JBJS | Characterization and Comparison of Wear Debris from Failed Total Hip Implants of Different Types*

The Journal of Bone & Joint Surgery, Volume 78, Issue 8

Articles | August 01, 1996

Characterization and Comparison of Wear Debris from Failed Total Hip Implants of Different Types*

KAZUO HIRAKAWA, M.D., PH.D.†; THOMAS W. BAUER, M.D., PH.D.†; BERNARD N. STULBERG, M.D.‡; ALAN H. WILDE, M.D.‡; MICHELLE SECIC, M.S.†, CLEVELAND, OHIO

View Disclosures and Other Information

Investigation performed at the Departments of Anatomic Pathology and Orthopaedic Surgery, and Biostatistics and Epidemiology, The Cleveland Clinic Foundation, and The Cleveland Center for Joint Reconstruction, Cleveland

The Journal of Bone & Joint Surgery. 1996; 78:1235-43

5 Recommendations (Recommend) | 3 Comments | Saved by 3 Users Save Case

text A A A

Abstract

Particles of wear debris have been associated with loosening of implants and with osteolysis, but few studies have examined the relationship between characteristics of the implant and clinical variables and the concentration of particles isolated from periarticular tissues. We isolated and quantified particles of wear debris from orthopaedic implants in 123 tissue samples that had been obtained adjacent to a failed total hip prosthesis from eighty-eight patients. The concentration of these particles in the tissue and the size of the particles were then analyzed in relation to patient and implant-related variables. The number of particles ranged from 8.5 x 10⁸ to 5.7 x 10¹¹ per gram of tissue (dry weight). More particles were found adjacent to failed titanium-alloy stems that had a cobalt-chromium-alloy modular head and failed titanium-alloy-backed cups than were found adjacent to all-cobalt-chromium-alloy prostheses. In addition, fewer particles were found adjacent to implants with a twenty-eight-millimeter femoral head than were found adjacent to implants with other femoral head sizes. Univariate analysis also showed correlations between a high concentration of particles and fixation without cement, an implant that had been in situ for a long duration, a young patient age, and an initial clinical diagnosis of avascular necrosis. Biopsy specimens from the proximal femoral membranes had higher concentrations than those from the joint capsules or the acetabular membranes. Although only five specimens were obtained directly from osteolytic lesions, the concentration of particles in those specimens was higher than that in biopsy specimens from other sites. Although many univariate correlations were identified, stepwise correlation regression analysis showed that the composition of the implant and the size of the modular femoral head were most strongly related to the concentration of debris in tissue.CLINICAL RELEVANCE: The results of this study show that most failed hip implants are associated with billions of debris particles and that the concentration of these particles in tissue is related to several factors that are thought to be associated with the extent of implant wear. There are probably many factors that influence the production of orthopaedic wear debris. Some are related to the design of the implant. Others, such as the level of activity, are related to the patient. Of the variables tested in this study, the composition of the implant and the size of the modular head appeared to be most closely linked with the production of particles. Our results do not necessarily reflect the extent of debris production by stable implants, but they suggest that the metallic composition of the femoral stem, the acetabular cup, and the modular head may have an important influence on the amount of wear debris.

Figures in this Article

Introduction

Introduction | Materials and Methods | Results | Discussion | References

The clinical result of total joint arthroplasty is usually excellent, but particles of wear debris have been associated with osteolytic lesions as well as with the general process of aseptic loosening^{1,4-6,8-16,18,22,24-26,28,33,34,37-39}. The importance of small particles of debris in the induction of the release of cytokines from macrophages, fibroblasts, and histiocytes has been suggested in several studies^{3,4,12,14,18,22,34}, but the size and number of particles associated with different types of implant materials and designs are unclear. Huo et al.¹⁵ used atomic absorption spectrophotometry to compare the amounts of debris associated with twelve cemented stems made of one of three different materials. They reported higher levels of metal in femoral osteolytic lesions associated with cobalt-chromium-alloy or stainless-steel implants than in the joint capsule or femoral pseudomembrane. The number of patients was small, but no such difference between the levels was identified in association with titanium stems. In addition, the levels of the particles were not associated with clinical variables or the duration of implantation. In a study of only four patients, Blumenthal et al.⁷ found no association between the concentration of metal ions and the site of the biopsy. Schmalzried et al.³² used a semiquantitative grading system and polarized light microscopy to compare the polyethylene debris adjacent to hip and knee implants that had failed. Tissues around failed knee implants showed a larger range of particle sizes, including more particles that were greater than two micrometers in diameter, but associations with other clinical or implant-related variables were not described.

Margevicius et al.²³ recently reported on a method to isolate and characterize particles of wear debris that were greater than approximately 0.5 micrometer in diameter and identifed billions of debris particles in tissues adjacent to twenty failed total joint prostheses. The purpose of the present study was to evaluate the relationship between the size and tissue concentration of the particles and the composition and design of the implant as well as other clinical factors in a larger group of samples.

*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.

†Departments of Anatomic Pathology and Orthopaedic Surgery (K. H. and T. W. B.) and Biostatistics and Epidemiology (M. S.), The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, Ohio 44195.

‡The Cleveland Center for Joint Reconstruction, 2322 East 22nd Street, Suite 102, Cleveland, Ohio 44115.

*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.

‡The Cleveland Center for Joint Reconstruction, 2322 East 22nd Street, Suite 102, Cleveland, Ohio 44115.

View Larger

TABLE I UNIVARIATE CORRELATION REGRESSION ANALYSIS OF THE RELATIONSHIP BETWEEN THE NUMBER OF PARTICLES AND CLINICAL VARIABLES*

*Sigma tests multiple observations from the same patient.†The test statistic is the correlated value between the variables.‡P value relative to Group 1.§P value relative to osteoarthrosis.¶P value relative to the femoral heads that had a diameter of twenty-two or twenty-six millimeters.

