41
Technical note
Revision of cemented fixation and cement-bone interface strength A Rosenstein, MD, W MacDonald, MPhil, A Iliadis, MD and P McLardySmith, FRCS Nuffield Orthopaedic Centre, Headington, Oxford Interfacial shear strength between poly(methy1 methacrylate) ( P M M A ) bone cement and cancellous bone was measured in bone samples from human proximal femora. Samples were prepared withfresh cement-bone, fresh cement inside a mantle of existing cement and with fresh cement-revised bone surfaces. Push-out tests to measure shear strength caused failure only at bone-cement interfaces; revised bone interfaces were 30 per cent weaker (P< 0.02)than primary interfaces. The clinical relevance is that revision of cemented joint arthroplasties may necessitate removal of components with sound cement-bone Jixation. The practice of removing all traces of P M M A cement may not yield the optimalfixation; adhesion of fresh cement to freshly prepared surfaces of the existing cement might also be considered where circumstances are favourable.
1 INTRODUCTION
Revision of cemented total joint arthroplasties is increasingly common, involving often difficult removal of components and cement. Many orthopaedic surgeons recommend complete removal of old cement before insertion of any new cement. In support of this, some studies have shown a 7-33 per cent reduction in shear strength where new cement interfaces with old (1-3). However, there do not appear to be any studies comparing cement-bone strength in primary and revision cementing, nor comparing the bond strength between new and old cement with revised cement-bone bonds.
eter indentor centered in the cement plug. The cortex of the specimen was supported on a hollow cylinder of appropriate dimensions resting on the platen of the Instron. Testing was performed at a cross-head speed of 2 mm/min and force and extension were recorded continuously. After push-out testing, the specimen and pushed out plug were photographed (Fig. 11, all the cement was removed with a chisel and the cementing and testing procedure was repeated with the newly formed cavity. The specimen was then rephotographed. 2.2 Group 2 (fourteen specimens: seven head, seven neck)
2 MATERIALS AND METHODS
Eleven fresh femoral heads and necks were obtained from total hip arthroplasty or autopsy, and were frozen immediately after removal. Using a saw guide and hacksaw, two slices of 15 mm nominal thickness were cut, one each from the head and neck region of the frozen specimen. Thickness was measured with vernier calipers at three locations equally spaced around the perimeter. Specimens were kept moist at all times, and paired head and neck slices were randomly allocated to two groups.
The specimens were drilled axially using a 12 mm diameter drill, and the cavity was filled with cement, as in group 1. After the 90 minute setting time, a 6 mm diameter hole was drilled axially through the centre of the cement plug. Water was poured into the cavity, the surfaces were tamped dry with paper towelling and the cavity was filled with newly mixed PMMA cement and pressurized. After a further 90 minutes setting time, a push-out test was performed with a 5 mm indentor centered in the second cement plug. When testing and photography
2.1 Group 1(eight specimens: four head, four neck)
A hole was drilled axially through the slice, using a drill press and 12 mm diameter drill. With the slice positioned on a non-adherent board, poly(methy1 methacrylate) (PMMA) bone cement (CMW Laboratories Limited, Exeter) was inserted into the cavity between two and three and a half minutes after mixing. Pressure was applied with the thumb, the pulp of the thumb sealing off the cavity. The slices were covered with saturated paper towelling while the cement was allowed to set. After 90 minutes, push-out tests were performed using a Universal Materials Testing System (Model 1122, Instron Limited, High Wycombe) with a 10 mm diamThe MS was received on I 1 September 1991 and was accepted for publication on 5 June 1992.
H03191
Fig. 1 Typical specimen photograph showing pushed-out cement-bone plug, millimetre scale in view
0954-4119/92 $3.00
+ .05
Downloaded from pih.sagepub.com at NANYANG TECH UNIV LIBRARY on June 15, 2015
Roc Instn Mech Engrs Vol 206
A ROSENSTEIN, W MacDONALD, A ILIADIS AND P McLARDY-SMITH
48
Table 1 Interfacial shear stress
Specimen Group I la lb 2a 2b 3a 3b
4a 4b Group 2 5a 5b 6a 6b 7a 7b 8a 8b 9a 9b 1Oa lob 1 la llb
Primary shear strength MPa
Revised shear strength MPa
34.1 38.7 8.2 58.8 34.4 37.7 14.2 63.3
31.5 32.8 15.5 34.9 24.5 11.8 20.9
20.4 17.9 20.5 38.1 22.6 15.2 41.7 23.4 21.5 43.3 29.3 46.8 24.7 32.9
21.9 11.8 11.2 8.6 21.4 6.8 24.1 17.9 17.9 35.9 26.2 28.1 21.0 7.8
~
had been completed, all cement was removed with a chisel, and the new cavity was cemented en bloc and retested.
