Abstract
Ten patients who had had a total hip replacement with a forged cobalt-chromium-molybdenum femoral prosthesis (Precoat or Precoat Plus) inserted with cement were seen with a fatigue fracture of the stem an average of fifty months (range, nineteen to seventy-four months) postoperatively. The average age of the patients was sixty-one years (range, forty-three to seventy-three years), and the average weight was ninety-six kilograms (range, seventy to 130 kilograms). Eight patients had had a primary total hip replacement, and two had had a revision; all of the acetabular components had been inserted without cement.Radiographs that had been made before the fracture were available for four of the eight hips that had had a primary replacement; all four had radiographic evidence of debonding of the cement mantle from the proximal end of the stem. This probably caused exaggerated cantilever bending stresses on the proximal aspect of the stem as the distal end of the stem was well fixed. The radiographs of both hips that had had a revision demonstrated a non-union of the greater trochanter, which had resulted in separation at the cement-bone interface at the proximal portion of the femur before the fracture. Scanning electron micrographs of five of the ten fractured prostheses demonstrated a fatigue fracture that began near the anterolateral corner of the prosthesis, through characters that had been etched on the implant with a laser. Metallurgical analysis indicated subsurface voids or inclusions, or both, immediately under the region that had been etched. This finding is consistent with thermal changes to the microstructure of the alloy that probably caused a focal reduction in the material strength.A high proportion (seven) of the ten stems had a poor cement mantle. Also, of the seven small stems that were used, six had been implanted in patients who weighed more than eighty kilograms, so there was relative undersizing of the prostheses. Early debonding of the proximal end of a Precoat femoral prosthesis from the cement mantle may occur as a result of a thin cement mantle, leading to loosening and possibly to early fatigue fracture of the stem if the distal portion of the stem remains solidly fixed in the distal portion of the cement column.On the basis of our experience, we recommend that patients who have radiographic evidence of a debonded Precoat femoral component should be informed of the risk of fatigue fracture of the stem and be followed closely even though there may be no symptoms of loosening of the femoral component.
Although the prevalence of fracture of stainless-steel femoral components that were used early in the history of total hip replacement with cement ranged from 0.23 to 0.67 per cent in series of as many as 6500 patients1,4-6, stems made of high-strength titanium or cobalt-chromium-molybdenum alloys rarely fracture7. To our knowledge, the recent literature contains only one report of a fracture of a femoral component composed of forged cobalt-chromium-molybdenum alloy9.
The purpose of the present study was to document and characterize an unusually high prevalence of fracture of forged cobalt-chromium-molybdenum stems inserted with modern cementing techniques.
*One or more of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was DePuy, Warsaw, Indiana.
†1220 University Drive, Suite 202, Menlo Park, California 94025.
‡Marshfield Clinic, 1000 North Oak Avenue, Marshfield, Wisconsin 54449.
§Montreal General Hospital, 1650 Cedar Avenue, Montreal, Quebec HG3 1A4, Canada.
#Department of Orthopaedic Surgery, Washington University Medical Center, 1 Barnes Hospital Plaza, Suite 11300, St. Louis, Missouri 63110.
Five hundred and sixty-four primary and revision total hip procedures were performed between 1985 and 1993 by three of us (S. T. W., J. P. M., and W. J. M.) with use of one of two similar femoral components (Precoat and Precoat Plus; Zimmer, Warsaw, Indiana) designed to be inserted with cement. The femoral prostheses were made of forged cobalt-chromium-molybdenum alloy (Zimaloy; Zimmer) and had a medial collar and a modular cobalt-chromium-molybdenum femoral head fixed with a Morse taper lock. The proximal third of each stem was precoated with a thin layer of polymethylmethacrylate. The original Precoat design was modified slightly in 1988 to include surface texturing of the precoated surface (Precoat Plus) for better fixation. Both implants had been labeled by the manufacturer, by etching with a laser, on the anterior aspect of the stem (when used in a left hip) at the junction of the middle and distal thirds.
The Precoat or Precoat Plus femoral component fractured in ten patients: eight who had had a primary hip replacement and two who had had a revision. (Since the completion of the present study, one of us [J. P. M.] has seen a fatigue fracture of three additional Precoat femoral components from the original series of 564. All three patients had had a primary hip replacement on the left and had a revision procedure after the fracture.)
