Revision Total Hip Arthroplasty Performed After Fracture of a Ceramic Femoral Head : A Multicenter Survivorship Study

Allain, Jérôme MD; Roudot-Thoraval, Françoise MD; Delecrin, Joel MD; Anract, Philippe MD; Migaud, Henri MD; Goutallier, Daniel MD

Journal of Bone & Joint Surgery - American Volume:
Scientific Article
Supplementary Content

Background: The alumina ceramic femoral head was introduced for total hip arthroplasty approximately thirty years ago. One of its main drawbacks was the risk of implant fracture. The aim of this study was to examine the results of revision total hip replacement performed specifically to treat a fracture of a ceramic femoral head and to identify technical factors that affected the outcomes.

Methods: One hundred and five surgical revisions to treat a fracture of a ceramic femoral head, performed at thirty-five institutions, were studied. The patients were examined clinically by the operating surgeon at the time of the last follow-up. The surgeon provided the latest follow-up radiographs, which were compared with the immediate postoperative radiographs. The success of the revisions was assessed with Kaplan-Meier survivorship analysis, with the need for repeat revision as the end point. We evaluated the complication rate and radiographic changes indicative of implant loosening. The average duration of follow-up between the index revision and the last clinical and radiographic review was 3.5 years (range, six months to twenty years).

Results: Following the revisions, radiographic evidence of cup loosening was seen in twenty-two hips (21%) and radiographic evidence of femoral loosening was seen in twenty-two (21%). One or several repeat revisions were necessary in thirty-three patients (31%) because of infection (four patients), implant loosening (twenty), osteolysis (eight), or fracture of the revision femoral head component (one). The survival rate at five years was 63% (95% confidence interval, 51% to 75%). The survival rate was significantly worse when the cup had not been changed, when the new femoral head was made of stainless steel, when a total synovectomy had not been done, and when the patient was less than fifty years old.

Conclusions: Fracture of a ceramic femoral head component is a rare but potentially serious event. A suitable surgical approach, including total synovectomy, cup exchange, and insertion of a cobalt-chromium or new ceramic femoral ball minimizes the chance of early loosening of the implants and the need for one or more repeat revisions.

Level of Evidence: Therapeutic study, Level IV (case series [no, or historical, control group]). See Instructions to Authors for a complete description of levels of evidence.

Author Information

Jérôme Allain, MD; Françoise Roudot-Thoraval, MD; Daniel Goutallier, MD; Service de Chirurgie Orthopédique (J.A. and D.G.) and Service de Santé Publique (F.R.-T.), Hôpital Henri Mondor, 51 avenue du Maréchal de Lattre de Tassigny, 94010 Créteil CEDEX, France. E-mail address for J. Allain:

Joel Delecrin, MD; Service de Chirurgie Orthopédique, Centre Hospitalo-Universitaire Saint-Jacques, 85 rue Saint-Jacques, 44035 Nantes CEDEX 1, France

Philippe Anract, MD; Service de Chirurgie Orthopédique, Centre Hospitalo-Universitaire Cochin, 27 rue du Faubourg Saint Jacques, 75014 Paris, France

Henri Migaud, MD; Service de Chirurgie Orthopédique B, Centre Hospitalo-Universitaire Roger Sallengro, 2 Avenue Oscar Lambrey, 59037 Lille CEDEX, France

Article Outline

Alumina ceramic femoral heads were introduced for total hip arthroplasty in the early 1970s 1,2 and, because of their theoretical wear characteristics 3,4, they are still used. To date, more than 2.5 million have been implanted 5. Because of its superior smoothness, another ceramic, zirconia 6, was introduced in the 1980s, but use of this implant had unpredictable results 7. Compared with metals, ceramic is weak in tension and is brittle 8-10, and fracture, especially of the early implants, is a well-recognized problem 11-15. The fracture rate has decreased, from approximately one in 2000 to one in 10,000 implants 5,16,17, but fractures are still encountered, especially because an unknown number of older implants still survive. Contamination of the joint by particulate ceramic debris is a recognized sequela of fracture 18-20, but, because of the lack of long-term follow-up, no common strategy for salvage of a hip with a fractured ceramic femoral head has been advanced, to our knowledge. This retrospective study was performed to compare the successes and complications of several different treatment alternatives.

