- Email: [email protected]
A comparison of the viscoelastic properties of bone grafts
Clinical Biomechanics 21 (2006) 761–766 www.elsevier.com/locate/clinbiomech
A comparison of the viscoelastic properties of bone grafts A. Datta *, S. Gheduzzi, A.W. Miles Centre for Orthopaedic Biomechanics, Department of Mechanical Engineering, University of Bath, United Kingdom Received 1 April 2005; accepted 22 March 2006
Abstract Background. The expansion in joint arthroplasty surgery in the 1970s has resulted in a large group of patients who require revision arthroplasty for aseptic loosening. Impaction bone grafting to deal with bone stock loss has become an increasingly popular procedure in revision hip surgery. The results of this revision surgery are variable and very much dependent on the grafting techniques adopted at the time of surgery. In vitro testing of impaction bone grafting methods can constitute an important tool to improve long-term clinical results. The increasing clinical demand for human allograft limits its availability for use in in vitro laboratory studies therefore suitable experimental alternatives are required. Methods. Human, porcine and ovine cancellous bone grafts were morsellised and prepared for in vitro laboratory testing following a standard operative technique. Each graft type was compressed using a die plunger test. After compression to a predetermined level the graft was left to relax for 120 s. A comparison of the compression moduli and of the relaxation characteristics for each graft preparation was performed. In addition, the effects of washing the graft and of cartilage removal from the graft mixture were investigated. Results. This study has demonstrated that there is no statistical difference in the compression modulus or relaxation percentage between human and ovine graft preparations. The effect of removing cartilage and washing the graft mixtures were inconsistent with regard to alterations in the viscoelastic properties of the grafts. Interpretation. On the basis of the experiments performed we recommend the use of ovine bone graft as a suitable substitute for human allograft for in vitro testing of impaction bone grafting methods. The properties of ovine graft were similar for both compression moduli and relaxation properties to human allograft. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Allograft; Impaction bone grafting
1. Introduction The expansion in joint arthroplasty surgery in the 1970s has resulted in a large group of patients, who require revision arthroplasty for aseptic loosening. These cases can be technically demanding as bone stock loss and large bony defects are often present. To allow for adequate fixation of the revision components, the areas of bone loss and bony defects can be filled with bone graft. This physiological solution is preferred to earlier techniques using large
*
Corresponding author. Address: Flat 4, 37 Portland Road, Edgbaston, Birmingham B16 9HS, United Kingdom. E-mail address: [email protected] (A. Datta). 0268-0033/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2006.03.009
volumes of bone cement alone (Amstuz et al., 1982). The morsellised graft, is able to fill irregular defects and cavities in the host bone and provide a mechanical and biological scaffold that can be incorporated into the host tissue as bone remodeling takes place (Gie et al., 1996). Impaction bone grafting was originally developed as a surgical technique for acetabular reconstruction and has been further refined to repair large femoral defects and compensate for bone stock loss at the time of revision surgery (Schreurs et al., 2001; Slooff et al., 1999). It involves filling the bone cavities and defects with morsellised graft, which is stabilised through compression using a hammer and impactors. Bone cement is then inserted directly onto the graft and the implant is inserted into the cement in a similar manner to primary cemented hip artroplasty (Slooff et al., 1996).
762
A. Datta et al. / Clinical Biomechanics 21 (2006) 761–766
The combination of a rough graft surface that allows cement interdigitation and tight impaction is designed to provide a stable mechanical and biological construct for prosthesis implantation. This technique has stood the test of time with good 20-year results (Schreurs et al., 2001). Impaction bone grafting of morsellised allograft is now widely used around the world for filling of both acetabular and femoral defects. The clinical results can vary widely (Gie et al., 1996; Eldridge et al., 1997) with the femoral results being most erratic. The major issue is subsidence and this may be due to inadequate impaction or inadequate containment of the graft. Over impaction can result in fracturing the femur during impaction. A number of technical refinements have been postulated to improve the initial stability of the graft and thereby improve the long-term results. To explore these technical improvements in more detail the preparation of the graft and impaction techniques need to be carefully investigated under controlled laboratory conditions. The clinical demand for human allograft is continuously increasing and the supply is limited, therefore alternative experimental materials must be sought for use in the laboratory. The aim of this study was to determine if porcine and ovine allograft have comparable viscoelastic properties to human bone graft. Comparable results would support the use of these xenografts for in vitro testing of impaction bone grafting methods. 2. Methods 2.1. Bone graft preparation
ence of washing and articular cartilage could also be explored. 2.2. Die plunger compression tests Die plunger compression testing was used to compare the stiffness and subsequent relaxation of the various bone graft specimens after compression. The die plunger mechanism consisted of a hollow cylinder for the die 80 mm in height with a polished bore 20 mm in diameter. The base of the die was closed with a 10 mm tall 20 mm diameter loose fitting plug and a 3 mm thick porous brass disc. The porous brass disc was utilised to allow extrusion of liquid from the graft with negligible resistance (Figs. 1 and 2). The various bone grafts were approximately measured as 10 cm3 volumes in measuring cylinders and subsequently weighed. The volumes were kept equal for all bone graft groups but due to different consistencies of bone, this produced a variable mass of graft (Table 1). Within each type of bone graft this was closely regulated by using electronic scales to determine the mass of graft used for each individual die plunger test. The die plunger mechanism was set up in an Instron screw driven test machine. The 10 cm3 samples were inserted into the hollow chamber and the plunger was manually used to settle the graft contents. The Instron series IX version software was set for the screw driven test machine to its default settings for a crosshead speed of 2.54 mm/ min. The output from the materials testing machine, load, displacement and time, was acquired at a frequency of 10 Hz and stored on a personal computer for further analysis.
