absolute magnitudes of mean outer fiber stress at fracture compared quite well to those cited by Lautenschlager et a1.6 They cited a mean value after immersion for 40 days in 37°C water of 45.92 MPa and a standard deviation of 4.96 MPa. There are few reports pointing towards a possible deterioration of methylmethacrylate in uiuo. Scales and Zarek,'O during a study of the causes of failure of the Judet prosthesis, found there was some evidence for change in the mechanical properties of acrylic resin in the course of time. Pautucekg stated that monomethylmethacrylate is broken down in a series of reactions starting with the hydroxylation of the double bond to pyruvic acid. Oppenheimer et a1.8showed that when labeled polymethylmethacrylate fibers were implanted in rats, traces of I4C could be recovered from the urine after 54 weeks. The extended interval from implantation to the appearance of urinary radioactivity indicated that the radioactivity could not be due to any residual monomer. Lee et al.7 stored test specimens in isotonic saline solution at 37°C and found an increase in compressive strength after seven days, but, tested after 6 and 12 months, a decrease of 7 and 81/2%, respectively, had occurred.
IN VZVO ENVIRONMENT ON BONE CEMENT
369
TABLE I1 t-Test Probability of Two Sets of Measurements Being Identical ComDarison of Periods
Sienificance”
1 with 2 1with 3 2 with 3 1 with 4 2 with 4 3 with 4
0.575 0.243 0.525 0.341 0.110 0.023 ~
~~
Significance a t 0.02 level means that there is a 2% chance that these two sets of numbers are identical. a
In our experiment there was a nonsignificant trend towards increasing strength in the first 12 months. However, there was a significant drop in strength between 12 and 26 months (Table 11). This finding, if confirmed by additional in uiuo experiments, would be of importance in the long-term performance of total joint replacement devices.
References 1. W. J. Astleford, M. A. Asher, V. S.Lindholm, and C. A. Rockwood, “Some Physical and Mechanical Factors Affecting the Simple Shear Strength of Methylmethacrylate,” Clin. Orthop. Rel. Res., 108, 145-148 (1975). 2. J. Charnley, Acrylic Cement in Orthopedic Surgery, Williams and Wilkins, Baltimore, 1970. 3. T. A. Freitag and S. L. Cannon, “Fracture Characteristics of Acrylic Bone Cements. 11. Fatigue,”J. Biomed. Matar. Res., 11,609-624 (1977). 4. C. A. Homsey, “Some Mechanical Aspects of Methylmethacrylate Prosthesis Seating Compound,” T h e H i p , Mosby, St.. Louis, 1973, pp. 156-163. 5. W. L. Jaffe, R. M. Rose, and E. L. Radin, “On the Stability of the Mechanical Properties of Self-curing Acrylic Bone Cement,” J . Bone Jt. Surg., 56A, 1711-1714 (1974). 6. E. P. Lautenschlager, J. J. Jacobs, G . W. Marshall, and P. R. Meyer, Jr., “Mechanical Properties of Bone Cements Containing Large Doses of Antibiotic Powders,” J . Biomed. Mater. Res., 10,929-938 (1976). 7. A.
INTRODUCTION The fixation of most human total-joint- prostheses to bone is achieved by the use of polymethylmethacrylate. This material acts as a grout providing a mechanical interlock and a force-transfer function between the device and the host bone. Accordingly, it must sustain the stresses generated by load transfer without cracking. Since the material is not ductile, stresses can only be sustained or relaxed by fracture. If the cement embedment were to crack once, it would crack repeatedly and this would lead to loosening of the device, one of the serious complications of total hip replacement. Several authors have dealt with the problems concerning the mechanical aspects of m e t h y l m e t h a ~ r y l a t e . Others ~ ~ ~ ~ ~have ~ ~ studied the effect of time upon the mechanical properties of polymethylmethacrylate. Jaffe et al.5 (1974) stored bone-cement samples in bovine serum at 37°C up to two years and found deterioration neither in static properties nor in compression-fatigue behavior. Freitag and Cannon3 (1977) found higher fracture toughness when testing was conducteu on cement specimens immersed in bovine serum at 37°C compared with specimens tested in air. Journal of Biomedical Materials Research, Vol. 13, 365-370 (1979) 0021-9~04/79/O01~~-0~65$01.00 Q 1979 John Wiley & Sons, Inc.