501 T
60
a
Primary
nmm
Revision
40
Group 1
Group 2
Fig. 2 Nomogram of shear strengths by group
interfaces (P < 0.02), strength being reduced by about 3 0 per cent (Fig. 2). This trend holds true independent of bone source area or procedure adopted (that is group). The results also demonstrate that the cementbone bond was 31 per cent stronger in specimens from the femoral head than in specimens from the femoral neck (P < 0.03) (Fig. 3).
3 ANALYSIS
For each slice, the thickness was determined as the mean of three measures. Scaled photographs were digitized and the perimeter was measured on both superior and inferior surfaces of the cortical annulus and the pushed-out plug. From the four perimeters thus measured, the mean interfacial perimeter was calculated and multiplied by the mean thickness to estimate the surface area of shear failure. Interfacial shear stress was calculated by dividing the break load by the surface area (Table 1). Analysis of variance (ANOVA) was performed using a computer software package (General Linear Model, SAS). 4 RESULTS
One specimen (3a) cracked through the cortex during the first push-out test and was unable to be recemented. In an initial analysis of variance, specimens were separated by group; the differences were shown to be insignificant statistically (P = 0.15) and the groups were combined. The analysis was repeated with specimens grouped by source (head or neck) and procedure (primary or revised cement-bone). In the first series of tests in group 2, failure always occurred at the cement-bone interface. None of these specimens failed by pushing out the plug of new cement from within the established cement mass. The statistical analysis indicated that these cement-bone strengths were not significantly different from the initial cementbone strengths of the specimens in group 1. The ANOVA results indicate that revised cementbone interfaces are significantly weaker than primary
5 DISCUSSION
Specimens were taken from femoral heads and necks so that different distributions of cortical and cancellous bone were represented in the test series. The statistical analysis shows that the cement-bone shear strength is significantly lower in the specimens from the femoral neck than those from the femoral head. This may be accounted for by the effect of cancellous structure. Since the cancellous bone in the femoral neck is more aligned with the neck axis, the stronger elements lie in that
:I T 40 I-
Primary
Revision
I
Heads
Necks
Fig. 3 Nomogram of shear strengths by specimen origin
Part H : Journal of Engineering in Medicine Downloaded from pih.sagepub.com at NANYANG TECH UNIV LIBRARY on June 15, 2015
REVISION OF CEMENTED FIXATION AND CEMENT-BONE INTERFACE STRENGTH
direction. As a result, in cylindrical surfaces parallel to the axis (such as encountered by the cement), the texture will be less open and the interlock with the cement will be less. Increased interfacial strength in the region with more cancellous bone (the head specimens) calls into question the practice of removing all cancellous bone from the trochanteric region as part of surgical preparation of the femur for total hip replacement. It would seem that improved fixation might result if the cancellous bone were retained in such areas. All specimens tested in the first test of group 2 failed at the boneecement interface, and not at the interface between the old and new cement. The interface between old and new cement took the form of a cylinder of 6 mm diameter, whereas that between old cement and bone was roughly a cylinder of 12 mm diameter; the latter interface covers about twice the area. Since the old cement-new cement interface was still intact when the bone-cement interface had failed, it can be assumed that it supported the same load, and hence twice the stress borne by the bone-cement interface. Previous studies of cement-cement adhesion have used specimen geometry more accurately, applying pure shear (1-3), and have found that the shear strength ’ of cementcement interfaces is strongly influenced by time of formation and moisture. Compared with the shear strength of unlaminated PMMA, reductions of between 7 (2) and 30 per cent (3) were reported. The improved interfacial strength for laminated cement samples measured in this study cannot be due to wedging of the newer cement in its hole in the old cement, for PMMA cement shrinks as it sets (4,s). In an attempt to mimic the clinical situation, water was deliberately applied to the interface and mopped up poorly, so it may also be assumed that the worst case of a wet interface has been modelled. The presence of blood or moisture at a forming interface was found to reduce the interfacial strength by 60-80 per cent (3,6). As the present test geometry more closely approximates the prosthetic situation than any previous tests of cementement adhesion, these results suggest that the recementing practice should be reappraised. It is implied that better fixation might be yielded at revision surgery if soundly attached PMMA were not removed but simply roughened and used as the basis for attachment of the new cement. The effect of moisture or blood at the forming interface should not be neglected, and care should be taken over this point. In all specimens, initial bone-cement fixation was 30 per cent stronger than when recemented. This cannot be due to ageing of the bone samples or cement, as precautions were taken to prevent drying of the bone. Sedlin and Hirsch (7) showed that bone strength increases with drying. Similarly, it has been shown that PMMA strength increases gradually with time (&11), so ageing of the PMMA cannot have been the cause.
