Secondary Logo

Journal Logo


Multidetector-CT Evaluation of Bone Substitutes Remodeling after Revision Hip Surgery

Nishii, Takashi MD; Sugano, Nobuhiko MD; Miki, Hidenobu MD; Koyama, Tsuyoshi MD; Yoshikawa, Hideki MD

Author Information
Clinical Orthopaedics and Related Research: January 2006 - Volume 442 - Issue - p 158-164
doi: 10.1097/01.blo.0000183740.88987.5f
  • Free


In revision hip surgery, a large bone defect often develops around the artificial implant because of osteolysis or removal of the cement layer. Various bone grafting procedures have been advocated for bone deficiencies or to augment implant fixation with autograft bone, allograft bone, or various types of bone substitutes including hydroxyapatites, calcium phosphates, and calcium sulfates. Surgeons have presumed the grafted materials would stimulate new bone formation and be resorbed and replaced by new bone, and that an increase in bone stock would ensure prolonged component fixation.5,9 This provides a potential advantage if additional revision surgery is required.

The major factors influencing bone graft incorporation and remodeling are biologic activity of grafted materials and the surrounding environment of the graft site.9,27 Various types of bone substitute material show different biologic activity in terms of osteoconductive and osteoinductive effects, and changes in their volume from bioresorption.7,13,19 The perigraft environment is important for supply of bone forming cells and blood vessels to the grafted regions.27 A damaged femoral environment with decreased cell and vascular viability in revision hip surgery may result in poor bone graft incorporation.3,6,15 Efficacy of bone or bone substitutes grafting has been examined in animal studies16,27 and human retrieval studies12,19,29; however, few studies have documented remodeling patterns of various bone substitutes and influence of the surrounding environment in living patients after revision hip surgery.4,9 Although radiographic examination has been widely used for assessing remodeling after grafting, reliable assessment has been difficult because of unclear images of bone and graft material after superimposing a three-dimensional structure on the two- dimensional image.21 Without a reliable assessment modality, clinical efficacy and appropriate indications for each graft material in revision hip surgery are not clear.

Computed tomography (CT) scans may be suitable for assessing grafted materials because of its high spatial resolution, its superior signal contrast between bone and soft tissue, and its production of three-dimensional images. The effectiveness of CT scans for evaluating the region around the hip implant has been limited because of scattering caused by metallic artifacts. Advances in spiral CT technology with multidetector arrays has enhanced its effectiveness by reducing the effects of metallic artifacts, and increased its accuracy in assessing periprosthetic osteolysis after total hip arthroplasty (THA).24-26

We compared remodeling patterns of two different types of bone substitutes in patients after revision THA by evaluating serial changes of graft volume using multidetector-row CT imaging. In addition, we determined whether factors relating to the perigraft bone environment influence subsequent bone remodeling after grafting of bone substitutes.


We prospectively studied 12 hips in 10 patients (two men and eight women) who had cementless revision surgery of the hip with bone substitute grafting around the femoral component. The average age of the patients at the time of revision surgery was 63 years (range, 48-75 years) (Table 1). The prior surgery was cemented THA or bipolar hemiarthroplasty in five patients (six hips) and cementless THA or bipolar hemiarthroplasty in five patients (six hips). The mean period between the previous surgery and the revision surgery was 20 years (range, 7-30 years). The reason for revision surgery was cup and/or stem loosening in nine patients (11 hips) and severe periprosthetic osteolysis in one hip of one patient. Femoral bone deficiency was evaluated using the classification of Paprosky et al 2 for radiographs obtained before revision surgery. Seven hips in seven patients were Grade 2A (absent calcar extending just below the intertrochanteric level), two hips in one patient were Grade 2B (anterolateral metaphyseal bone loss), one hip in one patient was Grade 2C (absent calcar with posteromedial metaphyseal bone loss), one hip in one patient was Grade 3A (2A plus diaphyseal bone loss), and one hip in one patient was Grade 3B (2B plus diaphyseal bone loss).

