- Email: [email protected]
Mechanisms for Pumping Fluid Through Cementless Acetabular Components With Holes
The Journal of Arthroplasty Vol. 20 No. 8 2005
Mechanisms for Pumping Fluid Through Cementless Acetabular Components With Holes William L. Walter, MBBS,* Jonathan Clabeaux,* Timothy M. Wright, PhD,* William Walsh, PhD,y William K. Walter, MBBS,z and Thomas P. Sculco, MD*
Abstract: The pumping of fluid and polyethylene wear debris from the joint space to the retroacetabular bone is implicated in the pathogenesis of osteolysis. Three possible mechanisms for this pumping: pressure gradients, diaphragm pumping, and piston pumping were studied in vitro in a laboratory model. The simulated activities of rising from a chair and climbing stairs produced high pressure gradients and high angles of loading that could pump fluid through the apical hole to the retroacetabular bone. A noncongruent liner acted as a diaphragm pump, producing pressures 6 times higher than that seen with a congruent liner. Pistoning motion of the liner produced pressures 8 times higher than when no pistoning occurs. These pumping mechanisms could be mitigated by the use of acetabular components without holes. Key words: polyethylene, acetabular component, holes, pumping. n 2005 Elsevier Inc. All rights reserved.
pressures within the hip joint force fluid and particles through the holes to the retroacetabular bone. Intracapsular pressure was measured in hips with prostheses [6,7]. In the supine patient with the hip extended, the average pressure was 26 mm Hg (range 0 -60 mm Hg) [6]. Peak joint pressures in the weight-bearing patient vary with different maneuvers, averaging 207 mm Hg (range 155-517 mm Hg) with walking and 517 mm Hg (range 362-776 mm Hg) with climbing a step or rising from a chair [7]. Little is known about the intraosseous pressure in the retroacetabular bone during these maneuvers, but presumably, the pressure would be considerably lower [8]. Therefore, providing there is a pathway, this pressure gradient could cause fluid to flow. A second possible mechanism is that relative motion between the polyethylene liner and the metallic shell could pump fluid through the holes [5,9,10]. Micromotion between liners and shells has been observed in vitro in laboratory experiments [11,12] and in vivo at revision surgery [2,4,13]. In vivo micromotion has also been deduced from careful inspection of retrieved liners [2,9]. Several authors have suggested that this
Retroacetabular osteolysis is a major cause of revision of cementless acetabular components [1,2]. Polyethylene wear debris has been identified in the granulomata of these retroacetabular osteolytic lesions [2 -5]. This wear debris is generated in the joint either on the articular surface or the backside of the polyethylene liner and may be pumped through the holes in the metallic shell to the retroacetabular bone. At least 2 possible mechanisms might cause this pumping. The most common belief is that there is a pressure gradient pumping mechanism [4,6], in which elevated From the *Hospital for Special Surgery, New York, New York; y Orthopaedic Research Laboratories, Prince of Wales Hospital, Randwick, NSW, Australia; and z Sydney Hip and Knee Surgeons, Waverton, NSW, Australia. Submitted July 31, 2003; accepted March 14, 2005. Benefits or funds were received in partial or total support of the research material described in this article. These benefits and/or support were received from Stryker Europe. Reprint requests: William L. Walter, MBBS, Sydney Hip and Knee Surgeons, 100 Bay Rd, 3rd Floor, Waverton, NSW 2060, Australia. n 2005 Elsevier Inc. All rights reserved. 0883-5403/05/1906-0004$30.00/0 doi:10.1016/j.arth.2005.03.039
1042
Pump Action Polyethylene ! Walter et al
1043
Fig. 1. Diaphragm pumping. A space exists between this noncongruent liner and the metallic shell containing joint fluid and wear debris. When the hip is loaded, the polyethylene liner deforms, the volume of the space is reduced, and the fluid and wear debris are forced through the apical hole to the retroacetabular bone.
relative motion might pump fluid and particles from the space between the liner and the shell through the holes to the retroacetabular bone [5,9,10]. The phenomenon of a dome hole obturator unscrewing in vivo has been cited as evidence for this pumping mechanism [9]. To examine these mechanisms, we developed a simple generic laboratory apparatus to model the acetabular component in vivo. Our aim was to document different pumping mechanisms in the hip including the 2 versions of the polyethylene pumping mechanism, namely, diaphragm pumping characterized by deformation of a noncongruent liner suspended at the rim of the shell (Fig. 1) and piston pumping characterized by pistoning of the liner in and out of the shell (Fig. 2). We have also analyzed the effect of activities such as of rising
from a chair and level walking by the pressure gradient mechanism.
Materials and Methods Three experiments were performed: a pressure gradient experiment, in which fluid flow through an apical hole in the shell was compared under loading protocols simulating walking and rising from a chair (Table 1); a diaphragm pumping experiment, in which the pressures in the apical hole produced by congruent compared with noncongruent liners were measured under the walking protocol; and a piston pumping experiment, in which the pressures in the apical hole produced by congruent liners were measured
Fig. 2. Piston pumping. A loose locking mechanism has allowed this liner to come partway out of the metallic shell. When the hip is loaded, the liner is pushed back into the shell, pumping fluid and wear debris through the apical hole.