49
Halawa et al. (3) found that the shear of the cementbone interface was related to the shear strength of the cancellous bone, which increased from the centre of the femoral neck towards the endosteal surface. On the basis of these findings, it could be expected that revision of our specimens, with exposure of cancellous bone closer to the endosteum, should yield stronger interfaces. 6 CONCLUSIONS
Chiselled and recemented cancellous bone-cement interfaces demonstrated 30 per cent lower interfacial shear strength than the primary bone-cement interfaces, which was statistically significant at the P = 0.02 level. Abutment of new cement against the existing cement mantle resulted in interfacial shear strength greater than that demonstrated at the primary bone-cement interface. @ Crown copyright 1992
REFERENCES Greenwald, A. S., Narten, N. C. and Wilde, A. H. Points in the technique of recementing in the revision of an implant arthroplasty. J . Bone J t Surg., 1978,60B, 107-110. Gruen, T. A., Markolf, K. L. and Amstutz, H. C. Effects of laminations and blood entrapment on the strength of acrylic bone cement. C h . Orthoo., 1976, 119,250-255. 3 Halawa, M., Lee, A. J. C, Ling, R. S. M. and Vangala, S. S. The shear strength of trabecular bone from the femur, and some factors affecting the shear strength of the cement-bone interface. Arch. Orthop. Traumat. Surg., 1978,92, 19-30. 4 Debrunner, H. U., Wettstein, A. and Hofer, P. The polymerisation of self-curing acrylic cements and problems due to the cement anchorage of joint prostheses. In Advances in artificial hip and knee joint technology (Eds M. Shaldach and D. Hohmann), 1976, pp. 294-324 (Springer Verlag, Berlin). 5 Haas S. S., Brauer, G. M. and Dickson, G. A characterisation of polymethyl-methacrylate bone cement. J . Bane Jt Surg., 1975,57A, 380-391. 6 Beaumont, P. W. R. and Plumpton, B. The strength of acrylic bone cements and acrylic cement-stainless steel interfaces. J . Mater. Sci., 1977, 12, 1853-1856. 7 Sedlin, E. D. and Hirsch, C. Factors affecting the determination of the physical properties of femoral cortical bone. Acta Orthop. Scand., 1966,37,29-48. 8 Lautenschlager, E. P, Marshall, G. W., Marks, K. E., Schwartz, J. and Nelson, C. J. Mechanical strength of acrylic bone cements impregnated with antibiotics. J . Biomed. Mater. Res., 1976, 10, 837-845. 9 Lautenschlager, E. P., Jacobs, J. J., Marshall, G. W. and Meyer, P. R. Mechanical properties of bone cements containing large doses of antibiotic powders. J . Biomed. Mater. Rex, 1976, 10, 929938. 10 Rostoker, W., Lereim, P. and Galante, J. 0. Effect of an in viuo environment on the strength of bone cement. J . Biomed. Mater. Res., 1979,13,365-370. 11 Treharne, R. W. and Brown, N. Factors influencing the creep behaviour of poly(methylmethacry1ate)cements. J . Biomed. Mater. Res. Symp., 1975,4 81-88.
