Cement porosity has been implicated in the early loosening of cemented total hip arthroplasties. 3,5,6,11,17,19,21 Improved hand mixing technique, 8 vacuum mixing, 7,23,24,25,37,45 and centrifugation 4,20 all have been successful in reducing bulk cement porosity. Numerous studies have shown that reduction of voids improves the mechanical strength of bulk cement. 4,5,30,31,43–45 However, James et al 18,19 reported that although centrifugation reduced porosity in bulk cement, it was not effective in preventing porosity at the cement-implant interface. In fact, they reported the porosity at the prosthesis interface was four to five times greater than the porosity in bulk cement after centrifugation. 18,19
There is evidence the cement-implant interface porosity results in biomechanical weakening and diminished longevity of cemented hip arthroplasties. 6,11,17 Several finite element model studies suggest that the cement-implant interface is the mechanical weak link of the cemented femoral prosthesis. 14,15,32 Improving the strength of this interface may increase the longevity of cemented implants. In a postmortem retrieval of 16 hip prostheses, Jasty et al 21 concluded that cement to prosthesis debonding occurred early, and cement mantle fractures extending from the cement-prosthesis interface commonly appeared later in the natural history of a hip implant. In contrast, bone to cement debonding was uncommon. Davies et al 6 found that decreasing the cement-prosthesis interface porosity resulted in increased cement to implant torsional strength. Investigators are looking for ways to reduce cementimplant interface porosity. Davies et al 6 recommended passage of the implant through a diaphragm to prevent air entrapment during implantation. Bishop et al 2 attempted to decrease cement shrinkage from the prosthesis by inserting a heated implant. Although factory precoating implants 6,12,41 with polymethylmethacrylate eliminates cement-prosthesis interface voids, these voids still appear at the new polymethylmethacrylate precoat interface.
This study evaluated whether vacuum mixing reduces cement-prosthesis interface porosity and improves interface mechanical properties. Four different cements were tested.
MATERIALS AND METHODS
Forty-eight blast finished steel rods of 13 mm diameter were implanted sequentially into tubes of 17 mm inner diameter that had been retrograde filled with polymethylmethacrylate cement, producing a 2-mm cement mantle. Vacuum mixed samples were prepared with an Osteobond cartridge vacuum mixing system with long extension (Zimmer, Warsaw, IN). Open bowl mixing was performed with a disposable mixing bowl and spatula (Zimmer). Retrograde filling of cement was done with a Miller Cement Delivery Cartridge and Miller cement injector (Zimmer). Custom-made proximal and distal rod centralizers were used. The tubes of polyvinyl chloride pipe were preheated to 37° C and maintained at 37° C for 15 minutes after cement mixing. There were eight cementing conditions. Six implantation teams simultaneously performed six replications of each cement condition, producing 48 constructs. Condition variables were cement type (Simplex, Howmedica, Rutherford, NJ; Palacos R, Smith and Nephew Richards, Memphis, TN; Osteobond, Zimmer, Warsaw, IN; or Dough Type, Zimmer) and mixing style (open bowl or vacuum). The rods randomly were assigned identification code numbers to blind the investigators to the exact condition of the specimen during processing and testing of the rods after implantation. During the implantation process, the room temperature remained between 22° C and 24° C.
The 48 samples were processed and analyzed as described to determine the effects of vacuum mixing and cement brand on interface porosity, rod pushout strength, and fatigue resistance. Each tube was cut transversely three times to obtain three test segments of approximately 2 cm in length and two end pieces. The specimens were cut using a water jet cutter with an abrasive as the cutting means. This eliminated any heat effects that would have been a factor in other cutting processes. The end pieces containing the centralizers were discarded. The proximal test piece was used for pushout testing. The middle test piece was used for fatigue analysis, and the distal test piece was used for cement-implant interface porosity analysis.
The cement mantle of the distal segment of the cement to implant construct was cut through longitudinally. The cement was pried from the rod. The two halves of cement mantle were painted with ink. The surface was sanded with fine paper, leaving dye in the interface pores. Leco Corporation (St Joseph, MI, 2001 Image Analysis System; Software, Main V1.10) was used to determine the percent surface area of the cement interface occupied by pores. The entire surface was analyzed by evaluating four quadrant surface areas. Any quadrant containing a large filling defect was excluded from analysis. An unusually strong adherence was found between the Zimmer Dough Type cement and the rod implants; none of the 12 cement mantles could be pried from the rod without first fracturing the cement. Thus, porosity data for the Zimmer Dough Type samples were obtained from the interfaces of the pushout samples. This was not the case with the other cements, which pried cleanly away from the rods. The cement interface of the Zimmer Dough Type pushout samples separated cleanly from the rod with no observed damage to the cement. This allowed for porosity data to be obtained on the Zimmer Dough Type.