1044 The Journal of Arthroplasty Vol. 20 No. 8 December 2005 Table 1. Pressure gradient experiment Outcome measured Protocol Load (N) Frequency (Hz) Angle of load Pressure gradient (mm Hg) Liner
Rate of fluid flow Walking Rising 50 to 2000 50 to 1600 1 0.25 238 468 200 500 Congruent Congruent
under both the standard walking protocol and a modified walking protocol that produced pistoning of the liner (Table 2). The apparatus consisted of an aluminum base, an aluminum fluid chamber representing the hip joint, a plastic collection cylinder representing the retroacetabular bone, and 2 removable aluminum cups representing cementless acetabular shells with apical holes (Figs. 3-5). The fluid chamber was bolted to the base, and the clear plastic collection cylinder was calibrated at 10-mL intervals and threaded directly into the aperture in the fluid chamber from below (this interface was sealed with a rubber ring gasket). The collection cylinder was then attached to a vacuum pump capable of generating negative pressures up to 760 mm Hg (oil-less piston-type pressure/vacuum pump, Gast Manufacturing, Benton Harbor, Mich). The difference between atmospheric pressure in the fluid chamber and negative pressure in the collection cylinder was intended to simulate the difference between positive pressure in the joint space and neutral pressure in the retroacetabular bone in vivo [6,7]. The aluminum cups had an apical hole (9-mm diameter) that connected to the collection cylinder creating a path for fluid to flow from the chamber to the collection cylinder. The aluminum cups were angled (Fig. 5): the first one at 238 to the axis of the applied load to represent the angle of the forces on the hip during walking and the second at 468 to the load axis to represent the angle of the forces on the
Fig. 3. Front view of the apparatus.
hip when rising from a chair [14,15]. For testing, a cup was inserted into the aperture in the fluid chamber from above and secured with a set screw. The interface between the cup and the fluid chamber was sealed with a rubber ring gasket so the only egress of fluid was via the apical hole in the cup. The integrity of this watertight seal was verified by pretesting the cups for 1000 cycles before drilling the apical holes. For the piston pumping and diaphragm pumping experiments, the collection cylinder was removed from below the fluid chamber. A pressure transducer (subminiature flush diaphragm, Sensotec, Columbus, Ohio) was threaded into the hole in the 238 aluminum cup from below leaving approximately 1 mL of space between the top of the transducer and the bottom of the polyethylene liner. The pressure transducer had a range from 0 to 10 343 mm Hg and an accuracy of 1%. Fluid pressure and load data were recorded at a rate of 10 Hz.
Table 2. Diaphragm pumping experiment Outcome measured Protocol Load (N) Frequency (Hz) Angle of load Liner
Piston pumping experiment
Pressure Walking 50 to 2000 1 238 Congruent
Pressure Walking 50 to 2000 1 238 Noncongruent
Walking 50 to 2000 1 238 Congruent
Pistoning +100 to 2000 1 238 Congruent
Pump Action Polyethylene ! Walter et al
1045
(Table 1). Five liners were tested in random sequence first by the rising protocol and then by the walking protocol. A preload of 50 N was applied to properly seat the liner in the cup. The fluid chamber was then filled with 200 mL of bovine serum (HyClone, Logan, Utah). The number of cycles to reach every 10 mL of fluid collected in the chamber was recorded, and each run was stopped after 1000 cycles or when 100 mL of fluid was collected, whichever occurred first. Between each test, the apparatus was removed, drained of serum, cleaned with soap and water, and dried thoroughly with compressed air. Diaphragm Pumping Experiment
Fig. 4. Top view of the apparatus.
Two types of liners were machined from HSS reference ultrahigh molecular weight polyethylene [16]. Congruent liners (n = 16) had a hemispherical internal geometry of 28mm diameter and a hemispherical external geometry of 42mm diameter abutting a rim. The noncongruent liners (n = 5) were identical to the congruent liners except that the centers of the hemispheres were displaced 0.3mm relative to the rim resulting in an incongruency of 0.3mm between the polyethylene liner and the aluminum cup at the dome. Both liner types were designed without locking mechanisms. The dimensional tolerance in manufacturing the liners was 0.127 mm. The apparatus was placed on a bearing plate that rested on the platen of an MTS uniaxial load frame (MTS Systems Corp, Eden Prairie, Minn). The bearing plate allowed the apparatus to selfcenter under the axial compressive load, which was delivered to the polyethylene liner through a 28mm metallic femoral head. Ambient conditions were recorded before each test: temperature (range 20.08C to 22.28C) and relative humidity (range 20% to 25%). Pressure Gradient Experiment Two loading protocols, based on published data [6,7,15,17], were used: one to simulate walking and a second to simulate rising from a chair
For this experiment, the walking protocol was used in testing 5 congruent liners and 5 noncongruent liners (Table 2). The apical hole was carefully filled with bovine serum using a syringe to prevent air bubbles, the fluid chamber was filled, and then the polyethylene liner was placed into the aluminum cup. We alternated a congruent liner with a noncongruent liner in random sequence on the first day, and then the order was reversed on the second day of testing (a minimum of 24 hours between tests). Each specimen was run for 100 cycles, and the pressure amplitude was recorded at 90 cycles.
Fig. 5. The cup on the left (shown with a polyethylene liner) is tilted at 238 to simulate the angle of the load on the acetabular component during walking. The cup on the right (shown without a polyethylene liner) is tilted at 468 to simulate the angle of the load on the acetabular component with rising from a chair.
1046 The Journal of Arthroplasty Vol. 20 No. 8 December 2005 Piston Pumping Experiment In this experiment, 2 loading conditions were examined: one with the normal walking protocol and the other walking with pistoning of the bearing. The apparatus was modified by adding a metallic bar supported by 2 threaded rods, positioned to limit upward load frame motion; the test could then be conducted in load control including both tensile and compressive loads. After filling the apparatus with bovine serum as in the diaphragm pumping experiment, a 50N load was applied, and the top bolts on the rods were adjusted to allow an upward displacement of 0.3mm as measured with calibrated shims. The load was then changed to 100 N of tensile force, and the test was run. On the first day of testing, each of 6 liners was tested in random sequence first by the normal walking protocol then by the pistoning protocol. On the second test day, the order of the specimens was reversed, as was the order of the protocols (again with 24 hours between tests). As in the diaphragm pumping experiment, each specimen was run for 100 cycles, and the pressure amplitude was recorded at 90 cycles. Statistical Analyses In the pressure gradient experiment, we compared the rates of flow between the 2 groups. In the diaphragm pumping and piston pumping experiments, we compared pressures. Data were analyzed for significant differences using a paired Student t test when comparing the 2 groups within an experiment and a nonpaired Student t test when comparing results between the diaphragm pump-
Fig. 7. Results from the piston pumping experiment showed significantly greater pressures with pistoning (5140 F 330 mm Hg) than without pistoning of the liners (650 F 300 mm Hg) ( P b .001).
ing and piston pumping experiments. A P value of less than .05 was considered significant.