Patient Demographics

The posterolateral approach without trochanteric osteotomy was used for revision surgery in all patients. A Wagner SL femoral component (Zimmer, Warsaw, IN) was implanted without cement in all patients. The Wagner SL femoral component was made of a titanium-aluminum-niobium alloy and had a conical stem with a distal taper angle of 2°.4,11 Eight longitudinal fins on the stem surface provided firm fixation at the femoral diaphysis, and little supportive strength was required in the proximal part of the femur. After inserting the femoral component, the bone defect space in the proximal part of the femur was filled with 5-10 g beta-tricalcium phosphate (β-TCP) granules (OS- ferion; Olympus, Tokyo, Japan). The β-TCP granules had a particle diameter of 1-3 mm, porosity of 75%, and pore size of 100-400 μm. Then, 6 mL of pastelike calcium phosphate cement (Biopex; Mitsubishi Materials, Tokyo, Japan) was injected onto the β-TCP granules. Biopex (Mitsubishi Materials) was produced by mixing cement powder consisting of α-tricalcium phosphate (75%), dicalcium phosphate dibasic (5%), and tetra- calcium phosphate (20%) with water containing sodium succinate (12%) and sodium chondroitin sulphate (5%). After injection, it hardened in a few minutes and formed a solid layer of carbonated apatite. This acted as a barrier against dispersion of the grafted β-TCP granules into the joint (Fig 1).

Fig 1
Fig 1:
An AP radiograph taken immediately after the revision surgery is shown. Beta-tricalcium phosphate granules were grafted in the proximal femoral bone defect around the Wagner SL stem Beta-tricalcium phosphate granules were grafted (white arrows), and a pastelike CPC was injected (black arrows) onto the β-TCP granules. Dispersion of grafted β-TCP granules was prevented by the CPC layer.

Three weeks postoperatively, scanning was completed using a four-detector or eight-detector-row CT system (LightSpeed; General Electric, Milwaukee, WI or Aquilion; Toshiba, Tokyo, Japan) after modifying the metal artifact minimizing protocol of Puri et al24 with the following settings: collimation, 1.25 mm; pitch value, 1; voltage, 120 kVp; and current, 250 mA. Axial images were reconstructed on a 512 × 512 matrix with an in- plane resolution of 0.4 mm and slice thickness of 1.25 mm. The initial grafted volumes of calcium phosphate cement and β-TCP around the femoral component were calculated on CT images obtained 3 weeks postoperatively by tracing the outlines of the grafted materials manually on each plane using image analysis software (Scion Image Software; Scion Corporation, Frederick, MD). Computed tomography was repeated 1 year postoperatively, and changes in calcium phosphate cement, β-TCP, and new bone volumes were calculated. One patient had an additional CT scan 6 months postoperatively. New bone formation was defined as an area with bony trabecular features and connection to the host bone 1 year postoperatively that was not seen on the corresponding CT image at 3 weeks postoperatively. A homogenous area with high signal density in the bone defect was defined as calcium phosphate cement grafting, and a granulated, heterogeneous area with slightly lower signal density than the calcium phosphate cement was defined as β-TCP grafting (Fig 2).

Fig 2A
Fig 2A:
F. (A) A grafted CPC area is shown (black arrow) on a CT image at the level of the greater trochanter 3 weeks postoperatively (Patient 3). (B) No remarkable volume change of the CPC area was found at 6 months postoperatively (black arrow). (C) No remarkable volume change of the CPC area was found at 1 year postoperatively (black arrow). (D) A grafted β-TCP area is shown (white arrows) on a CT image at the level of the lesser trochanter at 3 weeks postoperatively (Patient 3). (E) Six months postoperatively, a translucent change appeared in the whole β-TCP area (white arrows). (F) At 1 year postoperatively, a new area with bony density connected directly to the host bone (white arrows).

To determine the validity of volume measurements, intraob- server and interobserver variabilities were assessed from two measurements separated by a 2-week interval in the first six hips. These measurements were obtained by two independent observers. Mean intraobserver variations in calcium phosphate cement and β-TCP volume were 1.7% and 2.5%, respectively. Mean interobserver variations in calcium phosphate cement and β-TCP volume were 1.6% and 3%, respectively.

