Original Article
Impact of implant size on cement filling in hip resurfacing arthroplasty
Proc IMechE Part H: J Engineering in Medicine 2014, Vol 228(1) 3–10 Ó IMechE 2013 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0954411913507660 pih.sagepub.com
Roel de Haan1, Nico Buls2 and Thierry Scheerlinck3
Abstract Larger proportions of cement within femoral resurfacing implants might result in thermal bone necrosis. We postulate that smaller components are filled with proportionally more cement, causing an elevated failure rate. A total of 19 femoral heads were fitted with polymeric replicas of ReCap (Biomet) resurfacing components fixed with low-viscosity cement. Two specimens were used for each even size between 40 and 56 mm and one for size 58 mm. All specimens were imaged with computed tomography, and the cement thickness and bone density were analyzed. The average cement mantle thickness was 2.63 mm and was not correlated with the implant size. However, specimen with low bone density had thicker cement mantles regardless of size. The average filling index was 36.65% and was correlated to both implant size and bone density. Smaller implants and specimens with lower bone density contained proportionally more cement than larger implants. According to a linear regression model, bone density but not implant size influenced cement thickness. However, both implant size and bone density had a significant impact on the filling index. Large proportions of cement within the resurfacing head have the potential to generate thermal bone necrosis and implant failure. When considering hip resurfacing in patients with a small femoral head and/or osteoporotic bone, extra care should be taken to avoid thermal bone necrosis, and alternative cementing techniques or even cementless implants should be considered. This study should help delimiting the indications for hip resurfacing and to choose an optimal cementing technique taking implant size into account.
Keywords Hip resurfacing, bone cement, implant size, necrosis, bone density
Date received: 19 February 2013; accepted: 13 September 2013
Introduction Worldwide, in the young and active patients, metal-onmetal hip resurfacing arthroplasty (HRA) is considered as an alternative to classic total hip arthroplasty (THA). Although overall medium- and long-term results are considered good,1–4 McMinn et al.5 found the revision rate of HRA to be higher than that of traditional cemented THA. Especially, the increased risk of early failures is a concern.6–10 Generally, revision of a HRA is regarded as an easier operation than THA revision. However, high re-revision rates (10% at 2 years,11 11% at 5 years,12 up to 20% at 5 years after isolated acetabular revision13) demonstrate that the problem of early HRA failure is not solved yet. Generally, age, female gender, implant design, metallurgy and geometry of both components as well as the orientation of the acetabular component are thought to be the most important contributors to failure in HRA.11,13,14 However, besides the presence of pores or defects in the cured cement,15,16 femoral neck
fractures and necrosis of the reamed femoral head inside the resurfacing component have also been blamed.17,18 Femoral neck fractures could be due to microfractures caused by forceful femoral component impaction,19,20 damage to the blood supply at the time of surgery21 and/or thermal bone necrosis caused by heat generation inside the reamed femoral head during cement curing. That last mechanism has been studied extensively. Gill et al.22 used thermal probes to record 1
Department of Orthopaedic Surgery and Traumatology, Tergooi Ziekenhuizen, Blaricum, The Netherlands 2 Department of Radiology, University Hospital Brussels, Brussels, Belgium 3 Department of Orthopaedic Surgery and Traumatology, University Hospital Brussels, Brussels, Belgium Corresponding author: Roel de Haan, Department of Orthopaedic Surgery and Traumatology, Tergooi Ziekenhuizen, Rijksstraatweg 1, 1261 AN Blaricum, The Netherlands. Email: [email protected]
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
4
Proc IMechE Part H: J Engineering in Medicine 228(1)
temperatures close to the cement–bone interface during resurfacing arthroplasty, and recorded temperatures able to kill osteocytes. To limit thermal damage, these authors suggest inserting a suction cannula into the lesser trochanter, generous pulsed lavage and early reduction of the freshly resurfaced joint. Janssen et al.23 calculated, based on finite element analysis (FEA), that thermal bone necrosis can be significant when a cementing technique that pressurizes large amounts of cement into the reamed femoral head is combined with bone cements that generate much heat during curing. Overall, women are at higher risk of HRA revision compared to men.24–26 However, when adjusted for femoral component size, the difference between men and women becomes less obvious.13 This suggests that implant size could be the major cause of early revision in women.27 Moreover, Corten and MacDonald12 found an inverse relationship between the size of the femoral component and the risk of revision for HRA. There is also evidence that independent of the patient’s gender, HRA with a femoral component smaller than 44 mm has a fourfold increase in revision rate compared to those with a femoral component of 55 mm or more.13 Although the cause is unclear, smaller femoral resurfacing implants tend to fail more often than larger implants. We postulate that smaller femoral HRA components are filled with proportionally more cement than larger heads. The increased proportion of bone cement inside the reamed femoral head could favor thermal bone necrosis and implant failure. As such, we investigated the relation between implant size and the quantity of cement that can be pressurized into a reamed femoral head.
Materials and methods Specimens and implantation technique Out of 44 femoral heads retrieved during primary conventional THA, we selected 19 specimens that could be fitted with a ReCap (Biomet, Warsaw, IN, USA) resurfacing component. The selected femoral heads were retrieved from 11 male and 8 female patients, with an average age of 71.9 years (standard deviation (SD): 8.3 years, range: 55–86 years). These patients were not selected for HRA because of their age and the surgeon’s preference to use a THA. The femoral heads were chosen so that they could be fitted with an even resurfacing implant size ranging from 40 to 58 mm. For sizes 40–56 mm, two heads were available, whereas for size 58 mm only one head could be found. To avoid metallic X-ray scatter artifacts during computed tomography (CT) scanning, we used glass-filled nylon replicas (ARRK Product Development Group, Gloucester, UK) that were produced based on threedimensional computer-aided design (CAD) models of original ReCap femoral implants (Figure 1).
Figure 1. Polymeric replica of a ReCap hip resurfacing implant.
The selected femoral heads were stored at 274 °C until they could be machined with the original ReCap instruments according to the recommendations of the manufacturer. All femoral heads were provided with anchoring holes drilled with the original 3.0 mm ReCap drill bit. The number of anchoring holes ranged from 5 for the smallest heads (40 mm implant) to 12 for the largest head (58 mm implant). The number of holes was proportional to the area of the hemisphere of the reamed femoral head and represented on average 0.48% (SD: 0.02%, range: 0.46%–0.51%) of that area (Table 1). Prior to implantation, all reamed femoral heads were stored in water at 37.0 °C and cleaned with 2.0 L of water at room temperature (21.0 °C) and pulse lavage. An amount of 40 g of low-viscosity cement (Refobacin Bone Cement LV; Biomet Orthopaedics Switzerland, Ried b. Kerzers, Switzerland) was vacuum-mixed with a mixing and pressuring system (Optivac; Biomet Cementing Technologies, Sjo¨bo, Sweden) for 30 s at 0.1–0.3 bar (10–30 kPa). Within the timeframe between 1.0 and 2.5 min after initiating the mixing, exactly half the available volume inside each implant was filled with cement. The volume available inside each implant was calculated based on the weight increase when filled with water (Table 1). The mass of cement required to fill half the volume was calculated based on a bone cement density of 1.17 g/mL28 and was measured on an electronic scale (On Balance Trueweight-2 pocket scale 600 3 0.1 g). Once filled with cement, implants were rotated to cover their entire inner surface with cement, and 2.5 min after starting cement mixing, they were pressed manually onto the prepared femoral head until seated. The implants were pressurized during curing. All cement that was pressed out of the implant during seating was removed.
