REVIEW ARTICLE The use of cemented and cementless femoral components in revision total hip arthroplasty: a review of the literature

Transcription

1 Hong Kong Journal of Orthopaedic Surgery 2002;6(1): Femoral components in revision total hip arthroplasty REVIEW ARTICLE The use of cemented and cementless femoral components in revision total hip arthroplasty: a review of the literature Division of Joint Replacement Surgery, Department of Orthopaedic Surgery, The University of Hong Kong, Hong Kong. ABSTRACT The choices of fixation in revising failed femoral components remain controversial. The early to mid-term rerevision rates for aseptic loosening of the revised femoral components ranged from 2% to 14% with early cementing techniques. They dropped to 0% to 7% with improved cementing techniques, but the re-revision rates increased with time, and the rates of radiological loosening were still high. Early to mid-term re-revision rates ranged from 0% to 16% with proximally porous-coated femoral components, which also carried a significant intraoperative fracture risk that ranged from 12% to 46%. The most reliable option in femoral revision is probably the extensively porous-coated femoral component, the re-revision rates for which ranged from 0 to 6% after an average follow-up of 4 to 13 years. Key Words: Arthroplasty, replacement, hip; Femur/radiography; Hip prosthesis; Prosthesis failure!!"#$%&'()*+,-)*+./ !"#$%&'()*+,-./ %+:;&6<"#=>ob NQB!"!"#$%&'()*+,-%MB TB!"#$%&'()*+, u!"#$%&'!"#$%&'()*+,-./01mb NSB!"#$%&NOB NSB!"#$%&'!"Q NP!"#$%&MB SB INTRODUCTION As the number of primary total hip arthroplasties has increased and patients are living and functioning longer, orthopaedic surgeons will inevitably face more late failures that require revisions. The choices of fixation in revising failed femoral components still remain controversial. In this article, the results of cemented revision using early and improved cementing techniques, and those of proximally and extensively porous-coated cementless femoral components will be reviewed. During the review process, it was not uncommon to find that some authors inappropriately interpreted the data when they summarised the results of the earlier reports. There are several possible reasons for this. In some early reports, it was difficult or even impossible to figure out the failure rate of the femoral revision alone, because the failure rate of the revised hip but not the femoral side itself was reported. The indexed revision could have been an isolated acetabular revision; it was thus unclear whether the loose femoral component was inserted in the primary setting or whether it was a failure of the revision. The criteria for clinical failure also differed across reports. In the context of fixation, it was important to compare only the re-revi- Correspondence: Dr. K.Y. Chiu, Division of Joint Replacement Surgery, The University of Hong Kong, Queen Mary Hospital, Pokfulam, Hong Kong Hong Kong Orthopaedic Association & Hong Kong College of Orthopaedic Surgeons. 47

2 HKJOS sion rates from the aseptic loosening of the femoral components, and re-operations for other reasons, such as infection or dislocation, were not considered relevant. In this review, the searched articles are carefully scrutinised in order to calculate the real re-revision rates of aseptic loosening for femoral components. The criteria for diagnosing radiographic loosening were also not standardised in the earlier reports. In this review, the searched articles are thoroughly examined to ensure a fair comparison between the articles different radiographic loosening rates. Their descriptions were matched to well-recognised criteria that radiographically depict failures of the femoral components. Definite and probable loosening of cemented femoral components as defined in the criteria proposed by Harris 12,13 and unstable fixation of the cementless femoral components as defined in the criteria proposed by Engh et al 7 were considered radiographic failures. FEMORAL REVISION USING EARLY CEMENTING TECHNIQUES The early results of cemented revision were far from satisfactory. Hunter et al 17 reviewed 140 cemented revision total hip arthroplasties performed at the University of Western Ontario and University of Toronto. The outcomes were poor: 32% were infected, 22% needed to be converted to excision arthroplasty, and 51% were rated as fair or poor results according to the Harris hip score. The reviewers