Porcine and ovine humeral heads were obtained from local meat processing plants. Soft tissue remnants were excised and the articular cartilage was removed from half of the porcine samples. Human allograft bone chips 6– 10 mm in size (Regeneration Technologies Inc., Alachua, FL, USA) were donated from the out of date stock of the local orthopaedic unit. This processed bone graft contains no soft tissue remnants or articular cartilage. All bone was subsequently milled using a Norwich bone mill and one ovine and two porcine samples were subsequently washed. This resulted in seven types of bone graft available for testing. 1. 2. 3. 4. 5. 6. 7.
Milled human allograft 6–10 mm (Human). Unwashed ovine graft with cartilage (Sheep cu). Washed ovine graft with cartilage (Sheep cw). Unwashed porcine graft with cartilage (Pig cu). Washed porcine graft with cartilage (Pig cw). Unwashed porcine graft without cartilage (Pig u). Washed porcine graft without cartilage (Pig w).
A comparison of ovine and porcine bone graft properties with that of human graft was now possible. The influ-
Fig. 1. Die plunger mechanism.
A. Datta et al. / Clinical Biomechanics 21 (2006) 761–766
763
Fig. 3. Load–time curve.
is sufficient to observe the exponential relaxation curve. The relaxation percentage is a crude quantification of the viscoelasticity of the graft. 2.4. Statistical analysis All results are presented as mean ± standard deviation. Differences between the groups were examined with analysis of variance (ANOVA) using SPSS v11 statistical software (SPSS Inc., Chicago, IL, USA). Significant main effects determined by ANOVA were further investigated using the Games-Howell post hoc test. Differences between groups were considered statistically significant if p was less than 0.05.
Fig. 2. Schematic diagram of die plunger mechanism.
The samples were compressed to a peak force of 500 N. When this force was attained, the displacement was kept constant and the force was continuously measured for 120 s to determine the relaxation behaviour of the various grafts. A force–time curve was plotted for the compression and relaxation phases of each experiment. The resultant plot was used to calculate a compression modulus and the percentage relaxation for each sample. Any retained fluid was removed from the plunger and hollow cylinder walls between individual experiments.
3. Results 3.1. Compression analysis In general terms porcine grafts were the least stiff (washed without cartilage 3.01 MPa SD 0.36, washed with cartilage 3.23 MPa SD 0.36, unwashed without cartilage 3.65 MPa SD 0.85, unwashed without cartilage 5.49 MPa SD 0.43) followed by human graft (3.86 MPa SD 0.41) with ovine graft (unwashed with cartilage 4.58 MPa SD 1.49 and washed with cartilage 5.65 MPa SD 1.11) having the greatest compression moduli (Table 2). There was no difference between the compression modulus of human graft and that of ovine graft with retained cartilage, whether this was washed or unwashed (p > 0.05 in both cases). Unwashed porcine graft with retained cartilage had a significantly
2.3. Data evaluation The compression modulus of each graft was calculated by taking the tangent of the load deflection curve after the initial settling effects had been excluded. The slope was therefore calculated for loads between 200 N and 500 N and then divided by the specimen nominal cross sectional area to obtain the stress. Relaxation was measured as the drop in reaction force 120 s after compression was stopped at the peak load of 500 N (Fig. 3). This duration
Table 1 Mass (g) of bone graft used for each 10 cm3 sample Graft type Human Ovine with cartilage unwashed Ovine with cartilage washed Porcine with cartilage unwashed Porcine with cartilage washed Porcine without cartilage unwashed Porcine without cartilage washed
Mean (g) (SD) 3.8 5.4 5.8 5.3 4.1 5.6 5.2
3.9 5.1 5.7 5.4 3.9 5.8 5.1
4.0 5.3 5.8 5.4 4.0 5.7 5.0
3.9 5.4 5.8 5.3 4.1 5.8 5.4
3.9 5.2 5.6 5.3 3.9 5.7 5.2
4.1 5.3 5.6 5.4 3.9 5.8 5.1
3.9 5.3 5.7 5.4 4.0 5.7 5.2
(0.1) (0.1) (0.1) (0.1) (0.1) (0.1) (0.1)
764
A. Datta et al. / Clinical Biomechanics 21 (2006) 761–766
Table 2 Mean compression modulus (MPa) and relaxation percentages (%) for all samples Compression modulus MPa (SD)
Relaxation % (SD)
Human Sheep cu Sheep cw Pig cu Pig cw Pig u Pig w
3.86 4.58 5.65 5.49 3.23 3.65 3.01
34.3 41.4 36.9 37.7 28.7 29.7 27.9
(0.41) (1.49) (1.11) (0.43) (0.36) (0.85) (0.36)
(0.2) (5.6) (4.3) (2.5) (1.2) (2.6) (2.7)
ovine with cartilage unwashed ovine with cartilage washed
60
porcine with cartilage unwashed Relaxation (%)
Sample
human 70
50
porcine with cartilage washed
40
porcine without cartilage unwashed porcine without cartilage washed
30 20 10 0
