- Email: [email protected]
A comparison of 2 modern femoral cementing techniques
The Journal of Arthroplasty Vol. 15 No. 4 2000
A Comparison of 2 Modern Femoral Cementing Techniques Analysis by Cement–Bone Interface Pressure Measurements, Computerized Image Analysis, and Static Mechanical Testing A. D. Reading, FRCS (Tr and Orth),* A. W. McCaskie, MD, FRCS(Orth),† M. R. Barnes, BSc,‡ and P. J. Gregg, MD, FRCS†
Abstract: Modern cementing techniques aim to improve microinterlock and to reduce aseptic loosening. The Norwegian and Swedish Arthroplasty Registers have shown an increased risk of revision using reduced-viscosity cement. We have compared 2 modern cementing techniques using retrograde insertion of normalviscosity and reduced-viscosity cements. Laboratory-simulated arthroplasty was performed in paired human femora. Performance was evaluated by measuring pressures generated during cementation, cement penetration, and shear strength of the prosthesis–cement and bone–cement interfaces. Large differences exist between these 2 modern techniques. Despite no statistical differences between the pressure measurements with the 2 techniques, greater penetration of reduced-viscosity cement was found proximally, with a trend toward increased penetration of the more viscous cement distally. Areas of greater cement penetration with reduced-viscosity cement proximally produced higher values of ultimate shear strength. Both techniques showed a progressive increase in the shear strength as the level of the section progressed toward the tip of the prosthesis. There is a trend with both techniques for the distal fixation to be stronger. Key words: cement, viscosity, modern, hip, arthroplasty.
Aseptic loosening is an important long-term complication of total hip arthroplasty [1] and is, in part, responsible for the increasing numbers of patients
requiring revision surgery [2]. It remains poorly understood but involves biologic and mechanical processes [3–9]. Cementation is an important factor because it is under the direct control of the surgeon. Despite this importance, 2 surveys of consultants have shown a lack of consensus [10,11]. In part, this lack of consensus can be explained by a lack of clinical and biologic evidence. Modern cementing techniques [12] aim to improve microinterlock between the cement and cancellous bone. Cement is encouraged to penetrate to the region of the strongest cancellous bone within 3 mm of the corticocancellous junction [13]. There are several aspects of modern cementing techniques that when combined achieve this goal: i) improved surface preparation [14], ii) canal restriction, iii) retrograde insertion of cement [15], iv) sustained
From the *Department of Orthopaedics, Glasgow Royal Infirmary, Glasgow, Scotland, United Kingdom; †Department of Trauma and Orthopaedic Surgery, School of Surgical Sciences, The Medical School, University of Newcastle, Newcastle upon Tyne, England; and the ‡Department of Sports Medicine, Leicester General Hospital, Leicester. Submitted April 6, 1999; accepted December 10, 1999. Benefits were received in partial support of the research material described in this article. These benefits were as follows: Equipment supplied by various companies, including prostheses (DePuy International), cement (CMW), cementation equipment (Fry Surgical), cannulae (Portex). Reprint requests: A. D. Reading, FRCS (Tr and Orth), Department of Orthopaedics, Glasgow Royal Infirmary, 84 Castle Street, Glasgow, UK. Copyright r 2000 by Churchill Livingstonet doi:10.1054/arth.2000.5266 0883-5403/00/1504-0013$10.00/0
479
480 The Journal of Arthroplasty Vol. 15 No. 4 June 2000 pressurization [16,17], v) reduced viscosity of cement [25], and vi) correct timing of prosthesis insertion [18–20]. The flow of cement into cancellous bone is proportional to the pressure gradient. Several studies have confirmed that the higher pressure applied to the cement, the greater the penetration of cement into the cancellous bone [13,26]. Bleeding at the endosteal surface is an important factor in the creation of the cement–bone interface [14,21]. The bleeding pressure has been measured as high as 36 cm saline (27 mmHg) [22,23]. Experiments have shown that a column of blood 15 to 30 cm high can displace reduced-viscosity cement 7 minutes after the start of mixing [24]. For normal-viscosity cement, no displacement occurred after 3 minutes. Pressure applied to the cement above the simulated bleeding pressure caused the cement to flow in the opposite direction. Experiments in ox femora showed that the shear strength of the cement–bone interface for normal-viscosity cement was not affected by simulated bleeding; however, low-viscosity cement had significantly weaker interfaces in 50% of the samples [14]. Penetration depth of the cement was not affected. Many surgeons prefer to use normal-viscosity cement (53–74% [10,11]) for the insertion of the femoral component because it is easier to handle, and the timing is less critical, with no need to delay prosthetic entry after cement insertion. The Norwegian Arthroplasty Register reviewed 8,579 primary Charnley arthroplasties for primary osteoarthrosis in 7,922 patients with respect to the type of cement used [27]. The rate of revision of femoral components inserted with reduced-viscosity cement was 2.5 times greater than those inserted with normalviscosity cement. The Swedish Arthroplasty Register confirmed that the worst results were noted with low-viscosity cement [28]. The present study uses a modern cementing technique with retrograde insertion to compare normal-viscosity and reducedviscosity cements. The cement–bone interface pressures were measured during cementation and prosthesis insertion, cement penetration to cancellous bone was measured by computerized image analysis, and the shear strength of the prosthesis– cement and cement–bone interfaces was measured by static mechanical testing.
Leeds UK). Standard straight and curved reamers were used to prepare the femoral canal to accept a flanged 40 Charnley prosthesis. The canal was prepared using pressurized lavage (500 mL of saline) and power brushing. A correctly sized Elite (DePuy, Leeds United Kingdom) cement restrictor was inserted to a level 20 mm below the anticipated tip of the prosthesis (13 cm below the calcar). The inner surface of the prepared medullary canal was coated with 3 mL of human blood (to re-create the physical presence of blood, not the effect of backbleeding pressure). At the center of each of Gruen’s zones 1 through 7 (except 4) [29], a 3.5-mm drill hole was made into which the tapered end of a manometer line (Portex, Hythe, UK) was inserted and secured by interference fit (Fig. 1). The tip of the manometer line remained within the cancellous bone and was of a diameter sufficient to prevent cement blocking the catheter. The other end was connected to a pressure transducer (Elcomatic EM751; Elcomatic Ltd., Glasgow, Scotland, UK) and the line run through with saline to exclude all air. The 6 pressure lines were connected to a 6-channel chart recorder to allow simultaneous recording of the pressures during simulated arthroplasty. The pressure transducer had a limited range of activity (linearity ⫾ 0.1% of 4,000 mmHg/damaged pressure 8,000 mmHg), and an arbitrary limit of 5,000 mmHg (667 kPa) was set for data interpretation. Any reading above this level was assigned this value. Reduced-viscosity and normal-viscosity cements (CMW 3 and CMW 1, DePuy, Leeds, United Kingdom)
Materials and Methods Pairs of fresh frozen femora from 6 human donors were used in this study. Femora were defrosted before performing simulated arthroplasty using the standard Charnley system (DePuy International Ltd,
Fig. 1. Positioning of the pressure transducers at the center of Gruen zones.