did not report the length of average follow-up, the radiographic loosening rate, nor the re-revision rate of the femoral components. Pellici et al 35 studied 110 cemented revision total hip arthroplasties performed in the Hospital for Special Surgery, New York, and in Brigham and Women s Hospital, Boston. They reported that 12% of the revised femoral components had to be re-revised for aseptic loosening after an average follow-up of only 3.4 years. In a later report, the re-revision rate for aseptic loosening increased slightly to 14%, and as many as 37% of the femoral components showed definite or probable loosening in radiographs at an average 8.1 years follow-up. 36 The best results achieved in femoral revision using early cementing techniques were reported by Turner et al. 43 They studied 105 patients who had cemented femoral revisions performed at the New England Baptist Hospital in Boston. After an average follow-up of 6.7 years, 2.2% of the femoral components were re-revised for aseptic loosening, and only 6.5% showed definite or probable loosening in the radiographs. The reason for such good results could be related to the routine use of long-stem femoral components. The latter bypassed the deficient proximal femur and permitted cement fixation into normal bone distally. Retpen and Jensen 41 supported the finding that the use of long-stem femoral components led to significantly lower mechanical failure rates. Excluding the above-mentioned reports that represented the best and the worst of the relevant studies, femoral re-revision for aseptic loosening was needed in 3.1% to 8.8% of patients after an average follow-up of 2.1 to 8.9 years (Table 1). 1,5,21,22,28,38,40 The actual re-revision rates may be higher, given that several reports included some femoral revisions done with improved cementing techniques (Table 2). 28,38,40 The results of early and improved cementing techniques, however, were not separately analysed in these studies. Radiographic loosening was much more common. Most reports described that 20% to 40% of the femoral revisions were definitely or probably loose. This situation may reflect the difficulty in achieving cement interdigitation into the smooth, sclerotic endosteal surface after removal of the failed femoral component. The fact that the proximal femur was usually defective and that the femoral shaft was frequently perforated also made cement pressurisation more difficult. Such cases could progress to clinical failures with time. Nevertheless, there was no report on femoral revision using early cementing techniques with more than 10 years followup. FEMORAL REVISION USING IMPROVED CEMENTING TECHNIQUES It was hoped that improved cementing techniques might improve the results of femoral revisions. Callaghan et al 3 studied the results of revision total hip arthroplasty performed at the Hospital for Special Surgery, New York, from 1979 to The improved techniques included the use of an intramedullary plug and a cement gun. After an average follow-up of 3.6 years, 4.4% of femoral components were re-revised. This rate was much lower than the 11.8% re-revision rate after 3.4 years average follow-up reported by Pellici et al, 35 who used early cementing techniques from 1972 to However, radiographic loosening still occurred in 25.7% of femoral components with improved cement- 48

3 Femoral components in revision total hip arthroplasty Table 1 Results for revision femoral components inserted with early cemented techniques. Study Prosthesis used No. of Age, years Follow-up, Use of long Definite and probable Revised for aseptic patients years stem, % loosening, % loosening, % Amstutz et al, 1982 Mixed, 3 types Pellici et al, 1982 NA NA Kavanagh et al, 1985 NA NA Pellici et al, 1985 NA NA Turner et al, 1987 NA Engelbrecht et al, 1990 Mixed, 4 types NA Marti et al, 1990* Weber NA Kershaw et al, 1991 Mixed, 4 types Pierson & Harris, 1994* Mixed, 5 types Raut et al, 1996* 97% Charnley NA NA = not available *Improved cementing techniques were used in some patients. Failure of the hip, no separate figure for the failure of femoral component alone. AlloPro, Baar, Switzerland. Depuy, Leeds, UK. Table 2 Results from revision femoral components inserted with improved cemented techniques. Study Prosthesis used No. of Age, Follow-up, Use of long Definite and probable Revised for aseptic patients years years stem, % loosening, % loosening, % Callaghan et al, 1985 NA Rubash & Harris, % HD-2* % calcar replacement* Izquierdo & Northmore- 89% Charnley NA NA 0 Ball, 1994 Estok II & Harris, % HD-2* % calcar replacement* Pierson & Harris, 1995 Mixed, 4 types Mulroy & Harris, % HD-2* % calcar replacement* Katz et al, % Iowa % Charnley Eisler et al % Charnley Hultmark et al % Charnley % Spectron NA = not available *Howmedica (Rutherford, NJ). Depuy (Leeds, UK). Zimmer (Warsaw, IN). Manufacturer not mentioned. ing techniques, which was only marginally better than the 30% incidence in the earlier report. The best results achieved with improved cementing techniques were reported by Izquierdo and Northmore- Ball 19 from the Robert Jones and Agnes Hunt Orthopaedic Hospital in Oswestry, England. A plastic cement plug was used in every hip to allow for cement pressurisation. Of the 148 cemented revisions, Charnley femoral prosthesis (Depuy, Leeds, UK) was used in 89% of their patients, and 38% were one-stage revisions for previously infected hips. After an average follow-up of 6.5 years, one acetabular component was revised for loosening, and four hips were infected. There was no report of femoral revision for aseptic loosening. The reviewers did not mention how many of the femoral 49

4 HKJOS were shown to have subsided 0.2 to 5.5 mm when examined with roentgen stereophotogrammetric analysis. The amount of subsidence correlated well with the degree of bone loss before the revision. Figure 1 The outcomes of femoral revisions with early cementing techniques (open circles) and improved cementing techniques (close circles) are compared. The re-revision rates are plotted against the respective average follow-up periods. components used were long stemmed. Similar to the case in early cementing techniques, the use of long stems inserted with improved cementing techniques reduced the mechanical failure rates. Hultmark et al 15 reported that the cumulative survival rates at 10 years, using re-revision or radiographic loosening as end points, were 65% for standard stems and 91% for long stems. However, the use of cemented longstem femoral components makes any subsequent revision even more troublesome. Excluding the few reports with more than 10 years follow-up, 0% to 7% of femoral components were re-revised from average follow-up of 3.6 to 8.8 years. 4,15,39,42 These rates appeared to be better than those in the results of the studies that used early cementing techniques and had similar follow-up periods (Fig. 1). Although improved cementing techniques have reduced re-revision rates, as many as 35% of femoral components were still reported to become radiographically loose in less than 4 years time. 4 The cement fixation also deteriorated with time. In a series of articles from Massachusetts General Hospital in Boston, the femoral re-revision rate for aseptic loosening was 2% at 6 years. 42 It increased to 10.5% at 12 years and 20% at 15 years. 8,33 Franzen et al 9 also showed that cemented femoral revision was insecure, as 21 out of 24 components PROXIMALLY POROUS-COATED FEMORAL COMPONENTS Because of the encouraging results from the use of cementless femoral components in primary hip replacement, many surgeons advocated the use of proximally porous-coated femoral components in revision total hip arthroplasty. 10,11,14,16,27 These early reports had follow-up periods of less than 2 or 3 years. The midterm results, as shown in Table 3, were much less promising. Berry et al 2 evaluated the outcomes of 375 proximally porous-coated femoral components performed at Mayo Clinic, Rochester, Minnesota. Six different prostheses were used; they were BIAS and Harris-Galante (Zimmer, Warsaw, IN), Omnifit and Omnifit long stem (Osteonics, Stryker, Allendale, NJ), and PCA and PCA long stem (Howmedica, Rutherford, NJ). The mechanical failure rates ranged from 7.7% for the Omnifit long stem to 23.5% for the Harris-Galante femoral components after follow-up of less than 5 years. The cumulative survival rate at 8 years, using radiographic loosening or revision for aseptic loosening as end points, was only 20%. Three articles focused on the outcome of the BIAS femoral component in femoral revision. The BIAS femoral component was made of titanium alloy. It was noncircumferentially coated, with fibre-metal pads over the front and back sides of the proximal part. The stem was over 200 mm long, and incorporated an anterior bow to fit the medullary canal. Hussamy and Lachiewicz 18 studied 39 patients who had this femoral