All statistical analyses were done using StatView™ software (StatView version 5.0, SAS Institute, Cary, NC). Residual volumes of calcium phosphate cement and β-TCP at 1 year were expressed as a percentage of the initial grafted volume, and volume of new bone formation at 1 year was expressed as a percentage of the initial grafted volume of β-TCP. Residual volumes of calcium phosphate cement and β-TCP 1 year postoperatively were compared using the nonparametric Mann- Whitney U test. The influence of factors related to the perigraft bone environment to subsequent bone remodeling after grafting of bone substitutes was analyzed with multiple linear regression analysis. This analysis used the classification of Paprosky et al22 (Grade 2A or higher), cement use in a prior operation, and initial grafted volumes of calcium phosphate cement and β-TCP as factors related to perigraft bone environment (independent variables), and residual volumes of calcium phosphate cement and β-TCP and volume of new bone formation 1 year postoperatively as bone remodeling indices (dependent variables). Initial grafted volumes of calcium phosphate cement and β-TCP were used as surrogates of preoperative bone defect volume. A p value less than 0.05 indicated significance.


The average initial grafted volumes of calcium phosphate cement and β-TCP were 3900 mm3 (range, 1730-6054 mm3) and 8018 mm3 (range, 2961-15,296 mm3), respectively. At 1 year followup, β-TCP had decreased (p = 0.0001) more than calcium phosphate cement (residual volume, 30% of the original volume, range, 10-62%, versus 78%, range, 37-96%, respectively). New bone formed after absorbing β-TCP, with a volume of 34% (range, 11-76%) of the initial β-TCP volume. On CT images obtained 1 year postoperatively, most residual calcium phosphate cement was in direct contact with the surrounding host bone, and the β-TCP was replaced by new bone formation (Figs 2, 3). However, the newly formed bone was not in direct contact with the stem surface, and an intervening radiolucent zone was seen (Figs 2, 3).

Fig 3A
Fig 3A:
D. (A) A grafted CPC area is shown (black arrow) on a CT image at the intertrochanteric level at 3 weeks postoperatively (Patient 2). (B) At 1 year postoperatively, the CPC area shows a slight decrease and direct contact to the surrounding host bone without intervention of radiolucent zone (black arrow). (C) The grafted β-TCP area is shown (white arrows) on a CT image at the level of the lesser trochanter at 3 weeks postoperatively (Patient 2). (D) The β-TCP area is completely replaced by an area with bony trabecular features (white arrows).

Cement use in a prior operation explained 57% of residual volume of β-TCP seen 1 year postoperatively (r2 = 0.57, p = 0.003), and 39% of the rate of new bone formation seen 1 year postoperatively (r2 = 0.39, p = 0.018). At 1 year postoperatively, the residual volume of β-TCP was 43% (range, 19%-62%) in hips with prior cemented implants, and 17% (range, 10%-26%) in hips with prior cementless implants. The rate of new bone formation was 21% (range, 11%-37%) in hips with prior cemented implants, and 47% (range, 21%-76%) in hips with prior cementless implants.


It has been difficult to reliably evaluate remodeling of grafted bone or bone substitutes for graft union, resorptive changes, and new bone formation around an artificial component in living patients. Plain radiographs have been used because they are noninvasive.12,14,21 However, they are unreliable because of limitations such as rotational positions of the leg, inferior contrast among bone substitute materials, and intrinsic limitations of two-dimensional measurement.12,14,21 There have been studies of the reliability of scintigraphic examinations, including single photon emission tomography with technetium-99m- labeled diphosphonate17 and positron emission tomography with [18F]fluoride.23 These modalities have improved imaging resolution compared with conventional scinti- graphic examinations and have been shown to be effective for monitoring bone metabolism and revascularization around grafted materials. However, their effectiveness is limited by their inability to quantify morphologic features of bone or graft and the difficulty in differentiating among bone integration change, resorptive change, and new bone formation. To resolve limitations in evaluating bone remodeling around an artificial component, we used helical CT scanning with a metal artifact suppression protocol in patients having cementless revision surgery with bone substitute grafting.