greater modulus than human graft (5.49 MPa SD 0.43 and 3.86 MPa SD 0.41, respectively, p = 0.001), washed porcine graft with cartilage (3.23 MPa SD 0.36, p = 0.001) and porcine graft without cartilage independently of it having been washed (3.01 MPa SD 0.36, p < 0.001) or it being unwashed (3.65 MPa SD 0.85, p = 0.018). Washed porcine graft without cartilage exhibited lower modulus than human graft (3.01 MPa SD 0.36 and 3.86 MPa SD 0.41, p = 0.037), washed ovine graft with retained cartilage (5.65 MPa SD 1.11, p = 0.013) and unwashed porcine graft with cartilage (5.49 MPa SD 0.43, p < 0.001). There was no difference between the two types of ovine graft, washed and unwashed with retained cartilage. Washed porcine graft with retained cartilage had a lower compression modulus compared to unwashed porcine graft with retained cartilage (3.23 MPa SD 0.36 and 5.49 MPa SD 0.43, respectively). The difference between these two groups was significant (p < 0.001), while washing did not affect the compression modulus of porcine samples without cartilage (3.01 MPa SD 0.36 unwashed, 3.65 MPa SD 0.85 washed, p > 0.05) and ovine samples with cartilage (4.58 MPa SD 1.49 unwashed, 5.65 MPa SD 1.11 washed, p > 0.05). The removal of cartilage from the porcine samples reduced the compression modulus of the unwashed graft (unwashed porcine without cartilage 3.65 MPa SD 0.85, unwashed porcine with cartilage 5.49 MPa SD 0.43, p = 0.018), but it had no effect on the washed graft (washed porcine without cartilage 3.01 MPa SD 0.36, washed porcine with cartilage 3.23 MPa SD 0.36, p > 0.05) (Fig. 4).
human
Compression Modulus (MPa)
10
ovine with cartilage unwashed
9
ovine with cartilage washed
8
porcine with cartilage unwashed
7
porcine with cartilage washed
6
porcine without cartilage unwashed
5
porcine without cartilage washed
4 3 2 1 0
Fig. 4. Mean compression modulus (MPa) of graft samples.
Fig. 5. Mean relaxation (%) of graft samples.
3.2. Relaxation analysis In general terms porcine grafts showed the least relaxation (washed without cartilage 27.9% SD 2.7, washed with cartilage 28.7% SD 1.2, unwashed without cartilage 29.7% SD 2.6 and unwashed with cartilage 37.7% SD 2.5) followed by human graft (34.3% SD 0.2) with ovine graft (unwashed with cartilage 41.4% SD 5.6 and washed with cartilage 36.9% SD 4.3) having largest percentage relaxation. Human graft showed higher percentage relaxation than washed porcine graft with retained cartilage (p = 0.001) and washed porcine graft without cartilage (p = 0.001). There was no difference between human graft, ovine graft with retained cartilage (both washed and unwashed) and unwashed porcine graft (with or without cartilage). Washing the graft had no effect on ovine samples with cartilage nor on porcine samples without cartilage, but it decreased the percentage relaxation of porcine samples with retained cartilage. This last difference was statistically significant (p = 0.002) (Fig. 5). The effect of cartilage removal was assessed on porcine samples only. The removal of cartilage had no effect on the washed samples while it decreased the percentage relaxation in the case of unwashed graft (p = 0.007). 4. Discussion In vitro testing is an important method for improving the quality and ensuring consistent impaction bone grafting techniques. To allow porcine or ovine bone graft to be utilised in experimental studies as a substitute to human allograft it must be demonstrated that they exhibit similar compression and relaxation properties to human graft. The consistent results of compression modulus and relaxation values in the control group of pre-prepared human allograft used in this study demonstrates the reliability and reproducibility of the testing procedures adopted. The allograft samples were the most homogenous used in this experimental study as they had been uniformly prepared for clinical use and were obtained from a single traceable donor. The human graft was freeze dried, therefore there were no additional soft tissues or incompressible
A. Datta et al. / Clinical Biomechanics 21 (2006) 761–766
fluids in the mixture. The only source of variation was represented by the final milling performed in the laboratory at the time of testing. In this study, the freeze dried human allograft bone chips were used as an experimental control group against which to compare the mechanical and relaxation properties of the ovine and porcine samples. The results obtained in this study correlate well to previously published data by Gozzard et al. (2002) and Grimm et al. (2002), these authors report compression moduli for various human allograft samples of 3.67–3.85 MPa in comparison to 3.86 MPa obtained in this study. Corresponding results for ovine graft were 3.93–4.27 MPa in comparison to 4.58–5.65 MPa. The mean relaxation percentage in this study for human allograft was 34.3% this varied between 32.4% and 38.5% for same published data. Similarly the values obtained in this study for ovine graft were comparable to those obtained by Gozzard et al. (2002) and Grimm et al. (2002) (36.9–41.4% compared to 27.4–39.6%). This study has demonstrated that there is no statistical difference in the compression modulus or relaxation percentage between human and ovine graft preparations. The clinical importance of stiffness in the graft relates to the biofeedback to the surgeon at the time of impaction and the stability of the subsequent graft bed. It is difficult to assess if any particular surgeon would be able to differentiate between the stiffness of porcine, ovine and human graft whilst they are impacting the graft. The change in biofeedback caused by the differences in compression modulus of the three graft types may be too subtle for a surgeon to appreciate during the relatively crude impaction process. The percentage relaxation of the graft, on the other hand, is more crucial as this affects the compactness of the graft. Too much relaxation will lead to a less compact graft bed and this has the potential to cause problems with initial