component implanted. There was no re-revision after an average follow-up of 5 years, but 32% of femoral components subsided by more than 2 mm, and pedestal was seen in 29% of hips. Despite the fact that there was no re-revision, the authors stopped using the BIAS femoral component, and changed to a modified design that had more extensive coating. Onsten et al 33 reported that 2 out of 45 BIAS femoral components were re-revised for aseptic loosening after an average follow-up of 4 years. Significant hip pain still occurred in 20% of the surviving hips. These researchers also performed the roentgen stereophotogrammetric analysis in 13 hips, and 12 femoral components showed significant 50

5 Femoral components in revision total hip arthroplasty Table 3 Results of proximally coated femoral components in revision total hip arthroplasty. Study Prosthesis No. of Age, Follow-up, Use of Intraoperative femoral Unstable Revised for aseptic patients years years long stem, % fracture, % fixation, % loosening, % Berry et al, 1995 Mixed, 6 types Hussamy & BIAS* Lachiewicz, 1994 Onsten et al, BIAS* NA Peters et al, 1995 BIAS* Woolson & Harris-Galante* Delaney, 1995 Mulliken et al, Mallory-Head Malkani et al, Omnifit long-stem NA = not available *Zimmer (Warsaw, IN). Biomet (Warsaw, IN). Osteonics (Stryker, Allendale, NJ). migration. Peters et al 37 found that 45% of 49 BIAS femoral components subsided by 2 mm or more. Although only 4% of femoral components were revised for aseptic loosening after 5.5 years of follow-up, the cumulative survival rate was only 37% at 6 years, using radiographic loosening or revision as endpoints. There were two reports on the use of the other longstem femoral components with proximal porous coating. Malkani et al 26 reviewed the 2- to 5-year results of 69 Omnifit long-stem femoral components. The latter was made of cobalt-chrome alloy. The metaphyseal part was circumferentially coated and was designed to achieve maximal filling at that level. The distal part was curved. The radiographic results were far from satisfactory: 57% of femoral components subsided by 5 mm or more, progressive radiolucent lines were seen in zones 1 and 7 in 50%, and distal pedestal existed in 61% of hips. After an average follow-up of 3 years, 7.2% of femoral components were re-revised for aseptic loosening. Mulliken et al 31 reported results from the use of 52 Mallory-Head (Biomet, Warsaw, IN) femoral components. These were made of titanium alloy and had proximal circumferential plasma spray coating; the stem length was from 250 to 300 mm. After an average follow-up of less than 5 years, 10% of femoral components were re-revised because of loosening, and another 14% were loose in the radiographs. There was only one report on the use of standard-length femoral components. Woolson and Delaney 44 reported the results of 25 femoral revisions with the Harris- Galante prosthesis implanted. The Harris-Galante femoral component was made of cobalt-chrome alloy; it was straight, collared, and was 140 to 170 mm long. After an average follow-up of 5.5 years, 16% of the femoral components were revised for aseptic loosening, and an additional 36% of femoral components subsided 5 mm or more. Proximally porous-coated femoral component had a relative high failure rate, despite the use of long-stem femoral components noted in most reports. Apart from one study that had no re-revision, the others reported that 4% to 16% of the femoral components were re-revised after an average follow-up of 3 to 5.5 years. Unstable fixation occurred in 24% to 57% of the femoral components. Another major problem with the proximally porouscoated femoral component in revision total hip arthroplasty was the high incidence of intraoperative femoral fractures. These could have been due to the fact that the proximal bone stock was usually deficient, and that the surgeon always attempted to achieve maximal filling at this level. The reported fracture rates ranged from 12% to 46%. Of the seven studies, only two of them reported a fracture rate of less than 20%. It was also shown that the mechanical failure rate was higher if there was an intraoperative fracture