CT scanner and measurements After cement curing, all specimens were scanned with a Somatom Sensation 16 (Siemens, Erlangen, Germany)
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
de Haan et al.
5
Table 1. The number of anchoring holes drilled in each ReCap implant size, corresponding percentage of the area of the reamed femoral head, volume inside each ReCap implant size and mass of cement needed to fill half that volume. Implant size (mm) Number of anchoring holes Percentage of surface of hemisphere (%) Volume inside implant (mL) Mass of cement to fill half the implant (g)
40
42
44
46
48
50
52
54
56
58
5 0.47 17.3 10.1
6 0.51 20.5 12.0
6 0.45 24.0 14.0
7 0.48 27.9 16.3
8 0.50 32.8 19.2
8 0.45 36.8 21.5
9 0.47 42.7 25.0
10 0.48 47.4 27.7
11 0.49 54.0 31.6
12 0.49 61.0 35.7
Table 2. Correlation between cement mantle characteristics and femoral head parameters (Pearson’s correlation coefficient (p value)).
Cement thickness Filling index Implant size Bone density
Cement thickness
Filling index
Implant size
Bone density
X 0.889 (\ 0.001) 20.119 (0.628) 20.676 (0.001)
0.889 (\ 0.001) X 20.508 (0.026) 20.568 (0.011)
20.119 (0.628) 20.508 (0.026) X 0.003 (0.991)
20.676 (0.001) 20.568 (0.011) 0.003 (0.991) X
CT scanner according to a standardized scan protocol.15 The resurfaced femoral heads were scanned from the top of the dome to the base of the resurfacing head along their longitudinal axis with following settings: scan collimation, 6 3 0.75 mm; reconstructed slice thickness, 1.0 mm; reconstructed slice interval, 0.5 mm; reconstruction kernel, U80u very sharp; tube potential, 140 kVp; tube current, 200 mA; exposure time, 1 s and field of view, 75 mm. An adapted version of validated segmentation software15,29 developed in MATLAB 6.5 (The MathWorks, Natick, MA, USA) was used to analyze the CT data. The number of images per specimen varied between 41 and 65 depending on implant size (mean: 53.3, SD: 7.8). Four different elements were identified based on their CT values, expressed in Hounsfield units (HU): the polymeric stem and the outer shell of the implant, the cement-free cancellous bone and the cement mantle including cement pressurized into the cancellous bone. The cement thickness, consisting of cement pressurized into cancellous bone and pure cement, was measured in 72 segments of 5° within each CT scan image as described previously.15,29,30 The mean cement mantle thickness per specimen was calculated as the average of all cement thickness measurements within each specimen. The ‘‘filling index’’ of the cement was defined as the fraction cement volume of the total volume within each implant.15 The ‘‘bone density’’ of a specimen was defined as the average CT value of the cement-free cancellous bone (expressed in HU). For each specimen, one bone density measurement was performed in five different regions and those measurements were averaged. The average number of pixel per measurement area was 4449 (SD: 2058 pixels).
Statistical analysis Correlations between implant size, the cement mantle thickness, the cement filling index as well as the bone density were evaluated with Pearson’s correlation coefficient. To assess the contribution of bone density and implant size on the cement mantle thickness and the cement filling index, we used linear regression. Statistical analysis was performed in MATLAB 6.5 (The MathWorks) and SPSS 19 (SPSS Inc., Chicago, IL, USA); p values \0.05 were considered to be significant.
Results The overall average thickness of the cement mantle (pure cement + cement pressurized into the cancellous bone) was 2.63 mm (SD: 0.86 mm, range: 1.65–4.60 mm). No significant correlation was found between cement mantle thickness and implant size. However, there was a significant negative correlation between cement mantle thickness and bone density (Table 2). Specimen with a low bone density had thicker cement mantles regardless of implant size (Figure 2). On average, cement filled 36.65% of the volume within the femoral resurfacing implants, but large variations did exist between specimens (SD: 10.81%, range: 21.52%– 57.60%). There was a significant negative correlation between the filling index and implant size and between filling index and bone density (Table 2). Smaller implants and specimens with a lower bone density contained proportionally more cement than larger implants (Figure 2). The average bone density was 299 HU (SD: 161 HU, range: 247 to 569 HU). There was no significant correlation between bone density and implant size (Table 2 and Figure 2).
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
6
Proc IMechE Part H: J Engineering in Medicine 228(1)
Figure 2. (a) Bone density versus cement mantle thickness with a trend line, (b) size of the femoral components versus filling index with a trend line, (c) bone density versus filling index with a trend line, (d) size of the femoral components versus bone density with a trend line, (e) filling index versus cement mantle thickness with a trend line and (f) size of the femoral components versus cement mantle thickness with a trend line. HU: Hounsfield units.
Linear regression suggested that cement thickness was strongly influenced by bone density but not by implant size. On the contrary, both implant size and bone density had a significant impact on the filling index (p \ 0.001; R2 = 0.58) (Table 3). The influence of bone density was somewhat stronger than that of the implant size as reflected by the effect size (partial h2).
Discussion This in vitro experimental study investigated the impact of the average bone density and implant size on the
cement mantle thickness and cement filling index within a resurfacing head implanted using a ‘‘low-viscosity cement filling technique.’’ In these circumstances, femoral heads with a lower average bone density presented thicker cement mantles and contained a larger proportion of cement. On the contrary, implant size had no significant impact on the cement mantle thickness, but smaller implants contained proportionally more cement than larger ones. The effect of bone density on cement penetration into the cancellous bone seems obvious. The bone density in this study was measured by averaging the CT
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
de Haan et al.
7
Table 3. Impact of implant size and bone density on the cement thickness and cement filling index according to a linear regression model. Predictor Cement thickness Filling index
Size Bone density Size Bone density
Regression coefficient 20.018 20.004 20.969 20.038
95% CI 20.076, 20.041 20.006, 20.002 21.626, 20.312 20.061, 20.015
p value
Partial h2
0.529 0.002 0.006 0.003
0.025 0.936 0.379 0.433
CI: confidence interval.