366
ROSTOKER, LEREIM, AND GALANTE
Degradation processes are usually environment-related so that it is important to design an in uiuo experiment. Time is also an important factor and since degradation processes are generally not well understood, accelerated tests are usually of only speculative significance. The purpose of this study was to evaluate the bending strength of' polymethylmethacrylate beams implanted in rabbit muscle for periods up to 26 months.
MATERIALS A N D METHODS Seventy-one New Zealand male, white rabbits, 8-9 lb in weight, were used as hosts. Freshly molded bars of Simplex P bone cement* were implanted under sterile surgical conditions in the paravertebral muscles, six implants per animal, except one rabbit that received only five. In the expectation that some animals might die before their planned time of sacrifice, the number of animals was increased according to the length of the experiment. The number of unplanned deaths was not as large as anticipated so that the number of test specimens evaluated did increase with implant period. The specimens of bone cement were of the following dimensions: 11/2 X '/4 X l/4 in. (38.1 X 6.3 X 6.3 mm). They were molded at the time of implantation using a partitioned metal mold. Moderate hand pressures were applied to achieve proper fill. Removal from the mold was done as soon as polymerization was completed and immediate implantation was performed. After sacrifice the bone-cement bars were immersed in 0.9% NaCl solution until mechanical testing which was performed within 24 hr. Mechanical tests to fracture were performed in air in a three-point bending mode. The specimens were supported by 2.3-mm-diam polished steel rods at a distance of 31.7 mm. The mid-span contact line was engaged by a polished steel bar of 4.8 mm diam, to which the weights were attached. After each load increment, a time lapse of 30 sec was allowed before another increment was added. The dead-weight increments were small compared with the ultimate load, i.e., 0.4-2.3 kg in 31.8 kg. The outer fiber stress at fracture was calculated: c = 2Fl/bh2 * Howmedica, packages dated July-August, 1974.
I N VIVO ENVIRONMENT ON BONE CEMENT
:367
where F is the applied force; 1, the distance between supporting beams; 6 , the width; and h , the height of the specimen. 1 = 31.7 mm, b = h = 6.3 mm. A one-way analysis of variance was used to analyze statistically the fracture data. Histological examination was performed on the tissue surrounding the bone-cement bars. The tissue was fixed immediately in 10% buffered formaldehyde for at least 48 hr, then dehydrated through grades of alcohol. The tissue specimens were embedded in paraffin from which 6-pm-thick sections were cut. The plane of section contained the area where the bone-cement bars had been implanted. Six sections were obtained for each of the specimens examined. The sections were stained with hematoxylin eosin.
RESULTS The outer fiber stress at fracture was 7010 psi (48.3 MPa) one day after molding. Six and twelve months after implantation, it rose to 7080 psi (48.8 MPa) and 7370 psi (50.9 MPa), respectively. From 12 to 26 months after implantation, there was a significant drop in stress at fracture to 6760 psi (46.6 MPa) ( P = 0.023) as shown in Table I. A fibrous layer of variable thickness (30-300 pm; average, 100 Fm) developed around the bone-cement bars. Adjacent to the cement a discontinuous seam of cells, commonly macrophages, was seen. Between this layer and the muscle there was a zone of fatty, loose connective tissue of variable thickness. In a few samples, an increased number of cells was found which reflected a low-grade inflammatory reaction. There was no obvious foreign-body reaction, nor was there any sign of tissue necrosis in any of the samples. One-third of the samples showed bone formation in and around the fibrous layer-both woven bone and mature bone. There were no histological differences between samples removed at 6,12, and 24 months after implantation, TABLE I Time Dependence of Fracture Stress of Bone Cement Tested in Flexure Period 1 2 3
4
Time After EnvironNo. of Outer Fiber Stress a t Fracture Molding ment Specimens Mean, psi (MPa) Std. Dev., psi (MPa) 1 day 6months 12 months 26 months
air inuiuo in viuo
in uiuo
8 12 16 35
7010 (48.3) 7080 (48.8) 7370 (50.9) 6760 (46.6)
549 (3.8) 710 (4.9) 829 (5.7) 864 (6.0)
368
ROSTOKER, LEREIM, AND GALANTE
nor did the organs show changes which could be referred to the implants.