Proc Instn Mech Engrs Vol 206 Downloaded from pih.sagepub.com at NANYANG TECH UNIV LIBRARY on June 15, 2015
Technical note
Revision of cemented fixation and cement-bone interface strength A Rosenstein, MD, W MacDonald, MPhil, A Iliadis, MD and P McLardySmith, FRCS Nuffield Orthopaedic Centre, Headington, Oxford Interfacial shear strength between poly(methy1 methacrylate) ( P M M A ) bone cement and cancellous bone was measured in bone samples from human proximal femora. Samples were prepared withfresh cement-bone, fresh cement inside a mantle of existing cement and with fresh cement-revised bone surfaces. Push-out tests to measure shear strength caused failure only at bone-cement interfaces; revised bone interfaces were 30 per cent weaker (P< 0.02)than primary interfaces. The clinical relevance is that revision of cemented joint arthroplasties may necessitate removal of components with sound cement-bone Jixation. The practice of removing all traces of P M M A cement may not yield the optimalfixation; adhesion of fresh cement to freshly prepared surfaces of the existing cement might also be considered where circumstances are favourable.
1 INTRODUCTION
Revision of cemented total joint arthroplasties is increasingly common, involving often difficult removal of components and cement. Many orthopaedic surgeons recommend complete removal of old cement before insertion of any new cement. In support of this, some studies have shown a 7-33 per cent reduction in shear strength where new cement interfaces with old (1-3). However, there do not appear to be any studies comparing cement-bone strength in primary and revision cementing, nor comparing the bond strength between new and old cement with revised cement-bone bonds.
eter indentor centered in the cement plug. The cortex of the specimen was supported on a hollow cylinder of appropriate dimensions resting on the platen of the Instron. Testing was performed at a cross-head speed of 2 mm/min and force and extension were recorded continuously. After push-out testing, the specimen and pushed out plug were photographed (Fig. 11, all the cement was removed with a chisel and the cementing and testing procedure was repeated with the newly formed cavity. The specimen was then rephotographed. 2.2 Group 2 (fourteen specimens: seven head, seven neck)
2 MATERIALS AND METHODS
Eleven fresh femoral heads and necks were obtained from total hip arthroplasty or autopsy, and were frozen immediately after removal. Using a saw guide and hacksaw, two slices of 15 mm nominal thickness were cut, one each from the head and neck region of the frozen specimen. Thickness was measured with vernier calipers at three locations equally spaced around the perimeter. Specimens were kept moist at all times, and paired head and neck slices were randomly allocated to two groups.
The specimens were drilled axially using a 12 mm diameter drill, and the cavity was filled with cement, as in group 1. After the 90 minute setting time, a 6 mm diameter hole was drilled axially through the centre of the cement plug. Water was poured into the cavity, the surfaces were tamped dry with paper towelling and the cavity was filled with newly mixed PMMA cement and pressurized. After a further 90 minutes setting time, a push-out test was performed with a 5 mm indentor centered in the second cement plug. When testing and photography
2.1 Group 1(eight specimens: four head, four neck)
A hole was drilled axially through the slice, using a drill press and 12 mm diameter drill. With the slice positioned on a non-adherent board, poly(methy1 methacrylate) (PMMA) bone cement (CMW Laboratories Limited, Exeter) was inserted into the cavity between two and three and a half minutes after mixing. Pressure was applied with the thumb, the pulp of the thumb sealing off the cavity. The slices were covered with saturated paper towelling while the cement was allowed to set. After 90 minutes, push-out tests were performed using a Universal Materials Testing System (Model 1122, Instron Limited, High Wycombe) with a 10 mm diamThe MS was received on I 1 September 1991 and was accepted for publication on 5 June 1992.
H03191
Fig. 1 Typical specimen photograph showing pushed-out cement-bone plug, millimetre scale in view
0954-4119/92 $3.00
+ .05
Downloaded from pih.sagepub.com at NANYANG TECH UNIV LIBRARY on June 15, 2015
Roc Instn Mech Engrs Vol 206
A ROSENSTEIN, W MacDONALD, A ILIADIS AND P McLARDY-SMITH
48
Table 1 Interfacial shear stress
Specimen Group I la lb 2a 2b 3a 3b
4a 4b Group 2 5a 5b 6a 6b 7a 7b 8a 8b 9a 9b 1Oa lob 1 la llb
Primary shear strength MPa
Revised shear strength MPa
34.1 38.7 8.2 58.8 34.4 37.7 14.2 63.3
31.5 32.8 15.5 34.9 24.5 11.8 20.9
20.4 17.9 20.5 38.1 22.6 15.2 41.7 23.4 21.5 43.3 29.3 46.8 24.7 32.9
21.9 11.8 11.2 8.6 21.4 6.8 24.1 17.9 17.9 35.9 26.2 28.1 21.0 7.8
~
had been completed, all cement was removed with a chisel, and the new cavity was cemented en bloc and retested.