Figure 3. The same cement mantle thickness within an implant (gray area) will result in a smaller filling index when the implant is large (A) than when it is small (B).
values of the cement-free cancellous bone in five areas per specimen. As such, it reflects the degree of osteoporosis (proportion of trabeculae within the cancellous bone) and also the nature of the material between the bone trabeculae, that is, air, fat, blood or saline used during pressure lavage of the femoral heads. Less bone trabeculae and a larger proportion of air between these trabeculae are expected to facilitate cement penetration into the cancellous bone. However, a previous experimental study did not find that relation even when using the same implantation technique.15 This could be explained by the use of more homogenous, extensively cleaned, bovine cancellous bone (density: 2446 HU, SD: 99 HU, range: 2665 to 2194 HU) compared to the human bone samples used in this study. In fact, the specimens used in this study might match the bone quality encountered in clinical practice much better. The fact that smaller implants, independent from their bone density, contained proportionally more cement than larger implants is more difficult to explain. From this and previous studies, we know that cement penetration depth (cement thickness) depends on the bone density and the cementing technique.15,31 However, we found that using the same ‘‘cement filling technique,’’ the average cement mantle thickness was independent from the implant size. As such, the proportion of cement within a smaller reamed femoral head will be larger than that of a larger femoral head. Indeed, in a simplified model where the reamed femoral
head is represented as a hemisphere, a shell of that hemisphere with a fixed thickness will represent a larger fraction of the total volume when the hemisphere has a smaller diameter compared to when it has a larger diameter (Figure 3). Based on that simplified model, we found a significant correlation (Pearson’s correlation coefficient: 0.505, p = 0.027) between the filling indices that were measured and those calculated for an average cement mantle of 2.63 mm (Figure 4). The proportion of cement filling the inside of the resurfacing implant is of importance because large quantities of cement could cause thermal bone necrosis during the exothermal polymerization reaction. Such bone necrosis has been demonstrated in failed resurfacing implants22,32–34 and has been simulated with FEA models.23 To reduce the amount of thermal bone necrosis, several proposals have been formulated. Gill et al.22 suggest the use of a suction cannula into the lesser trochanter during cementing, early reduction of the resurfaced head into the acetabular component filled with cooled saline solution and extensive pressure lavage. Scheerlinck et al.15 suggested precooling the reamed femoral head with cooled pressure lavage fluid prior to cementing. Based on an FEA model by Janssen et al.,23 precooling of the femoral head with fluid at 10 °C could reduce the amount of thermal necrosis by a factor of 5. However, bone cement is sensitive to differences in temperature at interfacial surfaces during curing. This could lead to pore formation at the cooler surface35
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
8
Proc IMechE Part H: J Engineering in Medicine 228(1)
Figure 4. Filling index versus femoral implant size. The trend line between the measured filling indices corresponds almost perfectly with the calculated filling indices based on a fixed cement mantle thickness of 3.00 mm within a hemisphere. The dotted line represents the calculated filling index for the average cement mantle of our specimens (2.63 mm).
carrying with it the risk of failure due to voids and defects in the cement mantle at critical locations. Based on this study and previous work investigating cementing techniques,15 small resurfacing heads implanted with a ‘‘cement filling technique’’ on a reamed femoral head with a low bone density (wellcleaned, osteoporotic bone) are the most at risk of overfilling the femoral head with cement. In our study, the combination of small sizes and low bone density resulted in a filling index above 50% and an average cement mantle thickness above 3.5 mm in three cases (implant size: 40, 44 and 50 mm; average bone density: 282, 247 and 12 HU, respectively). It was possible to obtain a filling index above 50% despite the fact that we filled the implants only halfway because the cement mantle contained not only cement but also trabecular bone. The combination of small femoral heads and osteoporotic bone is most likely to be found in elderly women, a population in which hip resurfacing is known to be problematic.13,27 The effect of implant size should not be underestimated as it could, on its own, explain some of the differences in resurfacing survival between men and women in the Australian hip registry.13,27 As for all in vitro experimental studies, several limitations apply to this work. First, the femoral heads we used were retrieved from a general THA population. That population does not necessarily match to the ideal resurfacing candidates. However, we used human femoral heads presenting major osteoarthritis and some degree of deformity that would typically be encountered during hip resurfacing. Second, back bleeding during cementing was not simulated. However, during in vivo hip resurfacing, back bleeding is generally not a major issue. Indeed, it is often controlled by inserting a
suction cannula into the lesser trochanter, and femoral head vascularity is often compromised by the approach and the dislocated hip position during cementing. Third, in this study, water was used for lavage and storage instead of saline. This can affect the properties of the bone as well as the condition of exposed proteins on the surface of the bone. This in turn may influence the cement interfacial bonding, but we do not believe that it could influence the cement penetration into the trabecular bone during implantation. Fourth, in order to allow CT scanning without metallic artifacts, we used plastic replicas of original implants. These replicas could have dimensional mismatches compared to the original implants or could theoretically deform during cementing. However, the manufacturer of the original implants produced the replicas based on accurate CAD models. Moreover, no force was needed to insert them on the reamed femoral heads as they presented an inner radial clearance of 0.5–0.6 mm.36,37 Besides the dimensional mismatches of the plastic replicas, the plastic material is likely to have an effect on the heat dissipation of the cement during curing due to the large difference in thermal conductivity and specific heat of the plastic models we used, compared to cobalt-chrome implants used in clinical practice. The reduced ability of plastic materials to conduct heat away from the exothermic polymerization of cement will increase the curing speed and the viscosity of the cement. This could reduce the cement penetration into the bone. Nevertheless, we do not expect the use of plastic replicas to affect the cement mantles significantly, at most this could accelerate cement curing close to the implant in a region that is filled with cement anyway. Accelerating the curing speed in that region is unlikely to influence the overall
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
de Haan et al.
9
cement penetration in the specimen. Finally, reported results only apply to the ReCap resurfacing heads implanted with a ‘‘low-viscosity cement filling technique combined with anchoring holes.’’ Indeed, resurfacing heads with different inner clearances and different inner morphologies could result in very different cement mantles. Also, discarding anchoring holes and the use of different cementing techniques could have a major impact on cement penetration into the femoral head. As such, the conclusions drawn from this study cannot readily be extrapolated to other implants or other implantation techniques. In conclusion, small femoral resurfacing implants with the same cement mantle thickness contain proportionally more cement. When using a ‘‘low-viscosity cement filling technique combined with anchoring holes,’’ small resurfacing implants have the potential to be filled with larger quantities of cement. The risk increases when the reamed femoral head contains lowdensity bone, that is, osteoporotic bone that has been extensively cleaned. The large proportion of cement within the resurfacing head has the potential to generate thermal bone necrosis and implant failure. This further reinforces that patients with osteoporotic bone are a contraindication for HRA. Consequently, when considering hip resurfacing in patients with a small femoral head and/or osteoporotic bone, extra care should be taken to avoid thermal bone necrosis during cementing, and alternative cementing techniques or even cementless implants could be considered. Acknowledgements The authors would like to thank Ronald Buyl, Professor of Statistics, Department of Biostatistics and Information Technology, Faculty of Medicine and Pharmacy, Vrije Universiteit Brussel, Brussels, Belgium, for his advice on statistical issues. The Department of Radiology of the Universitair Ziekenhuis Brussel, Brussels, Belgium, provided the CT imaging data of the specimens, and Biomet Deutschland, Dieburg, Germany, provided the polymeric implant replicas, bone cement and cement mixing systems. The study protocol of this work was approved by the ethical committee of the Universitair Ziekenhuis Brussel, Brussels, Belgium. Declaration of conflicting interests The authors declare that there is no conflict of interest. Funding This study was funded by the Department of Orthopedic Surgery and Traumatology of the Universitair Ziekenhuis Brussel, Brussels, Belgium. References 1. De Smet KA. Belgium experience with metal-on-metal surface arthroplasty. Orthop Clin North Am 2005; 36(2): 203–213, ix.