Results In the pressure gradient experiment, no fluid flow occurred in any of the 5 runs with the walking protocol, but significant flow ( P b .001) at a rate of 127 F 16 mL/min was detected through the apical hole in all 5 runs with the rising protocol. In the diaphragm pumping experiment, the pressure produced by the noncongruent liners was 6 times the pressure produced by the congruent liners (Fig. 6). In the piston pumping experiment, the pressure produced by the pistoning liners was 8 times the pressure produced without pistoning (Fig. 7). No significant difference occurred between the congruent liners in the walking protocol in the piston pumping experiment and the congruent liners in the walking protocol in the diaphragm pumping experiment ( P = .9). Pressures produced by the pistoning liners were higher than pressures produced by the noncongruent liners, but this difference was not significant ( P = .064).
Discussion
Fig. 6. Results from the diaphragm pumping experiment showed significantly greater pressures (4030 F 1250 mm Hg) for the noncongruent liners than for the congruent liners (670 F 240 mm Hg) ( P = .006).
The pressure gradient experiment illustrates that a pressure gradient could pump fluid and wear debris from the hip joint through an apical hole to the retroacetabular bone. The higher pressures and higher load angles when rising from a chair and stair climbing probably make these activities more important in pumping fluid by this mechanism
Pump Action Polyethylene ! Walter et al
than level walking. Our model had a single apical hole but the number, size, and location of holes could all be important factors. The pressure gradient mechanism could also pump fluid and wear debris into the acetabular cancellous bone at the periphery of the shell causing peripheral acetabular osteolysis. Peripheral osteolytic lesions have been observed in association with acetabular components with and without holes [1,3,18], but apical osteolytic lesions typically occur only behind components with holes [18]. Which of these patterns of osteolysis develops in a particular hip is probably determined by which pathway offers the least resistance to the flow of fluid and debris. The diaphragm pumping and piston pumping experiments show 2 different mechanisms for the pumping action of polyethylene liners in cementless metal acetabular components with holes, both of which probably coexist in vivo. The relative importance of each will depend on mechanical factors such as the conformity of the liner to the shell, the ability of the locking mechanism to restrict movement of the liner within the shell, and the nature of the applied load. In diaphragm pumping, the liner is suspended at the rim and noncongruent at the dome. The volume of the space between the liner and shell reduces as the polyethylene deforms with loading and increases again with unloading (the so-called btrampoline effectQ [11]). Fluid may flow in and out through the hole or back past the locking mechanism into and out of the joint. Diaphragm pumping relies on a combination of peripheral capture and nonconformity between the liner and shell but can occur even if the peripheral capture locking mechanism does not allow any motion at the rim. Most locking mechanisms are variations of locking tabs or locking rings located at the periphery of the shell, designed to achieve peripheral capture of the liner. Nonconformity of the liner and shell exists to some degree in all modular acetabular components [10,19]. Apart from tolerance variations, radial clearance between the 2 parts is necessary to allow insertion of the liner [10]. In vitro studies measuring the micromotion between the dome of the shell and the liner in new components show that this motion decreases with time in the majority of designs [12], presumably because of creep of the polyethylene, which improves the congruency with time. Therefore, the diaphragm pumping effect should decrease with time. In piston pumping, the entire liner moves in and out of the shell with loading. Piston pumping relies on a locking mechanism that allows movement. Motion between the liner and the shell at the rim is
1047
invariably present in modular components, even when they are new and has been measured to be between 7.5 and 359 lm [11]. Previous in vitro work on liner micromotion [11,12] used axial loads in compression (sometimes in combination with rotation); as a result, the femoral head remained fully seated in the liner throughout the cycle, and the liner remained seated in the shell. However, in vivo fluoroscopic imaging of unconstrained total hip arthroplasties has shown an average of 1.2 mm of separation of the femoral head from the acetabulum during gait [20]. We are aware of no in vitro study of liner motion that has included separation; therefore, the liner motion that occurs in vivo has been underestimated. Pistoning motion of polyethylene liners in their metallic shells has been observed at revision surgery [2,3,13]. The ability of the locking mechanism to limit motion might deteriorate with time in service due to polyethylene wear and creep within the mechanism itself [21], thereby allowing more pistoning of the liner. Therefore, the piston pumping effect would be expected to increase with time. Our study has limitations, but we do not believe that they compromise our conclusions. For example, we simulated a retroacetabular osteolytic lesion (a cyst in bone) as a cavity in the metal apparatus. Given the greater stiffness of the aluminum than cancellous bone, the absolute pressures that we measured in the piston pumping and diaphragm pumping experiments will not be directly translatable to the in vivo situation, but the differences in the pressures between the noncongruent liners and the congruent liners and between the pistoning liners and the nonpistoning liners are important. Similarly, we used a single load direction and a constant pressure gradient because our primary aim was to show the presence of these mechanisms under physiological type loads. We would hypothesize that if we altered our apparatus to simulate the multidirectional loads and fluctuating pressure gradients that occur in vivo, the pumping would be even more severe. Osteolysis is widely thought to be caused by polyethylene wear debris. However, cyclic fluid pressure changes, perhaps in combination with wear debris, have been implicated in the pathogenesis of osteolysis [22-24]. The mechanisms of pumping that are shown in the current study could be important factors. Each hole in a metallic shell increases the beffective joint spaceQ [4] and creates the potential for particles and pressurized fluid to gain access to the retroacetabular bone, which may in turn increase the likelihood of retroacetabular osteolysis [5,10,25]. Holes in shells
1048 The Journal of Arthroplasty Vol. 20 No. 8 December 2005 are usually unnecessary. Screw holes can be avoided when screws are not used, and dome holes, which facilitate insertion of the component, can be avoided by minor alterations in component design and surgical instrumentation. A welldesigned acetabular component with no holes, an appropriately constraining locking mechanism, and a good quality surface used in conjunction with the press-fit insertion technique could mitigate these pumping mechanisms and reduce the potential for osteolysis.