The femoral component had been inserted with cement and a Harris-Galante acetabular component (Zimmer) had been inserted without cement in all ten patients. The index procedure had been performed between December 1987 and June 1991, and the fracture was seen between July 1992 and February 1995, an average of fifty months (range, nineteen to seventy-four months) postoperatively. All ten patients had a revision operation for replacement of the fractured femoral component, and nine of the ten fractured implants were analyzed in detail by two of us (J. D. B. and S. Y.) (five implants) or by the manufacturer (four implants). The detailed results of the analyses of the latter four implants were kept by the manufacturer.
There were eight men and two women, and the average age at the time of the index operation was sixty-one years (range, forty-three to seventy-three years). The average height of the patients was 174 centimeters (range, 163 to 183 centimeters), and the average weight was ninety-six kilograms (range, seventy to 130 kilograms). The diagnosis at the time of the primary hip replacement was osteoarthrosis in nine patients and rheumatoid arthritis in one patient. One of the two patients who had a revision operation had had an initial diagnosis of osteoarthrosis and the other patient, rheumatoid arthritis. Nine procedures were performed on the left hip and one, on the right. All of the primary procedures were performed through a posterolateral approach, and both revision procedures were done through a transtrochanteric approach. Each femoral component was used with a twenty-eight-millimeter-diameter femoral head. Seven femoral components were small, two were medium (one, used for a revision procedure, had a long stem [225 millimeters]), and one was large. Four modular heads had a medium-length neck, and six had a long neck. All of the femoral prostheses were inserted with so-called modern cementing techniques. The femoral canal was cleaned with pulsatile lavage and dried with sponges and suction before the cement was injected with a cement gun. The porosity of the cement was reduced by vacuum-mixing. A plug was inserted into the femoral canal in nine patients; four plugs were composed of bone cement, four were made of plastic, and one was made of bone. A femoral plug was not used with the one long-stemmed femoral component.
The quality of the cement mantle in the femur was assessed on the immediate postoperative radiographs according to the classification of Barrack et al. Grade A indicated that the cement mantle completely filled the femoral canal in all areas (a so-called whiteout), extended two centimeters distal to the tip of the prosthesis, and was at least two millimeters thick in all areas without focal voids. Grade B also indicated that the mantle completely filled the femoral canal, but areas of trabecular bone surrounding the stem were not completely filled with cement. Grade C1 indicated that there were voids or bubbles in the cement, and grade C2 meant that there were focal deficiencies in the cement so that small portions of the stem were in contact with bone. Grade D indicated that the cement mantle had multiple focal deficiencies, that there was no cement distal to the tip of the stem, or that there were radiolucent lines at the cement-prosthesis interface. The cement mantle at the proximal-medial aspect of the femoral component within three centimeters distal to the collar was graded separately, as completely or focally deficient, poor (a complete mantle that was less than two millimeters thick), or good (a complete mantle that was more than two millimeters thick)10. Each prosthesis was determined to be in a neutral, varus, or valgus position with respect to the longitudinal axis of the femoral intramedullary canal. The radiographs of each hip were assessed for the presence or absence of contact between the collar of the femoral prosthesis and the calcar.
The fracture surfaces of five of the retrieved femoral implants were analyzed with scanning electron microscopy and stereomicroscopy. In addition, each of the five implants was sectioned transversely, with use of an abrasive wheel, through a region of the laser-etched label; the cut surfaces were ground, polished, and etched to reveal the microstructure of the subsurface; and the surfaces were examined with scanning electron microscopy.
Radiographic Analysis
The immediate postoperative radiographs demonstrated that the cement mantle about the femoral component was grade A in one hip, grade B in two, grade C2 in six, and grade D in one. The cement mantle at the proximal-medial aspect of the femoral component was graded as good in four and as either completely or focally deficient in six. There was initial collar-calcar contact in seven hips and no such contact in three (including one in which the component was left proud by one centimeter from the calcar). Only one femoral component was in varus (3 degrees) in relation to the femoral intramedullary canal, seven were in neutral alignment, and two were in valgus (3 and 5 degrees).