Variable	Test Statistic†	P Value
Type of implant‡
Group2	-2.83	0.006
Group 3	-2.44	0.02
Group 4	-1.35	0.18
Sigma	3.07	0.003
Original diagnosis§
Avascular necrosis	2.75	0.007
Rheumatoid arthritis	-1.64	0.103
Sigma	3.16	0.002
Duration of implantation	4.12	0.001
Sigma	3.18	0.002
Age of patient	-2.37	0.02
Sigma	2.99	0.004
Fixation with cement	-4.88	<0.001
Sigma	4.26	<0.001
Diameter of femoral head¶
28 mm	-3.11	0.003
32 mm	-1.84	0.06
Sigma	3.07	0.003
Thickness of polyethylene	0.73	0.47
Sigma	3.79	<0.001

Add to My JBJS

TABLE I UNIVARIATE CORRELATION REGRESSION ANALYSIS OF THE RELATIONSHIP BETWEEN THE NUMBER OF PARTICLES AND CLINICAL VARIABLES*

Variable	Test Statistic†	P Value
Type of implant‡
Group2	-2.83	0.006
Group 3	-2.44	0.02
Group 4	-1.35	0.18
Sigma	3.07	0.003
Original diagnosis§
Avascular necrosis	2.75	0.007
Rheumatoid arthritis	-1.64	0.103
Sigma	3.16	0.002
Duration of implantation	4.12	0.001
Sigma	3.18	0.002
Age of patient	-2.37	0.02
Sigma	2.99	0.004
Fixation with cement	-4.88	<0.001
Sigma	4.26	<0.001
Diameter of femoral head¶
28 mm	-3.11	0.003
32 mm	-1.84	0.06
Sigma	3.07	0.003
Thickness of polyethylene	0.73	0.47
Sigma	3.79	<0.001

View Larger

TABLE II ANALYSIS OF PARTICLES BY LOCATION

*The numbers in parentheses represent the results when the five osteolytic lesions were excluded.†There were no significant differences among the groups, with the numbers available.‡Particles from the proximal femoral membrane had a larger estimated surface area than did those from the other sites (p < 0.05).

	Acetabular Membrane	Joint Capsule	Proximal Femoral Membrane
No. of particles per gram (x 10⁹)*
Group 1	14.9	12.90	77.9 (23.2)
Group 2	15.8	8.49	38.9
Group 3	13.6	8.73	30.6
Group 4	11.3	7.16	33.8
Average	13.9	9.32	45.3 (19.9)
Mode of diameter of particles (µm)†
Group 1	0.88	0.82	0.96
Group 2	0.91	0.64	0.70
Group 3	0.70	0.82	0.88
Group 4	0.67	0.69	0.87
Average	0.79	0.74	0.86
Mode of surface area of particles (µm²)‡
Group 1	6.13	4.09	7.81
Group 2	4.25	4.42	3.91
Group 3	4.54	5.66	5.95
Group 4	2.24	3.63	4.49
Average	4.29	4.65	5.54

Add to My JBJS

TABLE II ANALYSIS OF PARTICLES BY LOCATION

	Acetabular Membrane	Joint Capsule	Proximal Femoral Membrane
No. of particles per gram (x 10⁹)*
Group 1	14.9	12.90	77.9 (23.2)
Group 2	15.8	8.49	38.9
Group 3	13.6	8.73	30.6
Group 4	11.3	7.16	33.8
Average	13.9	9.32	45.3 (19.9)
Mode of diameter of particles (µm)†
Group 1	0.88	0.82	0.96
Group 2	0.91	0.64	0.70
Group 3	0.70	0.82	0.88
Group 4	0.67	0.69	0.87
Average	0.79	0.74	0.86
Mode of surface area of particles (µm²)‡
Group 1	6.13	4.09	7.81
Group 2	4.25	4.42	3.91
Group 3	4.54	5.66	5.95
Group 4	2.24	3.63	4.49
Average	4.29	4.65	5.54

View Larger

TABLE III STEPWISE CORRELATION REGRESSION ANALYSIS OF THE RELATIONSHIP BETWEEN THE NUMBER OF PARTICLES AND CLINICAL VARIABLES*

*Sigma tests multiple observations from the same patient.†The test statistic is the correlated value between the variables‡P value relative to Group 1.§P value relative to the femoral heads that had a diameter of twenty-six millimeters.

Variable	Test Statistic†	P Value
Type of implant‡
Group 2	-3.78	0.03
Group 3	-4.36	0.01
Group 4	-1.58	0.22
Duration of implantation	2.43	0.001
Sigma	3.26	0.002
Diameter of femoral head§
28 mm	-2.80	0.007
32 mm	1.36	0.18
Sigma	2.46	0.02

Add to My JBJS

TABLE III STEPWISE CORRELATION REGRESSION ANALYSIS OF THE RELATIONSHIP BETWEEN THE NUMBER OF PARTICLES AND CLINICAL VARIABLES*

Variable	Test Statistic†	P Value
Type of implant‡
Group 2	-3.78	0.03
Group 3	-4.36	0.01
Group 4	-1.58	0.22
Duration of implantation	2.43	0.001
Sigma	3.26	0.002
Diameter of femoral head§
28 mm	-2.80	0.007
32 mm	1.36	0.18
Sigma	2.46	0.02

View Large | Download Slide(.ppt)
Add to My JBJS

Figs. 1-A and 1-B: Specimen from Group 4 (implants that were made of only titanium alloy). Fig. 1-A: Scanning electron micrograph showing many irregularly shaped particles (x 6400).

View Large | Download Slide(.ppt)
Add to My JBJS

Energy-dispersive x-ray spectroscopy study confirming the titanium composition of most particles. The gold coating used in the scanning electron microscopy was also detected. Fewer filamentous particles, which are consistent with polyethylene particles, were also present.

View Large | Download Slide(.ppt)
Add to My JBJS

Figs. 2-A and 2-B: Specimen from Group 2 (implants that were made of only cobalt-chromium alloy). Fig. 2-A: Scanning electron micrograph showing a single particle of cobalt-chromium alloy surrounded by particles consistent with ultra-high molecular weight polyethylene (x 6400).

View Large | Download Slide(.ppt)
Add to My JBJS

Energy-dispersive x-ray spectroscopy study showing cobalt-chromium alloy.