2. Steffen RT, Pandit HP, Palan J, et al. The five-year results of the Birmingham hip resurfacing arthroplasty: an independent series. J Bone Joint Surg Br 2008; 90(4): 436–441. 3. Van Der Bracht H, Vander Eecken S, Vyncke D, et al. Clinical and functional outcome of the Birmingham hip resurfacing. Acta Orthop Belg 2011; 77(6): 771–776. 4. Gross TP and Liu F. Hip resurfacing with the Biomet Hybrid ReCap-Magnum system: 7-year results. J Arthroplasty 2012; 27(9): 1683–1689. 5. McMinn DJ, Snell KI, Daniel J, et al. Mortality and implant revision rates of hip arthroplasty in patients with osteoarthritis: registry based cohort study. BMJ 2012; 344: e3319. 6. Morlock MM, Bishop N, Zustin J, et al. Modes of implant failure after hip resurfacing: morphological and wear analysis of 267 retrieval specimens. J Bone Joint Surg Am 2008; 90(Suppl. 3): 89–95. 7. Corten K, Ganz R, Simon JP, et al. Hip resurfacing arthroplasty: current status and future perspectives. Eur Cell Mater 2011; 21: 243–258. 8. Della Valle CJ, Nunley RM, Raterman SJ, et al. Initial American experience with hip resurfacing following FDA approval. Clin Orthop Relat Res 2009; 467(1): 72–78. 9. Berend KR, Lombardi AV Jr, Adams JB, et al. Unsatisfactory surgical learning curve with hip resurfacing. J Bone Joint Surg Am 2011; 93(Suppl. 2): 89–92. 10. Van der Weegen W, Hoekstra HJ, Sijbesma T, et al. Hip resurfacing in a district general hospital: 6-year clinical results using the ReCap hip resurfacing system. BMC Musculoskelet Disord 2012; 13: 247. 11. De Haan R, Campbell PA, Su EP, et al. Revision of metal-on-metal resurfacing arthroplasty of the hip: the influence of malpositioning of the components. J Bone Joint Surg Br 2008; 90(9): 1158–1163. 12. Corten K and MacDonald SJ. Hip resurfacing data from national joint registries: what do they tell us? What do they not tell us? Clin Orthop Relat Res 2010; 468(2): 351–357. 13. Australian Orthopaedic Association National Joint Replacement Registry. Annual reports, https://aoanjrr. dmac.adelaide.edu.au/annual-reports (accessed 11 February 2013). 14. Shimmin AJ, Walter WL and Esposito C. The influence of the size of the component on the outcome of resurfacing arthroplasty of the hip: a review of the literature. J Bone Joint Surg Br 2010; 92(4): 469–476. 15. Scheerlinck T, Delport H and Kiewitt T. Influence of the cementing technique on the cement mantle in hip resurfacing: an in vitro computed tomography scanbased analysis. J Bone Joint Surg Am 2010; 92(2): 375–387. 16. Bitsch RG, Jager S, Lurssen M, et al. The influence of cementing technique in hip resurfacing arthroplasty on the initial stability of the femoral component. Int Orthop 2011; 35(12): 1759–1765. 17. Falez F, Favetti F, Casella F, et al. Results of hip resurfacing. Int Orthop 2011; 35(2): 239–243. 18. Carrothers AD, Gilbert RE, Jaiswal A, et al. Birmingham hip resurfacing: the prevalence of failure. J Bone Joint Surg Br 2010; 92(10): 1344–1350. 19. Schlegel UJ, Rothstock S, Siewe J, et al. Does impaction matter in hip resurfacing? A cadaveric study. J Arthroplasty 2011; 26(2): 296–302.
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
10
Proc IMechE Part H: J Engineering in Medicine 228(1)
20. Campbell P, Beaule PE, Ebramzadeh E, et al. The John Charnley Award: a study of implant failure in metal-onmetal surface arthroplasties. Clin Orthop Relat Res 2006; 453: 35–46. 21. Steffen RT, Athanasou NA, Gill HS, et al. Avascular necrosis associated with fracture of the femoral neck after hip resurfacing: histological assessment of femoral bone from retrieval specimens. J Bone Joint Surg Br 2010; 92(6): 787–793. 22. Gill HS, Campbell PA, Murray DW, et al. Reduction of the potential for thermal damage during hip resurfacing. J Bone Joint Surg Br 2007; 89(1): 16–20. 23. Janssen D, Srinivasan P, Scheerlinck T, et al. Effect of cementing technique and cement type on thermal necrosis in hip resurfacing arthroplasty—a numerical study. J Orthop Res 2012; 30(3): 364–370. 24. Della Valle CJ, Nunley RM and Barrack RL. When is the right time to resurface? Orthopedics 2008; 31(12 Suppl. 2), http://www.orthosupersite.com/view.asp?rID=37185 (accessed 11 February 2013). 25. Jameson SS, Langton DJ, Natu S, et al. The influence of age and sex on early clinical results after hip resurfacing: an independent center analysis. J Arthroplasty 2008; 23(6 Suppl. 1): 50–55. 26. Nunley RM, Della Valle CJ and Barrack RL. Is patient selection important for hip resurfacing? Clin Orthop Relat Res 2009; 467(1): 56–65. 27. McBryde CW, Theivendran K, Thomas AM, et al. The influence of head size and sex on the outcome of Birmingham hip resurfacing. J Bone Joint Surg Am 2010; 92(1): 105–112. 28. Skinner HB and Murray WR. Variations in the density of bone cement after centrifugation. Clin Orthop Relat Res 1986; 207: 263–269.
29. Scheerlinck T, de Mey J and Deklerck R. In vitro analysis of the cement mantle of femoral hip implants: development and validation of a CT-scan based measurement tool. J Orthop Res 2005; 23(4): 698–704. 30. Scheerlinck T, de Mey J, Deklerck R, et al. CT analysis of defects of the cement mantle and alignment of the stem: in vitro comparison of Charnley-Kerboul femoral hip implants inserted line-to-line and undersized in paired femora. J Bone Joint Surg Br 2006; 88(1): 19–25. 31. Chandler M, Kowalski RS, Watkins ND, et al. Cementing techniques in hip resurfacing. Proc IMechE, Part H: J Engineering in Medicine 2006; 220(2): 321–331. 32. Little CP, Ruiz AL, Harding IJ, et al. Osteonecrosis in retrieved femoral heads after failed resurfacing arthroplasty of the hip. J Bone Joint Surg Br 2005; 87(3): 320–323. 33. Morlock MM, Bishop N, Ruther W, et al. Biomechanical, morphological, and histological analysis of early failures in hip resurfacing arthroplasty. Proc IMechE, Part H: J Engineering in Medicine 2006; 220(2): 333–344. 34. Zustin J, Sauter G, Morlock MM, et al. Association of osteonecrosis and failure of hip resurfacing arthroplasty. Clin Orthop Relat Res 2009; 468(3): 756–761. 35. Pelletier MH, Lau AC, Smitham PJ, et al. Pore distribution and material properties of bone cement cured at different temperatures. Acta Biomater 2010; 6(3): 886–891. 36. Beaule PE, Matar WY, Poitras P, et al. 2008 Otto Aufranc Award: component design and technique affect cement penetration in hip resurfacing. Clin Orthop Relat Res 2009; 467(1): 84–93. 37. Grigoris P, Roberts P, Panousis K, et al. Hip resurfacing arthroplasty: the evolution of contemporary designs. Proc IMechE, Part H: J Engineering in Medicine 2006; 220(2): 95–105.