501 T
60
a
Primary
nmm
Revision
40
Group 1
Group 2
Fig. 2 Nomogram of shear strengths by group
interfaces (P < 0.02), strength being reduced by about 3 0 per cent (Fig. 2). This trend holds true independent of bone source area or procedure adopted (that is group). The results also demonstrate that the cementbone bond was 31 per cent stronger in specimens from the femoral head than in specimens from the femoral neck (P < 0.03) (Fig. 3).
3 ANALYSIS
For each slice, the thickness was determined as the mean of three measures. Scaled photographs were digitized and the perimeter was measured on both superior and inferior surfaces of the cortical annulus and the pushed-out plug. From the four perimeters thus measured, the mean interfacial perimeter was calculated and multiplied by the mean thickness to estimate the surface area of shear failure. Interfacial shear stress was calculated by dividing the break load by the surface area (Table 1). Analysis of variance (ANOVA) was performed using a computer software package (General Linear Model, SAS). 4 RESULTS
One specimen (3a) cracked through the cortex during the first push-out test and was unable to be recemented. In an initial analysis of variance, specimens were separated by group; the differences were shown to be insignificant statistically (P = 0.15) and the groups were combined. The analysis was repeated with specimens grouped by source (head or neck) and procedure (primary or revised cement-bone). In the first series of tests in group 2, failure always occurred at the cement-bone interface. None of these specimens failed by pushing out the plug of new cement from within the established cement mass. The statistical analysis indicated that these cement-bone strengths were not significantly different from the initial cementbone strengths of the specimens in group 1. The ANOVA results indicate that revised cementbone interfaces are significantly weaker than primary
5 DISCUSSION
Specimens were taken from femoral heads and necks so that different distributions of cortical and cancellous bone were represented in the test series. The statistical analysis shows that the cement-bone shear strength is significantly lower in the specimens from the femoral neck than those from the femoral head. This may be accounted for by the effect of cancellous structure. Since the cancellous bone in the femoral neck is more aligned with the neck axis, the stronger elements lie in that
:I T 40 I-
Primary
Revision
I
Heads
Necks
Fig. 3 Nomogram of shear strengths by specimen origin
Part H : Journal of Engineering in Medicine Downloaded from pih.sagepub.com at NANYANG TECH UNIV LIBRARY on June 15, 2015
REVISION OF CEMENTED FIXATION AND CEMENT-BONE INTERFACE STRENGTH
direction. As a result, in cylindrical surfaces parallel to the axis (such as encountered by the cement), the texture will be less open and the interlock with the cement will be less. Increased interfacial strength in the region with more cancellous bone (the head specimens) calls into question the practice of removing all cancellous bone from the trochanteric region as part of surgical preparation of the femur for total hip replacement. It would seem that improved fixation might result if the cancellous bone were retained in such areas. All specimens tested in the first test of group 2 failed at the boneecement interface, and not at the interface between the old and new cement. The interface between old and new cement took the form of a cylinder of 6 mm diameter, whereas that between old cement and bone was roughly a cylinder of 12 mm diameter; the latter interface covers about twice the area. Since the old cement-new cement interface was still intact when the bone-cement interface had failed, it can be assumed that it supported the same load, and hence twice the stress borne by the bone-cement interface. Previous studies of cement-cement adhesion have used specimen geometry more accurately, applying pure shear (1-3), and have found that the shear strength ’ of cementcement interfaces is strongly influenced by time of formation and moisture. Compared with the shear strength of unlaminated PMMA, reductions of between 7 (2) and 30 per cent (3) were reported. The improved interfacial strength for laminated cement samples measured in this study cannot be due to wedging of the newer cement in its hole in the old cement, for PMMA cement shrinks as it sets (4,s). In an attempt to mimic the clinical situation, water was deliberately applied to the interface and mopped up poorly, so it may also be assumed that the worst case of a wet interface has been modelled. The presence of blood or moisture at a forming interface was found to reduce the interfacial strength by 60-80 per cent (3,6). As the present test geometry more closely approximates the prosthetic situation than any previous tests of cementement adhesion, these results suggest that the recementing practice should be reappraised. It is implied that better fixation might be yielded at revision surgery if soundly attached PMMA were not removed but simply roughened and used as the basis for attachment of the new cement. The effect of moisture or blood at the forming interface should not be neglected, and care should be taken over this point. In all specimens, initial bone-cement fixation was 30 per cent stronger than when recemented. This cannot be due to ageing of the bone samples or cement, as precautions were taken to prevent drying of the bone. Sedlin and Hirsch (7) showed that bone strength increases with drying. Similarly, it has been shown that PMMA strength increases gradually with time (&11), so ageing of the PMMA cannot have been the cause.