Materials and Methods

Introduction | Materials and Methods | Results | Discussion | References

Identification of the Samples

We reviewed our pathology files and identified 123 samples of tissue that had been obtained from eighty-eight patients at the time of a revision total hip arthroplasty and for which adequate clinical information was available. The hip implants that had been removed at the revision arthroplasties were examined, and the specimens were grouped according to the composition of the implant. Group 1 comprised fifty-one specimens from patients who had had a titanium-alloy femoral stem with a cobalt-chromium-alloy modular head and a titanium-alloy-backed acetabular cup. Group 2 consisted of thirty-three specimens from patients who had had a cobalt-chromium-alloy femoral stem with a cobalt-chromium-alloy modular head and an acetabular cup backed with cobalt-chromium alloy. Group 3 included twenty-four specimens from patients who had had a cobalt-chromium-alloy femoral stem with a cobalt-chromium-alloy modular head and a titanium-alloy-backed acetabular cup. Group 4 comprised fifteen specimens from patients who had had a titanium-alloy femoral stem with a titanium-alloy head and a titanium-alloy-backed acetabular cup. Six of the Group-4 samples were from patients who had had a modular femoral head, and the remaining nine were from patients who had had a monolithic (fixed-head) femoral component.

In addition to the composition of the implant, several other design variables are thought to influence the rate of implant wear. These include the diameter of the femoral head, the presence or absence of bone cement, and the thickness of the acetabular polyethylene. Therefore, implants were also grouped according to each of these variables. Three categories of femoral head size—twenty-six, twenty-eight, and thirty-two millimeters—were used. The number of twenty-two-millimeter femoral heads was insufficient for statistical analysis. The implants were also grouped according to whether or not bone cement had been used for the acetabular component. Grouping by the thickness of the acetabular polyethylene was more arbitrary because in several implant designs the thickness varies according to the geometry. Therefore, the approximate thickness of the polyethylene was estimated as half the difference between the stated diameter of the outer cup and the diameter of the femoral head.

Isolation and Characterization of the Particles

Particles were isolated and characterized with use of the method described by Margevicius et al.²³. In brief, slides were viewed under light microscopy and portions of the fibrous membranes that appeared to be representative of the rest of the specimen were selected visually. The corresponding tissue was then obtained directly from the paraffin blocks. No attempt was made to select areas of maximum or minimum debris. Tissues were deparaffinized in 100 per cent toluene overnight at room temperature. The toluene was replaced with two changes of absolute ethanol, and the tissue was critical-point-dried with use of liquid carbon dioxide (SPC-900; Bomar, Tacoma, Washington). Approximately 0.02 to 0.03 gram of tissue was digested with filtered 70 per cent nitric acid, washed, and sonicated. For many of the specimens, this resulted in a sedimented pellet of dark particles and a floating layer of white particles. To confirm the composition of the floating particles, a single specimen was analyzed with Fourier transform infrared spectroscopy. The results confirmed the presence of fragments of ultra-high molecular weight polyethylene in that layer. The samples were then resuspended in acetone and were centrifuged at 11,600 revolutions per minute for fifteen minutes. The low density of acetone allows sedimentation of both metal and polyethylene debris particles. The particles were resuspended in prefiltered water with ten microliters of dispersant (Coulter IB; Coulter Electronics, Hialeah, Florida) and were divided into two aliquots. One portion was collected on 0.4-micrometer and ten-micrometer filters (Millipore, Bedford, Massachusetts) and was used to characterize particle content, as will be described. The other portion was sonicated and was diluted with twenty milliliters of filtered electrolyte solution (Isoton; Coulter Electronics). The particles between 0.5 and ten micrometers in size were electronically counted and sized with use of a multisizer particle analyzer (Coulter Electronics) with a twenty-micrometer-aperture tube. Counts for each sample were obtained in duplicate.

Electrical resistance particle quantitation, as previously described, is adequate for particles that are between 0.5 and ten micrometers but has a low sensitivity for larger particles. Therefore, to analyze the relatively large particles better, each sample was also collected on ten-micrometer filters. Each filter was cut in half, and one half was viewed with transmitted and polarized light while the other half was prepared for electron microscopy, as described in the following section. Three microscope fields, representing more than 50 per cent of the area of each filter, were evaluated with use of a light microscope. In addition, an interactive image analysis system (Bioquant-IVa; R and M Biometrics, Nashville, Tennessee) was used to determine the size of all particles that were larger than ten micrometers.

Composition of the Particles

All ten-micrometer and 0.4-micrometer filters were viewed with transmitted and polarized light. In addition, particles that were isolated on the second half of the ten-micrometer filter and those isolated on the 0.4-micrometer filter were gold-coated with standard techniques (Hummer II; Technics, Alexandria, Virginia) and viewed with scanning electron microscopy (EM400T; Phillips Industries, Eindhoven, The Netherlands). Energy dispersive x-ray spectroscopy (EDX 9100; EDAX, Mahwah, New Jersey) was used to identify metal particles as well as possible contaminants. A portion of the filter from a single representative specimen was also fixed with 3.75 per cent glutaraldehyde, post-fixed with 0.8-molar osmium tetroxide, embedded in epoxy resin, and evaluated with transmission electron microscopy (EM400T; Phillips Industries).

Control Tissues

In order to verify that the particles counted from the tissue represented orthopaedic wear debris rather than undigested biological materials, synovial tissues obtained from five patients who had rheumatoid arthritis and from eight patients who had osteoarthrosis as well as periarticular tissue collected arthroscopically from seven patients who had no implants were used as negative controls.

Statistical Analysis

The total number of particles per gram of tissue was analyzed with respect to both continuous variables (the duration of implantation, the age of the patient, and the thickness of the polyethylene) and categorical variables (the type of implant, the location of the biopsy, the original diagnosis, the use of fixation with cement, and the diameter of the femoral head). Because more than one tissue sample was obtained from some patients, both univariate and stepwise correlation regression techniques were used; p values of less than 0.05 were considered significant. The number of particles was transformed with use of the natural log to reduce skewness before statistical analysis, and the extent of correlation between observations from the same patient (sigma) was also determined. The use of univariate and stepwise correlation regression allowed recognition of variables of independent significance with respect to the total number of particles. We used a more sophisticated extension of regression analysis called correlated regression. The term correlated refers to the fact that the multiple observations per subject may be correlated with each other. There is no reference to a possible correlation among the set of variables. Although the modeling technique does not provide the common r values to assess the simple relationship between two continuous variables, it does provide model coefficients, standard errors for the coefficients, and test statistics in addition to p values.

Results

Introduction | Materials and Methods | Results | Discussion | References

Information on the Patients

Sixty-eight specimens were from women and fifty-five were from men. Primary arthroplasty had been done for osteoarthrosis (fifty-seven patients), avascular necrosis (thirteen patients), or rheumatoid arthritis (eighteen patients). The average age of the patients was fifty-six years (range, twenty-three to eighty-eight years). The implants had been in situ for an average of fifty months (range, three to 144 months).

