- Email: [email protected]
Uncemented acetabular components
The Journal of Arthroplasty Vol. 8 No. 2 1993
U n c e m e n t e d Acetabular C o m p o n e n t s Histologic Analysis of Retrieved Hydroxyapatite-coated and Porous Implants Thomas
W . B a u e r , M D , P h D , * B e r n a r d N. S t u l b e r g , MD,']- J i a n g M i n g , M D , * a n d R u d o l p h G. T. G e e s i n k , M D , PhD:]:
Abstract: Histologic sections of five hydroxyapatite-coated acetabular components retrieved at autopsy (2 dual geometry and 3 threaded cup designs) were analyzed, and the extent and pattem of bone apposition were compare d wit h that of an uncoated, porous (beaded) dual geometry acetabular cup that had been removed for repeated dislocations. The results show hydroxyapatite and bone apposition most prominent in areas of likely load transmission, primarily around the peripheral rim of the dualgeometry cups and at the peaks of the threadsof the screw cups. Hydroxyapatite and bone may have been removed by remodeling from the grooves between the threads. Although no significant wear debris was identified, the screw holes were incompletely covered by bone. The dual-geometry designs generally showed somewhat more bone apposition than the threaded cups, but all implants demonstrated less bone than might have been predicted from radiographs. Key words: hydroxyapatite, dual geometry, porous, bone apposition, threaded cup.
h u m a n clinical trials, 4-~ hydroxyapatite ( H A ) m a y be helpful in facilitating implant-fixation. The purpose of this study is to d o c u m e n t the extent and pattern of bone apposition of two ttA-coated dual-geometry cups and three HA-coated threaded cups retrieved at autopsy from clinically successful hip arthroplasties. For comparison, we also report a clinically successful porous (beaded) cup of similar design that was removed for causes other than fixation failure.
Uncemented acetabular fixation has enjoyed widespread use since its introduction in 1983. However, as with cemented acetabular components, early clinical success appears to be giving way to problems relating to design, materials, and fixation. Early migration was thought to be a technical problem related to the implant or its insertion, but late migration n o w suggests insufficient fixation of m a n y implants, with subsequent motion and deleterious effects on acetabular bone remodeling. Polyethylene wear debris may also contribute to loosening. ~~ Evolution of the design of u n c e m e n t e d acetabular components appears necessary to ensure predictable long-term fixation, and based on the results of animal studies and
Materials and Methods Five HA-coated acetabular components were retrieved from three patients at the time of autopsy. The evaluation of the corresponding retrieved HAcoated femoral components from these hips have beeo previously reported. ~The single porous component was retrieved at revision surgery performed for repeated dislocations, but the implant was felt to be stable and well fixed.
*From the Departments of Pathology and Orthopedic Surgery, The Cleveland Clinic Foundation, Cleveland, Ohio. tFrom the Centerfor JohTt Reconstruction, Cleveland, Ohio. $From the Department of Orthopedic Surgery, Unitersity Hospital, State University of Lhnburg, Maastricht, The Netherlands. Reprint requests: Thomas W. Bauer, MD, PhD, Department of Anatomic Pathology, L25, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195-5138.
167
168
The Journal of Arthroplasty Vol. 8 No. 2 April 1993
Specimen processing methods have been previously reported.' Briefly, after initial fixation in buffered formalin, each implant was rough cut into smaller blocks to facilitate further fixation and embedding. The autopsy specimens were sectioned serially, yielding five to six blocks for each implant, while the clinically retrieved specimen was sectioned radially, resulting in four blocks, one for each quadrant. The blocks were radiographed, dehydrated, embedded in plastic, ground, and polished to approximately 50 g m thickness. The final sections were stained with hematoxylin and eosin. Histomorphometry by interactive image analysis (R & M Biometrics, Nashville, TN) was used to measure the linear extent of HA and the linear extent of bone apposition (expressed as percent). For the beaded specimen, the percent of bone ingrowth was also calculated, with bone i n g r o w t h defined as any bone penetrating greater than one half of one bead diameter.' ' Results were ~xpressed as percent of surface length or area available for bone ingrowth. For all acetabular components, bone apposition or ingrowth was m e a s u r e d separately for the peripheral rim and the d o m e of the implant.