Back to Top | Article Outline

Materials and Methods

One hundred and five patients who had undergone revision surgery specifically to treat a fracture of a ceramic femoral head and who had been followed for a minimum of six months were included in this retrospective multicenter study. The short minimum duration of follow-up was due to cases of early repeat revision because of failure of the index revision. The cases were collected from thirty-five centers. The average age of the patients at the time of the index revision was 60.5 years (range, thirty-three to eighty-four years).

The fractures of the ceramic heads occurred at an average of 3.9 years (range, less than one to sixteen years) postoperatively. The prostheses had been implanted between 1977 and 1999. Eighteen types of cups from fourteen manufacturers were used. One hundred and two cups had a polyethylene liner, and three had an alumina liner. The cup was cemented in seventy-two hips and was inserted without cement in thirty-two; it was not known whether cement had been used for the cup in one hip. A monoblock cup was used in sixty-nine hips, the cup had a metal backing in thirty-five, and the type of cup was not specified for one hip. The femoral stem was made of titanium in eighty-nine hips and of cobalt-chromium in one; the type of material used for the femoral stem was not specified for fifteen. The stem was cemented in eighty-six hips and was implanted without cement in seventeen; it was not known whether cement had been used for the stem in two hips. The femoral head was made of alumina in ninety-eight hips and of zirconia in seven.

The patients were examined by the operating surgeon, who also provided radiographs made at the last follow-up visit. Either the surgeon or we interpreted the radiographs and reported the findings. Postoperative complications (fracture of the revision ceramic head, infection, or postoperative prosthetic dislocation) were noted. Either the surgeon or we compared the immediate postoperative and last radiographs of the index revision prosthesis to determine the position of the femoral and acetabular implants and to look for periprosthetic osteolysis and radiolucent lines between the implants and bone. The position of the cup relative to the pelvis and the position of the femoral implant relative to the lesser trochanter were also assessed 7. We assessed loosening of the femoral implant according to the criteria of Harris et al. 21, who defined definite loosening as evidence of migration, the appearance of a new radiolucent line at the stem-cement junction that had not been seen on the immediate postoperative radiographs, a discernible shift in the position of the femoral component and/or the cement mantle, or the appearance of a crack in the cement. We modified the criteria of Massin et al. 22 for cup loosening by considering a variation of >5 mm or >5° between the immediate postoperative and latest radiographs or the presence of a cement fracture as evidence of such loosening. The average duration of follow-up between the index revision for treatment of the ceramic head fracture and the last examination at which radiographs were made was 3.5 years (range, six months to twenty years).

The last radiographs before the thirty-three repeat revisions were analyzed, and the intraoperative findings during the repeat revisions were reported by the individual surgeons. Special attention was paid to the presence of metallosis, defined as a grossly visible black infiltration of the surrounding tissues. Specimens from eleven patients were analyzed histologically with use of Perls reaction 23, and the surfaces of the femoral head and acetabular implants from four hips were studied with scanning electron microscopy and a microprobe analysis of the energy-dispersive spectrum. The Morse taper was grossly inspected and was classified as normal, slightly scratched, or notched.

Back to Top | Article Outline
Statistical methods

The success of the index revision arthroplasties was assessed with Kaplan-Meier survivorship analysis, with at least one repeat revision as the end point. Continuous variables were reported as means and standard deviations, and p values for differences in these variables between groups were derived with the t test or the Mann-Whitney U test. Categorical variables were reported as proportions, and p values for differences between groups were derived with the chi-square or Fisher exact test. Cumulative survivorship curves were constructed with Kaplan-Meier methods, and comparisons of curves were performed with the log-rank test. A p value of <0.05 was considered significant. A multivariable analysis with a logistic regression test was performed to determine factors that were independently associated with the results. Odds ratios and their 95% confidence intervals are given for each factor.