Cycles to failure data of the middle segments was obtained from cyclic compression loading of the cement to implant segments using an MTS Model (Instron, Canton, MA) at 3 Hz and 450 pounds maximum load (R = 0.1) in a 37° C water bath 810 Material Testing System. Testing of each sample halted at mechanical failure or 500,000 loading cycles (Fig 1).
Shear strength determinations were calculated in pounds per square inch (psi) after recording the maximum load to failure of the implant from the cement mantle using an Instron Corporation (Canton, MA) Model SSR1125 Series IX Automated Materials Testing System. Sample lengths were obtained at four corners of the tube and averaged. The pushout tests were performed using the same setup as the fatigue test but performed on the Instron at a cross head speed of 0.2 inches per minute.
The Student’s t test and analysis of variance (ANOVA) were performed to identify significance between groups. Pearson’s correlation coefficient analysis also was performed on the data to uncover any positive correlation between decreased porosity and pushout or fatigue strength.
Rod pushout test data were obtained for each of the eight conditions. The mean pushout shear for all of the samples combined was 625 psi. There was no significant difference between the mean shear force for all open bowl mixed samples combined (621 ± 205 psi) and all vacuum mixed samples combined (629 ± 83 psi; p = 0.4154) (Fig 2). There also was no significant difference in the mean shear force of open bowl mixed and vacuum mixed Simplex samples (624 ± 96 versus 636 ± 78; p = 0.9362) or Osteobond samples (538 ± 35 versus 542 ± 23; p = 0.8102). Mean shear force for Palacos vacuum mixed samples was greater than for Palacos open bowl samples (634 ± 47 versus 423 ± 171; p < 0.0131). Mean shear force for Zimmer Dough Type open bowl samples was significantly greater than that for Zimmer Dough Type vacuum mixed samples (901 ± 71 versus 705 ± 82; p < 0.0082). Analysis of variance showed that the Zimmer Dough Type open mixed samples had a significantly greater pushout resistance than did any of the other conditions (p < 0.05) (Fig 2).
Image analysis showed no significant difference in the percent surface area occupied by voids at the cement-implant interfaces between vacuum mixed versus open bowl mixed samples when all cements are combined. Vacuum mixing reduced interface voids in Palacos R (4.9% ± 4.2% versus 12.5% ± 6.3%; p < 0.0453), and Osteobond (20.7 ± 2.8 versus 14.4 ± 1.8; p < 0.0082) but did not reduce interface voids in Simplex R (15.5 ± 4.7 versus 10.3 ± 4.9; p < 0.0656) or Zimmer Dough Type (21.5 ± 6.1 versus 19.5 ± 4.0; p = 0.6889) for vacuum mixed versus open bowl mixed samples, respectively (Fig 3).
Fatigue analysis of the 48 samples showed high intragroup variability and low intergroup variability. However, vacuum mixed Palacos survived cyclic testing to 500,000 cycles in four of five specimens. All other cement types had one survivor to 500,000 cycles, except Osteobond open bowl mixed, which had two survivors to 500,000 cycles.
Statistical analysis did not find a positive correlation between decreased number of pores and increased pushout strength.
Newer prosthetic materials, prosthesis designs, and modern cementing techniques have been cited as reasons for the dramatic improvement in femoral implant longevity. 29 The incidence of femoral loosening at 10 years is 3% or less. 22,23,29,30,35,36,46,47 Although some early studies showed femoral loosening rates that were 10 times greater than this, 40,42 Schulte et al 38 published 20-year results on the Charnley total hip prosthesis, implanted between 1970 and 1972, showing 7% femoral component loosening. It becomes important to question which changes in materials and cementing technique have been effective in reducing the incidence of femoral loosening. The current study evaluated the impact of vacuum mixing on four brands of commonly used bone cement on implant-cement interface porosity, pushout strength, and shear fatigue. Although numerous studies have examined the mechanical properties of bulk cement and the relative effect of internal voids, 3,4,7,20,23,24,25,31,34,43,45 the authors think the current cement to implant model more closely replicates the in vivo hip. The authors of the current study used a commercial cement injection gun to retrograde fill hollow tubes kept at 37° C with cement. The introduction of an implant into the cement filled tubes introduces turbulence, produces air voids, and allows the biomechanical testing of the cement-implant interface.