Comparison of Femoral Cementing Techniques ●
were mixed in the same manner by using a syringe vacuum mixing system (Mitab, Fry Surgical International Ltd, UK). The normal-viscosity cement was extruded into the femur, retrograde with a caulking gun, 2 minutes after the start of mixing. A complete seal of the femoral opening was achieved with a thumb and pressure exerted on the cement column. This digital pressurization continued until the prosthesis (a new flanged 40 Charnley prosthesis) was inserted at 2 minutes, 45 seconds. For the reducedviscosity cement, extrusion into the femur, retrograde with a caulking gun, was performed at 2 minutes, 15 seconds, followed by digital pressurization until prosthesis insertion at between 3 minutes, 45 seconds, and 4 minutes. The timings used for reduced-viscosity cement were as per manufacturer’s instructions, whereas for normal-viscosity cement they were modified from manufacturer’s instructions to allow for extrusion of the more viscous cement from the gun system. All cement was mixed by the same person and the cementation of the femora performed by 1 person to reduce errors resulting from differing operative techniques. The pressure recordings were divided into 3 time periods: the first during cement insertion, the second during digital pressurization, and the third during prosthesis insertion. The duration of each period was measured assuming the start time to be the initial pressure rise and the end point to be the first pressure rise of the next period. The end of the final period (prosthesis insertion) was taken to be the point at which the pressure recording returned to 0. During each time period, the maximum peak pressure and the area under the curve were recorded. The mean pressure for that period could then be calculated (area under curve divided by the duration). The time that the pressure recorded was less than 100 mmHg was also measured. The data were not normally distributed and were analyzed using the Wilcoxon signed rank test, against a null hypothesis of no difference. Significance was taken at P ⱕ .05; a Bonferroni correction was made for the multiple testing of pressure measurements. The 95% confidence intervals (CIs) for the difference between the 2 techniques were calculated. After simulated arthroplasty, the femora were sectioned at 90° to the long axis of the bone (Fig. 2). This sectioning was performed with a high-precision wafering system (Isomet 2000, Buehler, United Kingdom). The first cut was performed at the cut edge of the calcar, the bone was then advanced 7 mm, and the second cut was performed. This process produced 12 parallel-sided samples 6.4 mm thick per femur. A high-resolution faxitron radiograph was taken of each section.
Reading et al.
481
Fig. 2. The sections of femora and those sections used in each study.
Computerized Image Analysis Computerized image analysis was initially performed on each of the 144 fresh sections (Omnimet 3, Buehler, United Kingdom). The system determined the area of the cement mantle, prosthesis, bone, and whole section (Figs. 3 and 4) as well as the length of the perimeter of the prosthesis and cement–bone interface. Computerized image analysis of fresh sections evaluated the total cement mantle. To subdivide the mantle further into reamed and penetrated portions, faxitrons were used (Fig. 5). The data obtained were not normally distributed. The differences between the 2 techniques were compared within each pair and at each level. The differences were analyzed using the Wilcoxon signed rank test, against a null hypothesis of no difference. Significance was taken at a level of P ⱕ .05. The 95% CIs for the difference between the 2 techniques were calculated.
Fig. 3. Image captured by the image analysis equipment after enhancement to define the edges clearly.
482 The Journal of Arthroplasty Vol. 15 No. 4 June 2000 Static Mechanical Tests The testing rig consisted of a servohydraulic ram with a platform attached to it and a 10-kN load cell. A temperature-regulated water tank was attached to the platform to allow the samples to be submerged in normal saline at 37°C ⫾ 0.5°C. Results of force and displacement were recorded on a personal computer. Reverse push-out mechanical tests were performed; the force was applied in a caudal-tocranial direction to prevent destruction of the specimens [29,30]. Prosthesis–Cement Interface
Fig. 4. Image after manipulation on which calculations were performed. Area of prosthesis, cement, and bone. Perimeter of prosthesis–cement and cement–bone interfaces.
Four sections were used for this experiment—1, 4, 7, and 10 (Fig. 2). To evaluate the prosthesis– cement interface, the specimen was placed on a platform that supported the cement mantle. The irregular shape of the specimens meant that special inserts had to be made to fit into the platform to support them; these were molded from bone– cement. A small punch was then attached to the load cell so that only the prosthesis was loaded. The specimens to be tested were then hydrated before being placed on the platform and submerged in saline at 37°C. Specimens were placed so that the direction of the taper of the prosthesis was reversed. The platform was then programmed to rise at a constant displacement of 3.93 mm/min. The force produced on the load cell was recorded at a sampling rate of 20 Hz until failure occurred and the prosthesis was pushed out of the cement mantle. Cement–Bone Interface
Fig. 5. (A) Image captured by the image analysis equipment of the faxitron radiograph shows the penetrated and reamed portions of the cement mantle (magnified). (B) Line diagram of image in (A).
Eight sections were used for this part of the experiment—1, 2, 4, 5, 7, 8, 10, and 11 (Fig. 2). The test rig for the cement–bone interface was similar to that used for the prosthesis–cement interface (Fig. 6). The only change was in the specimen support ring, which was modified to provide support for the cortical bone, rather than the cement. A problem encountered in other studies performing similar tests was specimen destruction. To overcome this problem, the specimens were placed in a mold, and a ring of polymethyl methacrylate casting resin was placed around the outside to provide support, reinforced with a nylon cerclage band. To apply force evenly to the cement–bone interface, a custom-made mushroom of bone–cement was made for each specimen. This bone–cement sat in the defect left by the prosthesis and spread out to the edge of the cement–bone interface. The punch attached to the load cell was also larger to provide an even force over the whole specimen. The speci-
Comparison of Femoral Cementing Techniques ●
Fig. 6. The test apparatus for the cement–bone interface testing.
mens were then submerged as before in the saline bath, and the force displacement curve was recorded during the failure cycle. A force displacement curve was plotted for each static mechanical test. The force at which failure occurred was taken from these graphs. Failure was said to have occurred once the value for recorded force decreased with a continuing increase in displacement. The height of the specimen and the perimeter of the prosthesis–cement or cement–bone interface, obtained by image analysis, multiplied together produced a value for the area of the prosthesis–cement or cement–bone interface. Dividing the force at which failure occurred by the area of the interface gave the ultimate shear strength of the interface. The differences between the 2 paired femora and the 2 techniques were then calculated. The data were nonparametric, and the differences were analyzed using the Wilcoxon signed rank test against a null hypothesis of no difference; significance was taken at the level P ⱕ .05. The 95% CIs for the difference between the 2 techniques were calculated.
Results Cement–Bone Interface Pressure Studies Considering cement insertion first, for the mean and the maximum pressures recorded, there were no statistically significant results between the 2 cementing techniques. Both cement types produced increasingly higher pressures the further distally the readings were taken along the femoral canal (me-
Reading et al.
483
dian value of arithmetic mean pressure for normalviscosity cement proximally at transducer, 1 ⫽ 1.02 kPa [interquartile range, 0.11–11.67 kPa], and distally at transducer, 3 ⫽ 19.13 kPa [interquartile range, 13–59 kPa]). If digital pressurization is considered, there were no statistically significant results between the 2 methods. Both cements produced higher pressures proximally and medially (median value normal-viscosity cement transducer, 7 ⫽ 62.94 kPa [interquartile range, 35.19–69.32 kPa]), decreasing distally during this phase (median value normalviscosity cement transducer, 5 ⫽ 48.06 kPa [interquartile range, 29.69–62.97 kPa]). During prosthesis insertion, there were no statistically significant results. The trend for increasing pressures distally was noted for both techniques (median value normalviscosity cement proximally at transducer, 1 ⫽ 37.59 kPa [interquartile range, 31.49–44.81 kPa], and distally at transducer, 3 ⫽ 87.44 kPa [interquartile range, 71.4–121.5 kPa]). Measurements of length of time recorded with the pressure below the mean arterial blood pressure of 13.3 kPa (100 mmHg) showed no statistical differences during the cement insertion phase (Table 1). During digital pressurization, there was a trend for the reduced-viscosity cement to spend a longer time with the pressure less than arterial bleeding pressure. This trend reached statistical significance at transducer sites 2, 3, and 4 (P ⫽ .036). At all transducer sites, the 95% CIs of the difference between techniques excluded 0. During the phase of prosthesis insertion, there was a trend for the normal-viscosity cement to spend a longer time with the pressure ⬍13.3 kPa. This trend reached statistical significance at transducer number 3 (P ⫽ .036). The trend also showed that the laterally placed transducers spent longer at a lower pressure than those placed medially. This situation was confirmed by the fact that the 95% CI difference excluded 0 at transducers 1, 2, 3, and 5. Computerized Image Analysis Image analysis of the femoral sections provided data on the percentage area of each section occupied by cement. The differences between the 2 paired femora and the 2 techniques were then calculated. The proximal 4 levels below the calcar showed a statistically significant trend for the reduced-viscosity cement to produce a larger area of cement per section (median value normal-viscosity cement section 1, 24.8% [range, 21.5–28.6%], and reducedviscosity cement, 33.1% [range, 30.15–41.81] [P ⫽ .036]). Further distally at levels 8 and 11, normal-viscosity cement produced a significantly larger cement mantle (P ⫽ .036).