Case Reports Case 1: HA-coated Dual-geometry Component A 55-year-old man underwent staged bilateral hip arthroplasty in April 1989 (right) and August 1989 (left) for osteoarthritis. He received ttA-coated dualgeometry acetabular components bilaterally (Osteonics, Allendale, N J). These were inserted in uncemented fashion without the use of screw fixation. His clinical function was excellent up to the time of death from myocardial infarction in January 1990. Histologic sections of the retrieved femoral components showed HA coating on more than 80% of the circumference of the implant at four different levels, and bone apposition to the femoral components ranged
Table
Case NoJSlde
Type
1R 1L
DG DG
2R 2L 3 4
Screw Screw Screw* Beads
Fig. 1. Section radiograph of right acetabular component of case 1. Bone apposition and, perhaps, trabecular hypertrophy around the rim are evident.
from 50 to 75%, most prominent in areas of probable load transmission (anterior and medial surfaces of the femoral stem). ~ The right acetabular c o m p o n e n t had been in situ for 9 months. For this implant, bone apposition to the rim and dome are similar (Table 1), but qualitatively, there appears to be a greater trabecular bone area in the bone tissue adjacent to the rim than over the dome (Plate 1, Fig. 1). Best seen in the specimen radiograph (Fig. 1), a thin lucent line is present over a portion of the dome, probably corresponding to the m a x i m u m depth of reaming. This " g a p " does not appear to have been filled with bone in this 9 - m o n t h specimen. Higher magnification of the d o m e (Fig. 2) shows that this space is occupied by loose, fibrous connective tissue. Hydroxyapatite is present over approximately 65% of the dome. In other portions of the dome, bone apposition was present directly against the implant without evidence of a fibrous m e m b r a n e or osteolysis (Plate 2). No specific p a n e m of HA loss relative to implant geometry is recognized. The screw holes in this implant are incompletely covered with bone (Plates I, 3A). The left acetabular c o m p o n e n t had been in situ for
1. Comparison of Hydroxyapatite and Porous Implants Months Implanted
Apposition (%) Dome Rim
tlA Remaining (%) Dome Rim
Bone Ingrowth (%) Dome Rim
9
30
29
63
55
NA
NA
4.5 13 25 6 60
35 18 18 9 0
51 40 42 6 31
94 69 73 65 NA
92 90 89 65 NA
NA NA NA NA 0
NA NA NA NA 22
HA, hydroxyapatite; DG, dual geometry; NA, not applicable. "Used in revision.
Uncemented Acetabular Components
9 Bauer et al.
169
Fig. 3. Specimen radiograph of the left acetabular component (case 1) before sectioning. Although orientation is not optimal, no radiolucencies are identified.
.q
B
of one of the screw holes (Plate 3B). Polarized light revealed rare, very small particles consistent with polyethylene in the screw hole, although no cellular reaction is present and there is no osteo]ysis. Comparing the histologic sections with a specimen radiograph (Fig 3) suggests that the extent of bone apposition might be overestimated by evaluating the clinical or specimen radiograph alone.
Case 2: HA-coated Threaded Acetabular Components Fig. 2. Higher magnification of the dome of the right cup from case l. The gap between the metal implant (M) and trabecular bone (B) contains loose, fibrous tissue (F). The ttA coating is not apparent, but granules of IIA and probable metal debris are present within scattered macrophages (arrow). No polyethylene debris have been identified. Other areas of the dome showed intact HA (hematoxylin and eosin with partially polarized light).
4.5 months. Trabecular bone area adjacent to the rim again appears to be greater than that over the dome (Plate 4), and bone apposition is quantitatively greater on the rim (Table 1). The pattem of bone apposition to the rim appears to show a nonrandom distribution, with bone apposition most prominent at the "peaks," and least prominent in the "valleys" of this surface (Plates 5, 6). The distribution of HA coating on the rim parallels that of bone. Both bone and HA have been resorbed from the "valleys" of the rim (Plates 5, 6). The dome of this implant is histologically similar to the contralateral cup. Although HA can be identified over greater than 90% of the surface, bone apposition is approximately 35%. The screw holes are incompletely covered with bone, and loose, fibrous tissue is present in the well
A 37-year-old woman with steroid-treated autoimmune chronic liver disease underwent total hip arthroplasty for avascular necrosis in April 1988 (left) and March 1989 (right). She did well orthopedically, but died of complications of liver disehse in May 1990. ttistologic sections of the femoral components showed that HA was present over 65-95% of the circumference, with 35-75% bone apposition. ~ Low-magnification photomicrographs (Plate 7) of the left acetabular component (25 months) illustrate more prominent bone apposition on the rim compared with the dome. Higher magnification again shows nonrandom apposition, with bone present primarily on the "peaks" of the threads (Plate 8). Bone and HA are essentially absent in the "valleys." Hydroxyapatite and bone are focally present against the dome of the implant, but less prominent histologically than might have been suggested by clinical or specimen radiographs (Table 1). The right acetabular component (13 months) also shows the most prominent bone apposition to the peaks of the threads around the rim, with bone and HA-16S~"from tile valleys between the threads (Plate 9). Overall, bone appositi'on is more prominent around the rim than the dome (Table 1).