Back to Top | Article Outline


Intraoperative Observations and Therapeutic Choices During Index Revisions

Metallosis was found in forty-five hips during the index revisions. The cup was revised in eighty-nine hips and was left in place in fourteen; the treatment of the cup was not specified for two hips. Of the fourteen retained cups, ten had no intraoperative macroscopic abnormality, two had slight wear, and two were slightly scratched. A new polyethylene cup was implanted in eighty-seven hips, and an alumina cup was used in two. Grossly, the Morse taper was normal in twenty-five hips, slightly scratched in fifty-nine, and notched in fourteen; no comment was made concerning the condition of the taper in seven hips. The femoral stem was revised in fifty-five hips and left in place in fifty. All of the femoral stems with a notched Morse taper were revised, but seventeen with a slightly scratched Morse taper were left in place. The new femoral head was made of stainless steel in fifty-three hips, alumina in fifteen, nitrated titanium in twelve, reinforced stainless steel (Bionium; Biomécanique integrée) in twelve, cobalt-chromium in nine, and zirconia in three; the material of the new femoral head was not specified for one hip. The stem was left in place in five of the eighteen hips in which a new ceramic head was implanted, and it was changed to provide a new Morse taper in thirteen. Of the fifty cemented femoral stems that were removed, twelve were simply tapped out and thirty-five were revised with a procedure that involved cement removal; there was no description of three of the stem removals. No loose implant (stem or cup) was retained. A partial synovectomy was done during twenty-four of the revisions, whereas a complete as possible synovectomy was done in sixty-seven. The extent of the synovectomy was not specified for fourteen hips.

Back to Top | Article Outline

A ceramic head that had been used for the revision fractured in one hip in which a slightly scratched Morse taper had not been changed. There were four deep infections, and eight patients had postoperative prosthetic dislocations.

Back to Top | Article Outline
Survivorship of the Revision Total Hip Prostheses

Survival analysis showed that, at five years after the index revision, the survival rate was 63% (95% confidence interval, 51% to 75%) ( Fig. 1 ). A second revision had been performed in thirty-three hips (31%); a third, in eight of them; and a fourth, in two of them.

Back to Top | Article Outline
Indications for Repeat Revisions

Of the thirty-three repeat revisions, twenty were due to prosthetic loosening. Sixteen of the twenty were performed because of both femoral and acetabular loosening; three, because of acetabular loosening alone; and one, because of femoral loosening alone. Of the remaining repeat revisions, eight were done because of osteolysis associated with a distorted metallic femoral head ( Fig. 2 ); four, because of infection; and one, because of a fracture of the new ceramic head.

Back to Top | Article Outline
Intraoperative Findings During the Thirty-three Repeat Revisions

Macroscopic metallosis was found intraoperatively in twenty hips (61%) ( Fig. 3 ), and it was associated with osteolysis, always affecting both the acetabulum and the femur, in sixteen (48%) of the twenty. Particles of alumina were found in thirteen hips (39%), either in the surrounding tissues (seven hips) or embedded in the polyethylene cup (six). Twenty of the twenty-six stainless-steel femoral heads were distorted, evidencing metallic wear, at the repeat revision ( Fig. 4 ).

Back to Top | Article Outline
Radiographic Outcomes of the Implants After the Index Revision

Radiographically, loosening of the cup was seen in twenty-two hips (21%), and loosening of the femoral component was seen in twenty-two. Sixteen of these hips had loosening of both components. Periprosthetic osteolysis appeared on the latest radiographs of twenty-two hips.