Our study has several limitations. First, the small number of patients might lead to difficulty in detecting significant influential factors of graft remodeling other than use of cement in prior operations. Second, only the patients who had revision surgery using a single type of femoral component by cementless fixation were evaluated. Caution is required before extrapolating our findings to patients using other types of femoral components or cemented fixation. Third, patients who had revision surgery without grafting of bone or bone substitutes were not examined. On plain radiologic examinations, some investigators reported high frequency of bone recovery in the proximal femur after revision surgery using a Wagner SL femoral component without grafting of bone materials.11,28 True efficacy of grafting various types of bone materials in revision surgery will not be clarified without additional comparative studies between revisions with and without grafting of bone materials. Fourth, the accuracy of graft volume measurement on CT images was not evaluated. Some degree of error may exist partially because of manual tracing errors of grafted areas and residual metallic artifacts, despite using a metal artifact minimizing protocol.

Despite these limitations, helical CT scanning with a metal artifact suppression protocol was clinically useful in analyzing three-dimensional quantities of grafted volume and serial remodeling of grafted materials around artificial hip components. New CT technology with multidetector arrays has produced superior spatial resolution of the longitudinal axis,2 with reconstructed slice thickness of 1.25 mm, reduced partial volume effects on each CT image, and differentiated clearly between grafted materials and new bone formation.

From relatively low intraobserver and interobserver variations of graft measurement in our study, and high reproducibility and accuracy of measurements in previous studies using similar methods to detect and monitor osteolytic change around metallic components,24-26 we think graft volume measurement is sufficient to compare remodeling patterns of bone substitutes and influence of prior cemented fixation on bone remodeling. It was not certain whether new bone seen on CT images had enough viability and mechanical strength to facilitate support of ingrowth components for future revisions. Histologic studies of grafted materials and newly formed bone will be needed at the time of biopsy examinations or rerevision surgery.

Computed tomography examinations at 1 year followup showed that the grafted β-TCP was decreased 10%-62% of the initial grafted volume, and new bone formed with an average volume of 34% of the initial grafted β-TCP volume. Our results using CT were comparable with those of a previous study in which plain radiographic assessment was used after femoral revision with a Wagner SL component.4 Bohm et al used autogenous or allogeneic bone grafting around a femoral component, and there was complete or incomplete incorporation of the grafted bone in 82% of the patients and a 25%-47% increase of mean relative bone mass (the ratio of the width of the bone to the outside diameter) after a mean followup of 4.8 years.4 However, assessing bone remodeling on plain radiographs is unreliable, and some authors have suggested using CT examinations for improved assessment.4,9

The materials for grafting bone deficiencies during revision surgery include autograft bone, allograft bone, demineralized bone matrix, and various types of bone substitutes. Factors considered in the choice of material include the source of the autograft, risk of infection, volume and shape of recipient sites, cost, and biologic and mechanical requirements.1,7,13 In our patients, two types of bone substitute materials were used around the proximal part of the femoral component because of the complicated shape of the recipient site and the low requirement for mechanical strength. At 1-year followup, β-TCP volume decreased more than calcium phosphate cement volume. In an animal study, after the calcium phosphate cement hardened, it was converted to microporous hydroxyapatite with high osteoconductive activity and showed active bone formation.16 This led to direct contact between the material and bone without intervening soft tissue.16 The calcium phosphate cement showed some resorptive change but proceeded relatively slowly.16 Beta-tricalcium phosphate is a well known biodegradable and osteoconductive material.7,13,18 In one animal study, grafting OSferion (Olympus) into bone defects promoted abundant bone formation and subsequent remodeling into a mature bony trabecular structure with extensive resorption.20 In a report of a human proximal femur retrieved because of resection of a bone tumor 4 weeks after grafting of OSferion granules, vigorous bone formation was evident histologically around and inside the pores of the granules.19 A considerable number of osteoclast cells were seen around the granules.19 Changes in the appearance of grafted materials and surrounding bone formation seen during serial CT examinations and quantitative analysis of the grafted materials were consistent with the results of the experimental animal and human retrieval studies.16,19,20 Based on our results and the results of previous studies,16,19,20 calcium phosphate cement is suitable for enhancing implant fixation, with slow resorptive change for a relatively long period. Beta-tricalcium phosphate is suitable for facilitating recovered bone loss with active resorptive change where little mechanical strength is required.