All of the fractures occurred near the junction of the middle and distal thirds of the femoral stem, and the distal portion was solidly fixed by cement (Figs. 1-A and 1-B). The distal fragment consisted of 29 to 46 per cent (average, 36 per cent) of the entire length of the stem (from the collar to the tip). Radiographs made before the fracture of the stem were available for four of the eight hips that had had a primary replacement, and all demonstrated debonding of the cement mantle from the proximal-lateral or proximal-anterior aspect, or both aspects, of the femoral prosthesis. None of the radiographs of the eight fractured stems in the hips that had had a primary replacement demonstrated loosening of the cement mantle at the proximal-lateral aspect of the femur or radiolucency at the cement-bone interface in zone 1, according to the classification of Gruen et al. However, radiographs of both hips that had had a revision showed non-union of the greater trochanter with separation of the cement mantle from the proximal-lateral aspect of the femur.
Analysis of the Fractured Implants
Gross inspection revealed polymethylmethacrylate precoating on portions of the proximal fragment of each of the ten stems. This finding indicated that the cement inserted into the femoral canal at the index operation failed to wet and polymerize completely with the precoated layer of cement on the stem. Thus, the precoated layer of cement debonded from the doughy cement that had been injected.
Analysis of the fracture surfaces of five implants revealed that the fracture had begun on the surface of the stem, near the anterolateral corner of the junction of the middle and distal thirds. So-called tidemarks or beachmarks, characteristic features of progressive propagation of a crack with application of cyclic load, could be seen emanating from the site where the crack had begun in some stems (Fig. 2). The fracture surfaces of other stems were too smeared and polished from abrasive contact for us to distinguish beachmarks. All five fractures occurred within the region where the serial numbers of the component, the manufacturer's logo, and the information regarding the size of the stem had been etched with a laser (Fig. 3). Moreover, all five fractures occurred through an etched area rather than between letters or numbers.
Scanning electron microscopy revealed that the surfaces of the etched characters were smoother than the surrounding surface of the implant; this finding is consistent with localized melting from the high-temperature laser beam. For a distance of ten to twenty micrometers below the etched region, the microstructure of the forged cobalt-chromium-molybdenum alloy was characterized by intragranular porosity and intergranular defects, suggestive of localized heat transformation (Fig. 4).
Reoperations after Fracture of the Stems
All of the reoperations were difficult, as indicated by an average duration of 218 minutes (range, 175 to 255 minutes). All acetabular components were found to be well fixed to bone, and none were revised. The proximal portion of each fractured femoral component was easily removed, but the distal fragment was well fixed to the cement mantle and the cement mantle, in turn, was well fixed to the distal part of the femur. The distal fragment of the stem was removed through a cortical femoral window in two patients, a trephine was used to drill around the distal fragment in five patients, burrs or cement-removal chisels were used in combination with a standard or an extended trochanteric osteotomy in two patients, and a proximal femoral allograft was used to reconstruct the hip in the patient who had had a revision with a long-stemmed prosthesis.
The fractured component was replaced with a fully porous-coated femoral prosthesis (Solution prosthesis; DePuy, Warsaw, Indiana) in eight patients, an ABR prosthesis (Zimmer) without cement in one patient, and a Centralign prosthesis (Zimmer) with cement in one patient.
The prevalence of fracture (2 per cent; ten implants) associated with the 564 forged high-strength cobalt-chromium-molybdenum femoral components implanted by three of us was extremely high. The prevalence may be even greater as recent follow-up radiographs were not available for all patients. Therefore, we are concerned that other patients may sustain a fracture of a cemented Precoat prosthesis in the future, as it is a fatigue phenomenon.