Back to Top | Article Outline
Factors Influencing Results

Change of the cup during the index revision: A repeat revision was performed in eight (57%) of the fourteen hips in which the cup had been retained and in twenty-three (26%) of the eighty-nine in which it had been revised.

Change of the stem during the index revision: Twenty-two (44%) of the fifty hips in which the stem had been retained and eleven (20%) of the fifty-five in which the stem had been revised had a repeat revision.

Material of the new femoral head inserted during the index revision: The rate of repeat revision was twenty-five (47%) of fifty-three for the hips in which a stainless-steel femoral head had been used, one (11%) of nine for those in which a cobalt-chromium head had been used, and three (17%) of eighteen for those in which a ceramic head had been used.

Extent of synovectomy: A repeat revision was done in sixteen (67%) of the twenty-four hips with a partial synovectomy and in thirteen (19%) of the sixty-seven with a complete synovectomy.

Comparison of survival rates in relation to surgical techniques used during the revision: The survival rate of the index total hip prostheses at five years was 36% (95% confidence interval, 1% to 71%) when the cup had not been changed and 68% (95% confidence interval, 55% to 81%) when it had been changed (p = 0.018), the rate was 49% (95% confidence interval, 32% to 66%) when a stainless-steel femoral head had been implanted for the index revision and 76.5% (95% confidence interval, 59% to 94%) when a ceramic head had been used (p = 0.004), and the rate was 49.5% (95% confidence interval, 28% to 60%) when a partial synovectomy had been performed and 73% (95% confidence interval, 58% to 88%) when a complete synovectomy had been done (p = 0.005) (see Appendix).

At an average of sixty-five months (range, eight to 144 months) after the index revision, no repeat revision had been performed in any of the seventeen hips in which the cup had been changed, a new cobalt-chromium or ceramic head had been fitted onto a new femoral stem, and a total synovectomy had been performed. Repeat revisions were more frequent in younger patients, with the average age (and standard deviation) being 63.5 ± 2 years for the patients who did not have a repeat revision and 57 ± 4 years for those who did (p = 0.02).

Multivariable analysis showed that two factors were independently related to the absence of a need for repeat revision: total synovectomy (odds ratio, 0.12; 95% confidence interval, 0.04 to 0.37) and age (odds ratio, 0.94 per year of age; 95% confidence interval, 0.90 to 0.99).

Back to Top | Article Outline
Histological Findings

Histological examination of specimens of the surrounding tissues in eleven hips treated with repeat revision after implantation of a stainless-steel femoral head to replace the fractured ceramic head revealed intense, diffuse metallosis and infiltration of the neosynovial tissue by macrophages and giant multinucleated cells. Stainless-steel material was identified in the neosynovial tissue by histochemical reaction (hematoxylin, eosin, and safranin stain). Analysis of the surfaces of four femoral head implants with a scanning electron microscope revealed the presence of adherent particles. A microprobe analysis of the energy-dispersive spectrum demonstrated alumina in these particles. The same analysis was done on the polyethylene cups. The scanning electron microscope revealed the existence of a third, additional component that, according to the microprobe analysis, was made up of polyethylene wear debris and alumina particles. The analysis of the bottom of the cup demonstrated an accumulation of abrasive material lying on the polyethylene. The material was composed of alumina particles (200 to 600 μm in size) with some traces of metallic debris.

Back to Top | Article Outline


Fracture of a ceramic femoral head is rare today 5,16,17, but this complication can be disastrous, with the risk of several revisions. The operative technique for the treatment of a ceramic head fracture was very variable in our series. Three parameters appeared to be important: cup revision, the material of the femoral head, and total synovectomy.

We believe that the cup should be removed at the time of revision, even when it appears normal macroscopically, because microscopic ceramic particles may be embedded in it. A new ceramic femoral head can be inserted so that the toughness can be at least equal to that of the first femoral head implant, and to avoid wear of the femoral head, but a cobalt-chromium head also provides very satisfactory results. In contrast, a stainless-steel femoral head should not be used because of the abnormal wear that can occur. We believe that the wear found in our study resulted from alumina particles that were still present in the periprosthetic soft tissues and that amalgamated with the polyethylene wear debris. This led to the formation of a third abrasive component (polyethylene wear debris and alumina particles) at the interface between the cup and the femoral head, which promoted the wear of the stainless-steel femoral head.