484 The Journal of Arthroplasty Vol. 15 No. 4 June 2000 Table 1. Descriptive Statistics for the Time (Seconds) the Pressure Recorded Was ⬍13.3 kPa Normal Viscosity
Reduced Viscosity Median
Interquartile Range
Median
Interquartile Range
P Value
Cement Insertion 1 2 3 5 6 7
15.5 12.9 10.6 10.8 12.45 15.5
12.35–21.3 8.5–17.4 5.53–24.55 5.82–24.4 8.12–24.45 9.25–21.6
13.7 11.1 7.65 8.25 9.1 12.5
11.78–16.35 9.7–12.9 7.1–9.45 7.02–9.6 8–11.45 11.2–15.25
.281 .855 .529 .402 .345 .855
Digital Pressurization 1 2 3 5 6 7
11 11.2 10.45 10.4 9.7 7.3
9.5–17 6.65–15.5 5.9–15.35 5.75–15.35 5.25–15.55 3.5–13.45
35 33.7 34.55 34.05 30.05 24.4
30–36 29.25–35.45 27.13–36.35 28.95–36.65 25.38–34.9 18.7–28.25
.059† .100† .036† .036† .036† .100†
Prosthesis Insertion 1 2 3 5 6 7
25 18.6 14 13 12.5 12
17.45–31.2 10.25–36.2 9.5–28.25 9.75–24.5 4–21 10–14
15.55 7.2 10 9 9 8
4.55–25.12 5.6–15.45 6.75–16.25 5–13.5 4–13.5 5–19.5
.059† .100† .059† .036† .295 .584
Transducer
*The mean, interquartile range, and statistical significance for each different time period and cement are shown. †95% confidence interval of the difference excludes 0.
The technique used for preparation of the femora before insertion of the prostheses was standard for all the bones. To ensure that this technique produced consistent results for the size of the reamed medullary canal, calculations were performed on the faxitron radiographs. The results for the area of the reamed cement mantle were analyzed. There were no statistical differences between the paired femora. The differences noted in the areas of cement per section are due to the differences in the amount of cement penetrating the cancellous bone. Static Mechanical Testing The ultimate shear strength results are presented in Table 2. For the prosthesis–cement interface, there were no significant differences at any level. There was a trend toward greater values of shear strength as the level of the section moved distally. For the cement–bone interface, reduced-viscosity cement produced a statistically significant greater value of shear strength at the most proximal level under the calcar (P ⫽ .036). For all other levels of section, apart from level 7, there was a trend for reduced-viscosity cement to have the greater value of shear strength. The 95% CI of the difference excluded 0 at levels 1, 2, and 8. There was a trend for increasing values of the shear strength the further distal the level of the section, with the exception of level 7 for the normal-viscosity cement.
Discussion During cementation of the femoral component, it is considered desirable for the pressures to remain greater than the endosteal bleeding pressure of 36 cm saline [22,23] (27 mmHg) and preferably greater than the mean arterial blood pressure (13.3 kPa or 100 mmHg) to ensure no back-bleeding at the interface. From this study, it can be seen that the mean pressures remained greater than these levels throughout most of the cementation. The exceptions are during cement insertion for normalviscosity and reduced-viscosity cements, in which the most proximal levels recorded pressures less than 10 kPa. During digital pressurization, the reduced-viscosity cement spent a significantly longer time below the mean arterial blood pressure. In the clinical situation, this time could allow interface disruption by back-bleeding, but the extent of this effect cannot be quantified. A possible method of overcoming this problem would be to use a proximal femoral seal to prevent cement escaping proximally during digital pressurization. One weakness of the model used for this study is the underestimation of the dynamic flow of blood from the endosteal surface into the cement. Blood was placed on the endosteal surface before cementation to allow for any effect its physical presence may have, but
Comparison of Femoral Cementing Techniques ●
Reading et al.
485
Table 2. Ultimate Shear Strength (MPa) of the Prosthesis–Cement and Cement–Bone Interfaces at Each Level and for Each Cementing Technique* Normal Viscosity
Reduced Viscosity Median
Interquartile Range
Median
Interquartile Range
P Value
Prosthesis–Cement Interface 1 0.055 4 0.145 7 0.285 10 0.36
0.038–0.07 0.123–0.218 0.218–0.41 0.215–0.762
0.015 0.165 0.3 0.485
0.0–0.06 0.09–0.25 0.198–0.405 0.33–0.668
0.295 0.834 1.0 0.529
Cement–Bone Interface 1 0.4 2 0.489 4 1.05 5 1.128 7 2.28 8 1.189 10 1.595 11 2.196
0.323–0.87 0.437–0.822 0.687–1.553 0.894–1.908 1.51–2.81 0.76–1.992 1.178–2.855 1.531–2.454
0.925 0.74 1.395 1.446 1.545 1.583 1.9 2.091
0.673–1.293 0.539–1.512 1.062–1.602 1.071–3.069 1.08–2.787 1.512–2.022 1.182–2.742 1.035–3.206
0.036† 0.093† 0.295 0.463 0.093† 0.093† 0.834 0.834
Level
*The median, interquartile range, and statistical significance are shown. †95% confidence interval of the difference excludes 0.