49
Halawa et al. (3) found that the shear of the cementbone interface was related to the shear strength of the cancellous bone, which increased from the centre of the femoral neck towards the endosteal surface. On the basis of these findings, it could be expected that revision of our specimens, with exposure of cancellous bone closer to the endosteum, should yield stronger interfaces. 6 CONCLUSIONS
Chiselled and recemented cancellous bone-cement interfaces demonstrated 30 per cent lower interfacial shear strength than the primary bone-cement interfaces, which was statistically significant at the P = 0.02 level. Abutment of new cement against the existing cement mantle resulted in interfacial shear strength greater than that demonstrated at the primary bone-cement interface. @ Crown copyright 1992
REFERENCES Greenwald, A. S., Narten, N. C. and Wilde, A. H. Points in the technique of recementing in the revision of an implant arthroplasty. J . Bone J t Surg., 1978,60B, 107-110. Gruen, T. A., Markolf, K. L. and Amstutz, H. C. Effects of laminations and blood entrapment on the strength of acrylic bone cement. C h . Orthoo., 1976, 119,250-255. 3 Halawa, M., Lee, A. J. C, Ling, R. S. M. and Vangala, S. S. The shear strength of trabecular bone from the femur, and some factors affecting the shear strength of the cement-bone interface. Arch. Orthop. Traumat. Surg., 1978,92, 19-30. 4 Debrunner, H. U., Wettstein, A. and Hofer, P. The polymerisation of self-curing acrylic cements and problems due to the cement anchorage of joint prostheses. In Advances in artificial hip and knee joint technology (Eds M. Shaldach and D. Hohmann), 1976, pp. 294-324 (Springer Verlag, Berlin). 5 Haas S. S., Brauer, G. M. and Dickson, G. A characterisation of polymethyl-methacrylate bone cement. J . Bane Jt Surg., 1975,57A, 380-391. 6 Beaumont, P. W. R. and Plumpton, B. The strength of acrylic bone cements and acrylic cement-stainless steel interfaces. J . Mater. Sci., 1977, 12, 1853-1856. 7 Sedlin, E. D. and Hirsch, C. Factors affecting the determination of the physical properties of femoral cortical bone. Acta Orthop. Scand., 1966,37,29-48. 8 Lautenschlager, E. P, Marshall, G. W., Marks, K. E., Schwartz, J. and Nelson, C. J. Mechanical strength of acrylic bone cements impregnated with antibiotics. J . Biomed. Mater. Res., 1976, 10, 837-845. 9 Lautenschlager, E. P., Jacobs, J. J., Marshall, G. W. and Meyer, P. R. Mechanical properties of bone cements containing large doses of antibiotic powders. J . Biomed. Mater. Rex, 1976, 10, 929938. 10 Rostoker, W., Lereim, P. and Galante, J. 0. Effect of an in viuo environment on the strength of bone cement. J . Biomed. Mater. Res., 1979,13,365-370. 11 Treharne, R. W. and Brown, N. Factors influencing the creep behaviour of poly(methylmethacry1ate)cements. J . Biomed. Mater. Res. Symp., 1975,4 81-88.
Proc Instn Mech Engrs Vol 206 Downloaded from pih.sagepub.com at NANYANG TECH UNIV LIBRARY on June 15, 2015