170
The Journal of Arthroplasty Vol. 8 No. 2 April 1993
Case 3: HA-coated Threaded Acetabular Component With Uncemented Revision A 54-year-old man with a history of osteonecrosis developed aseptic loosening of his cemented femoral component. In June 1989, he undenvent revision arthroplasty, receiving uncemented, HA-coated femoral and screw-cup acetabular components. He was fully weight bearing at 4 months, but died 6 months after the arthroplasty from complications of a perforated intestinal diverticulum. Sections of the retrieved femoral component showed tlA over 80-98% of the surface; b o n e apposition was 35-40% proximally and greater than 60% near the distal aspect of the HA coating. ~ Histologic sections illustrate relatively little bone apposition (Plate 10). In a few areas of the rim, bone apposition shows the same pattern as described above. An area of bone loss is present adjacent to the threads, however, and this appears continuous with a gap over tile adjacent dome (Plate 10B). While this space is consistent with the concept of an "expanded joint space, ''m it might also reflect incomplete filling of the space Occupied by the previous cemented acetabular component. No particles of polyethylene were identified within the fibrous membrane, and n a histiocyte-mediated osteolysis was recognized. The space is occupied by a sparsely cellular fibrous membrane without giant cells, although a few particles of probable metal debris are present. Hydroxyapatite can be recognized over approximately 65% of the dome, but bone apposition is minimal (Table 1).
Fig. 4. Radiographs of a porous dual-geometry acetabular "component (case 4) show features suggestive of ingrowth especially along the superior aspect of the rim.
Case 4: Porous (Beaded) Dual-geometry Omnifit Acetabular Component An adult woman underwent total hip arthroplasty in 1987, but experienced dislocation several months after surgery. She subsequently experienced two to three dislocations a year, and the acetabular component was revised at 5 years. There was no evidence of infection and, except for frequent dislocations, the patient had been doing well (Fig. 4). At the time of revision, only the cup was removed with difficulty. Four representative histologie sections each demonstrate good bone apposition and ingrowth into the rim of the dual-geometry cup, with relatively little ingrowth over the dome (Fig. 5, Plate 11) (Table 1). In the rim there is extensive ingrowth into the first (superficial) layer of beads, but bone rarely penetrates the entire thickness of the bead layer to the substrate. Ingrowth is more prominent in the superior aspect of the rim than inferiorly. Overall bone. ingrowth in the rim is approximately 22%. There is no histologic evidence of polyethylene or metal wear debris. As in the previous cases, clinical and microra-
Fig. 5. Histologic section through the dome of the porous acetabular cup. No bone ingrowth is present in this microscopic field taken from the dome of the implant. There is no histologic evidence of infection or wear debris.
Uncemented Acetabular Components
diographs overestimated the degree of bone apposition, especiaIIy over the dome.