there was no pressurized flow into the femoral canal. The mean pressures created during digital pressurization show an interesting pattern. Higher pressures are created under the calcar and on the medial side of the femur, as recommended originally by Charnley [32]. One feature common to both techniques was the high mean and maximum pressures generated during prosthesis insertion. This feature has been reported before [20,31], but the almost identical values for both techniques are notable. This finding suggests that the manufacturers’ timings do provide for a similar viscosity of cement at the time of prosthesis insertion. During prosthesis insertion, pressure was maintained above the mean arterial blood pressure for longer on the medial side of the femur, especially at the calcar. This situation could be related to the design of the prosthesis, with the cobra neck flange and shape of the stem directing the cement to the calcar. The mean and maximum pressures applied during cement insertion and digital pressurization were not significantly different for the 2 types of cement. Image analysis of the femoral sections, however, showed a larger area of cement mantle for the reduced-viscosity cement at the proximal levels. This larger area was a function of viscosity and the availability of cancellous bone. Further distally, there was no significant difference between the sizes of the cement mantles between the cements. The reverse push-out tests, although not physiologic, provided a more accurate comparison of the interface strengths and have previously been re-
ported [29,30,33]. For the prosthesis–cement interface, there were no statistically significant results between the 2 cements. There was a trend with both cement types for the values of the shear strength to increase as the level of the section moved further distally. A similar trend was also evident for the pressures measured during the insertion of the prosthesis. High prosthesis insertion pressures may be associated with increase in the bond strength between the prosthesis and the cement. Static mechanical testing of the cement–bone interface shows that the shear strength is significantly greater for the reduced-viscosity cement proximally. As the levels of section move further distally, there was no significant difference between the 2 cements; however, there is a trend for the reduced-viscosity cement to achieve a higher value. There is 1 exception, at level 7, and this is difficult to explain. There were no differences in the pressures generated during cementation at this level between the 2 techniques. There was increased penetration of normal-viscosity cement noted at that level, although this was not a significant difference. At the proximal level, there was a significant increase in the penetration of and the shear strength at the cement–bone interface for the reducedviscosity cement. This increase suggests that the shear strength of the cement–bone interface is related to degree of penetration of cement achieved into the cancellous bone. The reduced-viscosity cement achieves greater penetration and greater
486 The Journal of Arthroplasty Vol. 15 No. 4 June 2000 shear strength in the areas where the proportion of cancellous bone is highest. The shear strength values again increase as the levels of the sections move further distally. This increase may be explained by examining the specimens. Distally, the cement penetrates through the cancellous bone to the cortical bone; proximally this is not usually the case. Examining the specimens after failure, it can be seen that failure occurs through the cancellous bone just beyond the cement mantle. To improve the mechanical strength of the cement–bone interface proximally, the cement would be required to penetrate to the cortex. Theoretically, this penetration could be encouraged in 1 of 2 ways: i) by the use of a proximal femoral silicone seal and reduced-viscosity cement; ii) by removing more cancellous bone from the proximal femur during preparation of the canal. This second action, however, could potentially reduce bone stock with effects on further revision surgery. These hypotheses require further study before reaching clinical relevance. The results of this study suggest that reducedviscosity cement used correctly under ideal conditions leads to a greater shear strength at the cement– bone interface. The results of the Norwegian Arthroplasty Register, however, suggest that reduced-viscosity cement used clinically may be related to a 2.5 times increase in the early revision rate for Charnley prostheses [27]. The difference may be related to the fact that the insertion of the femoral component with reduced-viscosity cement is more technically demanding. Charnley [32] wrote: ‘‘. . . one wonders whether to improve distal cement injection might not be counter productive; it could be argued that we need controlled subsidence of the prosthesis in the distal half of the femur to maintain load bearing contacts.’’ The pressure and mechanical profiles produced in this study would oppose just such an idea.
Acknowledgment The authors gratefully acknowledge the assistance and expertise of Mike Roberts, Chief Technician, Department of Orthopaedics, Leicester University, and Colin Morrison, Principle Experiment Officer, Department of Engineering, Leicester University.
References 1. Goldring SR, Clark CR, Wright TM: The problem in total joint arthroplasty: aseptic loosening (editorial). J Bone Joint Surg Am 75:799, 1993
2. Bulstrode CJK, Carr A, Murray D: Prediction of future work-load in total joint replacement. p 25. In Wallace WA (ed): The institution of mechanical engineers. Joint replacement in the 1990s: clinical studies, financial implications and marketing approaches. Bury St. Edmunds, Mechanical Engineering Publications, 1992 3. Lee AJC, Ling RSM: Loosening. p 137. In Ling RSM (ed): Complications of total hip replacement. Edinburgh, Churchill-Livingstone, 1982 4. Huiskes R: The current state and future of cemented and non-cemented total hip replacement. Post Graduate Lectures Efort Masson 1:52, 1993 5. Harris WH: Osteolysis and particle disease in hip replacement. Acta Orthop Scand 65:113, 1994 6. Taylor M, Tanner KE: Fatigue failure of cancellous bone: a possible cause of implant migration and loosening. J Bone Joint Surg Br 79:181, 1997 7. Hukkanen M, Corbett SA, Batten J, et al: Aseptic loosening of total hip replacement. J Bone Joint Surg Br 79:467, 1997 8. Apple AM, Sauder WG, Siverhuss W, et al: Prosthesis associated membrane induced bone resorption. Br J Rheumatol 29:32, 1990 9. Wang W, Ferguson DJP, Quinn JMW, et al: Biomaterial particle phagocytosis by bone-resorbing osteoclasts. J Bone Joint Surg Br 79:849, 1997 10. McCaskie AW, Harper WM, Gregg PJ: Femoral cementation technique: a survey of current UK Practice. J Bone Joint Surg Br 76(suppl):137, 1994 11. Hashemi-Nejad A, Birch NC, Goddard NJ: Current attitudes to cementing techniques in British hip surgery. Ann R Coll Surg Engl 76:396, 1994 12. Mulroy RD Jr, Harris WH: The effect of improved cementing techniques on component loosening in total hip replacement, an 11-year radiographic followup. J Bone Joint Surg Br 72:757, 1990 13. Halawa M, Lee AJC, Ling RSM, Vangala SS: The shear strength of trabecular bone from the femur, and some factors affecting the shear strength of the cementbone interface. Arch Orthop Trauma Surg 92:19, 1978 14. Majkowski RS, Miles AW, Bannister GC, et al: Bone surface preparation in cemented joint replacement. J Bone Joint Surg Br 75:459, 1993 15. Gruen TA, Markolf KL, Amstutz HC: Effects of laminations and blood entrapment on the strength of acrylic bone cement. Clin Orthop 119:250, 1976 16. Oh I, Carlson CE, Tomford WW, et al: Improved fixation of the femoral component after total hip replacement using a methacrylate intramedullary plug. J Bone Joint Surg Am 60:608, 1978 17. Oh I, Bourne RB, Harris WE: The femoral cement compactor. J Bone Joint Surg Am 65:1335, 1983 18. Oh I, Harris WE: A cement fixation system for total hip arthroplasty. Clin Orthop 164:221, 1982 19. Weber BG: Pressurised cement fixation in total hip arthroplasty. Clin Orthop 232:87, 1988 20. McCaskie AW, Barnes MR, Lin E, et al: Cement pressurisation during hip replacement. J Bone Joint Surg Br 79:379, 1997
Comparison of Femoral Cementing Techniques ● 21. Bannister GC, Miles AW: The influence of cementing technique strength of the bone-cement interface. Engineering in Medicine 17:131, 1988 22. Ling RSM: Cementing techniques in the femur. Tech Orthop 6:34, 1991 23. Heyes-Moore GH, Ling RSM: Current cementing techniques. In Morti R (ed): Progress in cemented total hip surgery and revision: proceedings of a symposium held in Amsterdam. Amsterdam, Excerpta Medica, 1982 24. Benjamin JB, Gie GA, Lee AJC, et al: Cementing techniques and the effect of bleeding. J Bone Joint Surg Br 69:620, 1987 25. Miller J, Krause WR, Krug WH, et al: Low viscosity cement. Orthop Trans 5:352, 1981 26. Markolf KL, Amstutz HC: In vitro measurement of bone-acrylic interface pressure during femoral component insertion. Clin Orthop 121:60, 1976 27. Havelin LI, Espehaug B, Vollset SE, et al: The effect of the type of cement on early revision of Charnley total hip prosthesis. J Bone Joint Surg Am 77:1543, 1995
Reading et al.