Histological Findings on Light Microscopy

Light microscopy showed particles of foreign materials in all specimens. Birefringence with polarized light suggested that many of the particles were polyethylene. Additional small particles of opaque material, probably metal, were also visible in most specimens. More specifically, coarse black particles were easily recognized in most tissues in which titanium had been used either as a bearing surface or elsewhere in the device (metal backing on an acetabular cup). Tissue from sites adjacent to implants of entirely cobalt-chromium-alloy composition (Group 2) contained occasional particles of metal, but the metal was more difficult to identify by light microscopy than it was in the other groups. Birefringent material consistent with ultra-high molecular weight polyethylene also was identified by light microscopy in most specimens. The polyethylene was usually most evident in the specimens from sites adjacent to Group-2 implants, but small particles of polyethylene were also present within histiocytes in all of the other groups. If polyethylene particles were ignored histologically, then tissue from sites adjacent to the implants of only titanium alloy (Group 4) appeared to have a greater total number of debris particles. Conversely, recognizing that the total debris load is the sum of metal and polyethylene, we found relatively good rank correlation between the over-all histological impression of the amount of debris and the results of particle quantitation (p < 0.01, r = 0.72).

Quantitation of Particles

Electronic particle quantitation showed a range of 8.5 x 10⁸ to 5.7 x 10¹¹ particles per gram of tissue (dry weight). Univariate correlation regression showed significant relationships between the number of particles and the type of implant, the original diagnosis, the duration that the implant had been in situ, the age of the patient, the use of cement, and the diameter of the femoral head (Table I). Group 1 had significantly more particles than did Groups 2 (p = 0.006) and 3 (p = 0.02).

With use of univariate correlation regression analysis, we found that biopsy specimens associated with prostheses that had been inserted without cement had significantly more particles than did those associated with cemented prostheses (p < 0.001) (Table I). Also, the specimens obtained from sites adjacent to prostheses that had a twenty-eight-millimeter femoral head had fewer particles than those near twenty-six-millimeter femoral heads (p = 0.003) and marginally fewer particles than those near thirty-two-millimeter femoral heads (p = 0.06).

When grouped by the location of the biopsy, the average concentration of particles was higher in specimens from sites near the femoral stems (45.3 x 10⁹ particles per gram) than in either those from the joint capsules (9.32 x 10⁹ particles per gram) or those from the acetabular membranes (13.9 x 10⁹ particles per gram) (p < 0.05) (Table II). Five of the biopsy specimens had been designated as representing osteolytic lesions. The small number of samples in this group precludes statistical analysis, but the high average value (1.65 x 10¹¹ particles per gram) suggests that osteolytic lesions may have a higher concentration of particles than the joint capsules or other adjacent tissues (p < 0.05). Control samples showed background counts ranging from 6.89 x 10⁷ to 1.76 x 10⁸ particles per gram. The counts, which represented a combination of electronic background and undigested tissue, were negligible compared with those from the samples from the patients.

With the numbers available, there was no significant correlation between the mode of the diameter of the particles and the type of implant (Table II). However, the mode of the surface area of the particles from the proximal femoral membranes was larger than that of the particles from the joint capsules (p < 0.05) (Table II).

The continuous variables of the duration that the implant had been in situ and the age of the patient showed significant univariate correlation with the number of particles. Specimens from sites adjacent to implants that had been in situ for a longer duration (p = 0.001) and those from patients with a younger age (p = 0.02) had significantly more particle debris (Table I). However, with the numbers available, we found no significant correlation between the thickness of the polyethylene and either the number of wear debris particles or the rate of particle accumulation (the number of particles per month). Interestingly, interclass correlation (sigma) was also significant (p < 0.001 to p = 0.004). This finding shows that, in statistical analyses of this type, it is important to account for multiple observations from each patient rather than to analyze them with use of traditional correlative methods on the assumption that they are independent.

On the basis of the factors that were found to be significant by univariate analysis, a predictor model was built in which independently significant predictors of particle debris were added in a stepwise fashion. This process is helpful for isolating variables of independent significance. The analysis showed that the type of the implant and the size of the femoral head were most strongly related to the concentration of particle debris. (Use of fixation with bone cement and the duration that the implant had been in situ as variables did not supply enough additional independent information to warrant inclusion in the model.) More specifically, there were more particles in Group 1 than in either Group 2 (p = 0.03) or Group 3 (p = 0.01) (Table III). With the numbers available, there was no significant difference between Group 1 and Group 4 (p = 0.22). Similarly, implants of all types that had a twenty-eight-millimeter femoral head were associated with fewer debris particles than those that had a twenty-six-millimeter head (p = 0.007). With the numbers available, the calculated thickness of the polyethylene was not of independent significance.

Although it was not the principal emphasis in the present study, we also compared the number of particles associated with monolithic (fixed-head) titanium femoral components and titanium femoral components with a modular titanium head. There were relatively few implants in each group (nine and six, respectively), but we found no significant difference in the total number of particles (p = 0.19) or the rate of particle production (number per month) (p = 0.15) between these groups.

Identification of the Particles

Several different techniques were used to help identify the composition of particles that were isolated on the filters. The filters were gold-coated and were viewed with the use of scanning electron microscopy with energy-dispersive x-ray spectroscopy. Energy-dispersive x-ray spectroscopy confirmed the presence of titanium, especially in the samples from Group 4 (Figs. 1-A and 1-B). Particles of titanium were also identified in samples from Groups 1 and 3. Particles consistent with cobalt-chromium alloy were identified in Group 2 (Figs. 2-A and 2-B).

We also used transmission electron microscopy with energy-dispersive x-ray spectroscopy to evaluate post-digestion residue. Ultra-high molecular weight polyethylene cannot be identified with certainty with energy-dispersive x-ray spectroscopy, but the morphology of most of the particles isolated on our filters was consistent with that of polyethylene and was similar to that of the polyethylene particles that have been described by other authors^25,30. These particles varied widely with regard to size and shape; they included granular, filamentous, flake, shard, and shredded forms. We could recognize no differences in the shapes of the polyethylene particles among the four groups. On the basis of the findings of energy-dispersive x-ray spectroscopy dot maps, debris from Group 4 had a greater proportion of metal (titanium) and a smaller proportion of polyethylene than that from the other groups.