We observed formation of a gap between new bone and the stem surface 1 year postoperatively. From the CT images, it was not clear if the gap was empty of tissue or occupied by fibrous tissue with or without osteogenic cells and sufficient vascular supply. The Wagner SL stem provided fixation around the femoral diaphysis, therefore, direct contact between new bone and the stem was not necessary to maintain stability of the femoral component. Greis et al10 did an experimental study of THAs in canines with overlying bone defects on the components. They found that grafting autogenetic bone or a composite of hydroxyapatite and tricalcium phosphate enhanced new bone formation at the defect site compared with no grafting in the bone defects, but did not promote bone ingrowth on the surface of the component.10 Heekin et al12 studied postmortem histologic specimens of patients who had acetabular revision surgery with morselized allograft. They found a thick layer of fibrous tissue at the graft-cup interface where a radiolucent line was evident.12 The biomechanical environment in which the component surface was exposed could affect bone ingrowth over the component.8 Longer followup with CT would be useful for examining whether bone ingrowth is the eventual result of new bone formation in living patients.

We observed considerable variations in rates of β-TCP resorption and new bone formation among our patients. Despite the small sample size, our results indicate that use of cement in prior operations was an unfavorable factor in graft remodeling after revision surgery. An important factor in bone graft incorporation was the perigraft environment in terms of vascularity and living bone forming cells.27 Unsatisfactory results of revision surgery after failure of cemented femoral components are partly because of residual fibrous tissue, lack of cancellous bone, and damaged intramedullary vascularity at the endosteal femoral bone.6,15 The perigraft bone environment may be damaged more severely after failure of cemented femoral components than after failure of cementless femoral components, resulting in inferior graft remodeling. However, additional studies with more patients are required to clarify the relevance of prior cement use.

Various bioactive bone substitute materials and tissues have been developed for clinical use around reconstructive artificial implants.1,7,13,19 Experimental studies have shown encouraging results for use of osteoinductive cytokines, such as bone morphogenetic proteins and fibroblast growth factor-2 for promoting bone healing and bone in- growth.5,30 Development of bone substitutes and osteoinductive cytokines have involved extensive preclinical studies that use cell culture or rodent models. However, the efficacy of those materials in vivo can be determined only in human clinical investigations because of the variability of bone metabolism among species. Multidetectorrow CT is promising for evaluating bone stock restoration and remodeling patterns of various bone substitutes and osteoinductive agents. It could help clinicians to ascertain the most appropriate graft materials and osteoinductive agents, based on individual biologic and mechanical factors, for use in revision surgery.


We thank Hisashi Tanaka, MD for technical support.