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
Impact of implant size on cement filling in hip resurfacing arthroplasty
Proc IMechE Part H: J Engineering in Medicine 2014, Vol 228(1) 3–10 Ó IMechE 2013 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0954411913507660 pih.sagepub.com
Roel de Haan1, Nico Buls2 and Thierry Scheerlinck3
Abstract Larger proportions of cement within femoral resurfacing implants might result in thermal bone necrosis. We postulate that smaller components are filled with proportionally more cement, causing an elevated failure rate. A total of 19 femoral heads were fitted with polymeric replicas of ReCap (Biomet) resurfacing components fixed with low-viscosity cement. Two specimens were used for each even size between 40 and 56 mm and one for size 58 mm. All specimens were imaged with computed tomography, and the cement thickness and bone density were analyzed. The average cement mantle thickness was 2.63 mm and was not correlated with the implant size. However, specimen with low bone density had thicker cement mantles regardless of size. The average filling index was 36.65% and was correlated to both implant size and bone density. Smaller implants and specimens with lower bone density contained proportionally more cement than larger implants. According to a linear regression model, bone density but not implant size influenced cement thickness. However, both implant size and bone density had a significant impact on the filling index. Large proportions of cement within the resurfacing head have the potential to generate thermal bone necrosis and implant failure. When considering hip resurfacing in patients with a small femoral head and/or osteoporotic bone, extra care should be taken to avoid thermal bone necrosis, and alternative cementing techniques or even cementless implants should be considered. This study should help delimiting the indications for hip resurfacing and to choose an optimal cementing technique taking implant size into account.
Keywords Hip resurfacing, bone cement, implant size, necrosis, bone density
Date received: 19 February 2013; accepted: 13 September 2013
Introduction Worldwide, in the young and active patients, metal-onmetal hip resurfacing arthroplasty (HRA) is considered as an alternative to classic total hip arthroplasty (THA). Although overall medium- and long-term results are considered good,1–4 McMinn et al.5 found the revision rate of HRA to be higher than that of traditional cemented THA. Especially, the increased risk of early failures is a concern.6–10 Generally, revision of a HRA is regarded as an easier operation than THA revision. However, high re-revision rates (10% at 2 years,11 11% at 5 years,12 up to 20% at 5 years after isolated acetabular revision13) demonstrate that the problem of early HRA failure is not solved yet. Generally, age, female gender, implant design, metallurgy and geometry of both components as well as the orientation of the acetabular component are thought to be the most important contributors to failure in HRA.11,13,14 However, besides the presence of pores or defects in the cured cement,15,16 femoral neck
fractures and necrosis of the reamed femoral head inside the resurfacing component have also been blamed.17,18 Femoral neck fractures could be due to microfractures caused by forceful femoral component impaction,19,20 damage to the blood supply at the time of surgery21 and/or thermal bone necrosis caused by heat generation inside the reamed femoral head during cement curing. That last mechanism has been studied extensively. Gill et al.22 used thermal probes to record 1
Department of Orthopaedic Surgery and Traumatology, Tergooi Ziekenhuizen, Blaricum, The Netherlands 2 Department of Radiology, University Hospital Brussels, Brussels, Belgium 3 Department of Orthopaedic Surgery and Traumatology, University Hospital Brussels, Brussels, Belgium Corresponding author: Roel de Haan, Department of Orthopaedic Surgery and Traumatology, Tergooi Ziekenhuizen, Rijksstraatweg 1, 1261 AN Blaricum, The Netherlands. Email: [email protected]
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
4
Proc IMechE Part H: J Engineering in Medicine 228(1)
temperatures close to the cement–bone interface during resurfacing arthroplasty, and recorded temperatures able to kill osteocytes. To limit thermal damage, these authors suggest inserting a suction cannula into the lesser trochanter, generous pulsed lavage and early reduction of the freshly resurfaced joint. Janssen et al.23 calculated, based on finite element analysis (FEA), that thermal bone necrosis can be significant when a cementing technique that pressurizes large amounts of cement into the reamed femoral head is combined with bone cements that generate much heat during curing. Overall, women are at higher risk of HRA revision compared to men.24–26 However, when adjusted for femoral component size, the difference between men and women becomes less obvious.13 This suggests that implant size could be the major cause of early revision in women.27 Moreover, Corten and MacDonald12 found an inverse relationship between the size of the femoral component and the risk of revision for HRA. There is also evidence that independent of the patient’s gender, HRA with a femoral component smaller than 44 mm has a fourfold increase in revision rate compared to those with a femoral component of 55 mm or more.13 Although the cause is unclear, smaller femoral resurfacing implants tend to fail more often than larger implants. We postulate that smaller femoral HRA components are filled with proportionally more cement than larger heads. The increased proportion of bone cement inside the reamed femoral head could favor thermal bone necrosis and implant failure. As such, we investigated the relation between implant size and the quantity of cement that can be pressurized into a reamed femoral head.
Materials and methods Specimens and implantation technique Out of 44 femoral heads retrieved during primary conventional THA, we selected 19 specimens that could be fitted with a ReCap (Biomet, Warsaw, IN, USA) resurfacing component. The selected femoral heads were retrieved from 11 male and 8 female patients, with an average age of 71.9 years (standard deviation (SD): 8.3 years, range: 55–86 years). These patients were not selected for HRA because of their age and the surgeon’s preference to use a THA. The femoral heads were chosen so that they could be fitted with an even resurfacing implant size ranging from 40 to 58 mm. For sizes 40–56 mm, two heads were available, whereas for size 58 mm only one head could be found. To avoid metallic X-ray scatter artifacts during computed tomography (CT) scanning, we used glass-filled nylon replicas (ARRK Product Development Group, Gloucester, UK) that were produced based on threedimensional computer-aided design (CAD) models of original ReCap femoral implants (Figure 1).
Figure 1. Polymeric replica of a ReCap hip resurfacing implant.
The selected femoral heads were stored at 274 °C until they could be machined with the original ReCap instruments according to the recommendations of the manufacturer. All femoral heads were provided with anchoring holes drilled with the original 3.0 mm ReCap drill bit. The number of anchoring holes ranged from 5 for the smallest heads (40 mm implant) to 12 for the largest head (58 mm implant). The number of holes was proportional to the area of the hemisphere of the reamed femoral head and represented on average 0.48% (SD: 0.02%, range: 0.46%–0.51%) of that area (Table 1). Prior to implantation, all reamed femoral heads were stored in water at 37.0 °C and cleaned with 2.0 L of water at room temperature (21.0 °C) and pulse lavage. An amount of 40 g of low-viscosity cement (Refobacin Bone Cement LV; Biomet Orthopaedics Switzerland, Ried b. Kerzers, Switzerland) was vacuum-mixed with a mixing and pressuring system (Optivac; Biomet Cementing Technologies, Sjo¨bo, Sweden) for 30 s at 0.1–0.3 bar (10–30 kPa). Within the timeframe between 1.0 and 2.5 min after initiating the mixing, exactly half the available volume inside each implant was filled with cement. The volume available inside each implant was calculated based on the weight increase when filled with water (Table 1). The mass of cement required to fill half the volume was calculated based on a bone cement density of 1.17 g/mL28 and was measured on an electronic scale (On Balance Trueweight-2 pocket scale 600 3 0.1 g). Once filled with cement, implants were rotated to cover their entire inner surface with cement, and 2.5 min after starting cement mixing, they were pressed manually onto the prepared femoral head until seated. The implants were pressurized during curing. All cement that was pressed out of the implant during seating was removed.