stability and cement penetration. This is especially relevant in the femoral canal. When using cemented total hip replacements, the thickness of the cement mantle is important. The ideal thickness is 2–4 mm too much relaxation will potentially provide too thin a cement mantle. It has been reported that bone graft obeys the laws of soil mechanics (Dunlop et al., 2003). In soil mechanics the importance of shear strength is widely recognised as it represents a measure of the amount of stress the aggregate can sustain prior to failure (Lambe and Whitman, 1969). In the case of morsellised bone graft the shear strength of the aggregate this is related to the internal friction of the graft, which is related to the angle at which the particles will slide and to the interlocking of the particles. The addition of small amounts of fluid to a dry aggregate in soil mechanics has been shown to increase shear strength. This is due to suction created by the fluid in aggregate pores (Smith, 1990) but too much fluid creates a supersaturated aggregate similar to quicksand that is characterised by a much lower shear strength. The major difference between soil and bone graft aggregates is related to the fat and bone marrow content of the latter. This vis-
765
cous fluid lies in the graft interstices and dampens the compaction energy, due to its incompressible nature. In addition the viscous fluids trapped within the graft can act as a lubricant. Washing the graft in saline removes the fat and associated soft tissue and the fluid is replaced by normal saline. This fluid is much less viscous and moves easily into the porous material on impaction. Dunlop et al. (2003) postulated that washed aggregates are characterised by increased shear strength due to tighter graft compaction as fat and marrow had been removed. This, they felt, would improve the initial stability of the bone graft. The results of this study were not conclusive with regard to the effects of washing and its relationship to viscoelastic properties of porcine and ovine bone grafts. A reduction in compression modulus and relaxation percentage was recorded in the case of porcine graft containing cartilage. The freeze dried human graft generally displayed less variable results than the other graft compositions. This difference was attributed to the fact that in this group all samples generated from a single donor, they did not contain soft tissues or cartilage and they were removed of all fat. 5. Conclusions This study compared the compression modulus and percentage relaxation of human freeze dried allograft, unwashed ovine graft with cartilage, washed ovine graft with cartilage, unwashed porcine graft with cartilage, washed porcine graft with cartilage, unwashed porcine graft without cartilage and washed porcine graft without cartilage. There was no difference between the compression moduli of human freeze dried allograft and the two ovine graft groups. A similar pattern was identified when the percentage relaxation of human allografts was compared to that of the two ovine graft groups investigated. Unwashed porcine graft (with or without cartilage) also displayed a similar relaxation behaviour to human allograft. In conclusion, on the basis of the experiments performed, the use of ovine bone graft is recommended as a suitable substitute for human allograft for in vitro testing of impaction bone grafting techniques when either compression of the graft or stress relaxation or both are important parameters to be replicated in the experimental model. References Amstuz, H.C., Ma, S.M., Jinnah, R.H., Mai, L., 1982. Revision of aseptic loose total hip arthroplasty. Clin. Orthop. 170, 21–33. Dunlop, D.G., Brewster, N.T., Madabhushi, S.P., Usmani, A.S., Pankaj, P., Howie, C.R., 2003. Techniques to improve the shear strength of impacted bone graft: the effect of particle size and washing of the graft. J. Bone Joint Surg. Am. 85-A (4), 639–646. Eldridge, J.D., Smith, E.J., Hubble, M.J., Whitehouse, S.L., Learmonth, I.D., 1997. Massive early subsidence following femoral impaction grafting. J. Arthroplasty 12, 535–540.
766
A. Datta et al. / Clinical Biomechanics 21 (2006) 761–766
Gie, G.A., Linder, L., Ling, R.S., Timperley, A.J., 1996. Contained morsellized allograft in revision hip arthroplasty: a minimum five year follow up. J. Bone J. Surg. Br. 78 (Suppl. 1), 71. Gozzard, C., Grimm, B., Miles, A.W., Learmonth, I.D., 2002. The effect of preparatory technique on the compressive properties of morsellised bone graft. Hip Int. 12 (2), 116–118. Grimm, B., Gozzard, C., Miles, A.W., Turner, I.G., 2002. Compression testing of bone grafts for impaction grafting. 48th Annual Meeting Orthop. Res. Soc. Lambe, T.W., Whitman, R.V., 1969. Soil Mechanics. Wiley, New York.
Schreurs, B.W., Slooff, T.J., Gardeniers, J.W., Buma, P., 2001. Acetabular reconstruction with bone impaction grafting and a cemented cup: 20 years experience. Clin. Orthop. (393), 202–215. Slooff, T.J., Buma, P., Schreurs, B.W., Schimmel, J., Huiskes, R., Gardeniers, J., 1996. Acetabular and femoral reconstruction with impacted grafts and cement. Clin. Orthop. 324, 108–116. Slooff, T.J., Schreurs, B.W., Buma, P., Gardeniers, J.W., 1999. Impaction morcellized allografting and cement. Instr. Course Lect. 48, 79–89. Smith, G.N., 1990. Elements of Soil Mechanics, 6th ed. Blackwell Science, Oxford.