Jasty et al 21 described the clinical importance of the cement-implant interface. In a postmortem retrieval study of 16 hip prostheses they reported that debonding occurred preferentially at the cement-prosthesis interface, particularly about thin mantle areas, cement voids, and adjacent to sharp prosthesis corners. Cement mantle and prosthesis loosening is thought to be a biomechanical issue. 1,4, 9,21,26,36 Shear forces and cyclic stresses cause cement fatigue and eventual fragmentation of the cement mantle. In contrast, cement-bone interface loosening is usually thought to be a biologic phenomena. As wear debris accumulate and migrate through the bonecement interface, a biologic scavenger response is activated. The cellular activity responding to the debris produce focal osteolysis and a bone-cement interface membrane. 10,13,28,33
Cement porosity has been cited as a key factor in aseptic loosening of cemented total hip arthroplasties. However, most of the mechanical data on cement reflect work on molded cement bars, rather than on stem and cement constructs. The relationship between fracture initiation and internal cement voids was shown experimentally by James et al. 17 Some in vitro studies have shown that high porosity cement is mechanically weaker than low porosity cement in static and fatigue testing. Eyerer and Jin 8 found that stirring after the initial mixing period significantly increased final cement porosity in Palacos R bulk samples. They recommended mixing for 30 seconds at 10 rpm, followed by a period of sitting and gentle kneading so additional air is not trapped in the cement. In 1983, Demarest et al 7 described the effective use of vacuum mixing cement to reduce cement porosity; in 1984, Burke et al 4 described the use of centrifugation of cement to reduce cement porosity, reporting that centrifugation caused a 24% increase in tensile strength, a 54% increase in ultimate tensile strain, and a 136% increase in fatigue life.
O’Connor et al 31 found that centrifuged Simplex P has a significantly higher fatigue life than does uncentrifuged cement. At low strain levels, they found none of the centrifuged samples fractured by 10 million cycles, whereas 70% of the uncentrifuged samples failed. Jasty et al 20 reported that Simplex P attained greater than 50% reduction in bulk porosity after 30 seconds of centrifugation and 25% reduction in porosity after 2 minutes. In contrast, centrifugation of Palacos R and low viscosity cements did not result in reduction of bulk porosity. Precooling the cement monomer to 0° C increased the porosity of Simplex P and was ineffective in reducing porosity in Palacos and low viscosity cements. Rimnac et al 34 found no improvement in ultimate strength when Zimmer and Palacos R cements were centrifuged during preparation. Schreurs et al 37 found centrifugation to be ineffective in reducing the porosity of Palacos R and Zimmer low viscosity cements. In addition, Skinner and Murray 39 reported that centrifugation of Simplex P leads to separation of barium and creation of a density gradient in the cement.
There are many in vitro studies on the effect of vacuum mixing cement to improve the mechanical properties of cement. Numerous authors 16,23,24,44,45 reported that vacuum mixing at 500 to 550 mm Hg effectively decreases bulk cement porosity and improves compression strength, tension strength, and fatigue strength. They described the importance of maintaining vacuum pressure below 600 mm Hg to prevent the monomer from boiling, which would increase the number of voids. 44,45 Wixson et al 44,45 found that, although vacuum mixing and centrifugation are effective means for reducing the porosity of Simplex P cement, vacuum mixing Simplex P cement is significantly more effective than is centrifugation. Wixson et al 44,45 reported that vacuum mixed Simplex P is 24% stronger than open bowl mixed Simplex P in compression, 23% stronger in diametral tension, 44% stronger in uniaxial tension, and significantly stronger in fatigue. Lidgren et al 24 found a significant increase in compressive and flexion strength with vacuum mixing of Palacos Refobacin and low viscosity cement. Cooling the monomer to 4° C did not reduce voids with vacuum mixing or hand mixing. 24 Kummer 23 reported that vacuum mixing results in decreased porosity, a 24% increase in flexural strength, and marked increase in the fatigue strength of bone cements. Lindén 25 found that vacuum mixing increased fatigue life threefold for Simplex cement. Lidgren et al 24 found that the fatigue life of Refobacin Palacos was 10 times greater with vacuum mixing, and the number of voids was reduced to ⅓.