DISCUSSION With regard to the loading system, it was found that fracture involved a small delayed-failure process: upon reaching a given load level, fracture was not immediate, but might occur as much as 30 sec later. Thus, any continuous loading procedure did not reflect the realities of the fracture process. Moreover, a loading rate did not easily translate into a strain rate in 3-point bending; the bending of a low-modulus material brought into question corrections for large deflections;ll the short beam geometry could also require corrections for surface stress. With all of these considerations the simple deadweight loading system described above was used. The breaking load converted to uncorrected surface tensile stress provided a measure which could serve in the present experiment to give statistical indication of a trend in change of strength. It was not intended that these fracture stresses be used for any design calculation. There are rather few comparable measurements in the literature. The absolute magnitudes of mean outer fiber stress at fracture compared quite well to those cited by Lautenschlager et a1.6 They cited a mean value after immersion for 40 days in 37°C water of 45.92 MPa and a standard deviation of 4.96 MPa. There are few reports pointing towards a possible deterioration of methylmethacrylate in uiuo. Scales and Zarek,'O during a study of the causes of failure of the Judet prosthesis, found there was some evidence for change in the mechanical properties of acrylic resin in the course of time. Pautucekg stated that monomethylmethacrylate is broken down in a series of reactions starting with the hydroxylation of the double bond to pyruvic acid. Oppenheimer et a1.8showed that when labeled polymethylmethacrylate fibers were implanted in rats, traces of I4C could be recovered from the urine after 54 weeks. The extended interval from implantation to the appearance of urinary radioactivity indicated that the radioactivity could not be due to any residual monomer. Lee et al.7 stored test specimens in isotonic saline solution at 37°C and found an increase in compressive strength after seven days, but, tested after 6 and 12 months, a decrease of 7 and 81/2%, respectively, had occurred.
IN VZVO ENVIRONMENT ON BONE CEMENT
369
TABLE I1 t-Test Probability of Two Sets of Measurements Being Identical ComDarison of Periods
Sienificance”
1 with 2 1with 3 2 with 3 1 with 4 2 with 4 3 with 4
0.575 0.243 0.525 0.341 0.110 0.023 ~
~~
Significance a t 0.02 level means that there is a 2% chance that these two sets of numbers are identical. a
In our experiment there was a nonsignificant trend towards increasing strength in the first 12 months. However, there was a significant drop in strength between 12 and 26 months (Table 11). This finding, if confirmed by additional in uiuo experiments, would be of importance in the long-term performance of total joint replacement devices.
References 1. W. J. Astleford, M. A. Asher, V. S.Lindholm, and C. A. Rockwood, “Some Physical and Mechanical Factors Affecting the Simple Shear Strength of Methylmethacrylate,” Clin. Orthop. Rel. Res., 108, 145-148 (1975). 2. J. Charnley, Acrylic Cement in Orthopedic Surgery, Williams and Wilkins, Baltimore, 1970. 3. T. A. Freitag and S. L. Cannon, “Fracture Characteristics of Acrylic Bone Cements. 11. Fatigue,”J. Biomed. Matar. Res., 11,609-624 (1977). 4. C. A. Homsey, “Some Mechanical Aspects of Methylmethacrylate Prosthesis Seating Compound,” T h e H i p , Mosby, St.. Louis, 1973, pp. 156-163. 5. W. L. Jaffe, R. M. Rose, and E. L. Radin, “On the Stability of the Mechanical Properties of Self-curing Acrylic Bone Cement,” J . Bone Jt. Surg., 56A, 1711-1714 (1974). 6. E. P. Lautenschlager, J. J. Jacobs, G . W. Marshall, and P. R. Meyer, Jr., “Mechanical Properties of Bone Cements Containing Large Doses of Antibiotic Powders,” J . Biomed. Mater. Res., 10,929-938 (1976). 7. A.