487
28. Malchau H, Herberts P: Prognosis of total hip replacement. Scientific Exhibition 63rd Annual Meeting of American Academy of Orthopaedic Surgeons, Atlanta, GA, 1996 29. Ward AJ, Smith EJ, Barlow JW, et al: The in vitro measurement of bone-cement interface pressures and shear strengths in the femur: a comparison of two cementation methods. Hip Int 5:124, 1995 30. MacDonald BE, Swarts E, Beaver R: Penetration and shear strength of cement-bone interfaces in vivo. Clin Orthop 286:283, 1993 31. Gruen TA, McNeice GM, Amstutz HC: ‘‘Modes of failure’’ of cemented stem type femoral components: a radiographic analysis of loosening. Clin Orthop 141:17, 1979 32. Charnley J: Low friction arthroplasty of the hip theory and practice. Berlin, Springer-Verlag, 1979 33. McCaskie AW: Studies related to the cemented fixation of the femoral component at total hip replacement. MD Thesis, University of Leicester, Leicester, 1995
A Comparison of 2 Modern Femoral Cementing Techniques Analysis by Cement–Bone Interface Pressure Measurements, Computerized Image Analysis, and Static Mechanical Testing A. D. Reading, FRCS (Tr and Orth),* A. W. McCaskie, MD, FRCS(Orth),† M. R. Barnes, BSc,‡ and P. J. Gregg, MD, FRCS†
Abstract: Modern cementing techniques aim to improve microinterlock and to reduce aseptic loosening. The Norwegian and Swedish Arthroplasty Registers have shown an increased risk of revision using reduced-viscosity cement. We have compared 2 modern cementing techniques using retrograde insertion of normalviscosity and reduced-viscosity cements. Laboratory-simulated arthroplasty was performed in paired human femora. Performance was evaluated by measuring pressures generated during cementation, cement penetration, and shear strength of the prosthesis–cement and bone–cement interfaces. Large differences exist between these 2 modern techniques. Despite no statistical differences between the pressure measurements with the 2 techniques, greater penetration of reduced-viscosity cement was found proximally, with a trend toward increased penetration of the more viscous cement distally. Areas of greater cement penetration with reduced-viscosity cement proximally produced higher values of ultimate shear strength. Both techniques showed a progressive increase in the shear strength as the level of the section progressed toward the tip of the prosthesis. There is a trend with both techniques for the distal fixation to be stronger. Key words: cement, viscosity, modern, hip, arthroplasty.
Aseptic loosening is an important long-term complication of total hip arthroplasty [1] and is, in part, responsible for the increasing numbers of patients
requiring revision surgery [2]. It remains poorly understood but involves biologic and mechanical processes [3–9]. Cementation is an important factor because it is under the direct control of the surgeon. Despite this importance, 2 surveys of consultants have shown a lack of consensus [10,11]. In part, this lack of consensus can be explained by a lack of clinical and biologic evidence. Modern cementing techniques [12] aim to improve microinterlock between the cement and cancellous bone. Cement is encouraged to penetrate to the region of the strongest cancellous bone within 3 mm of the corticocancellous junction [13]. There are several aspects of modern cementing techniques that when combined achieve this goal: i) improved surface preparation [14], ii) canal restriction, iii) retrograde insertion of cement [15], iv) sustained
From the *Department of Orthopaedics, Glasgow Royal Infirmary, Glasgow, Scotland, United Kingdom; †Department of Trauma and Orthopaedic Surgery, School of Surgical Sciences, The Medical School, University of Newcastle, Newcastle upon Tyne, England; and the ‡Department of Sports Medicine, Leicester General Hospital, Leicester. Submitted April 6, 1999; accepted December 10, 1999. Benefits were received in partial support of the research material described in this article. These benefits were as follows: Equipment supplied by various companies, including prostheses (DePuy International), cement (CMW), cementation equipment (Fry Surgical), cannulae (Portex). Reprint requests: A. D. Reading, FRCS (Tr and Orth), Department of Orthopaedics, Glasgow Royal Infirmary, 84 Castle Street, Glasgow, UK. Copyright r 2000 by Churchill Livingstonet doi:10.1054/arth.2000.5266 0883-5403/00/1504-0013$10.00/0
479
480 The Journal of Arthroplasty Vol. 15 No. 4 June 2000 pressurization [16,17], v) reduced viscosity of cement [25], and vi) correct timing of prosthesis insertion [18–20]. The flow of cement into cancellous bone is proportional to the pressure gradient. Several studies have confirmed that the higher pressure applied to the cement, the greater the penetration of cement into the cancellous bone [13,26]. Bleeding at the endosteal surface is an important factor in the creation of the cement–bone interface [14,21]. The bleeding pressure has been measured as high as 36 cm saline (27 mmHg) [22,23]. Experiments have shown that a column of blood 15 to 30 cm high can displace reduced-viscosity cement 7 minutes after the start of mixing [24]. For normal-viscosity cement, no displacement occurred after 3 minutes. Pressure applied to the cement above the simulated bleeding pressure caused the cement to flow in the opposite direction. Experiments in ox femora showed that the shear strength of the cement–bone interface for normal-viscosity cement was not affected by simulated bleeding; however, low-viscosity cement had significantly weaker interfaces in 50% of the samples [14]. Penetration depth of the cement was not affected. Many surgeons prefer to use normal-viscosity cement (53–74% [10,11]) for the insertion of the femoral component because it is easier to handle, and the timing is less critical, with no need to delay prosthetic entry after cement insertion. The Norwegian Arthroplasty Register reviewed 8,579 primary Charnley arthroplasties for primary osteoarthrosis in 7,922 patients with respect to the type of cement used [27]. The rate of revision of femoral components inserted with reduced-viscosity cement was 2.5 times greater than those inserted with normalviscosity cement. The Swedish Arthroplasty Register confirmed that the worst results were noted with low-viscosity cement [28]. The present study uses a modern cementing technique with retrograde insertion to compare normal-viscosity and reducedviscosity cements. The cement–bone interface pressures were measured during cementation and prosthesis insertion, cement penetration to cancellous bone was measured by computerized image analysis, and the shear strength of the prosthesis– cement and cement–bone interfaces was measured by static mechanical testing.
Leeds UK). Standard straight and curved reamers were used to prepare the femoral canal to accept a flanged 40 Charnley prosthesis. The canal was prepared using pressurized lavage (500 mL of saline) and power brushing. A correctly sized Elite (DePuy, Leeds United Kingdom) cement restrictor was inserted to a level 20 mm below the anticipated tip of the prosthesis (13 cm below the calcar). The inner surface of the prepared medullary canal was coated with 3 mL of human blood (to re-create the physical presence of blood, not the effect of backbleeding pressure). At the center of each of Gruen’s zones 1 through 7 (except 4) [29], a 3.5-mm drill hole was made into which the tapered end of a manometer line (Portex, Hythe, UK) was inserted and secured by interference fit (Fig. 1). The tip of the manometer line remained within the cancellous bone and was of a diameter sufficient to prevent cement blocking the catheter. The other end was connected to a pressure transducer (Elcomatic EM751; Elcomatic Ltd., Glasgow, Scotland, UK) and the line run through with saline to exclude all air. The 6 pressure lines were connected to a 6-channel chart recorder to allow simultaneous recording of the pressures during simulated arthroplasty. The pressure transducer had a limited range of activity (linearity ⫾ 0.1% of 4,000 mmHg/damaged pressure 8,000 mmHg), and an arbitrary limit of 5,000 mmHg (667 kPa) was set for data interpretation. Any reading above this level was assigned this value. Reduced-viscosity and normal-viscosity cements (CMW 3 and CMW 1, DePuy, Leeds, United Kingdom)
Materials and Methods Pairs of fresh frozen femora from 6 human donors were used in this study. Femora were defrosted before performing simulated arthroplasty using the standard Charnley system (DePuy International Ltd,
Fig. 1. Positioning of the pressure transducers at the center of Gruen zones.