A comparison of the viscoelastic properties of bone grafts A. Datta *, S. Gheduzzi, A.W. Miles Centre for Orthopaedic Biomechanics, Department of Mechanical Engineering, University of Bath, United Kingdom Received 1 April 2005; accepted 22 March 2006
Abstract Background. The expansion in joint arthroplasty surgery in the 1970s has resulted in a large group of patients who require revision arthroplasty for aseptic loosening. Impaction bone grafting to deal with bone stock loss has become an increasingly popular procedure in revision hip surgery. The results of this revision surgery are variable and very much dependent on the grafting techniques adopted at the time of surgery. In vitro testing of impaction bone grafting methods can constitute an important tool to improve long-term clinical results. The increasing clinical demand for human allograft limits its availability for use in in vitro laboratory studies therefore suitable experimental alternatives are required. Methods. Human, porcine and ovine cancellous bone grafts were morsellised and prepared for in vitro laboratory testing following a standard operative technique. Each graft type was compressed using a die plunger test. After compression to a predetermined level the graft was left to relax for 120 s. A comparison of the compression moduli and of the relaxation characteristics for each graft preparation was performed. In addition, the effects of washing the graft and of cartilage removal from the graft mixture were investigated. Results. This study has demonstrated that there is no statistical difference in the compression modulus or relaxation percentage between human and ovine graft preparations. The effect of removing cartilage and washing the graft mixtures were inconsistent with regard to alterations in the viscoelastic properties of the grafts. Interpretation. On the basis of the experiments performed we recommend the use of ovine bone graft as a suitable substitute for human allograft for in vitro testing of impaction bone grafting methods. The properties of ovine graft were similar for both compression moduli and relaxation properties to human allograft. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Allograft; Impaction bone grafting
1. Introduction The expansion in joint arthroplasty surgery in the 1970s has resulted in a large group of patients, who require revision arthroplasty for aseptic loosening. These cases can be technically demanding as bone stock loss and large bony defects are often present. To allow for adequate fixation of the revision components, the areas of bone loss and bony defects can be filled with bone graft. This physiological solution is preferred to earlier techniques using large
*
Corresponding author. Address: Flat 4, 37 Portland Road, Edgbaston, Birmingham B16 9HS, United Kingdom. E-mail address: [email protected] (A. Datta). 0268-0033/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2006.03.009
volumes of bone cement alone (Amstuz et al., 1982). The morsellised graft, is able to fill irregular defects and cavities in the host bone and provide a mechanical and biological scaffold that can be incorporated into the host tissue as bone remodeling takes place (Gie et al., 1996). Impaction bone grafting was originally developed as a surgical technique for acetabular reconstruction and has been further refined to repair large femoral defects and compensate for bone stock loss at the time of revision surgery (Schreurs et al., 2001; Slooff et al., 1999). It involves filling the bone cavities and defects with morsellised graft, which is stabilised through compression using a hammer and impactors. Bone cement is then inserted directly onto the graft and the implant is inserted into the cement in a similar manner to primary cemented hip artroplasty (Slooff et al., 1996).
762
A. Datta et al. / Clinical Biomechanics 21 (2006) 761–766
The combination of a rough graft surface that allows cement interdigitation and tight impaction is designed to provide a stable mechanical and biological construct for prosthesis implantation. This technique has stood the test of time with good 20-year results (Schreurs et al., 2001). Impaction bone grafting of morsellised allograft is now widely used around the world for filling of both acetabular and femoral defects. The clinical results can vary widely (Gie et al., 1996; Eldridge et al., 1997) with the femoral results being most erratic. The major issue is subsidence and this may be due to inadequate impaction or inadequate containment of the graft. Over impaction can result in fracturing the femur during impaction. A number of technical refinements have been postulated to improve the initial stability of the graft and thereby improve the long-term results. To explore these technical improvements in more detail the preparation of the graft and impaction techniques need to be carefully investigated under controlled laboratory conditions. The clinical demand for human allograft is continuously increasing and the supply is limited, therefore alternative experimental materials must be sought for use in the laboratory. The aim of this study was to determine if porcine and ovine allograft have comparable viscoelastic properties to human bone graft. Comparable results would support the use of these xenografts for in vitro testing of impaction bone grafting methods. 2. Methods 2.1. Bone graft preparation
ence of washing and articular cartilage could also be explored. 2.2. Die plunger compression tests Die plunger compression testing was used to compare the stiffness and subsequent relaxation of the various bone graft specimens after compression. The die plunger mechanism consisted of a hollow cylinder for the die 80 mm in height with a polished bore 20 mm in diameter. The base of the die was closed with a 10 mm tall 20 mm diameter loose fitting plug and a 3 mm thick porous brass disc. The porous brass disc was utilised to allow extrusion of liquid from the graft with negligible resistance (Figs. 1 and 2). The various bone grafts were approximately measured as 10 cm3 volumes in measuring cylinders and subsequently weighed. The volumes were kept equal for all bone graft groups but due to different consistencies of bone, this produced a variable mass of graft (Table 1). Within each type of bone graft this was closely regulated by using electronic scales to determine the mass of graft used for each individual die plunger test. The die plunger mechanism was set up in an Instron screw driven test machine. The 10 cm3 samples were inserted into the hollow chamber and the plunger was manually used to settle the graft contents. The Instron series IX version software was set for the screw driven test machine to its default settings for a crosshead speed of 2.54 mm/ min. The output from the materials testing machine, load, displacement and time, was acquired at a frequency of 10 Hz and stored on a personal computer for further analysis.