CT scanner and measurements After cement curing, all specimens were scanned with a Somatom Sensation 16 (Siemens, Erlangen, Germany)
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
de Haan et al.
5
Table 1. The number of anchoring holes drilled in each ReCap implant size, corresponding percentage of the area of the reamed femoral head, volume inside each ReCap implant size and mass of cement needed to fill half that volume. Implant size (mm) Number of anchoring holes Percentage of surface of hemisphere (%) Volume inside implant (mL) Mass of cement to fill half the implant (g)
40
42
44
46
48
50
52
54
56
58
5 0.47 17.3 10.1
6 0.51 20.5 12.0
6 0.45 24.0 14.0
7 0.48 27.9 16.3
8 0.50 32.8 19.2
8 0.45 36.8 21.5
9 0.47 42.7 25.0
10 0.48 47.4 27.7
11 0.49 54.0 31.6
12 0.49 61.0 35.7
Table 2. Correlation between cement mantle characteristics and femoral head parameters (Pearson’s correlation coefficient (p value)).
Cement thickness Filling index Implant size Bone density
Cement thickness
Filling index
Implant size
Bone density
X 0.889 (\ 0.001) 20.119 (0.628) 20.676 (0.001)
0.889 (\ 0.001) X 20.508 (0.026) 20.568 (0.011)
20.119 (0.628) 20.508 (0.026) X 0.003 (0.991)
20.676 (0.001) 20.568 (0.011) 0.003 (0.991) X
CT scanner according to a standardized scan protocol.15 The resurfaced femoral heads were scanned from the top of the dome to the base of the resurfacing head along their longitudinal axis with following settings: scan collimation, 6 3 0.75 mm; reconstructed slice thickness, 1.0 mm; reconstructed slice interval, 0.5 mm; reconstruction kernel, U80u very sharp; tube potential, 140 kVp; tube current, 200 mA; exposure time, 1 s and field of view, 75 mm. An adapted version of validated segmentation software15,29 developed in MATLAB 6.5 (The MathWorks, Natick, MA, USA) was used to analyze the CT data. The number of images per specimen varied between 41 and 65 depending on implant size (mean: 53.3, SD: 7.8). Four different elements were identified based on their CT values, expressed in Hounsfield units (HU): the polymeric stem and the outer shell of the implant, the cement-free cancellous bone and the cement mantle including cement pressurized into the cancellous bone. The cement thickness, consisting of cement pressurized into cancellous bone and pure cement, was measured in 72 segments of 5° within each CT scan image as described previously.15,29,30 The mean cement mantle thickness per specimen was calculated as the average of all cement thickness measurements within each specimen. The ‘‘filling index’’ of the cement was defined as the fraction cement volume of the total volume within each implant.15 The ‘‘bone density’’ of a specimen was defined as the average CT value of the cement-free cancellous bone (expressed in HU). For each specimen, one bone density measurement was performed in five different regions and those measurements were averaged. The average number of pixel per measurement area was 4449 (SD: 2058 pixels).
Statistical analysis Correlations between implant size, the cement mantle thickness, the cement filling index as well as the bone density were evaluated with Pearson’s correlation coefficient. To assess the contribution of bone density and implant size on the cement mantle thickness and the cement filling index, we used linear regression. Statistical analysis was performed in MATLAB 6.5 (The MathWorks) and SPSS 19 (SPSS Inc., Chicago, IL, USA); p values \0.05 were considered to be significant.
Results The overall average thickness of the cement mantle (pure cement + cement pressurized into the cancellous bone) was 2.63 mm (SD: 0.86 mm, range: 1.65–4.60 mm). No significant correlation was found between cement mantle thickness and implant size. However, there was a significant negative correlation between cement mantle thickness and bone density (Table 2). Specimen with a low bone density had thicker cement mantles regardless of implant size (Figure 2). On average, cement filled 36.65% of the volume within the femoral resurfacing implants, but large variations did exist between specimens (SD: 10.81%, range: 21.52%– 57.60%). There was a significant negative correlation between the filling index and implant size and between filling index and bone density (Table 2). Smaller implants and specimens with a lower bone density contained proportionally more cement than larger implants (Figure 2). The average bone density was 299 HU (SD: 161 HU, range: 247 to 569 HU). There was no significant correlation between bone density and implant size (Table 2 and Figure 2).
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
6
Proc IMechE Part H: J Engineering in Medicine 228(1)
Figure 2. (a) Bone density versus cement mantle thickness with a trend line, (b) size of the femoral components versus filling index with a trend line, (c) bone density versus filling index with a trend line, (d) size of the femoral components versus bone density with a trend line, (e) filling index versus cement mantle thickness with a trend line and (f) size of the femoral components versus cement mantle thickness with a trend line. HU: Hounsfield units.
Linear regression suggested that cement thickness was strongly influenced by bone density but not by implant size. On the contrary, both implant size and bone density had a significant impact on the filling index (p \ 0.001; R2 = 0.58) (Table 3). The influence of bone density was somewhat stronger than that of the implant size as reflected by the effect size (partial h2).
Discussion This in vitro experimental study investigated the impact of the average bone density and implant size on the
cement mantle thickness and cement filling index within a resurfacing head implanted using a ‘‘low-viscosity cement filling technique.’’ In these circumstances, femoral heads with a lower average bone density presented thicker cement mantles and contained a larger proportion of cement. On the contrary, implant size had no significant impact on the cement mantle thickness, but smaller implants contained proportionally more cement than larger ones. The effect of bone density on cement penetration into the cancellous bone seems obvious. The bone density in this study was measured by averaging the CT
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
de Haan et al.
7
Table 3. Impact of implant size and bone density on the cement thickness and cement filling index according to a linear regression model. Predictor Cement thickness Filling index
Size Bone density Size Bone density
Regression coefficient 20.018 20.004 20.969 20.038
95% CI 20.076, 20.041 20.006, 20.002 21.626, 20.312 20.061, 20.015
p value
Partial h2
0.529 0.002 0.006 0.003
0.025 0.936 0.379 0.433
CI: confidence interval.