Acknowledgments
10.
11.
12.
13.
The authors thank Eric Charriere and Robert Streicher of Stryker Clinical and Scientific Affairs Europe for their support of this work.
14.
References
16.
1. Maloney WJ, Galante JO, Anderson M, et al. Fixation, polyethylene wear, and pelvic osteolysis in primary total hip replacement. Clin Orthop 1999; 369:157. 2. Astion DJ, Saluan P, Stulberg BN, et al. The porouscoated anatomic total hip prosthesis: failure of the metal-backed acetabular component. J Bone Joint Surg Am 1996;78:755. 3. Schmalzried TP, Harris WH. The Harris-Galante porous-coated acetabular component with screw fixation. Radiographic analysis of eighty-three primary hip replacements at a minimum of five years. J Bone Joint Surg Am 1992;74:1130. 4. Schmalzried TP, Jasty M, Harris WH. Periprosthetic bone loss in total hip arthroplasty. Polyethylene wear debris and the concept of the effective joint space. J Bone Joint Surg Am 1992;74:849. 5. Huk OL, Bansal M, Betts F, et al. Polyethylene and metal debris generated by non-articulating surfaces of modular acetabular components. J Bone Joint Surg Br 1994;76:568. 6. Robertsson O, Wingstrand H, Kesteris U, et al. Intracapsular pressure and loosening of hip prostheses. Preoperative measurements in 18 hips. Acta Orthop Scand 1997;68:231. 7. Hendrix RW, Wixson RL, Rana NA, et al. Arthrography after total hip arthroplasty: a modified technique used in the diagnosis of pain. Radiology 1983; 148:647. 8. Downey DJ, Simkin PA, Taggart R. The effect of compressive loading on intraosseous pressure in the femoral head in vitro. J Bone Joint Surg Am 1988; 70:871. 9. Scott DL, Campbell PA, McClung CD, et al. Factors contributing to rapid wear and osteolysis in hips
15.
17.
18.
19.
20.
21.
22.
23.
24.
25.
with modular acetabular bearings made of hylamer. J Arthroplasty 2000;15:35. Young AM, Sychterz CJ, Hopper Jr RH, et al. Effect of acetabular modularity on polyethylene wear and osteolysis in total hip arthroplasty. J Bone Joint Surg Am 2002;84:58. Fehring TK, Smith SE, Braun ER, et al. Motion at the modular acetabular shell and liner interface. A comparative study. Clin Orthop 1999;367:306. Williams II VG, Whiteside LA, White SE, et al. Fixation of ultrahigh-molecular-weight polyethylene liners to metal-backed acetabular cups. J Arthroplasty 1997;12:25. Walter WL, Walter WK, O’Sullivan M. The pumping of fluid in cementless cups with holes. J Arthroplasty 2004;19:230. Berme N, Paul JP. Load actions transmitted by implants. J Biomed Eng 1979;1:268. Bergmann G, Deuretzbacher G, Heller M, et al. Hip contact forces and gait patterns from routine activities. J Biomech 2001;34:859. Bennett AP, Wright WT, Li S. Global reference UHMWPE: characterization and comparison to commercial UHMWPE. 42nd Annual Meeting of Orthopaedic Research Society, Atlanta, Georgia, 1996. Pedersen DR, Brand RA, Davy DT. Pelvic muscle and acetabular contact forces during gait. J Biomech 1997;30:959. Claus AM, Sychterz CJ, Hopper Jr RH, et al. Pattern of osteolysis around two different cementless metalbacked cups: retrospective, radiographic analysis at minimum 10-year follow-up. J Arthroplasty 2001; 16:177. Kurtz SM, Ochoa JA, White CV, et al. Backside nonconformity and locking restraints affect liner/ shell load transfer mechanisms and relative motion in modular acetabular components for total hip replacement. J Biomech 1998;31:431. Lombardi Jr AV, Mallory TH, Dennis DA, et al. An in vivo determination of total hip arthroplasty pistoning during activity. J Arthroplasty 2000;15:702. Collier JP, Mayor MB, Jensen RE, et al. Mechanisms of failure of modular prostheses. Clin Orthop 1992; 285:129. Ferrier GM, McEvoy A, Evans CE. The effect of cyclic pressure on human monocyte-derived macrophages in vitro. J Bone Joint Surg Br 2000;82:755. Skripitz R, Aspenberg P. Pressure-induced periprosthetic osteolysis: a rat model. J Orthop Res 2000; 18:481. Van Der Vis HM, Marti RK, Tigchelaar W, et al. Benign cellular responses in rats to different wear particles in intra-articular and intramedullary environments. J Bone Joint Surg Br 1997;79:837. Kilgus JTJ. Retro-acetabular osteolysis and polyethylene access: a radiographic analysis of uncemented acetabular cups with 2-14 year follow-up. 2001 [San Francisco].