Porcine and ovine humeral heads were obtained from local meat processing plants. Soft tissue remnants were excised and the articular cartilage was removed from half of the porcine samples. Human allograft bone chips 6– 10 mm in size (Regeneration Technologies Inc., Alachua, FL, USA) were donated from the out of date stock of the local orthopaedic unit. This processed bone graft contains no soft tissue remnants or articular cartilage. All bone was subsequently milled using a Norwich bone mill and one ovine and two porcine samples were subsequently washed. This resulted in seven types of bone graft available for testing. 1. 2. 3. 4. 5. 6. 7.
Milled human allograft 6–10 mm (Human). Unwashed ovine graft with cartilage (Sheep cu). Washed ovine graft with cartilage (Sheep cw). Unwashed porcine graft with cartilage (Pig cu). Washed porcine graft with cartilage (Pig cw). Unwashed porcine graft without cartilage (Pig u). Washed porcine graft without cartilage (Pig w).
A comparison of ovine and porcine bone graft properties with that of human graft was now possible. The influ-
Fig. 1. Die plunger mechanism.
A. Datta et al. / Clinical Biomechanics 21 (2006) 761–766
763
Fig. 3. Load–time curve.
is sufficient to observe the exponential relaxation curve. The relaxation percentage is a crude quantification of the viscoelasticity of the graft. 2.4. Statistical analysis All results are presented as mean ± standard deviation. Differences between the groups were examined with analysis of variance (ANOVA) using SPSS v11 statistical software (SPSS Inc., Chicago, IL, USA). Significant main effects determined by ANOVA were further investigated using the Games-Howell post hoc test. Differences between groups were considered statistically significant if p was less than 0.05.
Fig. 2. Schematic diagram of die plunger mechanism.
The samples were compressed to a peak force of 500 N. When this force was attained, the displacement was kept constant and the force was continuously measured for 120 s to determine the relaxation behaviour of the various grafts. A force–time curve was plotted for the compression and relaxation phases of each experiment. The resultant plot was used to calculate a compression modulus and the percentage relaxation for each sample. Any retained fluid was removed from the plunger and hollow cylinder walls between individual experiments.
3. Results 3.1. Compression analysis In general terms porcine grafts were the least stiff (washed without cartilage 3.01 MPa SD 0.36, washed with cartilage 3.23 MPa SD 0.36, unwashed without cartilage 3.65 MPa SD 0.85, unwashed without cartilage 5.49 MPa SD 0.43) followed by human graft (3.86 MPa SD 0.41) with ovine graft (unwashed with cartilage 4.58 MPa SD 1.49 and washed with cartilage 5.65 MPa SD 1.11) having the greatest compression moduli (Table 2). There was no difference between the compression modulus of human graft and that of ovine graft with retained cartilage, whether this was washed or unwashed (p > 0.05 in both cases). Unwashed porcine graft with retained cartilage had a significantly
2.3. Data evaluation The compression modulus of each graft was calculated by taking the tangent of the load deflection curve after the initial settling effects had been excluded. The slope was therefore calculated for loads between 200 N and 500 N and then divided by the specimen nominal cross sectional area to obtain the stress. Relaxation was measured as the drop in reaction force 120 s after compression was stopped at the peak load of 500 N (Fig. 3). This duration
Table 1 Mass (g) of bone graft used for each 10 cm3 sample Graft type Human Ovine with cartilage unwashed Ovine with cartilage washed Porcine with cartilage unwashed Porcine with cartilage washed Porcine without cartilage unwashed Porcine without cartilage washed
Mean (g) (SD) 3.8 5.4 5.8 5.3 4.1 5.6 5.2
3.9 5.1 5.7 5.4 3.9 5.8 5.1
4.0 5.3 5.8 5.4 4.0 5.7 5.0
3.9 5.4 5.8 5.3 4.1 5.8 5.4
3.9 5.2 5.6 5.3 3.9 5.7 5.2
4.1 5.3 5.6 5.4 3.9 5.8 5.1
3.9 5.3 5.7 5.4 4.0 5.7 5.2
(0.1) (0.1) (0.1) (0.1) (0.1) (0.1) (0.1)
764
A. Datta et al. / Clinical Biomechanics 21 (2006) 761–766
Table 2 Mean compression modulus (MPa) and relaxation percentages (%) for all samples Compression modulus MPa (SD)
Relaxation % (SD)
Human Sheep cu Sheep cw Pig cu Pig cw Pig u Pig w
3.86 4.58 5.65 5.49 3.23 3.65 3.01
34.3 41.4 36.9 37.7 28.7 29.7 27.9
(0.41) (1.49) (1.11) (0.43) (0.36) (0.85) (0.36)
(0.2) (5.6) (4.3) (2.5) (1.2) (2.6) (2.7)
ovine with cartilage unwashed ovine with cartilage washed
60
porcine with cartilage unwashed Relaxation (%)
Sample
human 70
50
porcine with cartilage washed
40
porcine without cartilage unwashed porcine without cartilage washed
30 20 10 0