Comparison of Femoral Cementing Techniques ●
were mixed in the same manner by using a syringe vacuum mixing system (Mitab, Fry Surgical International Ltd, UK). The normal-viscosity cement was extruded into the femur, retrograde with a caulking gun, 2 minutes after the start of mixing. A complete seal of the femoral opening was achieved with a thumb and pressure exerted on the cement column. This digital pressurization continued until the prosthesis (a new flanged 40 Charnley prosthesis) was inserted at 2 minutes, 45 seconds. For the reducedviscosity cement, extrusion into the femur, retrograde with a caulking gun, was performed at 2 minutes, 15 seconds, followed by digital pressurization until prosthesis insertion at between 3 minutes, 45 seconds, and 4 minutes. The timings used for reduced-viscosity cement were as per manufacturer’s instructions, whereas for normal-viscosity cement they were modified from manufacturer’s instructions to allow for extrusion of the more viscous cement from the gun system. All cement was mixed by the same person and the cementation of the femora performed by 1 person to reduce errors resulting from differing operative techniques. The pressure recordings were divided into 3 time periods: the first during cement insertion, the second during digital pressurization, and the third during prosthesis insertion. The duration of each period was measured assuming the start time to be the initial pressure rise and the end point to be the first pressure rise of the next period. The end of the final period (prosthesis insertion) was taken to be the point at which the pressure recording returned to 0. During each time period, the maximum peak pressure and the area under the curve were recorded. The mean pressure for that period could then be calculated (area under curve divided by the duration). The time that the pressure recorded was less than 100 mmHg was also measured. The data were not normally distributed and were analyzed using the Wilcoxon signed rank test, against a null hypothesis of no difference. Significance was taken at P ⱕ .05; a Bonferroni correction was made for the multiple testing of pressure measurements. The 95% confidence intervals (CIs) for the difference between the 2 techniques were calculated. After simulated arthroplasty, the femora were sectioned at 90° to the long axis of the bone (Fig. 2). This sectioning was performed with a high-precision wafering system (Isomet 2000, Buehler, United Kingdom). The first cut was performed at the cut edge of the calcar, the bone was then advanced 7 mm, and the second cut was performed. This process produced 12 parallel-sided samples 6.4 mm thick per femur. A high-resolution faxitron radiograph was taken of each section.
Reading et al.
481
Fig. 2. The sections of femora and those sections used in each study.
Computerized Image Analysis Computerized image analysis was initially performed on each of the 144 fresh sections (Omnimet 3, Buehler, United Kingdom). The system determined the area of the cement mantle, prosthesis, bone, and whole section (Figs. 3 and 4) as well as the length of the perimeter of the prosthesis and cement–bone interface. Computerized image analysis of fresh sections evaluated the total cement mantle. To subdivide the mantle further into reamed and penetrated portions, faxitrons were used (Fig. 5). The data obtained were not normally distributed. The differences between the 2 techniques were compared within each pair and at each level. The differences were analyzed using the Wilcoxon signed rank test, against a null hypothesis of no difference. Significance was taken at a level of P ⱕ .05. The 95% CIs for the difference between the 2 techniques were calculated.
Fig. 3. Image captured by the image analysis equipment after enhancement to define the edges clearly.
482 The Journal of Arthroplasty Vol. 15 No. 4 June 2000 Static Mechanical Tests The testing rig consisted of a servohydraulic ram with a platform attached to it and a 10-kN load cell. A temperature-regulated water tank was attached to the platform to allow the samples to be submerged in normal saline at 37°C ⫾ 0.5°C. Results of force and displacement were recorded on a personal computer. Reverse push-out mechanical tests were performed; the force was applied in a caudal-tocranial direction to prevent destruction of the specimens [29,30]. Prosthesis–Cement Interface
Fig. 4. Image after manipulation on which calculations were performed. Area of prosthesis, cement, and bone. Perimeter of prosthesis–cement and cement–bone interfaces.
Four sections were used for this experiment—1, 4, 7, and 10 (Fig. 2). To evaluate the prosthesis– cement interface, the specimen was placed on a platform that supported the cement mantle. The irregular shape of the specimens meant that special inserts had to be made to fit into the platform to support them; these were molded from bone– cement. A small punch was then attached to the load cell so that only the prosthesis was loaded. The specimens to be tested were then hydrated before being placed on the platform and submerged in saline at 37°C. Specimens were placed so that the direction of the taper of the prosthesis was reversed. The platform was then programmed to rise at a constant displacement of 3.93 mm/min. The force produced on the load cell was recorded at a sampling rate of 20 Hz until failure occurred and the prosthesis was pushed out of the cement mantle. Cement–Bone Interface
Fig. 5. (A) Image captured by the image analysis equipment of the faxitron radiograph shows the penetrated and reamed portions of the cement mantle (magnified). (B) Line diagram of image in (A).
Eight sections were used for this part of the experiment—1, 2, 4, 5, 7, 8, 10, and 11 (Fig. 2). The test rig for the cement–bone interface was similar to that used for the prosthesis–cement interface (Fig. 6). The only change was in the specimen support ring, which was modified to provide support for the cortical bone, rather than the cement. A problem encountered in other studies performing similar tests was specimen destruction. To overcome this problem, the specimens were placed in a mold, and a ring of polymethyl methacrylate casting resin was placed around the outside to provide support, reinforced with a nylon cerclage band. To apply force evenly to the cement–bone interface, a custom-made mushroom of bone–cement was made for each specimen. This bone–cement sat in the defect left by the prosthesis and spread out to the edge of the cement–bone interface. The punch attached to the load cell was also larger to provide an even force over the whole specimen. The speci-
Comparison of Femoral Cementing Techniques ●
Fig. 6. The test apparatus for the cement–bone interface testing.
mens were then submerged as before in the saline bath, and the force displacement curve was recorded during the failure cycle. A force displacement curve was plotted for each static mechanical test. The force at which failure occurred was taken from these graphs. Failure was said to have occurred once the value for recorded force decreased with a continuing increase in displacement. The height of the specimen and the perimeter of the prosthesis–cement or cement–bone interface, obtained by image analysis, multiplied together produced a value for the area of the prosthesis–cement or cement–bone interface. Dividing the force at which failure occurred by the area of the interface gave the ultimate shear strength of the interface. The differences between the 2 paired femora and the 2 techniques were then calculated. The data were nonparametric, and the differences were analyzed using the Wilcoxon signed rank test against a null hypothesis of no difference; significance was taken at the level P ⱕ .05. The 95% CIs for the difference between the 2 techniques were calculated.
Results Cement–Bone Interface Pressure Studies Considering cement insertion first, for the mean and the maximum pressures recorded, there were no statistically significant results between the 2 cementing techniques. Both cement types produced increasingly higher pressures the further distally the readings were taken along the femoral canal (me-
Reading et al.
483
dian value of arithmetic mean pressure for normalviscosity cement proximally at transducer, 1 ⫽ 1.02 kPa [interquartile range, 0.11–11.67 kPa], and distally at transducer, 3 ⫽ 19.13 kPa [interquartile range, 13–59 kPa]). If digital pressurization is considered, there were no statistically significant results between the 2 methods. Both cements produced higher pressures proximally and medially (median value normal-viscosity cement transducer, 7 ⫽ 62.94 kPa [interquartile range, 35.19–69.32 kPa]), decreasing distally during this phase (median value normalviscosity cement transducer, 5 ⫽ 48.06 kPa [interquartile range, 29.69–62.97 kPa]). During prosthesis insertion, there were no statistically significant results. The trend for increasing pressures distally was noted for both techniques (median value normalviscosity cement proximally at transducer, 1 ⫽ 37.59 kPa [interquartile range, 31.49–44.81 kPa], and distally at transducer, 3 ⫽ 87.44 kPa [interquartile range, 71.4–121.5 kPa]). Measurements of length of time recorded with the pressure below the mean arterial blood pressure of 13.3 kPa (100 mmHg) showed no statistical differences during the cement insertion phase (Table 1). During digital pressurization, there was a trend for the reduced-viscosity cement to spend a longer time with the pressure less than arterial bleeding pressure. This trend reached statistical significance at transducer sites 2, 3, and 4 (P ⫽ .036). At all transducer sites, the 95% CIs of the difference between techniques excluded 0. During the phase of prosthesis insertion, there was a trend for the normal-viscosity cement to spend a longer time with the pressure ⬍13.3 kPa. This trend reached statistical significance at transducer number 3 (P ⫽ .036). The trend also showed that the laterally placed transducers spent longer at a lower pressure than those placed medially. This situation was confirmed by the fact that the 95% CI difference excluded 0 at transducers 1, 2, 3, and 5. Computerized Image Analysis Image analysis of the femoral sections provided data on the percentage area of each section occupied by cement. The differences between the 2 paired femora and the 2 techniques were then calculated. The proximal 4 levels below the calcar showed a statistically significant trend for the reduced-viscosity cement to produce a larger area of cement per section (median value normal-viscosity cement section 1, 24.8% [range, 21.5–28.6%], and reducedviscosity cement, 33.1% [range, 30.15–41.81] [P ⫽ .036]). Further distally at levels 8 and 11, normal-viscosity cement produced a significantly larger cement mantle (P ⫽ .036).