Mechanisms for Pumping Fluid Through Cementless Acetabular Components With Holes William L. Walter, MBBS,* Jonathan Clabeaux,* Timothy M. Wright, PhD,* William Walsh, PhD,y William K. Walter, MBBS,z and Thomas P. Sculco, MD*
Abstract: The pumping of fluid and polyethylene wear debris from the joint space to the retroacetabular bone is implicated in the pathogenesis of osteolysis. Three possible mechanisms for this pumping: pressure gradients, diaphragm pumping, and piston pumping were studied in vitro in a laboratory model. The simulated activities of rising from a chair and climbing stairs produced high pressure gradients and high angles of loading that could pump fluid through the apical hole to the retroacetabular bone. A noncongruent liner acted as a diaphragm pump, producing pressures 6 times higher than that seen with a congruent liner. Pistoning motion of the liner produced pressures 8 times higher than when no pistoning occurs. These pumping mechanisms could be mitigated by the use of acetabular components without holes. Key words: polyethylene, acetabular component, holes, pumping. n 2005 Elsevier Inc. All rights reserved.
pressures within the hip joint force fluid and particles through the holes to the retroacetabular bone. Intracapsular pressure was measured in hips with prostheses [6,7]. In the supine patient with the hip extended, the average pressure was 26 mm Hg (range 0 -60 mm Hg) [6]. Peak joint pressures in the weight-bearing patient vary with different maneuvers, averaging 207 mm Hg (range 155-517 mm Hg) with walking and 517 mm Hg (range 362-776 mm Hg) with climbing a step or rising from a chair [7]. Little is known about the intraosseous pressure in the retroacetabular bone during these maneuvers, but presumably, the pressure would be considerably lower [8]. Therefore, providing there is a pathway, this pressure gradient could cause fluid to flow. A second possible mechanism is that relative motion between the polyethylene liner and the metallic shell could pump fluid through the holes [5,9,10]. Micromotion between liners and shells has been observed in vitro in laboratory experiments [11,12] and in vivo at revision surgery [2,4,13]. In vivo micromotion has also been deduced from careful inspection of retrieved liners [2,9]. Several authors have suggested that this
Retroacetabular osteolysis is a major cause of revision of cementless acetabular components [1,2]. Polyethylene wear debris has been identified in the granulomata of these retroacetabular osteolytic lesions [2 -5]. This wear debris is generated in the joint either on the articular surface or the backside of the polyethylene liner and may be pumped through the holes in the metallic shell to the retroacetabular bone. At least 2 possible mechanisms might cause this pumping. The most common belief is that there is a pressure gradient pumping mechanism [4,6], in which elevated From the *Hospital for Special Surgery, New York, New York; y Orthopaedic Research Laboratories, Prince of Wales Hospital, Randwick, NSW, Australia; and z Sydney Hip and Knee Surgeons, Waverton, NSW, Australia. Submitted July 31, 2003; accepted March 14, 2005. Benefits or funds were received in partial or total support of the research material described in this article. These benefits and/or support were received from Stryker Europe. Reprint requests: William L. Walter, MBBS, Sydney Hip and Knee Surgeons, 100 Bay Rd, 3rd Floor, Waverton, NSW 2060, Australia. n 2005 Elsevier Inc. All rights reserved. 0883-5403/05/1906-0004$30.00/0 doi:10.1016/j.arth.2005.03.039
1042
Pump Action Polyethylene ! Walter et al
1043
Fig. 1. Diaphragm pumping. A space exists between this noncongruent liner and the metallic shell containing joint fluid and wear debris. When the hip is loaded, the polyethylene liner deforms, the volume of the space is reduced, and the fluid and wear debris are forced through the apical hole to the retroacetabular bone.
relative motion might pump fluid and particles from the space between the liner and the shell through the holes to the retroacetabular bone [5,9,10]. The phenomenon of a dome hole obturator unscrewing in vivo has been cited as evidence for this pumping mechanism [9]. To examine these mechanisms, we developed a simple generic laboratory apparatus to model the acetabular component in vivo. Our aim was to document different pumping mechanisms in the hip including the 2 versions of the polyethylene pumping mechanism, namely, diaphragm pumping characterized by deformation of a noncongruent liner suspended at the rim of the shell (Fig. 1) and piston pumping characterized by pistoning of the liner in and out of the shell (Fig. 2). We have also analyzed the effect of activities such as of rising
from a chair and level walking by the pressure gradient mechanism.
Materials and Methods Three experiments were performed: a pressure gradient experiment, in which fluid flow through an apical hole in the shell was compared under loading protocols simulating walking and rising from a chair (Table 1); a diaphragm pumping experiment, in which the pressures in the apical hole produced by congruent compared with noncongruent liners were measured under the walking protocol; and a piston pumping experiment, in which the pressures in the apical hole produced by congruent liners were measured
Fig. 2. Piston pumping. A loose locking mechanism has allowed this liner to come partway out of the metallic shell. When the hip is loaded, the liner is pushed back into the shell, pumping fluid and wear debris through the apical hole.