greater modulus than human graft (5.49 MPa SD 0.43 and 3.86 MPa SD 0.41, respectively, p = 0.001), washed porcine graft with cartilage (3.23 MPa SD 0.36, p = 0.001) and porcine graft without cartilage independently of it having been washed (3.01 MPa SD 0.36, p < 0.001) or it being unwashed (3.65 MPa SD 0.85, p = 0.018). Washed porcine graft without cartilage exhibited lower modulus than human graft (3.01 MPa SD 0.36 and 3.86 MPa SD 0.41, p = 0.037), washed ovine graft with retained cartilage (5.65 MPa SD 1.11, p = 0.013) and unwashed porcine graft with cartilage (5.49 MPa SD 0.43, p < 0.001). There was no difference between the two types of ovine graft, washed and unwashed with retained cartilage. Washed porcine graft with retained cartilage had a lower compression modulus compared to unwashed porcine graft with retained cartilage (3.23 MPa SD 0.36 and 5.49 MPa SD 0.43, respectively). The difference between these two groups was significant (p < 0.001), while washing did not affect the compression modulus of porcine samples without cartilage (3.01 MPa SD 0.36 unwashed, 3.65 MPa SD 0.85 washed, p > 0.05) and ovine samples with cartilage (4.58 MPa SD 1.49 unwashed, 5.65 MPa SD 1.11 washed, p > 0.05). The removal of cartilage from the porcine samples reduced the compression modulus of the unwashed graft (unwashed porcine without cartilage 3.65 MPa SD 0.85, unwashed porcine with cartilage 5.49 MPa SD 0.43, p = 0.018), but it had no effect on the washed graft (washed porcine without cartilage 3.01 MPa SD 0.36, washed porcine with cartilage 3.23 MPa SD 0.36, p > 0.05) (Fig. 4).
human
Compression Modulus (MPa)
10
ovine with cartilage unwashed
9
ovine with cartilage washed
8
porcine with cartilage unwashed
7
porcine with cartilage washed
6
porcine without cartilage unwashed
5
porcine without cartilage washed
4 3 2 1 0
Fig. 4. Mean compression modulus (MPa) of graft samples.
Fig. 5. Mean relaxation (%) of graft samples.
3.2. Relaxation analysis In general terms porcine grafts showed the least relaxation (washed without cartilage 27.9% SD 2.7, washed with cartilage 28.7% SD 1.2, unwashed without cartilage 29.7% SD 2.6 and unwashed with cartilage 37.7% SD 2.5) followed by human graft (34.3% SD 0.2) with ovine graft (unwashed with cartilage 41.4% SD 5.6 and washed with cartilage 36.9% SD 4.3) having largest percentage relaxation. Human graft showed higher percentage relaxation than washed porcine graft with retained cartilage (p = 0.001) and washed porcine graft without cartilage (p = 0.001). There was no difference between human graft, ovine graft with retained cartilage (both washed and unwashed) and unwashed porcine graft (with or without cartilage). Washing the graft had no effect on ovine samples with cartilage nor on porcine samples without cartilage, but it decreased the percentage relaxation of porcine samples with retained cartilage. This last difference was statistically significant (p = 0.002) (Fig. 5). The effect of cartilage removal was assessed on porcine samples only. The removal of cartilage had no effect on the washed samples while it decreased the percentage relaxation in the case of unwashed graft (p = 0.007). 4. Discussion In vitro testing is an important method for improving the quality and ensuring consistent impaction bone grafting techniques. To allow porcine or ovine bone graft to be utilised in experimental studies as a substitute to human allograft it must be demonstrated that they exhibit similar compression and relaxation properties to human graft. The consistent results of compression modulus and relaxation values in the control group of pre-prepared human allograft used in this study demonstrates the reliability and reproducibility of the testing procedures adopted. The allograft samples were the most homogenous used in this experimental study as they had been uniformly prepared for clinical use and were obtained from a single traceable donor. The human graft was freeze dried, therefore there were no additional soft tissues or incompressible
A. Datta et al. / Clinical Biomechanics 21 (2006) 761–766
fluids in the mixture. The only source of variation was represented by the final milling performed in the laboratory at the time of testing. In this study, the freeze dried human allograft bone chips were used as an experimental control group against which to compare the mechanical and relaxation properties of the ovine and porcine samples. The results obtained in this study correlate well to previously published data by Gozzard et al. (2002) and Grimm et al. (2002), these authors report compression moduli for various human allograft samples of 3.67–3.85 MPa in comparison to 3.86 MPa obtained in this study. Corresponding results for ovine graft were 3.93–4.27 MPa in comparison to 4.58–5.65 MPa. The mean relaxation percentage in this study for human allograft was 34.3% this varied between 32.4% and 38.5% for same published data. Similarly the values obtained in this study for ovine graft were comparable to those obtained by Gozzard et al. (2002) and Grimm et al. (2002) (36.9–41.4% compared to 27.4–39.6%). This study has demonstrated that there is no statistical difference in the compression modulus or relaxation percentage between human and ovine graft preparations. The clinical importance of stiffness in the graft relates to the biofeedback to the surgeon at the time of impaction and the stability of the subsequent graft bed. It is difficult to assess if any particular surgeon would be able to differentiate between the stiffness of porcine, ovine and human graft whilst they are impacting the graft. The change in biofeedback caused by the differences in compression modulus of the three graft types may be too subtle for a surgeon to appreciate during the relatively crude impaction process. The percentage relaxation of the graft, on the other hand, is more crucial as this affects the compactness of the graft. Too much relaxation will lead to a less compact graft bed and this has the potential to cause problems with initial