6 HKJOS Table 4 Results for extensively coated femoral components in revision total hip arthroplasty. Study Prosthesis No. of Age, Follow-up, Thigh pain, Severe stress Unstable fixation, Revised for aseptic patients years years % shielding, % % loosening, % Engh et al, % AML* NA NA 1 0 Lawrence et al, Mixed, 3 types NA NA Lawrence et al, 77% AML* NA NA NA Moreland & AML/Solution* Bernstein, 1995 Krishnamurthy AML/Solution* et al, 1997 Paprosky et al, AML/Solution* Moreland & AML/Solution* Moreno 2001 NA = not available *Depuy (Warsaw, IN). For bone ingrown stems only. For all stems. EXTENSIVELY POROUS-COATED FEMORAL COMPONENTS Three surgeons from United States have reported their experiences with extensively porous-coated femoral components. They are Charles Engh Sr from Anderson Orthopaedic Clinics and Research Institute in Arlington, Virginia, John Moreland from UCLA Medical Center, Los Angeles, and Wayne Paprosky from Rush-Presbyterian-St. Luke s Hospital, Chicago. Table 4 presents results of extensively coated femoral components in revision total hip arthroplasty. Engh et al 6 studied 160 cementless revision total hip arthroplasties. The mean age of the patients was 55.8 years. The average follow-up was 4.4 years. In 127 patients, the femoral components were revised with one of the three porous-coated designs. An extensively coated, straight, and cylindrical (AML, Depuy, Warsaw, IN) femoral component was used in 93 patients; none was revised for aseptic loosening. There were eight AML femoral components that were only proximally coated; one component was loose despite the fact that initial press-fit was achieved. In 26 patients, a fully porous-coated, tapered femoral component with a more rectangular cross section was used; four components (15%) were loose. The authors recommended the use of extensively porous-coated, canal-filling femoral component in achieving reliable distal fixation in femoral revision. Lawrence et al 24 reviewed 174 extensively porous-coated femoral components used in surgery performed at the same institution after an average followup of 7.4 years. Most of the components used were the AML designs, but the exact proportion was not mentioned. Re-revision for aseptic loosening was needed in 3.4% of hips. In another report, Lawrence et al 25 reported the 5-year minimum follow-up of 83 revisions with extensively porous-coated femoral components. The AML prosthesis was used in 77% of cases, but the results for the different prostheses were not analysed separately. After an average follow-up of 9 years, 6% of femoral components were re-revised for aseptic loosening. Moreland and Bernstein 29 reviewed 175 femoral revisions with extensively porous-coated, cementless femoral components. All the revisions were performed by Dr. John Moreland. An extensively porous-coated straight femoral component (AML or Solution, Depuy) was used in every patient. There were 137 standard length and 38 longer femoral components. After an average follow-up of 5 years (range 2-10), three femoral components were loose, and two of them were revised. Five other femoral components that were not loose were revised; two were painful, two were removed as part of resection arthroplasty for severe acetabular defects, and one was removed for deep sepsis. Bone ingrowth was achieved in 83% of femoral components; it was 88% for 52

7 Femoral components in revision total hip arthroplasty components with complete canal fill and only 71% for components with incomplete fill. Significant thigh pain occurred in 4% of bone-ingrown femoral components and more than 18% of stable fibrous femoral components. Severe stress shielding affected less than 8% of bone-ingrown femoral components and was related to preoperative osteoporosis and prosthesis with a large diameter. Moreland and Boreno 30 reviewed the 5-year minimal follow-up results of the same series of patients. There were 134 patients, and the mean follow-up was 9.3 years. There was no new failure after the 5-year follow-up report. The incidence of severe stress shielding increased to 20%, but it has not caused any failure. Krishnamurthy et al 23 reported Paprosky s minimal 5- year results (mean follow-up of 8.3 years) for 297 femoral revisions. A 6-inch femoral component was used in 37%, a 7-inch calcar-replacement femoral component or an 8-inch femoral component was used in 49%, and a 10-inch femoral component was used in 14% of patients. With regard to the diameters of the femoral components, 17% were 13.5 mm or smaller, and 22% were 18 mm or larger. Bone ingrowth was achieved in 82% of patients, 15.6% had stable fibrous fixation, and only 2.4% had unstable fixation. Of the seven femoral components that were unstable, two were asymptomatic and the other