In order to provide additional confirmation of the composition of the debris, one sample was analyzed with infrared spectroscopy. Confirming the observations of Shanbhag et al.³⁵, this sample showed a spectral pattern that matched the reference for ultra-high molecular weight polyethylene. Products of corrosion (those that contained chromium in combination with other elements)³⁶ were not identified with certainty in any of the specimens.

Discussion

Introduction | Materials and Methods | Results | Discussion | References

Particles of wear debris have been strongly linked to osteolysis and to loosening of implants in many studies^{3,6,8,10-16,18,19,22,24-27,29,31}. Several previous reports have described the sizes and types of debris particles associated with small groups of retrieved implants. Light microscopy alone³⁷ or with semiquantitative grading²⁴ is relatively easy to perform, but it is limited by the optical resolution of the light microscope. In studies in which polarized light was not used, particles of polyethylene may not have been recognized. Horowitz et al.¹⁴ used transmission electron microscopy to evaluate tissues adjacent to cemented hip prostheses, and they identified polymethylmethacrylate particles that ranged in size from about one to seventeen micrometers. Lee et al.²⁰ reported metal and polyethylene particles that ranged in size from 0.8 to 2.0 micrometers and from 0.3 to 0.7 micrometer, respectively. Shanbhag et al.³⁵ used electron microscopy and other methods to identify particles adjacent to eleven failed titanium-alloy hip replacements that had been inserted without cement. They reported most particles to be smaller than one micrometer in diameter. Calcium phosphates, ions suggestive of stainless steel, silica, and other elements were also identified, but no attempt was made to associate the size or number of the particles with clinical variables.

Relatively few investigators have attempted to find an association between the number or physical characteristics of debris particles and the materials or design of the implant or the location of the biopsy. In a study involving multiple biopsies performed on only four patients, Blumenthal et al.⁷ measured debris particles and tissue ions with use of light microscopy and atomic absorption spectroscopy, respectively. They reported no association between the extent of the debris and the location of the biopsy. Lee et al.²⁰ used light microscopy to estimate the size of metal and polyethylene particles from thirty patients who had a failed hip prosthesis. They reported no notable difference in the size of metal particles on the basis of the implant material, but they noted larger polyethylene particles from the titanium-alloy implants compared with those from either the cobalt-chromium-alloy or the stainless-steel implants. Lee et al.²⁰ did not attempt to quantify the number of particles or to find a relationship between particle parameters and clinical variables. Salvati et al.³⁰ attempted to quantify metal particles in tissues and suggested that implants composed of titanium alloy were associated with more debris particles than were cobalt-chromium-alloy implants, but the methods for collection of particles used in that study did not allow effective quantitation of polyethylene particles. On the basis of the results of atomic absorption spectrophotometry of specimens from twelve patients, Huo et al.¹⁵ reported higher concentrations of metal in regions of femoral endosteal lysis than in either the joint capsule or the femoral membranes associated with implants constructed of stainless steel. However, they found no significant differences when they studied the regions adjacent to titanium-alloy or cobalt-chromium-alloy implants. In their small group of twelve patients, they were unable to identify a significant association between the level of metal and the duration of implantation. Schmalzried et al.³² used a semiquantitative grading system and polarized light microscopy to compare polyethylene debris found adjacent to nineteen failed knee implants with that found adjacent to twenty-four failed hip implants. Particles that were smaller than one micrometer were common in both the hip and the knee specimens, but tissues associated with failed knee implants showed a larger range of particle sizes, including more particles that were greater than two micrometers in diameter.

In the present study, we attempted to evaluate the influence of the location of the biopsy, the construction of the implant, and other clinical factors on the concentration of particles around failed hip implants. Our results suggest that the concentration of particles may be higher in samples retrieved from sites around the proximal part of the femoral component than in the joint capsule or the acetabular membrane. Although the number of samples designated as representing osteolytic lesions was too small for statistical analysis, such specimens demonstrated especially high concentrations of particles (average, 1.65 x 10¹¹ particles per gram of tissue).

We evaluated the influence of implant materials on the production of particles and used stepwise correlation regression techniques to adjust for multiple biopsy specimens from some of the same patients as well as to help identify variables of independent significance (Table II). We found high concentrations of particles in all groups, but Group 1 (tissues associated with failed titanium-alloy stems, cobalt-chromium-alloy modular heads, and titanium-alloy-backed cups) had significantly more particles than did Group 2 (all-cobalt-chromium-alloy constructs) (p = 0.03) and Group 3 (cobalt-chromium-alloy stems with cobalt-chromium-alloy modular heads and titanium-alloy-backed cups) (p = 0.01). Although it is largely speculation, we believe that these data raise the possibility that the production of debris is more closely linked to the composition of the metal backing of the acetabular component than to the composition of the femoral component.

We also evaluated the impact of the original diagnosis on the over-all extent of the debris. To our knowledge, few other studies have demonstrated a difference in the rate of wear or the amount of debris on the basis of the original diagnosis. Jasty et al.¹⁷, for example, measured the percentage of the surface of retrieved modular heads that showed evidence of abrasive wear and reported a greater extent of surface scratches in implants that had been inserted without cement than in those that had been inserted with cement or both with and without cement (hybrid fixation). Although it was not specifically discussed by Jasty et al.¹⁷, a comparison of their results for the eight patients who had osteonecrosis and those for the thirty-four patients who had osteoarthrosis suggests that there was no significant association between abrasion of the modular head and the original diagnosis. Univariate correlation regression analysis of our own results suggested that patients who had had a primary arthroplasty for avascular necrosis had more periarticular wear debris than did those who had had the operation for osteoarthrosis (p = 0.007). When studied with multivariate analysis, however, this factor was no longer significant, suggesting that wear in this group may be linked with either the duration of implantation or the type of implant chosen for this population of patients.