stability and cement penetration. This is especially relevant in the femoral canal. When using cemented total hip replacements, the thickness of the cement mantle is important. The ideal thickness is 2–4 mm too much relaxation will potentially provide too thin a cement mantle. It has been reported that bone graft obeys the laws of soil mechanics (Dunlop et al., 2003). In soil mechanics the importance of shear strength is widely recognised as it represents a measure of the amount of stress the aggregate can sustain prior to failure (Lambe and Whitman, 1969). In the case of morsellised bone graft the shear strength of the aggregate this is related to the internal friction of the graft, which is related to the angle at which the particles will slide and to the interlocking of the particles. The addition of small amounts of fluid to a dry aggregate in soil mechanics has been shown to increase shear strength. This is due to suction created by the fluid in aggregate pores (Smith, 1990) but too much fluid creates a supersaturated aggregate similar to quicksand that is characterised by a much lower shear strength. The major difference between soil and bone graft aggregates is related to the fat and bone marrow content of the latter. This vis-
765
cous fluid lies in the graft interstices and dampens the compaction energy, due to its incompressible nature. In addition the viscous fluids trapped within the graft can act as a lubricant. Washing the graft in saline removes the fat and associated soft tissue and the fluid is replaced by normal saline. This fluid is much less viscous and moves easily into the porous material on impaction. Dunlop et al. (2003) postulated that washed aggregates are characterised by increased shear strength due to tighter graft compaction as fat and marrow had been removed. This, they felt, would improve the initial stability of the bone graft. The results of this study were not conclusive with regard to the effects of washing and its relationship to viscoelastic properties of porcine and ovine bone grafts. A reduction in compression modulus and relaxation percentage was recorded in the case of porcine graft containing cartilage. The freeze dried human graft generally displayed less variable results than the other graft compositions. This difference was attributed to the fact that in this group all samples generated from a single donor, they did not contain soft tissues or cartilage and they were removed of all fat. 5. Conclusions This study compared the compression modulus and percentage relaxation of human freeze dried allograft, unwashed ovine graft with cartilage, washed ovine graft with cartilage, unwashed porcine graft with cartilage, washed porcine graft with cartilage, unwashed porcine graft without cartilage and washed porcine graft without cartilage. There was no difference between the compression moduli of human freeze dried allograft and the two ovine graft groups. A similar pattern was identified when the percentage relaxation of human allografts was compared to that of the two ovine graft groups investigated. Unwashed porcine graft (with or without cartilage) also displayed a similar relaxation behaviour to human allograft. In conclusion, on the basis of the experiments performed, the use of ovine bone graft is recommended as a suitable substitute for human allograft for in vitro testing of impaction bone grafting techniques when either compression of the graft or stress relaxation or both are important parameters to be replicated in the experimental model. References Amstuz, H.C., Ma, S.M., Jinnah, R.H., Mai, L., 1982. Revision of aseptic loose total hip arthroplasty. Clin. Orthop. 170, 21–33. Dunlop, D.G., Brewster, N.T., Madabhushi, S.P., Usmani, A.S., Pankaj, P., Howie, C.R., 2003. Techniques to improve the shear strength of impacted bone graft: the effect of particle size and washing of the graft. J. Bone Joint Surg. Am. 85-A (4), 639–646. Eldridge, J.D., Smith, E.J., Hubble, M.J., Whitehouse, S.L., Learmonth, I.D., 1997. Massive early subsidence following femoral impaction grafting. J. Arthroplasty 12, 535–540.
766
A. Datta et al. / Clinical Biomechanics 21 (2006) 761–766
Gie, G.A., Linder, L., Ling, R.S., Timperley, A.J., 1996. Contained morsellized allograft in revision hip arthroplasty: a minimum five year follow up. J. Bone J. Surg. Br. 78 (Suppl. 1), 71. Gozzard, C., Grimm, B., Miles, A.W., Learmonth, I.D., 2002. The effect of preparatory technique on the compressive properties of morsellised bone graft. Hip Int. 12 (2), 116–118. Grimm, B., Gozzard, C., Miles, A.W., Turner, I.G., 2002. Compression testing of bone grafts for impaction grafting. 48th Annual Meeting Orthop. Res. Soc. Lambe, T.W., Whitman, R.V., 1969. Soil Mechanics. Wiley, New York.
Schreurs, B.W., Slooff, T.J., Gardeniers, J.W., Buma, P., 2001. Acetabular reconstruction with bone impaction grafting and a cemented cup: 20 years experience. Clin. Orthop. (393), 202–215. Slooff, T.J., Buma, P., Schreurs, B.W., Schimmel, J., Huiskes, R., Gardeniers, J., 1996. Acetabular and femoral reconstruction with impacted grafts and cement. Clin. Orthop. 324, 108–116. Slooff, T.J., Schreurs, B.W., Buma, P., Gardeniers, J.W., 1999. Impaction morcellized allografting and cement. Instr. Course Lect. 48, 79–89. Smith, G.N., 1990. Elements of Soil Mechanics, 6th ed. Blackwell Science, Oxford.