Discussion One purpose of evaluating retrieved implants is to determine whether the devices have functioned in vivo in the manner in which they were designed. For implants designed to attach directly to bone, the histology can provide direct evidence of the degree of attachment, and may allow some extrapolation of that attachment to mechanical performance. Relatively few studies have reported the histologic analysis of retrieved threaded acetabular components. Bobyn and co-workers 2 analyzed a single threaded type of acetabular component (MECRING) obtained at autopsy from a 49-year-old man who died of liver failure 27 months after revision arthroplasw. The clinical and radiographic performance of the implant up to the time of death was reported to be excellent, and specimen radiographs suggested good bone apposition to the grooves of the threads. Microradiographs of Bobyn's case suggested less bone apposV tion than expected, however, and histologic sections reveals that the interface was entirely fibrous with bone separated from the implant by a circumferential fibrous membrane of 0.5-1.0 mm minimum thickness. No ultra-high molecular weight polyethylene wear debris was detected around the dome of the implant. Our results with the three HA-coated threaded cups demonstrated bone apposition with persistent HA over the tips of the threads, but less bone in the grooves of the threads or over the dome. This suggests bone remodeling with trabecular hypertrophy in response to load, with possible stress shielding elsewhere. The difficulty of obtaining circumferential contact and uniform stress transfer with this design is apparent. The extent of ingrowth into clinically satisfactory porous acetabular components has also been incompletely documented. Cook 3 summarized the histologic findings of 42 uncemented porous acetabular components that had been retrieved for reasons other than clinical loosening. Bone ingrowth was absent in 33% of the implants studied, there was less than 2% ingrowth in 29%, and 2-5% ingrowth in 24%. Only 14% demonstrated bone ingrowth in greater than 5% of the available area. Specific patterns of ingrowth was most frequent around the pegs and screws, with substantially less bone seen at the dome. Tooke and co-workers ~2 compared threaded and porous titanium acetabular cup designs in dogs at 2 and 6 months. One of the three 6-month threaded
9
Bauer et al.
171
cups showed 9% bone apposition, while the other threaded implants were covered by only a fibrous membrane. The porous cups showed 0-70% ingrowth at 6 months. Ingrowth did not correspond consistently with anatomic location.X2 In another canine study, Harris and co-workers 9 evaluated uncemented hemispherical acetabular components with two to four layers of chromiumcobalt alloy beads. Microradiography from components retrieved at 6 and 8 months following implantation suggested that if a substantial gap existed between the dome of the component and bone at the time of insertion, bone was unlikely to fill the gap. Overall, these authors demonstrated that "the most extensive ingrowth often occurred at the periphery of the acetabulum, where dense trabeculae frequently were visible." These observations are consistent with our findings, in which periphei-al attachment was common. This was especially prominent in the dualgeometry cups (both HA-coated and porous) for which 0ptimal insertion includes peripheral interference fit. Oui" results also demonstrated focal loss of ttA, especially from the dome of the implants, but the mechanism of HA loss is unclear. Although osteoelastic resorption associated with bone remodeling is likely, especially in the shielded areas of the rim, micromotion of the implant or bone could have been associated with focal abrasion. Although small fragments of HA and metal were present within scattered histiocytes, the histologic appearance of the adjacent bone and fibrous tissue did not suggest a proliferative histiocytic response with secondary osteolysis. Clinical and retrieval results with HA-coated femoral components suggest that it is possible to obtain excellent initial bone apposition to HA-coated devices. We previously hypothesized that active bone remodeling occurs, and that the mechanical properties Of the device, coupled with the magnitude and direction of transmitted loads, ultimately determine the site and extent of bone attachment. 12 The results of this study appear to support that hypothesis. Bone apposition oriented relative to the textured surface at the corners of the dual-geometry component suggests preferential contact and loading in all three of these related designs, regardless of the specific surface modifications. The degree of bone attachment at the periphery of the ItA-coated devices suggests that, as with the femoral components, HA is highly conducive to bone apposition, but may be subject to active biologic and perhaps mechanical compromise. No ad~er'~6 consequences of the resorbed HA are apparent in these specimens. For both the femoraP and acetabular components,
172
The Journal of Arthroplasty Vol. 8 No. 2 April 1993
!:.
....... ~T.: : ~ " ' ( " : - ~ - ) . : .,&-~ _.,~ - - . _~ > ~ . _ ~ . ~ i > ~ j
.
~
-..--
i : ~ i . ~ Z ~ 9
~
J
'
5
-
", -....,,:, -
. .
,
. . . . _ , ._-. :. - ~ , ~ - : . ~ r
:. ;
....... - 7 - - - : ,. ,i..
:. : - ~- ~'." ,'?-,~i.~7,/., ,',',
. "
-'-.:,..,'a~,..:'~,J
,'.Jar
~'-]:.
.
9
'
"
~. ~.~.r
"..~..
.
",7-:,,~
i "1
Plate 1
Plate 2
.
Plate 3A
.