James et al 19 first described cement-prosthesis interface porosity and reported that centrifugation, which reduces porosity of bulk Simplex P cement, 4 was not effective in reducing interface porosity. In their samples, voids occupied 20% to 50% of the cement-prosthesis interface surface area, whereas voids occupied only 5% to 16% of the surface area of slices from bulk cement specimens. James et al 19 also thought that the cement-prosthesis interface voids are created when air is dragged into the cement during insertion of the prosthesis (rheologic behavior), and they discounted the idea that interface voids are created by volatilization of the monomer. They also found that precoating the stem with polymethylmethacrylate resulted in significant interface porosity between the precoated cement and the new cement. Although polymethylmethacrylate precoated specimens have been shown to have improved mechanical properties at the cement-prosthesis interface, Gardiner and Hozack 12 published a report of early precoat prosthesis loosening occurring at the cement-bone interface.
The current study is the first to evaluate the effect of vacuum mixing cement on the mechanical properties of the cement-implant construct with four different cement types, simulating the clinical situation. Unlike the constructs used in most investigations, which conducted mechanical tests on bulk cement bars, the stem cement construct used in the current study closely replicates the clinical situation. The data show that vacuum mixing is not effective in reducing interface porosity. These data correlate well with the observations of James et al, 18,19 who found that centrifugation is not effective in reducing cement-implant interface porosity. James et al 17 examined the effects of bulk cement porosity on fatigue life. Interfacial porosity was not included in the experimental design. In 1993, James et al 18 characterized the porosity at the cement-implant interface in in vitro and in vivo samples. No mechanical analysis was performed. 18 Even more striking is the observation that vacuum mixing is ineffective in improving implant pushout strength and fatigue strength. Despite that vacuum mixing effectively improves the mechanical properties of isolated molded cement cylinders, vacuum mixing is not effective in improving the mechanical characteristics of the cement-implant construct. This observation implies that in the clinical situation, vacuum mixing may not improve the strength of fixation or longevity of a cemented femoral prosthesis. The type of cement used had a large impact on the results of the current study. Zimmer Dough Type had a significantly greater pushout strength than did the other cements. Because of the unusually strong adherence of the Zimmer Dough Type cement for the blast rod, the current authors were unable to separate doughy cement from the blast rod for porosity analysis without fracturing the cement mantle. As a result, the pushout samples were used to obtain porosity data for the Zimmer Dough Type constructs. The pushout samples characteristically had a clean separation of the implant from the rod and provided a good source for obtaining porosity data that otherwise would be unavailable.
The study was unable to show a relationship between interface porosity and pushout or fatigue strength when all cement types were examined. However, Palacos R cement and Osteobond had a decrease in porosity when vacuum mixing was used, and Palacos R was the only cement to have an improvement in pushout and fatigue strength when vacuum mixing was used. It is unclear why vacuum mixing had a significant effect only on Palacos R cement. Davies et al 6 has used a diaphragm to prevent air from being dragged into the cement during implantation. Perhaps the high viscosity of Palacos prevents air from being brought into the cement interface, allowing the effects of vacuum mixing to be seen. Palacos is not an easy cement to vacuum mix. Because of its high viscosity, it is difficult to drive out of a syringe. The Palacos monomer was chilled before mixing to retain it in a lower viscosity state during vacuum mixing. Lidgren et al 24 showed that chilling Palacos monomer may improve the toughness and fatigue life of Palacos R. In contrast, Davies et al 5 found that chilling Simplex P monomer significantly decreased the fatigue life of this cement. Although the current review and another study by Davies et al 6 suggest that decreases in interfacial porosity do result in greater interface strength, more work is needed. The model used in the current experiment involved a blast surface implant. These findings are preliminary data on matte finish stems. They cannot be extrapolated to polished, smooth, or precoated implants. Additional work with other surface treatments is needed because other surface treatments for hip stems may behave differently than do matte finishes.