1. Bauer TW, Smith ST. Bioactive materials in orthopaedic surgery: Overview and regulatory considerations. Clin Orthop Relat Res. 2002;395:11-22.
2. Berland LL, Smith JK. Multidetector-array CT: Once again, technology creates new opportunities. Radiology. 1998;209:327-329.
3. Berman AT, Reid JS, Yanicko Jr DR, et al. Thermally induced bone necrosis in rabbits: Relation to implant failure in humans. Clin Orthop Relat Res. 1984;186:284-292.
4. Böhm P, Bischel O. Femoral revision with the Wagner SL revision stem: Evaluation of one hundred and twenty-nine revisions followed for a mean of 4.8 years. J Bone Joint Surg. 2001;83A:1023-1031.
5. Cook SD, Barrack RL, Patron LP, Salkeld SL. Osteoinductive agents in reconstructive hip surgery: A look forward. Clin Orthop Relat Res. 2003;417:195-202.
6. Dohmae Y, Bechtold JE, Sherman RE, et al. Reduction in cement- bone interface shear strength between primary and revision arthroplasty. Clin Orthop Relat Res. 1988;236:214-220.
7. Finkemeier CG. Bone-grafting and bone-graft substitutes. J Bone Joint Surg. 2002;84A:454-464.
8. Flatley TJ, Lynch KL, Benson M. Tissue response to implants of calcium phosphate ceramic in the rabbit spine. Clin Orthop Relat Res. 1983;179:246-252.
9. Gamradt SC, Lieberman JR. Bone graft for revision hip arthroplasty: Biology and future applications. Clin Orthop Relat Res. 2003;417:183-194.
10. Greis PE, Kang JD, Silvaggio V, Rubash HE. A long-term study on defect filling and bone ingrowth using a canine fiber metal total hip model. Clin Orthop Relat Res. 1992;274:47-59.
11. Grünig R, Morscher E, Ochsner PE. Three-to 7-year results with the uncemented SL femoral revision prosthesis. Arch Orthop Trauma Surg. 1997;16:187-197.
12. Heekin RD, Engh CA, Vinh T. Morselized allograft in acetabular reconstruction: A postmortem retrieval analysis. Clin Orthop Relat Res. 1995;319:184-190.
13. Hollinger JO, Brekke J, Gruskin E, Lee D. Role of bone substitutes. Clin Orthop Relat Res. 1996;324:55-65.
14. Hooten Jr JP, Engh CA, Heekin RD, Vinh TN. Structural bulk allografts in acetabular reconstruction: Analysis of two grafts retrieved at post-mortem. J Bone Joint Surg. 1996;78B:270-275.
15. Hultmark P, Kärrholm J, Strömberg C, et al. 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;15:551-561.
16. Kurashina K, Kurita H, Kotani A, et al. In vivo study of a calcium phosphate cement consisting of alpha-tricalcium phosphate/dicalcium phosphate dibasic/tetracalcium phosphate monoxide. Biomaterials. 1997;18:147-151.
17. Mazhar Tokgözogğlu A, Aydin M, Atilla B, Caner B. Scintigraphic evaluation of impaction grafting for total hip arthroplasty revision. Arch Orthop Trauma Surg. 2000;120:416-419.
18. Metsger DS, Driskell TD, Paulsrud JR. Tricalcium phosphate ceramic: A resorbable bone implant. Review and current status. J Am Dent Assoc. 1982;105:1035-1038.
19. Ogose A, Hotta T, Hatano H, et al. Histological examination of beta-tricalcium phosphate graft in human femur. J Biomed Mater Res. 2002;63:601-604.
20. Ozawa M. Experimental study of bone conductivity and absorbability of the pure β-TCP. J Jpn Soc Biomater. 1995;13:167-175.
21. Ozonek W, Rapala K, Obrebski M. The application of allogenic bone grafts for replantation of a hip prosthesis: An attempt at radiological evaluation of graft rebuilding. Ann Transplant. 1999;4:52-58.
22. Paprosky WG, Lawrence J, Cameron H. Femoral defect classification: Clinical application. Orthop Rev. 1990;19 (Suppl):9-15.
23. Piert M, Winter E, Becker GA, et al. Allogenic bone graft viability after hip revision arthroplasty assessed by dynamic [18F]fluoride ion positron emission tomography. Eur J Nucl Med. 1999;26:615-624.
24. Puri L, Wixson RL, Stern SH, et al. Use of helical computed tomography for the assessment of acetabular osteolysis after total hip arthroplasty. J Bone Joint Surg. 2002;84A:609-614.
25. Schwarz EM, Campbell D, Totterman S, et al. Use of volumetric computerized tomography as a primary outcome measure to evaluate drug efficacy in the prevention of peri-prosthetic osteolysis: A 1-year clinical pilot of etanercept vs. placebo. J Orthop Res. 2003;21:1049-1055.
26. Stamenkov R, Howie D, Taylor J, et al. Measurement of bone defects adjacent to acetabular components of hip replacement. Clin Orthop Relat Res. 2003;412:117-124.
27. Stevenson S, Emery SE, Goldberg VM. Factors affecting bone graft incorporation. Clin Orthop Relat Res. 1996;324:66-74.
28. Weber M, Hempfing A, Orler R, Ganz R. Femoral revision using the Wagner stem: Results at 2-9 years. Int Orthop. 2002;26:36-39.
29. Won YY, Dorr LD, Bloebaum R, Wan Z. Assessment of bone graft substitutes in autopsy implant retrievals of the acetabulum in total hip arthroplasty. J Arthroplasty. 2004;19:116-122.
30. Yoon ST, Boden SD. Osteoinductive molecules in orthopaedics: Basic science and preclinical studies. Clin Orthop Relat Res. 2002;395:33-43.
© 2006 Lippincott Williams & Wilkins, Inc.