The intent of our study design was to focus on the influence of the composition of the implant on the production of debris particles. However, we also grouped implants by several other variables that are thought to influence wear, including the use of bone cement, the size of the femoral head, the approximate thickness of the polyethylene, and the use of a modular femoral head compared with a monolithic (fixed-head) femoral component. Although relatively few monolithic implants were studied, with the numbers available we found no significant correlation between the use of a modular head and the number of particles. Similarly, we found no significant correlation between the estimated thickness of the polyethylene and the number of particles. Other investigators^2,21 have provided radiographic evidence of favorable wear characteristics for twenty-eight-millimeter femoral heads, especially when compared with thirty-two-millimeter femoral heads. Although too few twenty-two-millimeter femoral heads were available for comparison in our study, our univariate correlation regression tests as well as stepwise multivariate regression analysis suggested that fewer debris particles were associated with the twenty-eight-millimeter femoral heads than with either the twenty-six-millimeter or the thirty-two-millimeter femoral heads. In agreement with other investigators^2,21, we identified a univariate correlation between the use of bone cement and a low number of debris particles, but the findings of our stepwise correlation regression analysis suggested that the use of bone cement did not provide independent information beyond that provided by the size of the femoral head and the composition of the implant.

The results of the present investigation should be interpreted with caution for several reasons. First, the number of debris particles associated with failed implants is likely to be higher than the number associated with successful implants. Thus, these results do not necessarily reflect patterns of wear for well fixed total joint replacements. Similarly, patterns of referral of patients as well as of attainment of retrieved implants are complex. Many of the implants evaluated in the present study were retrieved at our hospital, and others were mailed to us, specifically for evaluation of wear, from other hospitals. Compared with the other groups, Group 1 had more implants from other hospitals, which suggests that we do not have a random sample of implants in each group. Although our results showed that more particles were associated with the implants from Group 1, this could have been strongly influenced by the selection of patients with high rates of wear who had been referred by other surgeons. Finally, our assay method had several disadvantages. The lower limit of detection of 0.5 micrometer is only partially overcome by parallel scanning electron microscopy. It is likely that we overestimated the size and underestimated the number of debris particles in all of the experimental groups.

With these limitations in mind, our results support those of Shanbhag et al.³⁵ and suggest that all failed total hip prostheses are associated with billions of particles of debris, most of which are smaller than one micrometer in diameter, in the adjacent tissue. Although metal debris was relatively prominent in the specimens associated with all-titanium-alloy constructs (Group 4), polyethylene was the most frequent type of particle identified in all groups and was especially prominent in the specimens associated with the all-cobalt-chromium-alloy implants (Group 2). Additional, preferably prospective, studies are needed to help clarify the influence of implant materials and design on production of debris, osteolysis, and aseptic loosening.

NOTE: The authors appreciate the help of Fredrick Van Lente, Ph.D., who performed Fourier transform infrared spectroscopy, and of James T. McMahon, Ph.D., who assisted with transmission electron microscopy, at The Cleveland Clinic Foundation.

References

Introduction | Materials and Methods | Results | Discussion | References

Amstutz, H. C.; Campbell, P.; Kossovsky, N.; and |and |Clarke, I. C.: Mechanism and clinical significance of wear debris-induced osteolysis. Clin. Orthop.,276: 7-18, 1992.2767 1992 [PubMed]

Bartel, D. L.; Bicknell, V. L.; and |and |Wright, T. M.: The effect of comformity, thickness, and material on stresses in ultra-high molecular weight components for total joint replacement. J. Bone and Joint Surg.,68-A: 1041-1051, Sept. 1986.68-A1041 1986

Bauer, T. W.: Polyethylene osteolysis [abstract]. J. Japanese Orthop. Assn.,67: 524, 1993.67524 1993

Bauer, T. W.; Saltarelli, M.; McMahon, J. T.; and |and |Wilde, A. H.: Regional dissemination of wear debris from a total knee prosthesis. A case report. J. Bone and Joint Surg.,75-A: 106-111, Jan. 1993.75-A106 1993

Betts, F.; Wright, T.; Salvati, E. A.; Boskey, A.; and |and |Bansal, M.: Cobalt-alloy metal debris in periarticular tissues from total hip revision arthroplasties: metal contents and associated histologic findings. Clin. Orthop.,276: 75-82, 1992.27675 1992 [PubMed]

Bischoff, U. W.; Freeman, M. A. R.; Smith, D.; Tuke, M. A.; and |and |Gregson, P. J.: Wear induced by motion between bone and titanium or cobalt-chrome alloys. J. Bone and Joint Surg.,76-B(5): 713-716, 1994.76-B(5)713 1994

Blumenthal, N. C.; Cosma, V.; Jaffe, W.; and |and |Stuchin, S.: A new technique for quantitation of metal particulates and metal reaction products in tissue near implants. J. Appl. Biomater.,5: 191-193, 1994.5191 1994 [PubMed][CrossRef]

Cates, H. E.; Faris, P. M.; Keating, E. M.; and |and |Ritter, M. A.: Polyethylene wear in cemented metal-backed acetabular cups. J. Bone and Joint Surg.,75-B(2): 249-253, 1993.75-B(2)249 1993

Feller, J. A.; Kay, P. R.; Hodgkinson, J. P.; and |and |Wroblewski, B. M.: Activity and socket wear in the Charnley low-friction arthroplasty. J. Arthroplasty,9: 341-345, 1994.9341 1994 [PubMed][CrossRef]

Goldring, S. R.; Clark, C. R.; and |and |Wright, T. M.: The problem in total joint arthroplasty: aseptic loosening [editorial]. J. Bone and Joint Surg.,75-A: 799-801, June 1993.75-A799 1993

Goodman, S. B.: The cellular processes of aseptic loosening of joint arthroplasties. Sem. Arthroplasty,4: 205-214, 1993.4205 1993

Haynes, D. R.; Rogers, S. D.; Hay, S.; Pearcy, M. J.; and |and |Howie, D. W.: The differences in toxicity and release of bone-resorbing mediators induced by titanium and cobalt-chromium-alloy wear particles. J. Bone and Joint Surg.,75-A: 825-834, June 1993.75-A825 1993

Hernandez, J. R.; Keating, E. M.; Faris, P. M.; Meding, J. B.; and |and |Ritter, M. A.: Polyethylene wear in uncemented acetabular components. J. Bone and Joint Surg.,76-B(2): 263-266, 1994.76-B(2)263 1994

Horowitz, S. M.; Doty, S. B.; Lane, J. M.; and |and |Burstein, A. H.: Studies of the mechanism by which the mechanical failure of polymethylmethacrylate leads to bone resorption. J. Bone and Joint Surg.,75-A: 802-813, June 1993.75-A802 1993