Figure 3. The same cement mantle thickness within an implant (gray area) will result in a smaller filling index when the implant is large (A) than when it is small (B).
values of the cement-free cancellous bone in five areas per specimen. As such, it reflects the degree of osteoporosis (proportion of trabeculae within the cancellous bone) and also the nature of the material between the bone trabeculae, that is, air, fat, blood or saline used during pressure lavage of the femoral heads. Less bone trabeculae and a larger proportion of air between these trabeculae are expected to facilitate cement penetration into the cancellous bone. However, a previous experimental study did not find that relation even when using the same implantation technique.15 This could be explained by the use of more homogenous, extensively cleaned, bovine cancellous bone (density: 2446 HU, SD: 99 HU, range: 2665 to 2194 HU) compared to the human bone samples used in this study. In fact, the specimens used in this study might match the bone quality encountered in clinical practice much better. The fact that smaller implants, independent from their bone density, contained proportionally more cement than larger implants is more difficult to explain. From this and previous studies, we know that cement penetration depth (cement thickness) depends on the bone density and the cementing technique.15,31 However, we found that using the same ‘‘cement filling technique,’’ the average cement mantle thickness was independent from the implant size. As such, the proportion of cement within a smaller reamed femoral head will be larger than that of a larger femoral head. Indeed, in a simplified model where the reamed femoral
head is represented as a hemisphere, a shell of that hemisphere with a fixed thickness will represent a larger fraction of the total volume when the hemisphere has a smaller diameter compared to when it has a larger diameter (Figure 3). Based on that simplified model, we found a significant correlation (Pearson’s correlation coefficient: 0.505, p = 0.027) between the filling indices that were measured and those calculated for an average cement mantle of 2.63 mm (Figure 4). The proportion of cement filling the inside of the resurfacing implant is of importance because large quantities of cement could cause thermal bone necrosis during the exothermal polymerization reaction. Such bone necrosis has been demonstrated in failed resurfacing implants22,32–34 and has been simulated with FEA models.23 To reduce the amount of thermal bone necrosis, several proposals have been formulated. Gill et al.22 suggest the use of a suction cannula into the lesser trochanter during cementing, early reduction of the resurfaced head into the acetabular component filled with cooled saline solution and extensive pressure lavage. Scheerlinck et al.15 suggested precooling the reamed femoral head with cooled pressure lavage fluid prior to cementing. Based on an FEA model by Janssen et al.,23 precooling of the femoral head with fluid at 10 °C could reduce the amount of thermal necrosis by a factor of 5. However, bone cement is sensitive to differences in temperature at interfacial surfaces during curing. This could lead to pore formation at the cooler surface35
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
8
Proc IMechE Part H: J Engineering in Medicine 228(1)
Figure 4. Filling index versus femoral implant size. The trend line between the measured filling indices corresponds almost perfectly with the calculated filling indices based on a fixed cement mantle thickness of 3.00 mm within a hemisphere. The dotted line represents the calculated filling index for the average cement mantle of our specimens (2.63 mm).
carrying with it the risk of failure due to voids and defects in the cement mantle at critical locations. Based on this study and previous work investigating cementing techniques,15 small resurfacing heads implanted with a ‘‘cement filling technique’’ on a reamed femoral head with a low bone density (wellcleaned, osteoporotic bone) are the most at risk of overfilling the femoral head with cement. In our study, the combination of small sizes and low bone density resulted in a filling index above 50% and an average cement mantle thickness above 3.5 mm in three cases (implant size: 40, 44 and 50 mm; average bone density: 282, 247 and 12 HU, respectively). It was possible to obtain a filling index above 50% despite the fact that we filled the implants only halfway because the cement mantle contained not only cement but also trabecular bone. The combination of small femoral heads and osteoporotic bone is most likely to be found in elderly women, a population in which hip resurfacing is known to be problematic.13,27 The effect of implant size should not be underestimated as it could, on its own, explain some of the differences in resurfacing survival between men and women in the Australian hip registry.13,27 As for all in vitro experimental studies, several limitations apply to this work. First, the femoral heads we used were retrieved from a general THA population. That population does not necessarily match to the ideal resurfacing candidates. However, we used human femoral heads presenting major osteoarthritis and some degree of deformity that would typically be encountered during hip resurfacing. Second, back bleeding during cementing was not simulated. However, during in vivo hip resurfacing, back bleeding is generally not a major issue. Indeed, it is often controlled by inserting a
suction cannula into the lesser trochanter, and femoral head vascularity is often compromised by the approach and the dislocated hip position during cementing. Third, in this study, water was used for lavage and storage instead of saline. This can affect the properties of the bone as well as the condition of exposed proteins on the surface of the bone. This in turn may influence the cement interfacial bonding, but we do not believe that it could influence the cement penetration into the trabecular bone during implantation. Fourth, in order to allow CT scanning without metallic artifacts, we used plastic replicas of original implants. These replicas could have dimensional mismatches compared to the original implants or could theoretically deform during cementing. However, the manufacturer of the original implants produced the replicas based on accurate CAD models. Moreover, no force was needed to insert them on the reamed femoral heads as they presented an inner radial clearance of 0.5–0.6 mm.36,37 Besides the dimensional mismatches of the plastic replicas, the plastic material is likely to have an effect on the heat dissipation of the cement during curing due to the large difference in thermal conductivity and specific heat of the plastic models we used, compared to cobalt-chrome implants used in clinical practice. The reduced ability of plastic materials to conduct heat away from the exothermic polymerization of cement will increase the curing speed and the viscosity of the cement. This could reduce the cement penetration into the bone. Nevertheless, we do not expect the use of plastic replicas to affect the cement mantles significantly, at most this could accelerate cement curing close to the implant in a region that is filled with cement anyway. Accelerating the curing speed in that region is unlikely to influence the overall
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
de Haan et al.
9
cement penetration in the specimen. Finally, reported results only apply to the ReCap resurfacing heads implanted with a ‘‘low-viscosity cement filling technique combined with anchoring holes.’’ Indeed, resurfacing heads with different inner clearances and different inner morphologies could result in very different cement mantles. Also, discarding anchoring holes and the use of different cementing techniques could have a major impact on cement penetration into the femoral head. As such, the conclusions drawn from this study cannot readily be extrapolated to other implants or other implantation techniques. In conclusion, small femoral resurfacing implants with the same cement mantle thickness contain proportionally more cement. When using a ‘‘low-viscosity cement filling technique combined with anchoring holes,’’ small resurfacing implants have the potential to be filled with larger quantities of cement. The risk increases when the reamed femoral head contains lowdensity bone, that is, osteoporotic bone that has been extensively cleaned. The large proportion of cement within the resurfacing head has the potential to generate thermal bone necrosis and implant failure. This further reinforces that patients with osteoporotic bone are a contraindication for HRA. Consequently, when considering hip resurfacing in patients with a small femoral head and/or osteoporotic bone, extra care should be taken to avoid thermal bone necrosis during cementing, and alternative cementing techniques or even cementless implants could be considered. Acknowledgements The authors would like to thank Ronald Buyl, Professor of Statistics, Department of Biostatistics and Information Technology, Faculty of Medicine and Pharmacy, Vrije Universiteit Brussel, Brussels, Belgium, for his advice on statistical issues. The Department of Radiology of the Universitair Ziekenhuis Brussel, Brussels, Belgium, provided the CT imaging data of the specimens, and Biomet Deutschland, Dieburg, Germany, provided the polymeric implant replicas, bone cement and cement mixing systems. The study protocol of this work was approved by the ethical committee of the Universitair Ziekenhuis Brussel, Brussels, Belgium. Declaration of conflicting interests The authors declare that there is no conflict of interest. Funding This study was funded by the Department of Orthopedic Surgery and Traumatology of the Universitair Ziekenhuis Brussel, Brussels, Belgium. References 1. De Smet KA. Belgium experience with metal-on-metal surface arthroplasty. Orthop Clin North Am 2005; 36(2): 203–213, ix.