'
,
Plate 3B
P l a t e 1. Low-magnification photomicrograph of section through the fight acetabular cup of case 1. Hydroxyapatite cannot be recognized at this magnification, but there appears to be more trabecular bone near the rim than near the dome of the implant. This screw hole is only partially bridged by bone. P l a t e 2. A different area of the dome from the right acetabular component of case 1. Bone apposition to the metal implant (M) has occurred without a fibrous membrane. P l a t e 3. Photomicrographs of screw wells through the metal backing of case I: (A) right, (B) left. Pale stained fibrous tissue is present in each well, and although there is focal good bone apposition to HA (ttA), holes are only partially sealed with bone. Very rare, small particles of metal and polyethylene (not visible at this magnification) are present in the fibrous tissue.
Uncemented Acetabular Components
9
Bauer et al.
173
I
.
.4
Plate 4
Plate 5
Plate 6
P l a t e 4. Low magnification of the left acetabular component from case 1. Bone apposition is more prominent against the rim than the dome (absent bone from the right side is an artifact of slide preparation). P l a t e 5. Bone apposition is most evident on the "peaks" of the knurled~surface of the rim with relatively less bone in the "valley." P l a t e 6. At a higher magnification, both bone and HA (HA) are present over the "peaks," with a remodeling canal (RC) near the "'valleys." Remodeling appears to be associated with loss of both bone and IIA.
174
The Journal of Arthroplasty Vol. 8 NO. 2 April 1993
!
I
i 9 / f
~
Plate 7B
Plate 7A
Plate 8
P l a t e 7. Low-magnification photomicrographs from the left (25-months) acetabular component of case 2. Bone apposition is most prominent on the "peaks" of the threads, with less bone in the "valleys" of the threads or the dome. P l a t e 8. Higher magnification of the threaded portion of the left acetabular component. Both bone and ItA are present on the "peaks" of the threads with loss of both bone and ttA from the "valleys."
Uncemented Acetabular Components
9
Bauer et al.
175
L Plate 9B
Plate 9A
Plate 10A
i -
k
.
.
.
.
.-,~
..
.
Plate 10B
Plate 10C
P l a t e 9. Low-magnification photomicrographs of the right acetabular component of case 2. Although overall bone apposition is relatively low, remodeling appears non,random, with definite orientation relative to the peaks of the threads. P l a t e 10. Histologic sections and radiograph of a rough-cut block of uncemented revision threaded cup (case 3). Although focal bone apposition is present in the area of the threads, overall bone_apposition is relatively low. A fibrous membrane is present over a portion of the threads (F) and much of the dome. It is unclear whether this reflects motion or is simply a consequence of the previous cemented acetabular component. The apparent difference in metal thickness in the sections reflects increasingly tangenital cuts near the edge of a curved implant rather than true differences in thickness.
176
The Journal of Arthroplasty Vol. 8 No. 2 April 1993
.
-
Plate 11
Plate 11. Histologic section of the rim of the dual-geometry implant of case 4. There is good bone ingrowth through the outer layer of beads, although bone rarely reaches the metal substrate beneath the inner bead layer.
Uncemented Acetabular Components
bone apposition appears to be most prominent in areas of transmitted load, and bone apposition is least prominent where load is not transmitted, or in areas of relatively poor initial fit (eg, revision arthroplasty). Although remodeling of bone according to Wolfe's law was evident in both acetabular and femoral components, the magnitude of b o n e apposition was s o m e w h a t greater in the femoral stems than the corresponding acetabular cups. Because areas of load transmission usually correspond to contact between the implant and endosteal cortex, the differences in magnitude of b o n e apposition between these types of implants might reflect, in part, the greater extent of endosteal cortex available in the f e m u r c o m p a r e d to the primarily trabecular bed of the acetabulum. Finally, as noted by Bobyn and co-workers, 2 radiographic studies are likely to substantially overestim a t e the amoufit of bone apposition achieved by an acetabular c o m p o n e n t . This relates to attempting to use two-dimensional radiographs to detect apposition over a three-dimensional curved surface. Thus, one should anticipate that even in clinically satisfactory cases, a proportion of the bone-acetabular implant interface m a y be fibrous.
Acknowledgments The authors gratefully acknowledge the following surgeons for submitting specimens with radiographs for analysis: Richard Z i m m e r m a n , MD, Portland Orthopedic Clinic, Portland, Oregon, and J o h n Cardea, MD, Medical College of Virginia, Richmond, Virginia.
9
Bauer et al.