Huo, M. H.; Salvati, E. A.; Lieberman, J. R.; Betts, F.; and |and |Bansal, M.: Metallic debris in femoral endosteolysis in failed cemented total hip arthroplasties. Clin. Orthop.,276: 157-168, 1992.276157 1992 [PubMed]

Jasty, M.; Estok, D.; and |and |Harris, W. H.: The mechanisms involved in the failure of fixation of components in total hip arthroplasty. Sem. Arthroplasty,4: 238-247, 1993.4238 1993

Jasty, M.; Bragdon, C. R.; Lee, K.; Hanson, A.; and |and |Harris, W. H.: Surface damage to cobalt-chrome femoral head prostheses. J. Bone and Joint Surg.,76-B(1): 73-77, 1994.76-B(1)73 1994

Jiranek, W. A.; Machado, M.; Jasty, M.; Jevsevar, D.; Wolfe, H. J.; Goldring, S. R.; Goldberg, M. J.; and |and |Harris, W. H.: Production of cytokines around loosened cemented acetabular components. Analysis with immunohistochemical techniques and in situ hybridization. J. Bone and Joint Surg.,75-A: 863-879, June 1993.75-A863 1993

Kabo, J. M.; Gebhard, J. S.; Loren, G.; and |and |Amstutz, H. C.: In vivo wear of polyethylene acetabular components. J. Bone and Joint Surg.,75-B(2): 254-258, 1993.75-B(2)254 1993

Lee, J. M.; Salvati, E. A.; Betts, F.; DiCarlo, E. F.; Doty, S. B.; and |and |Bullough, P. G.: Size of metallic and polyethylene debris particles in failed cemented total hip replacements. J. Bone and Joint Surg.,74-B(3): 380-384, 1992.74-B(3)380 1992

Livermore, J.; Ilstrup, D.; and |and |Morrey, B.: Effect of femoral head size on wear of the polyethylene acetabular component. J. Bone and Joint Surg.,72-A: 518-528, April 1990.72-A518 1990

Maloney, W. J.; Smith, R. L.; Castro, F.; and |and |Shurman, D. J.: Fibroblast response to metallic debris in vitro. Enzyme induction, cell proliferation, and toxicity. J. Bone and Joint Surg.,75-A: 835-844, June 1993.75-A835 1993

Margevicius, K. J.; Bauer, T. W.; McMahon, J. T.; Brown, S. A.; and |and |Merritt, K.: Isolation and characterization of debris in membranes around total joint prostheses. J. Bone and Joint Surg.,76-A: 1664-1675, Nov. 1994.76-A1664 1994

Meldrum, R. D.; Bloebaum, R. D.; and |and |Dorr, L. D.: Metal ion concentrations in retrieved polyethylene total hip inserts and implications for artifactually high readings in tissue. J. Biomed. Mater. Res.,27: 1349-1355, 1993.271349 1993 [PubMed][CrossRef]

Owen, T. D.; Moran, C. G.; Smith, S. R.; and |and |Pinder, I. M.: Results of uncemented porous-coated anatomic total hip replacement. J. Bone and Joint Surg.,76-B(2): 258-262, 1994.76-B(2)258 1994

Revell, P. A.; Weightman, B.; Freeman, M. A. R.; and |and |Roberts, B. V.: The production and biology of polyethylene wear debris. Arch. Orthop. and Trauma Surg.,91: 167-181, 1978.91167 1978 [CrossRef]

Rogers, S. D.; Pearcy, M. J.; Hay, S. J.; Haynes, D. R.; Bramley, A.; and |and |Howie, D. W.: A method for production and characterization of metal prosthesis wear particles. J. Orthop. Res.,11: 856-864, 1993.11856 1993 [PubMed][CrossRef]

Rose, R. M., and |and |Radin, E. L.: Wear of polyethylene in the total hip prosthesis. Clin. Orthop.,170: 107-115, 1982.170107 1982 [PubMed]

Rose, R. M.; Nusbaum, H. J.; Schneider, H.; Ries, M.; Paul, I.; Crugnola, A.; Simon, S. R.; and |and |Radin, E. L.: On the true wear rate of ultra high-molecular-weight polyethylene in the total hip prosthesis. J. Bone and Joint Surg.,62-A: 537-549, June 1980.62-A537 1980

Salvati, E. A.; Betts, F.; and |and |Doty, S. B.: Particulate metallic debris in cemented total hip arthroplasty. Clin. Orthop.,293: 160-173, 1993.293160 1993 [PubMed]

Savio, J. A., III; Overcamp, L. M.; and |and |Black, J.: Size and shape of biomaterial wear debris. Clin. Mater.,15: 101-147, 1994.15101 1994 [PubMed][CrossRef]

Schmalzried, T. P.; Jasty, M.; Rosenberg, A.; and |and |Harris, W. H.: Polyethylene wear debris and tissue reactions in knee as compared to hip replacement prostheses. J. Appl. Biomater.,5: 185-190, 1994.5185 1994 [PubMed][CrossRef]

Schmalzried, T. P.; Kwong, L. M.; Jasty, M.; Sedlacek, R. C.; Haire, T. C.; O'Connor, D. O.; Bragdon, C. R.; Kabo, J. M.; Malcolm, A. J.; and |and |Harris, W. H.: The mechanism of loosening of cemented acetabular components in total hip arthroplasty. Analysis of specimens retrieved at autopsy. Clin. Orthop.,274: 60-78, 1992.27460 1992 [PubMed]

Shanbhag, A. S.; Jacobs, J. J.; Black, J.; Galante, J. O.; and |and |Glant, T. T.: Macrophage/particle interactions: effect of size, composition and surface area. J. Biomed. Mater. Res.,28: 81-90, 1994.2881 1994 [PubMed][CrossRef]

Shanbhag, A. S.; Jacobs, J. J.; Glant, T. T.; Gilbert, J. L.; Black, J.; and |and |Galante, J. O.: Composition and morphology of wear debris in failed uncemented total hip replacement. J. Bone and Joint Surg.,76-B(1): 60-67, 1994.76-B(1)60 1994

Urban, R. M.; Jacobs, J. J.; Gilbert, J. L.; and |and |Galante, J. O.: Migration of corrosion products from modular hip prostheses. Particle microanalysis and histopathological findings. J. Bone and Joint Surg.,76-A: 1345-1359, Sept. 1994.76-A1345 1994