2. Steffen RT, Pandit HP, Palan J, et al. The five-year results of the Birmingham hip resurfacing arthroplasty: an independent series. J Bone Joint Surg Br 2008; 90(4): 436–441. 3. Van Der Bracht H, Vander Eecken S, Vyncke D, et al. Clinical and functional outcome of the Birmingham hip resurfacing. Acta Orthop Belg 2011; 77(6): 771–776. 4. Gross TP and Liu F. Hip resurfacing with the Biomet Hybrid ReCap-Magnum system: 7-year results. J Arthroplasty 2012; 27(9): 1683–1689. 5. McMinn DJ, Snell KI, Daniel J, et al. Mortality and implant revision rates of hip arthroplasty in patients with osteoarthritis: registry based cohort study. BMJ 2012; 344: e3319. 6. Morlock MM, Bishop N, Zustin J, et al. Modes of implant failure after hip resurfacing: morphological and wear analysis of 267 retrieval specimens. J Bone Joint Surg Am 2008; 90(Suppl. 3): 89–95. 7. Corten K, Ganz R, Simon JP, et al. Hip resurfacing arthroplasty: current status and future perspectives. Eur Cell Mater 2011; 21: 243–258. 8. Della Valle CJ, Nunley RM, Raterman SJ, et al. Initial American experience with hip resurfacing following FDA approval. Clin Orthop Relat Res 2009; 467(1): 72–78. 9. Berend KR, Lombardi AV Jr, Adams JB, et al. Unsatisfactory surgical learning curve with hip resurfacing. J Bone Joint Surg Am 2011; 93(Suppl. 2): 89–92. 10. Van der Weegen W, Hoekstra HJ, Sijbesma T, et al. Hip resurfacing in a district general hospital: 6-year clinical results using the ReCap hip resurfacing system. BMC Musculoskelet Disord 2012; 13: 247. 11. De Haan R, Campbell PA, Su EP, et al. Revision of metal-on-metal resurfacing arthroplasty of the hip: the influence of malpositioning of the components. J Bone Joint Surg Br 2008; 90(9): 1158–1163. 12. Corten K and MacDonald SJ. Hip resurfacing data from national joint registries: what do they tell us? What do they not tell us? Clin Orthop Relat Res 2010; 468(2): 351–357. 13. Australian Orthopaedic Association National Joint Replacement Registry. Annual reports, https://aoanjrr. dmac.adelaide.edu.au/annual-reports (accessed 11 February 2013). 14. Shimmin AJ, Walter WL and Esposito C. The influence of the size of the component on the outcome of resurfacing arthroplasty of the hip: a review of the literature. J Bone Joint Surg Br 2010; 92(4): 469–476. 15. Scheerlinck T, Delport H and Kiewitt T. Influence of the cementing technique on the cement mantle in hip resurfacing: an in vitro computed tomography scanbased analysis. J Bone Joint Surg Am 2010; 92(2): 375–387. 16. Bitsch RG, Jager S, Lurssen M, et al. The influence of cementing technique in hip resurfacing arthroplasty on the initial stability of the femoral component. Int Orthop 2011; 35(12): 1759–1765. 17. Falez F, Favetti F, Casella F, et al. Results of hip resurfacing. Int Orthop 2011; 35(2): 239–243. 18. Carrothers AD, Gilbert RE, Jaiswal A, et al. Birmingham hip resurfacing: the prevalence of failure. J Bone Joint Surg Br 2010; 92(10): 1344–1350. 19. Schlegel UJ, Rothstock S, Siewe J, et al. Does impaction matter in hip resurfacing? A cadaveric study. J Arthroplasty 2011; 26(2): 296–302.
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
10
Proc IMechE Part H: J Engineering in Medicine 228(1)
20. Campbell P, Beaule PE, Ebramzadeh E, et al. The John Charnley Award: a study of implant failure in metal-onmetal surface arthroplasties. Clin Orthop Relat Res 2006; 453: 35–46. 21. Steffen RT, Athanasou NA, Gill HS, et al. Avascular necrosis associated with fracture of the femoral neck after hip resurfacing: histological assessment of femoral bone from retrieval specimens. J Bone Joint Surg Br 2010; 92(6): 787–793. 22. Gill HS, Campbell PA, Murray DW, et al. Reduction of the potential for thermal damage during hip resurfacing. J Bone Joint Surg Br 2007; 89(1): 16–20. 23. Janssen D, Srinivasan P, Scheerlinck T, et al. Effect of cementing technique and cement type on thermal necrosis in hip resurfacing arthroplasty—a numerical study. J Orthop Res 2012; 30(3): 364–370. 24. Della Valle CJ, Nunley RM and Barrack RL. When is the right time to resurface? Orthopedics 2008; 31(12 Suppl. 2), http://www.orthosupersite.com/view.asp?rID=37185 (accessed 11 February 2013). 25. Jameson SS, Langton DJ, Natu S, et al. The influence of age and sex on early clinical results after hip resurfacing: an independent center analysis. J Arthroplasty 2008; 23(6 Suppl. 1): 50–55. 26. Nunley RM, Della Valle CJ and Barrack RL. Is patient selection important for hip resurfacing? Clin Orthop Relat Res 2009; 467(1): 56–65. 27. McBryde CW, Theivendran K, Thomas AM, et al. The influence of head size and sex on the outcome of Birmingham hip resurfacing. J Bone Joint Surg Am 2010; 92(1): 105–112. 28. Skinner HB and Murray WR. Variations in the density of bone cement after centrifugation. Clin Orthop Relat Res 1986; 207: 263–269.
29. Scheerlinck T, de Mey J and Deklerck R. In vitro analysis of the cement mantle of femoral hip implants: development and validation of a CT-scan based measurement tool. J Orthop Res 2005; 23(4): 698–704. 30. Scheerlinck T, de Mey J, Deklerck R, et al. CT analysis of defects of the cement mantle and alignment of the stem: in vitro comparison of Charnley-Kerboul femoral hip implants inserted line-to-line and undersized in paired femora. J Bone Joint Surg Br 2006; 88(1): 19–25. 31. Chandler M, Kowalski RS, Watkins ND, et al. Cementing techniques in hip resurfacing. Proc IMechE, Part H: J Engineering in Medicine 2006; 220(2): 321–331. 32. Little CP, Ruiz AL, Harding IJ, et al. Osteonecrosis in retrieved femoral heads after failed resurfacing arthroplasty of the hip. J Bone Joint Surg Br 2005; 87(3): 320–323. 33. Morlock MM, Bishop N, Ruther W, et al. Biomechanical, morphological, and histological analysis of early failures in hip resurfacing arthroplasty. Proc IMechE, Part H: J Engineering in Medicine 2006; 220(2): 333–344. 34. Zustin J, Sauter G, Morlock MM, et al. Association of osteonecrosis and failure of hip resurfacing arthroplasty. Clin Orthop Relat Res 2009; 468(3): 756–761. 35. Pelletier MH, Lau AC, Smitham PJ, et al. Pore distribution and material properties of bone cement cured at different temperatures. Acta Biomater 2010; 6(3): 886–891. 36. Beaule PE, Matar WY, Poitras P, et al. 2008 Otto Aufranc Award: component design and technique affect cement penetration in hip resurfacing. Clin Orthop Relat Res 2009; 467(1): 84–93. 37. Grigoris P, Roberts P, Panousis K, et al. Hip resurfacing arthroplasty: the evolution of contemporary designs. Proc IMechE, Part H: J Engineering in Medicine 2006; 220(2): 95–105.
Downloaded from pih.sagepub.com at UNIV OF OKLAHOMA LIBRARIES on February 5, 2015