five (1.7%) were symptomatic and were re-revised. The unstable femoral components were attributed to the use of undersized femoral components and severe bony defects. Paprosky et al 34 reported his minimum 10-year follow-up results of 170 patients. After an average follow-up of 13.2 years, the total mechanical failure rate was only 4.1%. The outcomes for extensively porous-coated femoral components in revision total hip arthroplasty appear to be promising. The re-revision rate for aseptic loosening ranged from 0 to 6% after follow-up periods from 4.4 to 13.2 years. Unstable fixation occurred in only 1% to 4.6% of femoral components, which could be attributed to the fact that fixation of such components occurred in the diaphysis, which was usually better preserved in revision total hip arthroplasty. DISCUSSION This review focuses on the results of cemented and cementless fixation in femoral revisions. Special techniques, such as cementing into allograft (impaction bone grafting techniques, massive proximal femoral allograft), and special cementless femoral components Figure 2 A 77-year-old man had a cemented femoral component inserted 18 years earlier. It became loose and then shifted into varus position (left). Cemented femoral revision was performed with improved techniques (middle). The femoral component was well fixed 4 years after the revision (right). (highly modular femoral components: for example, S- ROM prosthesis, Johnson and Johnson, Warsaw, IN; conical femoral components with longitudinal fins: eg, Wagner prosthesis, Protek, Berne, Switzerland) were not covered. The remaining bone stock affects the outcomes of femoral revisions. Most of the articles reviewed reported on the severity of femoral bone deficiency, and several authors also showed that the results of the femoral revisions were better if the bone deficiencies were less severe. However, it was extremely difficult, if not possible to compare the different series directly, because the methods used to describe the degree of femoral bone deficiency were not standardised. Readers must understand this limitation when they interpret the data presented in this review article. Figure 1 is a diagrammatic representation of the re-revision rates of all the series related to cemented femoral revisions. The early to mid-term re-revision rates for aseptic loosening of revised femoral components ranged from 2% to 14% with early cementing techniques, and they dropped to 0% to 7% with improved cementing techniques. However, the re-revision rates with improved cementing techniques increased with time, and the radiological loosening rates were still high. Cemented femoral revision is still an option in a patient who is relatively old and inactive if adequate cancellous bone is present for cement interdigitation (Fig. 2). The use of cemented long-stem femoral components 53

8 HKJOS Figure 3 The outcomes of femoral revisions with proximally porous-coated (open squares) and extensively porous-coated (closed squares) femoral components are compared. The re-revision rates are plotted against the respective average follow-up periods. Figure 4 A 63-year-old woman suffered from chronic osteomyelitis around a cemented femoral component (left). The latter was removed, and a temporary prosthesis made of antibiotic impregnated cement was inserted (middle); a long-stem, extensively porous-coated femoral component was inserted after 3 months. Three years after the revision, the patient was asymptomatic, and the femoral component was well fixed with distal spot weld and remodeling of the sclerotic proximal bone (right). produced good results, but it made the subsequent revision extremely difficult due to the difficulties in retrieving cement way down in the medullary canal. Figure 3 is a diagrammatic representation of the re-revision rates of all the series related to cementless femoral revisions. The early to mid-term re-revision rates ranged from 0% to 16% with proximally porous-coated femoral components, and they were 0% to 6% with extensively porous-coated femoral components. The follow-up periods were also longer in the reports on extensively porous-coated femoral components. Use of the proximally porous-coated femoral component was, therefore, not a reliable means of femoral revision. It also carried a significant intraoperative fracture risk that ranged from 12% to 46%. The most reliable option in femoral revision is probably the extensively porouscoated