Our study did not allow us to determine the best approach with regard to the Morse taper. If a new ceramic femoral head is used, a modification in the geometry of the femoral Morse taper may lead to a mismatch between the bore of the head and the metal taper, and this can cause high stress concetrations and predispose the ceramic femoral head to refracture 5,24. In our series, of the five revision ceramic heads that were implanted without exchange of the Morse taper, one fractured. Therefore, a ceramic femoral head probably should not be used on an existing Morse taper at the time of a revision arthroplasty 25. This concept necessitates removal of the femoral stem, which can be a very complex surgical procedure when the stem is stable. The use of a new head made of cobalt-chromium avoids this problem and provides satisfactory clinical results. Finally, a synovectomy has to be performed as extensively as possible to remove as much of the ceramic debris as possible. Use of the above approach in the rare cases of ceramic femoral head fracture that we identified led to satisfactory medium-term survivorship.

Back to Top | Article Outline


Survivorship curves related to exchange of the cup at revision, the material of the new femoral head, and the performance of synovectomy are available in the electronic versions of this article, on our web site at (go to the article citation and click on Supplementary Material) and on our quarterly CD-ROM (call our subscription department, at 781-449-9780, to order the CD-ROM).

Note: The authors thank C. Argenson, J.M. Artigou, B. Balay, Z. Bassot, J. Bejui-Hugues, M. Bercovy, P. Biancharelli, J.P. Blanchard, P. Boileau, J.M. Bonnet, D. Burgot, J.C. Cartillier, M. Chanzy, A. Dambreville, A. Duquennoy, M. Fleuriel, J.L. Guillamon, D. Huten, X. Hy, J.Y. Jenny, I. Kempf, J.C. Lambotte, F. Langlais, P. Le Couteur, F. Lecuire, J.L. Lerat, J. Letenneur, J.P. Levai, J.H. Marotte, B. Melchior, J.Y. Nordin, C. Nourissat, C. Pages, N. Passuti, J.M. Postel, J. Preaut, H. Robert, H. Robin, E. Roland, G. Saillant, M. Serrault, B. Tayon, and S. Terver for their contributions to this study.

Investigation performed at Service de Chirurgie Orthopédique, Hôpital Henri Mondor, Créteil, France

The authors did not receive grants or outside funding in support of their research or preparation of this manuscript. They did not receive payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated.