484 The Journal of Arthroplasty Vol. 15 No. 4 June 2000 Table 1. Descriptive Statistics for the Time (Seconds) the Pressure Recorded Was ⬍13.3 kPa Normal Viscosity
Reduced Viscosity Median
Interquartile Range
Median
Interquartile Range
P Value
Cement Insertion 1 2 3 5 6 7
15.5 12.9 10.6 10.8 12.45 15.5
12.35–21.3 8.5–17.4 5.53–24.55 5.82–24.4 8.12–24.45 9.25–21.6
13.7 11.1 7.65 8.25 9.1 12.5
11.78–16.35 9.7–12.9 7.1–9.45 7.02–9.6 8–11.45 11.2–15.25
.281 .855 .529 .402 .345 .855
Digital Pressurization 1 2 3 5 6 7
11 11.2 10.45 10.4 9.7 7.3
9.5–17 6.65–15.5 5.9–15.35 5.75–15.35 5.25–15.55 3.5–13.45
35 33.7 34.55 34.05 30.05 24.4
30–36 29.25–35.45 27.13–36.35 28.95–36.65 25.38–34.9 18.7–28.25
.059† .100† .036† .036† .036† .100†
Prosthesis Insertion 1 2 3 5 6 7
25 18.6 14 13 12.5 12
17.45–31.2 10.25–36.2 9.5–28.25 9.75–24.5 4–21 10–14
15.55 7.2 10 9 9 8
4.55–25.12 5.6–15.45 6.75–16.25 5–13.5 4–13.5 5–19.5
.059† .100† .059† .036† .295 .584
Transducer
*The mean, interquartile range, and statistical significance for each different time period and cement are shown. †95% confidence interval of the difference excludes 0.
The technique used for preparation of the femora before insertion of the prostheses was standard for all the bones. To ensure that this technique produced consistent results for the size of the reamed medullary canal, calculations were performed on the faxitron radiographs. The results for the area of the reamed cement mantle were analyzed. There were no statistical differences between the paired femora. The differences noted in the areas of cement per section are due to the differences in the amount of cement penetrating the cancellous bone. Static Mechanical Testing The ultimate shear strength results are presented in Table 2. For the prosthesis–cement interface, there were no significant differences at any level. There was a trend toward greater values of shear strength as the level of the section moved distally. For the cement–bone interface, reduced-viscosity cement produced a statistically significant greater value of shear strength at the most proximal level under the calcar (P ⫽ .036). For all other levels of section, apart from level 7, there was a trend for reduced-viscosity cement to have the greater value of shear strength. The 95% CI of the difference excluded 0 at levels 1, 2, and 8. There was a trend for increasing values of the shear strength the further distal the level of the section, with the exception of level 7 for the normal-viscosity cement.
Discussion During cementation of the femoral component, it is considered desirable for the pressures to remain greater than the endosteal bleeding pressure of 36 cm saline [22,23] (27 mmHg) and preferably greater than the mean arterial blood pressure (13.3 kPa or 100 mmHg) to ensure no back-bleeding at the interface. From this study, it can be seen that the mean pressures remained greater than these levels throughout most of the cementation. The exceptions are during cement insertion for normalviscosity and reduced-viscosity cements, in which the most proximal levels recorded pressures less than 10 kPa. During digital pressurization, the reduced-viscosity cement spent a significantly longer time below the mean arterial blood pressure. In the clinical situation, this time could allow interface disruption by back-bleeding, but the extent of this effect cannot be quantified. A possible method of overcoming this problem would be to use a proximal femoral seal to prevent cement escaping proximally during digital pressurization. One weakness of the model used for this study is the underestimation of the dynamic flow of blood from the endosteal surface into the cement. Blood was placed on the endosteal surface before cementation to allow for any effect its physical presence may have, but
Comparison of Femoral Cementing Techniques ●
Reading et al.
485
Table 2. Ultimate Shear Strength (MPa) of the Prosthesis–Cement and Cement–Bone Interfaces at Each Level and for Each Cementing Technique* Normal Viscosity
Reduced Viscosity Median
Interquartile Range
Median
Interquartile Range
P Value
Prosthesis–Cement Interface 1 0.055 4 0.145 7 0.285 10 0.36
0.038–0.07 0.123–0.218 0.218–0.41 0.215–0.762
0.015 0.165 0.3 0.485
0.0–0.06 0.09–0.25 0.198–0.405 0.33–0.668
0.295 0.834 1.0 0.529
Cement–Bone Interface 1 0.4 2 0.489 4 1.05 5 1.128 7 2.28 8 1.189 10 1.595 11 2.196
0.323–0.87 0.437–0.822 0.687–1.553 0.894–1.908 1.51–2.81 0.76–1.992 1.178–2.855 1.531–2.454
0.925 0.74 1.395 1.446 1.545 1.583 1.9 2.091
0.673–1.293 0.539–1.512 1.062–1.602 1.071–3.069 1.08–2.787 1.512–2.022 1.182–2.742 1.035–3.206
0.036† 0.093† 0.295 0.463 0.093† 0.093† 0.834 0.834
Level
*The median, interquartile range, and statistical significance are shown. †95% confidence interval of the difference excludes 0.