Vacuum mixing does not improve mechanical pushout or fatigue strength at the blast surface implant-cement interface except in Palacos R. Although vacuum mixing reduces porosity of isolated cement cylinders, it does not reduce interface porosity except in Palacos R and Osteobond. The added expense of using a vacuum mixing system in the authors’ hospital is $106.71 per case. This amounts to approximately $40,000 in additional cost per year at the authors’ institution.
1. Ahmed AM, Raab S, Miller JE: Metal cement interface strength in cemented stem fixation. J Orthop Res 2:105–118, 1984.
2. Bishop NE, Ferguson S, Tepic S: Porosity reduction in bone cement at the cement-stem interface. J Bone Joint Surg 78B:349–356, 1996.
3. Bragdon CR, Biggs S, Mulroy WF, Harris WH: Defects in the cement mantle: A fatal flaw in cemented femoral stems for THR. Trans Orthop Res Soc 20:710, 1995.
4. Burke DW, Gates EI, Harris WH: Centrifugation as a method of improving tensile and fatigue properties of acrylic bone cement. J Bone Joint Surg 66A: 1265–1273, 1984.
5. Davies JP, Jasty M, O’Connor DO, et al: The effect of centrifuging bone cement. J Bone Joint Surg 71B:39–42, 1989.
6. Davies JP, Kawate K, Harris WH: Effect of interfacial porosity on the torsional strength of the cement-metal interface. Trans Orthop Res Soc 20:713, 1995.
7. Demarest VA, Lautenschlager EP, Wixson RL: Vacuum mixing of acrylic bone cement. Transactions of the Ninth Annual Meeting of the Society for Biomaterials. Birmingham, AL 37, 1983.
8. Eyerer P, Jin R: Influence of mixing technique on some properties of PMMA bone cement. J Biomed Mater Res 20:1057–1094, 1986.
9. Fornasier VL, Cameron HU: The femoral stem/cement interface in total hip replacement. Clin Orthop 116:248–252, 1976.
10. Freeman MAR, Bradley GW, Revell PA: Observations upon the interface between bone and polymethylmethacrylate cement. J Bone Joint Surg 64B:489–493, 1982.
11. Freitag TA, Cannon SL: Fracture characteristics of acrylic bone cements. II. Fatigue. J Biomed Mater Res 11:609–624, 1977.
12. Gardiner RC, Hozack WJ: A consequence of strengthening the cement-prosthesis interface? Failure of the cement-bone interface. J Bone Joint Surg 76B:49–52, 1994.
13. Goldring SR, Schiller AL, Roelke M, et al: The synovial-like membrane at the bone-cement interface in loose total hip replacements and its proposed role in bone lysis. J Bone Joint Surg 65A:75–584, 1983.
14. Harrigan TP, Harris WH: A three-dimensional non-linear finite element study of the effect of cement-prosthesis debonding in cemented femoral total hip components. J Biomech 24:1047–1058, 1991.
15. Harrigan TP, Karch JA, O’Connor DO, Burke DW, Harris WH: A finite element study of the initiation of failure of fixation in cemented femoral total hip components. J Orthop Res 10:134–144, 1992.
16. Horowitz SM, Doty SB, Lane JM, Burstein AH: Studies of the mechanism by which the mechanical failure of polymethylmethacrylate leads to bone resorption. J Bone Joint Surg 75A:802–813, 1993.
17. James SP, Jasty M, Davies J, Piehler H, Harris WH: A fractographic investigation of PMMA bone cement focusing on the relationship between porosity reduction and increased fatigue life. J Biomed Mater Res 26:651–662, 1992.
18. James SP, Karydas D, McGarry FJ, Harris WH: Reduction of the extensive porosity in the cement at the femoral component/bone cement interface. Trans Orthop Res Soc 18:520, 1993.
19. James SP, Schmalzried TP, McGarry FJ, Harris WH: Extensive porosity at the cement-femoral prosthesis interface: A preliminary study. J Biomed Mater Res 27:71–78, 1993.
20. Jasty M, Jensen JF, Burke DW, Harrigan TP, Harris WH: Porosity measurement in commercial bone cement preparation and the effect of centrifugation of porosity reduction. Trans Orthop Res Soc 10:239, 1985.
21. Jasty M, Maloney WJ, Bragdon CR, et al: The initiation of failure in cemented femoral components of hip arthroplasties. J Bone Joint Surg 73B:551–558, 1991.