The mechanism of failure was determined to be initial debonding of the proximal end of the femoral component from the cement mantle in four of the eight patients who had had a primary replacement. The debonded proximal portion of the stem, which was loose within the cement mantle, was subjected to torsional cantilever bending stresses when the patient rose from a chair or climbed stairs. These stresses were concentrated on the anterolateral aspect of the distal third of the stem. We assume that this also was the mechanism of failure of the other four stems that had been inserted during primary replacements, although radiographs made before the fractures of those stems were not available. Moreover, as radiographs made after those stems had fractured showed no evidence of loosening of the cement from the proximal part of the femur, it is probable that these prostheses also became loose within the cement mantle as opposed to the cement mantle becoming loose within the femur. In the two hips for which the index procedure was a revision, a non-union of the greater trochanter caused loosening at the cement-bone interface at the proximal-lateral and proximal-anterior portions of the stem, resulting in the same stresses on the stem. In nine hips (all of the left hips), once the proximal portion of the stem had loosened, a crack was initiated at the junction of the middle and distal thirds of the stem surface, the region where a stress-riser had been produced by laser-etching.
The analysis of five fractured stems demonstrated that initiation and propagation of the crack occurred in the area that had been etched with a laser. In general, cracks initiate under tensile stresses. It is well known that the anterior and lateral surfaces of femoral prostheses are subjected to high tensile stresses during walking and stair-climbing2. These stresses are additive and generally peak at or near the anterolateral corner of the stem. Historically, analysis has shown that fractures most commonly emanate from the anterolateral corner in the middle third of the stem5,6.
Nine of the ten fractured stems were in left hips, so the laser-etching was on the anterior, or tensile, aspect of the stem, and many of the letters and numbers were large enough to extend close to the anterolateral edge. Although the details of the manufacturer's analysis of four fractured stems were not available, our general findings of fracture through the etched characters and microstructural changes in the subsurface of all five stems that we analyzed were corroborated by the manufacturer. The localized changes in the microstructure under the laser-etching were probably heat-induced as the process involves temperatures higher than the melting point of the alloy (approximately 1350 degrees Celsius). It is reasonable to assume that the microstructural changes caused a localized reduction in the strength of the alloy, effectively creating a stress-riser that precipitated the initiation of a crack. In the orthopaedic implant industry, laser-etching on crucial areas of an implant has been supplanted by the use of an electrochemical process that does not affect the fatigue strength of the alloy. The manufacturer of the components in the present study no longer uses a laser to label its femoral implants.
Although the cracks began at the site of a stress-riser caused by the laser-etching, it is perhaps more distressing that the proximal portion of the stem debonded from the cement mantle, exaggerating the cantilever bending stresses on the distal end of the stem, as early as nineteen months postoperatively. This fact is especially worrisome because it should have been possible to optimize the cementing technique during primary hip replacement. We believe that the cement mantle at the proximal portion of the stem was deficient because the femoral rasp was undersized. The precoated prosthesis was designed to improve the bond between the prosthesis and the injected cement and to increase the tensile strength of the cement-prosthesis interface.
The cement mantle was classified as grade A or B in two of the eight hips that had had a primary replacement, and it was classified as grade C2 or D in six hips because of focal voids at the proximal-medial aspect of the stem. Although the cement mantle was deficient at the proximal portion of six of these eight stems, it was excellent, providing strong fixation, at the distal end in all eight hips. The thin cement mantle at the proximal portion of the stem was probably inadequate to withstand weight-bearing and out-of-plane stresses and it fractured, resulting in anterior and lateral debonding. Once the proximal end of the stem had loosened, the bending forces generated in these relatively heavy patients, combined with the stress-riser caused by the laser-etching, resulted in fatigue failure.
Although the mechanism of proximal debonding at the cement-prosthesis interface in most of these patients was thought to be due to an insufficiently thick or uniform cement mantle in that region, we have seen early debonding of Precoat femoral components even when the cement mantle was excellent (Figs. 5-A and 5-B). In a consecutive series of 121 hybrid total hip replacements involving a Precoat prosthesis, reported in 1996 by one of us (S. T. W.) and Haber10, six femoral components (5 per cent), including two in the present study, had either a fatigue fracture or definite radiographic loosening; four of the six were subsequently revised. The cement mantle at the proximal-medial aspect of four stems was initially graded as poor (less than two millimeters thick or containing focal voids). Overall, the cement mantle was classified as grade C2 or D in twenty-four (20 per cent) of the 121 hips, and the cement mantle at the proximal-medial aspect of the stem was thin or deficient in forty hips (33 per cent). We believe that, in both the previous10 and the current series, the cement mantle was deficient because the rasp that corresponded to each size of implant was not oversized enough proximally to create sufficient space for a two-millimeter-thick cement mantle.