there was no pressurized flow into the femoral canal. The mean pressures created during digital pressurization show an interesting pattern. Higher pressures are created under the calcar and on the medial side of the femur, as recommended originally by Charnley [32]. One feature common to both techniques was the high mean and maximum pressures generated during prosthesis insertion. This feature has been reported before [20,31], but the almost identical values for both techniques are notable. This finding suggests that the manufacturers’ timings do provide for a similar viscosity of cement at the time of prosthesis insertion. During prosthesis insertion, pressure was maintained above the mean arterial blood pressure for longer on the medial side of the femur, especially at the calcar. This situation could be related to the design of the prosthesis, with the cobra neck flange and shape of the stem directing the cement to the calcar. The mean and maximum pressures applied during cement insertion and digital pressurization were not significantly different for the 2 types of cement. Image analysis of the femoral sections, however, showed a larger area of cement mantle for the reduced-viscosity cement at the proximal levels. This larger area was a function of viscosity and the availability of cancellous bone. Further distally, there was no significant difference between the sizes of the cement mantles between the cements. The reverse push-out tests, although not physiologic, provided a more accurate comparison of the interface strengths and have previously been re-
ported [29,30,33]. For the prosthesis–cement interface, there were no statistically significant results between the 2 cements. There was a trend with both cement types for the values of the shear strength to increase as the level of the section moved further distally. A similar trend was also evident for the pressures measured during the insertion of the prosthesis. High prosthesis insertion pressures may be associated with increase in the bond strength between the prosthesis and the cement. Static mechanical testing of the cement–bone interface shows that the shear strength is significantly greater for the reduced-viscosity cement proximally. As the levels of section move further distally, there was no significant difference between the 2 cements; however, there is a trend for the reduced-viscosity cement to achieve a higher value. There is 1 exception, at level 7, and this is difficult to explain. There were no differences in the pressures generated during cementation at this level between the 2 techniques. There was increased penetration of normal-viscosity cement noted at that level, although this was not a significant difference. At the proximal level, there was a significant increase in the penetration of and the shear strength at the cement–bone interface for the reducedviscosity cement. This increase suggests that the shear strength of the cement–bone interface is related to degree of penetration of cement achieved into the cancellous bone. The reduced-viscosity cement achieves greater penetration and greater
486 The Journal of Arthroplasty Vol. 15 No. 4 June 2000 shear strength in the areas where the proportion of cancellous bone is highest. The shear strength values again increase as the levels of the sections move further distally. This increase may be explained by examining the specimens. Distally, the cement penetrates through the cancellous bone to the cortical bone; proximally this is not usually the case. Examining the specimens after failure, it can be seen that failure occurs through the cancellous bone just beyond the cement mantle. To improve the mechanical strength of the cement–bone interface proximally, the cement would be required to penetrate to the cortex. Theoretically, this penetration could be encouraged in 1 of 2 ways: i) by the use of a proximal femoral silicone seal and reduced-viscosity cement; ii) by removing more cancellous bone from the proximal femur during preparation of the canal. This second action, however, could potentially reduce bone stock with effects on further revision surgery. These hypotheses require further study before reaching clinical relevance. The results of this study suggest that reducedviscosity cement used correctly under ideal conditions leads to a greater shear strength at the cement– bone interface. The results of the Norwegian Arthroplasty Register, however, suggest that reduced-viscosity cement used clinically may be related to a 2.5 times increase in the early revision rate for Charnley prostheses [27]. The difference may be related to the fact that the insertion of the femoral component with reduced-viscosity cement is more technically demanding. Charnley [32] wrote: ‘‘. . . one wonders whether to improve distal cement injection might not be counter productive; it could be argued that we need controlled subsidence of the prosthesis in the distal half of the femur to maintain load bearing contacts.’’ The pressure and mechanical profiles produced in this study would oppose just such an idea.
Acknowledgment The authors gratefully acknowledge the assistance and expertise of Mike Roberts, Chief Technician, Department of Orthopaedics, Leicester University, and Colin Morrison, Principle Experiment Officer, Department of Engineering, Leicester University.
References 1. Goldring SR, Clark CR, Wright TM: The problem in total joint arthroplasty: aseptic loosening (editorial). J Bone Joint Surg Am 75:799, 1993
2. Bulstrode CJK, Carr A, Murray D: Prediction of future work-load in total joint replacement. p 25. In Wallace WA (ed): The institution of mechanical engineers. Joint replacement in the 1990s: clinical studies, financial implications and marketing approaches. Bury St. Edmunds, Mechanical Engineering Publications, 1992 3. Lee AJC, Ling RSM: Loosening. p 137. In Ling RSM (ed): Complications of total hip replacement. Edinburgh, Churchill-Livingstone, 1982 4. Huiskes R: The current state and future of cemented and non-cemented total hip replacement. Post Graduate Lectures Efort Masson 1:52, 1993 5. Harris WH: Osteolysis and particle disease in hip replacement. Acta Orthop Scand 65:113, 1994 6. Taylor M, Tanner KE: Fatigue failure of cancellous bone: a possible cause of implant migration and loosening. J Bone Joint Surg Br 79:181, 1997 7. Hukkanen M, Corbett SA, Batten J, et al: Aseptic loosening of total hip replacement. J Bone Joint Surg Br 79:467, 1997 8. Apple AM, Sauder WG, Siverhuss W, et al: Prosthesis associated membrane induced bone resorption. Br J Rheumatol 29:32, 1990 9. Wang W, Ferguson DJP, Quinn JMW, et al: Biomaterial particle phagocytosis by bone-resorbing osteoclasts. J Bone Joint Surg Br 79:849, 1997 10. McCaskie AW, Harper WM, Gregg PJ: Femoral cementation technique: a survey of current UK Practice. J Bone Joint Surg Br 76(suppl):137, 1994 11. Hashemi-Nejad A, Birch NC, Goddard NJ: Current attitudes to cementing techniques in British hip surgery. Ann R Coll Surg Engl 76:396, 1994 12. Mulroy RD Jr, Harris WH: The effect of improved cementing techniques on component loosening in total hip replacement, an 11-year radiographic followup. J Bone Joint Surg Br 72:757, 1990 13. Halawa M, Lee AJC, Ling RSM, Vangala SS: The shear strength of trabecular bone from the femur, and some factors affecting the shear strength of the cementbone interface. Arch Orthop Trauma Surg 92:19, 1978 14. Majkowski RS, Miles AW, Bannister GC, et al: Bone surface preparation in cemented joint replacement. J Bone Joint Surg Br 75:459, 1993 15. Gruen TA, Markolf KL, Amstutz HC: Effects of laminations and blood entrapment on the strength of acrylic bone cement. Clin Orthop 119:250, 1976 16. Oh I, Carlson CE, Tomford WW, et al: Improved fixation of the femoral component after total hip replacement using a methacrylate intramedullary plug. J Bone Joint Surg Am 60:608, 1978 17. Oh I, Bourne RB, Harris WE: The femoral cement compactor. J Bone Joint Surg Am 65:1335, 1983 18. Oh I, Harris WE: A cement fixation system for total hip arthroplasty. Clin Orthop 164:221, 1982 19. Weber BG: Pressurised cement fixation in total hip arthroplasty. Clin Orthop 232:87, 1988 20. McCaskie AW, Barnes MR, Lin E, et al: Cement pressurisation during hip replacement. J Bone Joint Surg Br 79:379, 1997
Comparison of Femoral Cementing Techniques ● 21. Bannister GC, Miles AW: The influence of cementing technique strength of the bone-cement interface. Engineering in Medicine 17:131, 1988 22. Ling RSM: Cementing techniques in the femur. Tech Orthop 6:34, 1991 23. Heyes-Moore GH, Ling RSM: Current cementing techniques. In Morti R (ed): Progress in cemented total hip surgery and revision: proceedings of a symposium held in Amsterdam. Amsterdam, Excerpta Medica, 1982 24. Benjamin JB, Gie GA, Lee AJC, et al: Cementing techniques and the effect of bleeding. J Bone Joint Surg Br 69:620, 1987 25. Miller J, Krause WR, Krug WH, et al: Low viscosity cement. Orthop Trans 5:352, 1981 26. Markolf KL, Amstutz HC: In vitro measurement of bone-acrylic interface pressure during femoral component insertion. Clin Orthop 121:60, 1976 27. Havelin LI, Espehaug B, Vollset SE, et al: The effect of the type of cement on early revision of Charnley total hip prosthesis. J Bone Joint Surg Am 77:1543, 1995
Reading et al.
487
28. Malchau H, Herberts P: Prognosis of total hip replacement. Scientific Exhibition 63rd Annual Meeting of American Academy of Orthopaedic Surgeons, Atlanta, GA, 1996 29. Ward AJ, Smith EJ, Barlow JW, et al: The in vitro measurement of bone-cement interface pressures and shear strengths in the femur: a comparison of two cementation methods. Hip Int 5:124, 1995 30. MacDonald BE, Swarts E, Beaver R: Penetration and shear strength of cement-bone interfaces in vivo. Clin Orthop 286:283, 1993 31. Gruen TA, McNeice GM, Amstutz HC: ‘‘Modes of failure’’ of cemented stem type femoral components: a radiographic analysis of loosening. Clin Orthop 141:17, 1979 32. Charnley J: Low friction arthroplasty of the hip theory and practice. Berlin, Springer-Verlag, 1979 33. McCaskie AW: Studies related to the cemented fixation of the femoral component at total hip replacement. MD Thesis, University of Leicester, Leicester, 1995