femoral component (Fig. 4). Its theoretical problem is stress shielding, but that has not been reported to lead to clinical failure so far. It can also be argued that such components carry the same problem as cemented long-stem femoral components, that of making the next revision more difficult. This is not so problematic, since it is not necessary to retrieve any cement down in the medullary canal if an extensively porous-coated femoral component becomes loose. In addition, the loosening is usually attributed to sub-optimal canal filling along with the use of a femoral components that were too small; the re-revision simply demands the insertion of a bigger femoral component in most cases. There was no long-term report with over 10 years of follow-up for femoral revisions using early cementing techniques and proximally porous-coated femoral components. Katz et al 20 reported the 12-year follow-up of femoral revisions with improved cementing techniques; 5.4% were re-revised for aseptic loosening and another 9.7% showed radiographic loosening. Mulroy and Harris 32 reported on a 15-year average follow-up of femoral revisions with improved cementing techniques; 20% of those femoral components were rerevised. These results were not as favourable compared with the 3.5% re-revision rate after 13.2 years average follow-up for extensively porous-coated femoral components that was reported by Paprosky et al. 34 CONCLUSION After failure of a cemented total hip arthroplasty, the proximal femur is either mechanically (the bone support is weak) or biologically (the bone may not be viable) compromised. The use of a cemented femoral component or a proximally porous-coated femoral component is often not favourable. The use of an extensively coated femoral component, which aims at diaphyseal fixation, appears to offer durable fixation in a reliable manner. 54

9 Femoral components in revision total hip arthroplasty REFERENCES 1. Amstutz HC, Steven M, Jinnah RH, Mai L. Revision of aseptic loose total hip arthroplasties. Clin Orthop 1982;170: Berry DJ, Harmsen WS, Ilstrup D, Lewallen DG, Cabanela ME. Survivorship of uncemented proximally porous-coated femoral components. Clin Orthop 1995;319: Callaghan JJ, Salvati EA, Pellici PM, Wilson PD, Ranawat CS. Results of revision for mechanical failure after cemented total hip arthroplasty, : A two to five year follow-up. J Bone Joint Surg Am 1985;67: Eisler T, Svensson O, Iyer V, Wejkner B, Schmalholz, Larsson H, Elmstedt E. Revision total hip arthroplasty using third generation cementing technique. J Arthroplasty 2000;15: Engelbrecht DJ, Weber FA, Sweet MB, Jakim I. Long-term results of revision total hip arthroplasty. J Bone Joint Surg Br 1990;72: Engh CA, Glassman AH, Griffin WL, Mayer JG. Results of cementless revision for failed cemented total hip arthroplasty. Clin Orthop 1988;235: Engh CA, Glassman AH, Suthers KE. The case for porouscoated hip implants. The femoral side. Clin Orthop 1990;261: Estok DM II, Harris WH. Long-term results of cemented femoral revision surgery using second generation cementing techniques: an average 11.7 year follow-up evaluation. Clin Orthop 1994;299: Franzen H, Mjoberg B, Onnerfalt R. Early loosening of femoral components after cemented revision. J Bone Joint Surg Br 1992;74: Gustilo RB, Pasternak HS. Revision total hip arthroplasty with titanium ingrowth prosthesis and bone grafting for failed cemented femoral component loosening. Clin Orthop 1988; 235: Harris WH, Kraushell RJ, Galante JO. Results of cementless revisions of total hip arthroplasties using the Harris-Galante prosthesis. Clin Orthop 1988;235: Harris WH, McCarthy JC, O Neill DA. Femoral component loosening using contemporary techniques of femoral cement fixation. J Bone Joint Surg Am 1982;64: Harris WH, McGann WA. Loosening of the femoral component after use of the medullary-plug cementing techniques. Follow-up note with a minimum five-year follow-up. J Bone Joint Surg Am 1986;68: Hedley AK, Gruen TA, Ruoff DP. Revision of failed total hip arthroplasties with uncemented porous-coated anatomic components. Clin Orthop 1988;235: Hultmark P, Karrholm J, Stromberg C, Herberts P, Mose CH, Malchau H. Cemented first-time revisions of the femoral component. Prospective 7 to 13 years follow-up using second-generation and third-generation technique. J Arthroplasty 2000;15: Hungerford DS, Jones LC. The rationale of cementless revision of cemented arthroplasty failures. Clin