1. Boutin P. Alumina and its use in surgery of the hip. (Experimental study). Presse Med , 1971;79: 639-40. French.
2. Boutin P, Christel P, Dorlot JM, Meunier A, de Roquancourt A, Blanquaert D, Herman S, Sedel L, Witvoet J. The use of dense alumina-alumina ceramic combination in total hip replacement. J Biomed Mater Res , 1988;22: 1203-32.
3.     Boutin PM. T.H.R. using alumina-alumina sliding and a metallic stem: 1330 cases and an 11 years follow-up. In: Oonishi H, Ooi Y, editors. Orthopaedic ceramic implants. Volume 1. Proceedings of the Japanese Society of Orthopaedic Ceramic Implants, 1981. p 11-8.
4. Boehler M, Knahr K, Plenk H Jr, Walter A, Salzer M, Schreiber V. Long-term results of uncemented alumina acetabular implants. J Bone Joint Surg Br , 1994;76: 53-9.
5. Willmann G. Ceramic femoral head retrieval data. Clin Orthop , 2000;379: 22-8.
6. Garvie RC, Urbani C, Kennedy DR, McNeuer JC. Biocompatibility of magnesia-partially stabilized zirconia (Mg-PSZ ceramics). J Mater Sci , 1984;19: 3224-8.
7. Allain J, Le Mouel S, Goutallier D, Voisin MC. Poor eight-year survival of cemented zirconia-polyethylene total hip replacement. J Bone Joint Surg Br , 1999;81: 835-42.
8. Christel P, Meunier A, Heller M, Torre JP, Peille CN. Mechanical properties and short-term in-vivo evaluation of yttrium-oxide-partially-stabilized zirconia. J Biomed Mater Res , 1989;23: 45-61.
9. Willert HG, Puls P. [The reaction of bone to bone-cement in the replacement of the hip joint]. Arch Orthop Unfall Chir , 1972;72: 33-71. German.
10. Christel PS. Zirconia: the second generation of ceramics for total hip replacement. Bull Hosp Jt Dis Orthop Inst , 1989;49: 170-7.
11. Hummer CD 3rd, Rothman RH, Hozack WJ. Catastrophic failure of modular zirconia-ceramic femoral head components after total hip arthroplasty. J Arthroplasty , 1995;10: 848-50.
12. Kempf I, Semlitsch M. Massive wear of a steel ball head by ceramic fragments in the polyethylene acetabular cup after revision of a total hip prosthesis with fractured ceramic ball. Arch Orthop Trauma Surg , 1990;109: 284-7.
13. Mangione P, Pascarel X, Vinciguerra B, Honton JL. [Fracture of the ceramic head in total hip prosthesis. Apropos of 2 cases]. Int Orthop , 1994;18: 359-62. French.
14. Otsuka NY, Schatzker J. A case of fracture of a ceramic head in total hip arthroplasty. Arch Orthop Trauma Surg , 1994;113: 81-2.
15. Peiro A, Pardo J, Navarrete R, Rodriguez-Alonso L, Martos F. Fracture of the ceramic head in total hip arthroplasty. Report of two cases. J Arthroplasty , 1991;6: 371-4.
16. Fritsch EW, Gleitz M. Ceramic femoral head fractures in total hip arthroplasty. Clin Orthop , 1996;328: 129-36.
17. Sedel L. Evolution of alumina-on-alumina implants: a review. Clin Orthop , 2000;379: 48-54.
18. Allain J, Goutallier D, Voisin MC, Lemouel S. Failure of a stainless-steel femoral head of a revision total hip arthroplasty performed after a fracture of a ceramic femoral head. A case report. J Bone Joint Surg Am , 1998;80: 1355-60.
19. Griss P, Heimke G. Five years experience with ceramic-metal-composite hip endoprostheses. I. clinical evaluation. Arch Orthop Trauma Surg , 1981;98: 157-64.
20. Higuchi F, Shiba N, Inoue A, Wakebe I. Fracture of an alumina ceramic head in total hip arthroplasty. J Arthroplasty , 1995;10: 851-4.
21. Harris WH, McCarthy JC Jr, O'Neill DA. Femoral component loosening using contemporary techniques of femoral cement fixation. J Bone Joint Surg Am , 1982;64: 1063-7.
22. Massin P, Schmidt L, Engh CA. Evaluation of cementless acetabular component migration. An experimental study. J Arthroplasty , 1989;4: 245-51.
23.     Stevens A, Chalk BD. Perls Prussian reaction for ferric iron. In: Bancroft JD, Stevens A, editors. Theory and practice of histological techniques. 4th ed. Edinburgh: Churchill Livingstone: 1996. p 245.
24. Callaway GH, Flynn W, Ranawat CS, Sculco TP. Fracture of the femoral head after ceramic-on-polyethylene total hip arthroplasty. J Arthroplasty , 1995;10: 855-9.
25. Pulliam IT, Trousdale RT. Fracture of a ceramic femoral head after a revision operation. A case report. J Bone Joint Surg Am , 1997;79: 118-21.
Copyright 2003 by The Journal of Bone and Joint Surgery, Incorporated