1044 The Journal of Arthroplasty Vol. 20 No. 8 December 2005 Table 1. Pressure gradient experiment Outcome measured Protocol Load (N) Frequency (Hz) Angle of load Pressure gradient (mm Hg) Liner
Rate of fluid flow Walking Rising 50 to 2000 50 to 1600 1 0.25 238 468 200 500 Congruent Congruent
under both the standard walking protocol and a modified walking protocol that produced pistoning of the liner (Table 2). The apparatus consisted of an aluminum base, an aluminum fluid chamber representing the hip joint, a plastic collection cylinder representing the retroacetabular bone, and 2 removable aluminum cups representing cementless acetabular shells with apical holes (Figs. 3-5). The fluid chamber was bolted to the base, and the clear plastic collection cylinder was calibrated at 10-mL intervals and threaded directly into the aperture in the fluid chamber from below (this interface was sealed with a rubber ring gasket). The collection cylinder was then attached to a vacuum pump capable of generating negative pressures up to 760 mm Hg (oil-less piston-type pressure/vacuum pump, Gast Manufacturing, Benton Harbor, Mich). The difference between atmospheric pressure in the fluid chamber and negative pressure in the collection cylinder was intended to simulate the difference between positive pressure in the joint space and neutral pressure in the retroacetabular bone in vivo [6,7]. The aluminum cups had an apical hole (9-mm diameter) that connected to the collection cylinder creating a path for fluid to flow from the chamber to the collection cylinder. The aluminum cups were angled (Fig. 5): the first one at 238 to the axis of the applied load to represent the angle of the forces on the hip during walking and the second at 468 to the load axis to represent the angle of the forces on the
Fig. 3. Front view of the apparatus.
hip when rising from a chair [14,15]. For testing, a cup was inserted into the aperture in the fluid chamber from above and secured with a set screw. The interface between the cup and the fluid chamber was sealed with a rubber ring gasket so the only egress of fluid was via the apical hole in the cup. The integrity of this watertight seal was verified by pretesting the cups for 1000 cycles before drilling the apical holes. For the piston pumping and diaphragm pumping experiments, the collection cylinder was removed from below the fluid chamber. A pressure transducer (subminiature flush diaphragm, Sensotec, Columbus, Ohio) was threaded into the hole in the 238 aluminum cup from below leaving approximately 1 mL of space between the top of the transducer and the bottom of the polyethylene liner. The pressure transducer had a range from 0 to 10 343 mm Hg and an accuracy of 1%. Fluid pressure and load data were recorded at a rate of 10 Hz.
Table 2. Diaphragm pumping experiment Outcome measured Protocol Load (N) Frequency (Hz) Angle of load Liner
Piston pumping experiment
Pressure Walking 50 to 2000 1 238 Congruent
Pressure Walking 50 to 2000 1 238 Noncongruent
Walking 50 to 2000 1 238 Congruent
Pistoning +100 to 2000 1 238 Congruent
Pump Action Polyethylene ! Walter et al
1045
(Table 1). Five liners were tested in random sequence first by the rising protocol and then by the walking protocol. A preload of 50 N was applied to properly seat the liner in the cup. The fluid chamber was then filled with 200 mL of bovine serum (HyClone, Logan, Utah). The number of cycles to reach every 10 mL of fluid collected in the chamber was recorded, and each run was stopped after 1000 cycles or when 100 mL of fluid was collected, whichever occurred first. Between each test, the apparatus was removed, drained of serum, cleaned with soap and water, and dried thoroughly with compressed air. Diaphragm Pumping Experiment
Fig. 4. Top view of the apparatus.
Two types of liners were machined from HSS reference ultrahigh molecular weight polyethylene [16]. Congruent liners (n = 16) had a hemispherical internal geometry of 28mm diameter and a hemispherical external geometry of 42mm diameter abutting a rim. The noncongruent liners (n = 5) were identical to the congruent liners except that the centers of the hemispheres were displaced 0.3mm relative to the rim resulting in an incongruency of 0.3mm between the polyethylene liner and the aluminum cup at the dome. Both liner types were designed without locking mechanisms. The dimensional tolerance in manufacturing the liners was 0.127 mm. The apparatus was placed on a bearing plate that rested on the platen of an MTS uniaxial load frame (MTS Systems Corp, Eden Prairie, Minn). The bearing plate allowed the apparatus to selfcenter under the axial compressive load, which was delivered to the polyethylene liner through a 28mm metallic femoral head. Ambient conditions were recorded before each test: temperature (range 20.08C to 22.28C) and relative humidity (range 20% to 25%). Pressure Gradient Experiment Two loading protocols, based on published data [6,7,15,17], were used: one to simulate walking and a second to simulate rising from a chair
For this experiment, the walking protocol was used in testing 5 congruent liners and 5 noncongruent liners (Table 2). The apical hole was carefully filled with bovine serum using a syringe to prevent air bubbles, the fluid chamber was filled, and then the polyethylene liner was placed into the aluminum cup. We alternated a congruent liner with a noncongruent liner in random sequence on the first day, and then the order was reversed on the second day of testing (a minimum of 24 hours between tests). Each specimen was run for 100 cycles, and the pressure amplitude was recorded at 90 cycles.
Fig. 5. The cup on the left (shown with a polyethylene liner) is tilted at 238 to simulate the angle of the load on the acetabular component during walking. The cup on the right (shown without a polyethylene liner) is tilted at 468 to simulate the angle of the load on the acetabular component with rising from a chair.
1046 The Journal of Arthroplasty Vol. 20 No. 8 December 2005 Piston Pumping Experiment In this experiment, 2 loading conditions were examined: one with the normal walking protocol and the other walking with pistoning of the bearing. The apparatus was modified by adding a metallic bar supported by 2 threaded rods, positioned to limit upward load frame motion; the test could then be conducted in load control including both tensile and compressive loads. After filling the apparatus with bovine serum as in the diaphragm pumping experiment, a 50N load was applied, and the top bolts on the rods were adjusted to allow an upward displacement of 0.3mm as measured with calibrated shims. The load was then changed to 100 N of tensile force, and the test was run. On the first day of testing, each of 6 liners was tested in random sequence first by the normal walking protocol then by the pistoning protocol. On the second test day, the order of the specimens was reversed, as was the order of the protocols (again with 24 hours between tests). As in the diaphragm pumping experiment, each specimen was run for 100 cycles, and the pressure amplitude was recorded at 90 cycles. Statistical Analyses In the pressure gradient experiment, we compared the rates of flow between the 2 groups. In the diaphragm pumping and piston pumping experiments, we compared pressures. Data were analyzed for significant differences using a paired Student t test when comparing the 2 groups within an experiment and a nonpaired Student t test when comparing results between the diaphragm pump-
Fig. 7. Results from the piston pumping experiment showed significantly greater pressures with pistoning (5140 F 330 mm Hg) than without pistoning of the liners (650 F 300 mm Hg) ( P b .001).
ing and piston pumping experiments. A P value of less than .05 was considered significant.