Willert, H. G.; Bertram, H.; and |and |Buchhorn, G. H.: Osteolysis in alloarthroplasty of the hip: the role of ultra-high molecular weight polyethylene wear particles. Clin. Orthop.,258: 95-107, 1990.25895 1990 [PubMed]

Wright, T. M.; Rimnac, C. M.; Stulberg, S. D.; Mintz, L.; Tsao, A. K.; Klein, R. W.; and |and |McCrae, C.: Wear of polyethylene in total joint replacements: observations from retrieved PCA knee implants. Clin. Orthop.,276: 126-134, 1992.276126 1992 [PubMed]

Wroblewski, B. M.: Wear of high-density polyethylene on bone and cartilage. J. Bone and Joint Surg.,61-B(4): 498-500, 1979.61-B(4)498 1979

Topics

View Large | Download Slide(.ppt)
Add to My JBJS

Figs. 1-A and 1-B: Specimen from Group 4 (implants that were made of only titanium alloy). Fig. 1-A: Scanning electron micrograph showing many irregularly shaped particles (x 6400).

View Large | Download Slide(.ppt)
Add to My JBJS

Energy-dispersive x-ray spectroscopy study showing cobalt-chromium alloy.

View Larger

TABLE I UNIVARIATE CORRELATION REGRESSION ANALYSIS OF THE RELATIONSHIP BETWEEN THE NUMBER OF PARTICLES AND CLINICAL VARIABLES*

Variable	Test Statistic†	P Value
Type of implant‡
Group2	-2.83	0.006
Group 3	-2.44	0.02
Group 4	-1.35	0.18
Sigma	3.07	0.003
Original diagnosis§
Avascular necrosis	2.75	0.007
Rheumatoid arthritis	-1.64	0.103
Sigma	3.16	0.002
Duration of implantation	4.12	0.001
Sigma	3.18	0.002
Age of patient	-2.37	0.02
Sigma	2.99	0.004
Fixation with cement	-4.88	<0.001
Sigma	4.26	<0.001
Diameter of femoral head¶
28 mm	-3.11	0.003
32 mm	-1.84	0.06
Sigma	3.07	0.003
Thickness of polyethylene	0.73	0.47
Sigma	3.79	<0.001

Add to My JBJS

TABLE I UNIVARIATE CORRELATION REGRESSION ANALYSIS OF THE RELATIONSHIP BETWEEN THE NUMBER OF PARTICLES AND CLINICAL VARIABLES*

Variable	Test Statistic†	P Value
Type of implant‡
Group2	-2.83	0.006
Group 3	-2.44	0.02
Group 4	-1.35	0.18
Sigma	3.07	0.003
Original diagnosis§
Avascular necrosis	2.75	0.007
Rheumatoid arthritis	-1.64	0.103
Sigma	3.16	0.002
Duration of implantation	4.12	0.001
Sigma	3.18	0.002
Age of patient	-2.37	0.02
Sigma	2.99	0.004
Fixation with cement	-4.88	<0.001
Sigma	4.26	<0.001
Diameter of femoral head¶
28 mm	-3.11	0.003
32 mm	-1.84	0.06
Sigma	3.07	0.003
Thickness of polyethylene	0.73	0.47
Sigma	3.79	<0.001

View Larger

TABLE II ANALYSIS OF PARTICLES BY LOCATION

	Acetabular Membrane	Joint Capsule	Proximal Femoral Membrane
No. of particles per gram (x 10⁹)*
Group 1	14.9	12.90	77.9 (23.2)
Group 2	15.8	8.49	38.9
Group 3	13.6	8.73	30.6
Group 4	11.3	7.16	33.8
Average	13.9	9.32	45.3 (19.9)
Mode of diameter of particles (µm)†
Group 1	0.88	0.82	0.96
Group 2	0.91	0.64	0.70
Group 3	0.70	0.82	0.88
Group 4	0.67	0.69	0.87
Average	0.79	0.74	0.86
Mode of surface area of particles (µm²)‡
Group 1	6.13	4.09	7.81
Group 2	4.25	4.42	3.91
Group 3	4.54	5.66	5.95
Group 4	2.24	3.63	4.49
Average	4.29	4.65	5.54

Add to My JBJS

TABLE II ANALYSIS OF PARTICLES BY LOCATION

	Acetabular Membrane	Joint Capsule	Proximal Femoral Membrane
No. of particles per gram (x 10⁹)*
Group 1	14.9	12.90	77.9 (23.2)
Group 2	15.8	8.49	38.9
Group 3	13.6	8.73	30.6
Group 4	11.3	7.16	33.8
Average	13.9	9.32	45.3 (19.9)
Mode of diameter of particles (µm)†
Group 1	0.88	0.82	0.96
Group 2	0.91	0.64	0.70
Group 3	0.70	0.82	0.88
Group 4	0.67	0.69	0.87
Average	0.79	0.74	0.86
Mode of surface area of particles (µm²)‡
Group 1	6.13	4.09	7.81
Group 2	4.25	4.42	3.91
Group 3	4.54	5.66	5.95
Group 4	2.24	3.63	4.49
Average	4.29	4.65	5.54

View Larger

TABLE III STEPWISE CORRELATION REGRESSION ANALYSIS OF THE RELATIONSHIP BETWEEN THE NUMBER OF PARTICLES AND CLINICAL VARIABLES*

Variable	Test Statistic†	P Value
Type of implant‡
Group 2	-3.78	0.03
Group 3	-4.36	0.01
Group 4	-1.58	0.22
Duration of implantation	2.43	0.001
Sigma	3.26	0.002
Diameter of femoral head§
28 mm	-2.80	0.007
32 mm	1.36	0.18
Sigma	2.46	0.02

Add to My JBJS

TABLE III STEPWISE CORRELATION REGRESSION ANALYSIS OF THE RELATIONSHIP BETWEEN THE NUMBER OF PARTICLES AND CLINICAL VARIABLES*

Variable	Test Statistic†	P Value
Type of implant‡
Group 2	-3.78	0.03
Group 3	-4.36	0.01
Group 4	-1.58	0.22
Duration of implantation	2.43	0.001
Sigma	3.26	0.002
Diameter of femoral head§
28 mm	-2.80	0.007
32 mm	1.36	0.18
Sigma	2.46	0.02