177
References 1. Bauer TW, Geesink RGT, Zimmerman R, McMahon JT: Hydroxyapatite coated femoral stems. J Bone Joint Surg 73A:1439, 1991 2. Bobyn JD, Engh CA, Glassman AII: Radiography and histology of a threaded acetabular implant. J Bone Joint Surg 70B:302, 1988 3. Cook SD: Clinical, radiographic and histologic evaluation of retrieved human non-cement porous coated implants. J Long Term Effects Med Implants 1:11, 1991 4. Cook SD, Thomas ttA, Kay JF, Jarcho M: ttydroxyapatite-coated porous titanium for use as an orthopedic biologic attachment system. Clin Orthop 230:303, 1988 5. D'Antonio JA, Capello WN, Jaffe WL: tIydroxyapatire-coated hip implants: multicenter three year clinical and radiographic results. Clin Orthop 285:102, 1992 6. Geesink RGT: Hydroxyapatite-coatcd total hip prostheses: two-year clinical and roentgenographic results of 100 cases. Clin Orthop 261:39, 1990 7. Geesink RGT, De Groot K, Klein CPAT: Chemical implant fixation using hydroxyapatite coatings. Clin Orthop 225:147, 1987 8. Geesink RGT, De Groot K, Klein CPAT: Bonding of bone to hydroxyapatite-coated implants. J Bone Joint Surg 70B:17, 1988 9. Harris WH, White RE, McCarthy JC et ah Bony ingrowth fixation of the acetabular component in canine hip joint arthroplasty. Clin Orlhop 176:7, 1983 10. Schmalzried TP, Kwong LM, Jasty M et ah The mecha9nism of loosening Of cemented acetabular components in total hip arthroplasty. Clin Orthop 274:60, 1992 1 I. Sumner DR, Jacobs J J, Turner TM et al: The amount and distribution of bone ingrowth in tibial components retrieved from human patients. Trans Orthop Res Soc 14:375, 1989 12. Tooke SM, Nugent PJ, Chotivichit Aet ah Comparison of in vivo cementless acetabular fixation. Clin Orthop 235:253, 1988
U n c e m e n t e d Acetabular C o m p o n e n t s Histologic Analysis of Retrieved Hydroxyapatite-coated and Porous Implants Thomas
W . B a u e r , M D , P h D , * B e r n a r d N. S t u l b e r g , MD,']- J i a n g M i n g , M D , * a n d R u d o l p h G. T. G e e s i n k , M D , PhD:]:
Abstract: Histologic sections of five hydroxyapatite-coated acetabular components retrieved at autopsy (2 dual geometry and 3 threaded cup designs) were analyzed, and the extent and pattem of bone apposition were compare d wit h that of an uncoated, porous (beaded) dual geometry acetabular cup that had been removed for repeated dislocations. The results show hydroxyapatite and bone apposition most prominent in areas of likely load transmission, primarily around the peripheral rim of the dualgeometry cups and at the peaks of the threadsof the screw cups. Hydroxyapatite and bone may have been removed by remodeling from the grooves between the threads. Although no significant wear debris was identified, the screw holes were incompletely covered by bone. The dual-geometry designs generally showed somewhat more bone apposition than the threaded cups, but all implants demonstrated less bone than might have been predicted from radiographs. Key words: hydroxyapatite, dual geometry, porous, bone apposition, threaded cup.
h u m a n clinical trials, 4-~ hydroxyapatite ( H A ) m a y be helpful in facilitating implant-fixation. The purpose of this study is to d o c u m e n t the extent and pattern of bone apposition of two ttA-coated dual-geometry cups and three HA-coated threaded cups retrieved at autopsy from clinically successful hip arthroplasties. For comparison, we also report a clinically successful porous (beaded) cup of similar design that was removed for causes other than fixation failure.
Uncemented acetabular fixation has enjoyed widespread use since its introduction in 1983. However, as with cemented acetabular components, early clinical success appears to be giving way to problems relating to design, materials, and fixation. Early migration was thought to be a technical problem related to the implant or its insertion, but late migration n o w suggests insufficient fixation of m a n y implants, with subsequent motion and deleterious effects on acetabular bone remodeling. Polyethylene wear debris may also contribute to loosening. ~~ Evolution of the design of u n c e m e n t e d acetabular components appears necessary to ensure predictable long-term fixation, and based on the results of animal studies and
Materials and Methods Five HA-coated acetabular components were retrieved from three patients at the time of autopsy. The evaluation of the corresponding retrieved HAcoated femoral components from these hips have beeo previously reported. ~The single porous component was retrieved at revision surgery performed for repeated dislocations, but the implant was felt to be stable and well fixed.