Results In the pressure gradient experiment, no fluid flow occurred in any of the 5 runs with the walking protocol, but significant flow ( P b .001) at a rate of 127 F 16 mL/min was detected through the apical hole in all 5 runs with the rising protocol. In the diaphragm pumping experiment, the pressure produced by the noncongruent liners was 6 times the pressure produced by the congruent liners (Fig. 6). In the piston pumping experiment, the pressure produced by the pistoning liners was 8 times the pressure produced without pistoning (Fig. 7). No significant difference occurred between the congruent liners in the walking protocol in the piston pumping experiment and the congruent liners in the walking protocol in the diaphragm pumping experiment ( P = .9). Pressures produced by the pistoning liners were higher than pressures produced by the noncongruent liners, but this difference was not significant ( P = .064).
Discussion
Fig. 6. Results from the diaphragm pumping experiment showed significantly greater pressures (4030 F 1250 mm Hg) for the noncongruent liners than for the congruent liners (670 F 240 mm Hg) ( P = .006).
The pressure gradient experiment illustrates that a pressure gradient could pump fluid and wear debris from the hip joint through an apical hole to the retroacetabular bone. The higher pressures and higher load angles when rising from a chair and stair climbing probably make these activities more important in pumping fluid by this mechanism
Pump Action Polyethylene ! Walter et al
than level walking. Our model had a single apical hole but the number, size, and location of holes could all be important factors. The pressure gradient mechanism could also pump fluid and wear debris into the acetabular cancellous bone at the periphery of the shell causing peripheral acetabular osteolysis. Peripheral osteolytic lesions have been observed in association with acetabular components with and without holes [1,3,18], but apical osteolytic lesions typically occur only behind components with holes [18]. Which of these patterns of osteolysis develops in a particular hip is probably determined by which pathway offers the least resistance to the flow of fluid and debris. The diaphragm pumping and piston pumping experiments show 2 different mechanisms for the pumping action of polyethylene liners in cementless metal acetabular components with holes, both of which probably coexist in vivo. The relative importance of each will depend on mechanical factors such as the conformity of the liner to the shell, the ability of the locking mechanism to restrict movement of the liner within the shell, and the nature of the applied load. In diaphragm pumping, the liner is suspended at the rim and noncongruent at the dome. The volume of the space between the liner and shell reduces as the polyethylene deforms with loading and increases again with unloading (the so-called btrampoline effectQ [11]). Fluid may flow in and out through the hole or back past the locking mechanism into and out of the joint. Diaphragm pumping relies on a combination of peripheral capture and nonconformity between the liner and shell but can occur even if the peripheral capture locking mechanism does not allow any motion at the rim. Most locking mechanisms are variations of locking tabs or locking rings located at the periphery of the shell, designed to achieve peripheral capture of the liner. Nonconformity of the liner and shell exists to some degree in all modular acetabular components [10,19]. Apart from tolerance variations, radial clearance between the 2 parts is necessary to allow insertion of the liner [10]. In vitro studies measuring the micromotion between the dome of the shell and the liner in new components show that this motion decreases with time in the majority of designs [12], presumably because of creep of the polyethylene, which improves the congruency with time. Therefore, the diaphragm pumping effect should decrease with time. In piston pumping, the entire liner moves in and out of the shell with loading. Piston pumping relies on a locking mechanism that allows movement. Motion between the liner and the shell at the rim is
1047
invariably present in modular components, even when they are new and has been measured to be between 7.5 and 359 lm [11]. Previous in vitro work on liner micromotion [11,12] used axial loads in compression (sometimes in combination with rotation); as a result, the femoral head remained fully seated in the liner throughout the cycle, and the liner remained seated in the shell. However, in vivo fluoroscopic imaging of unconstrained total hip arthroplasties has shown an average of 1.2 mm of separation of the femoral head from the acetabulum during gait [20]. We are aware of no in vitro study of liner motion that has included separation; therefore, the liner motion that occurs in vivo has been underestimated. Pistoning motion of polyethylene liners in their metallic shells has been observed at revision surgery [2,3,13]. The ability of the locking mechanism to limit motion might deteriorate with time in service due to polyethylene wear and creep within the mechanism itself [21], thereby allowing more pistoning of the liner. Therefore, the piston pumping effect would be expected to increase with time. Our study has limitations, but we do not believe that they compromise our conclusions. For example, we simulated a retroacetabular osteolytic lesion (a cyst in bone) as a cavity in the metal apparatus. Given the greater stiffness of the aluminum than cancellous bone, the absolute pressures that we measured in the piston pumping and diaphragm pumping experiments will not be directly translatable to the in vivo situation, but the differences in the pressures between the noncongruent liners and the congruent liners and between the pistoning liners and the nonpistoning liners are important. Similarly, we used a single load direction and a constant pressure gradient because our primary aim was to show the presence of these mechanisms under physiological type loads. We would hypothesize that if we altered our apparatus to simulate the multidirectional loads and fluctuating pressure gradients that occur in vivo, the pumping would be even more severe. Osteolysis is widely thought to be caused by polyethylene wear debris. However, cyclic fluid pressure changes, perhaps in combination with wear debris, have been implicated in the pathogenesis of osteolysis [22-24]. The mechanisms of pumping that are shown in the current study could be important factors. Each hole in a metallic shell increases the beffective joint spaceQ [4] and creates the potential for particles and pressurized fluid to gain access to the retroacetabular bone, which may in turn increase the likelihood of retroacetabular osteolysis [5,10,25]. Holes in shells
1048 The Journal of Arthroplasty Vol. 20 No. 8 December 2005 are usually unnecessary. Screw holes can be avoided when screws are not used, and dome holes, which facilitate insertion of the component, can be avoided by minor alterations in component design and surgical instrumentation. A welldesigned acetabular component with no holes, an appropriately constraining locking mechanism, and a good quality surface used in conjunction with the press-fit insertion technique could mitigate these pumping mechanisms and reduce the potential for osteolysis.