22. Joshi AB, Porter ML, Trail IA, et al: Long term results of Charnley low-friction arthroplasty in young patients. J Bone Joint Surg 75B:616–623, 1993.
23. Kummer FJ: Improved mixing of bone cements. Trans Orthop Res Soc 10:238, 1985.
24. Lidgren L, Bodelind B, Möller J: Bone cement improved by vacuum mixing and chilling. Acta Orthop Scand 57:27–32, 1987.
25. Lindén U: Fatigue properties of bone cement. Comparison of mixing techniques. Acta Orthop Scand 60:431–433, 1989.
26. Mann KA, Bartel DL, Wright TM: Cement stresses in a femoral hip component with coulomb friction at the stem-cement interface. Trans Orthop Res Soc 16:107, 1991.
27. McCoy TH, Salvati EA, Ranawat CS, Wilson Jr PD: A fifteen-year follow-up study of one hundred Charnley low-friction arthroplasties. Orthop Clin North Am 19:467–476, 1988.
28. Mirra JM, Marder RA, Amstutz HC: The pathology of failed total joint arthroplasty. Clin Orthop 170:174–183, 1982.
29. Mulroy Jr RD, Harris WH: The effect of improved cementing techniques on component loosening in total hip replacement. J Bone Joint Surg 72B:757–760, 1990.
30. Noble PC, Mirkovic S, Davidson JP: The distribution of porosity in acrylic bone cement. Trans Orthop Res Soc 10:242, 1985.
31. O’Connor DO, Burke DW, Davies JP, Harris WH: S-N curve for centrifuged and uncentrifuged PMMA. Trans Orthop Res Soc 10:325, 1985.
32. Raab S, Ahmed AM, Provan JW: The quasistatic and fatigue performance of the implant/bone cement interface. J Biomed Mater Res 15:159–182, 1981.
33. Radin EL, Rubin CT, Thrasher EL, et al: Changes in the bone-cement interface after total hip replacement. J Bone Joint Surg 64A:1188–1200, 1982.
34. Rimnac CE, Wright TM, McGill DL: The effect of centrifugation on the fracture properties of acrylic bone cements. J Bone Joint Surg 68A:366–371, 1987.
35. Russotti GM, Coventry MB, Stauffer RN: Cemented total hip arthroplasty with contemporary techniques. Clin Orthop 235:141–147, 1988.
36. Schmalzried TP, Kwong LM, Jasty M, et al: The mechanism of loosening of cemented acetabular components in total hip arthroplasty. Clin Orthop 274:60–78, 1992.
37. Schreurs W, Spierings PTJ, Huiskes R, Slooff TJJH: Effects of preparation techniques on the porosity of acrylic cements. Acta Orthop Scand 59:403–409, 1988.
38. Schulte KR, Callaghan JJ, Kelley SS, Johnston RC: The outcome of Charnley total hip arthroplasty with cement after a minimum twenty-year follow-up. J Bone Joint Surg 75A:961–975, 1993.
39. Skinner HB, Murray WR: Density gradient in bone cement after centrifugation. Trans Orthop Res Soc 10:243, 1985.
40. Stauffer RN: Ten-year follow-up study of total hip replacement. J Bone Joint Surg 64A:983–990, 1982.
41. Stone MH, Wilkinson R, Stother IG: Some factors affecting the strength of the cement-metal interface. J Bone Joint Surg 71B:217–221, 1989.
42. Sutherland CJ, Wilde AB, Borden LS, Marks KE: A ten-year follow-up of one hundred consecutive Müller curved-stem total hip-replacement arthroplasties. J Bone Joint Surg 64A:970–982, 1982.
43. Topoleski LD, Ducheyne P, Cuckler JM: A fractographic analysis of in-vivo poly (methylmethacrylate) bone cement failure mechanisms. J Biomed Mater Res 24:135–154, 1990.
44. Wixson RL, Lautenschlager EP, Novak M: Vacuum mixing of methylmethacrylate bone cement. Trans Orthop Res Soc 10:327, 1985.
45. Wixson RL, Lautenschlager EP, Novak MA: Vacuum mixing of acrylic bone cement. J Arthroplasty 2:141–149, 1987.
46. Wroblewski BM, Siney PD: Charnley low-friction arthroplasty in the young patient. Clin Orthop 285:45–47, 1992.
47. Reference not provided.