Orthop 1988;235: Hunter GA, Welsch RP, Cameron HU, Bailey WH. The results of revision of total hip arthroplasty. J Bone Joint Surg Br 1979;61: Hussamy O, Lachiewicz PF. Revision total hip arthroplasty with the BIAS femoral component. J Bone Joint Surg Am 1994; 76A: Izquierdo RJ, Northmore-Ball MD. Long-term results of revision hip arthroplasty: survival analysis with special reference to the femoral component. J Bone Joint Surg Br 1994; 76: Katz RP, Callaghan JJ, Sullivan PM, Johnston RC. Long-term results of revision total hip arthroplasty with improved cementing technique. J Bone Joint Surg Br 1997;79: Kavanagh BF, Ilstrup DM, Fitzgerald RH. Revision total hip arthroplasty. J Bone Joint Surg Am 1985;67: Kershaw CJ, Atkins RM, Dodd CA, Bulstrode CJ. Revision total hip arthroplasty for aseptic failure: a review of 276 cases. J Bone Joint Surg Br 1991;73: Krishnamurthy AB, MacDonald SJ and Paprosky WG. 5- to 13-year follow-up study on cementless femoral components in revision surgery. J Arthroplasty 1997;12: Lawrence JM, Engh CA, Macalino GE. Revision total hip arthroplasty. Long-term results without cement. Orthop Clin North Am 1993;24: Lawrence JM, Engh CA, Macalino GE, Lauro GR. Outcome of revision hip arthroplasty done without cement. J Bone Joint Surg Am 1994;76: Malkani AL, Lewallen DG, Cabanela ME, Wallrichs SL. Femoral component revision using an uncemented, proximally coated, long-stem prosthesis. J Arthroplasty 1996;11: Mallory TH. Preparation of the proximal femur in cementless total hip revision. Clin Orthop 1988;235: Marti RK, Schuller HM, Besselaar PP, Van Frank Haasnoot EL. Results of revision of hip arthroplasty with cement: a five to fourteen year follow-up study. J Bone Joint Surg Am 1990; 72: Moreland JR, Bernstein ML. Femoral revision hip arthroplasty with uncemented, porous-coated stems. Clin Orthop 1995;319: Moreland JR, Moreno MA. Cementless femoral revision arthroplasty of the hip: minimum 5 years follow-up. Clin Orthop 2001;393: Mulliken BD, Rorabeck CH, Bourne RB. Uncemented revision total hip arthroplasty. Clin Orthop 1996;325: Mulroy WF, Harris WH. Revision total hip arthroplasty with use of so-called second generation cementing techniques for aseptic loosening of the femoral component. A fifteen year average follow-up study. J Bone Joint Surg Am 1996;78: Onsten I, Sanzen L, Carlsson A, Besjakov. Migration of uncemented, long-stem femoral components in revision hip arthroplasty. Acta Orthop Scand 1995;66: Paprosky WG, Greidanus NV, Antoniou J. Minimum 10-yearresults of extensively porous-coated stems in revision hip arthroplasty. Clin Orthop 1999;369: Pellici PM, Wilson PD Jr, Sledge CB, Salvati EA, Ranawat CS, Poss R. Revision total hip arthroplasty. Clin Orthop 1982; 170: Pellici PM, Wilson PD Jr, Sledge CB, Salvati EA, Ranawat CS, Poss R, Callaghan JJ. Long-term results of revision total hip replacement: A follow-up report. J Bone Joint Surg Am 1985;67: Peters CL, Rivero DP, Kull LR, Jacobs JJ, Rosenberg AG, Galante JO. Revision total hip arthroplasty without cement: subsidence of proximally porous-coated femoral components. 55

10 HKJOS J Bone Joint Surg Am 1995;77: Pierson JL, Harris WH. Cemented revision for femoral osteolysis in cemented arthroplasties: results in 29 hips after a mean 8.5-year follow-up. J Bone Joint Surg Br 1994;76: Pierson JL, Harris WH. Effect of improved cementing techniques on the longevity of fixation in revision cemented femoral arthroplasties. Average 8.8-year follow-up period. J Arthroplasty 1995;10: Raut VV, Siney PD, Wroblewski BM. Outcome of revision for mechanical stem failure using the cemented Charnley stem: a study of 399 cases. J Arthroplasty 1996;11: Retpen JB, Jensen JS. Risk factors for recurrent aseptic loosening of the femoral component after cemented revision. J Arthroplasty 1993;8: Rubash HE, Harris WH. Revision of non-septic, loose, cemented femoral components using modern cement techniques. J Arthroplasty 1988;3: Turner RH, Mattingly DA, Scheller A. Femoral total hip arthroplasty using a long stem femoral component: Clinical and radiographic analysis. J Arthroplasty 1987;2: Woolson ST, Delaney TJ. Failure of a proximally coated femoral prosthesis in revision total hip arthroplasty. J Arthroplasty 1995;10(Suppl):22-8. The Author CHIU Kwong-Yuen, FRCS (Edin), FHKAM (Orth Surg), Chief, Division of Joint Replacement Surgery and Associate Professor, Department of Orthopaedic Surgery, The University of Hong Kong, Hong Kong. 56