Acknowledgments
10.
11.
12.
13.
The authors thank Eric Charriere and Robert Streicher of Stryker Clinical and Scientific Affairs Europe for their support of this work.
14.
References
16.
1. Maloney WJ, Galante JO, Anderson M, et al. Fixation, polyethylene wear, and pelvic osteolysis in primary total hip replacement. Clin Orthop 1999; 369:157. 2. Astion DJ, Saluan P, Stulberg BN, et al. The porouscoated anatomic total hip prosthesis: failure of the metal-backed acetabular component. J Bone Joint Surg Am 1996;78:755. 3. Schmalzried TP, Harris WH. The Harris-Galante porous-coated acetabular component with screw fixation. Radiographic analysis of eighty-three primary hip replacements at a minimum of five years. J Bone Joint Surg Am 1992;74:1130. 4. Schmalzried TP, Jasty M, Harris WH. Periprosthetic bone loss in total hip arthroplasty. Polyethylene wear debris and the concept of the effective joint space. J Bone Joint Surg Am 1992;74:849. 5. Huk OL, Bansal M, Betts F, et al. Polyethylene and metal debris generated by non-articulating surfaces of modular acetabular components. J Bone Joint Surg Br 1994;76:568. 6. Robertsson O, Wingstrand H, Kesteris U, et al. Intracapsular pressure and loosening of hip prostheses. Preoperative measurements in 18 hips. Acta Orthop Scand 1997;68:231. 7. Hendrix RW, Wixson RL, Rana NA, et al. Arthrography after total hip arthroplasty: a modified technique used in the diagnosis of pain. Radiology 1983; 148:647. 8. Downey DJ, Simkin PA, Taggart R. The effect of compressive loading on intraosseous pressure in the femoral head in vitro. J Bone Joint Surg Am 1988; 70:871. 9. Scott DL, Campbell PA, McClung CD, et al. Factors contributing to rapid wear and osteolysis in hips
15.
17.
18.
19.
20.
21.
22.
23.
24.
25.
with modular acetabular bearings made of hylamer. J Arthroplasty 2000;15:35. Young AM, Sychterz CJ, Hopper Jr RH, et al. Effect of acetabular modularity on polyethylene wear and osteolysis in total hip arthroplasty. J Bone Joint Surg Am 2002;84:58. Fehring TK, Smith SE, Braun ER, et al. Motion at the modular acetabular shell and liner interface. A comparative study. Clin Orthop 1999;367:306. Williams II VG, Whiteside LA, White SE, et al. Fixation of ultrahigh-molecular-weight polyethylene liners to metal-backed acetabular cups. J Arthroplasty 1997;12:25. Walter WL, Walter WK, O’Sullivan M. The pumping of fluid in cementless cups with holes. J Arthroplasty 2004;19:230. Berme N, Paul JP. Load actions transmitted by implants. J Biomed Eng 1979;1:268. Bergmann G, Deuretzbacher G, Heller M, et al. Hip contact forces and gait patterns from routine activities. J Biomech 2001;34:859. Bennett AP, Wright WT, Li S. Global reference UHMWPE: characterization and comparison to commercial UHMWPE. 42nd Annual Meeting of Orthopaedic Research Society, Atlanta, Georgia, 1996. Pedersen DR, Brand RA, Davy DT. Pelvic muscle and acetabular contact forces during gait. J Biomech 1997;30:959. Claus AM, Sychterz CJ, Hopper Jr RH, et al. Pattern of osteolysis around two different cementless metalbacked cups: retrospective, radiographic analysis at minimum 10-year follow-up. J Arthroplasty 2001; 16:177. Kurtz SM, Ochoa JA, White CV, et al. Backside nonconformity and locking restraints affect liner/ shell load transfer mechanisms and relative motion in modular acetabular components for total hip replacement. J Biomech 1998;31:431. Lombardi Jr AV, Mallory TH, Dennis DA, et al. An in vivo determination of total hip arthroplasty pistoning during activity. J Arthroplasty 2000;15:702. Collier JP, Mayor MB, Jensen RE, et al. Mechanisms of failure of modular prostheses. Clin Orthop 1992; 285:129. Ferrier GM, McEvoy A, Evans CE. The effect of cyclic pressure on human monocyte-derived macrophages in vitro. J Bone Joint Surg Br 2000;82:755. Skripitz R, Aspenberg P. Pressure-induced periprosthetic osteolysis: a rat model. J Orthop Res 2000; 18:481. Van Der Vis HM, Marti RK, Tigchelaar W, et al. Benign cellular responses in rats to different wear particles in intra-articular and intramedullary environments. J Bone Joint Surg Br 1997;79:837. Kilgus JTJ. Retro-acetabular osteolysis and polyethylene access: a radiographic analysis of uncemented acetabular cups with 2-14 year follow-up. 2001 [San Francisco].