The seven small prostheses used in this series were implanted in patients who weighed an average of ninety-one kilograms (201 pounds); thus, the stems were relatively undersized. The small stems were used because the intramedullary canals were small and were not reamed to accommodate a larger stem. Once the small stem debonded, it was subjected to excessive cantilever bending stresses that are not normally borne by a stem that has been sized proportionally for the weight of the patient. We believe that patients who are relatively young (less than sixty-five years old) and heavy (ninety kilograms or more) and who have a narrow femoral intramedullary canal are good candidates for a porous-coated femoral component inserted without cement. The use of an implant without cement allows for reaming of the intramedullary canal and fitting of the canal with a larger-diameter implant, which is less likely to fracture than a smaller-diameter stem inserted with cement.
Proximal debonding at the cement-prosthesis interface or a non-union of the greater trochanter occurred before the fracture of the prosthesis in all ten patients; therefore, these radiographic signs, combined with a well fixed distal tip within a solid cement mantle, should warn the surgeon that a fatigue fracture of a Precoat stem may occur. When radiographic evidence of proximal debonding of a Precoat prosthesis is identified in an active patient who weighs more than eighty kilograms and has the replacement on the left side, the patient should be followed closely and should be counseled about considering a prophylactic revision before the prosthesis fractures, even if there are no symptoms. We recommend a prophylactic revision because it is easier to revise an intact, proximally debonded, femoral prosthesis than it is to revise a fractured prosthesis, as it is extremely difficult to remove the distal fragment of the stem embedded in its cement mantle.
Amstutz, H. C.; Markolf, K. L.; McNeice, G. M.; and Gruen, T. A.: Loosening of total hip components: cause and prevention. In The Hip. Proceedings of the Fourth Open Scientific Meeting of The Hip Society, pp. 102-116. St. Louis, C. V. Mosby, 1976.
Andriacchi, T. P.; Galante, J. O.; Belytschko, T. B.; and Hampton, S.: A stress analysis of the femoral stem in total hip prostheses. J. Bone and Joint Surg.,58-A: 618-624, July 1976.58-A618
1976
Barrack, R. L.; Mulroy, R. D., Jr.; and Harris, W. H.: Improved cementing techniques and femoral component loosening in young patients with hip arthroplasty. A 12-year radiographic review. J. Bone and Joint Surg.,74-B(3): 385-389, 1992.74-B(3)385
1992
Carlsson, A. S.; Gentz, C. F.; and Stenport, J.: Fracture of the femoral prosthesis in total hip replacement according to Charnley. Acta Orthop. Scandinavica,48: 650-655, 1977.48650
1977
Chao, E. Y. S., and Coventry, M. B.: Fracture of the femoral component after total hip replacement. An analysis of fifty-eight cases. J. Bone and Joint Surg.,63-A: 1078-1094, Sept. 1981.63-A1078
1981
Charnley, J.: Fracture of femoral prostheses in total hip replacement. A clinical study. Clin. Orthop.,111: 105-120, 1975.111105
1975
[PubMed]
Gilbert, J. L.; Buckley, C. A.; Jacobs, J. J.; Bertin, K. C.; and Zernich, M. R.: Intergranular corrosion-fatigue failure of cobalt-alloy femoral stems. A failure analysis of two implants. J. Bone and Joint Surg.,76-A: 110-115, Jan. 1994.76-A110
1994
Gruen, T. A.; McNeice, G. M.; and Amstutz, H. C.: "Modes of failure" of cemented stem-type femoral components. A radiographic analysis of loosening. Clin. Orthop.,141: 17-27, 1979.14117
1979
[PubMed]
Miller, E. H.; Shastri, R.; and Shih, C.-I.: Fracture failure of a forged vitallium prosthesis. A case report. J. Bone and Joint Surg.,64-A: 1359-1363, Dec. 1982.64-A1359
1982
Woolson, S. T., and Haber, D. F.: Primary total hip replacement with insertion of an acetabular component without cement and a femoral component with cement. Follow-up study at an average of six years. J. Bone and Joint Surg.,78-A: 698-705, May 1996.78-A698
1996