*From the Departments of Pathology and Orthopedic Surgery, The Cleveland Clinic Foundation, Cleveland, Ohio. tFrom the Centerfor JohTt Reconstruction, Cleveland, Ohio. $From the Department of Orthopedic Surgery, Unitersity Hospital, State University of Lhnburg, Maastricht, The Netherlands. Reprint requests: Thomas W. Bauer, MD, PhD, Department of Anatomic Pathology, L25, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195-5138.
167
168
The Journal of Arthroplasty Vol. 8 No. 2 April 1993
Specimen processing methods have been previously reported.' Briefly, after initial fixation in buffered formalin, each implant was rough cut into smaller blocks to facilitate further fixation and embedding. The autopsy specimens were sectioned serially, yielding five to six blocks for each implant, while the clinically retrieved specimen was sectioned radially, resulting in four blocks, one for each quadrant. The blocks were radiographed, dehydrated, embedded in plastic, ground, and polished to approximately 50 g m thickness. The final sections were stained with hematoxylin and eosin. Histomorphometry by interactive image analysis (R & M Biometrics, Nashville, TN) was used to measure the linear extent of HA and the linear extent of bone apposition (expressed as percent). For the beaded specimen, the percent of bone ingrowth was also calculated, with bone i n g r o w t h defined as any bone penetrating greater than one half of one bead diameter.' ' Results were ~xpressed as percent of surface length or area available for bone ingrowth. For all acetabular components, bone apposition or ingrowth was m e a s u r e d separately for the peripheral rim and the d o m e of the implant.
Case Reports Case 1: HA-coated Dual-geometry Component A 55-year-old man underwent staged bilateral hip arthroplasty in April 1989 (right) and August 1989 (left) for osteoarthritis. He received ttA-coated dualgeometry acetabular components bilaterally (Osteonics, Allendale, N J). These were inserted in uncemented fashion without the use of screw fixation. His clinical function was excellent up to the time of death from myocardial infarction in January 1990. Histologic sections of the retrieved femoral components showed HA coating on more than 80% of the circumference of the implant at four different levels, and bone apposition to the femoral components ranged
Table
Case NoJSlde
Type
1R 1L
DG DG
2R 2L 3 4
Screw Screw Screw* Beads
Fig. 1. Section radiograph of right acetabular component of case 1. Bone apposition and, perhaps, trabecular hypertrophy around the rim are evident.
from 50 to 75%, most prominent in areas of probable load transmission (anterior and medial surfaces of the femoral stem). ~ The right acetabular c o m p o n e n t had been in situ for 9 months. For this implant, bone apposition to the rim and dome are similar (Table 1), but qualitatively, there appears to be a greater trabecular bone area in the bone tissue adjacent to the rim than over the dome (Plate 1, Fig. 1). Best seen in the specimen radiograph (Fig. 1), a thin lucent line is present over a portion of the dome, probably corresponding to the m a x i m u m depth of reaming. This " g a p " does not appear to have been filled with bone in this 9 - m o n t h specimen. Higher magnification of the d o m e (Fig. 2) shows that this space is occupied by loose, fibrous connective tissue. Hydroxyapatite is present over approximately 65% of the dome. In other portions of the dome, bone apposition was present directly against the implant without evidence of a fibrous m e m b r a n e or osteolysis (Plate 2). No specific p a n e m of HA loss relative to implant geometry is recognized. The screw holes in this implant are incompletely covered with bone (Plates I, 3A). The left acetabular c o m p o n e n t had been in situ for
1. Comparison of Hydroxyapatite and Porous Implants Months Implanted
Apposition (%) Dome Rim
tlA Remaining (%) Dome Rim
Bone Ingrowth (%) Dome Rim
9
30
29
63
55
NA
NA
4.5 13 25 6 60
35 18 18 9 0
51 40 42 6 31
94 69 73 65 NA
92 90 89 65 NA
NA NA NA NA 0
NA NA NA NA 22
HA, hydroxyapatite; DG, dual geometry; NA, not applicable. "Used in revision.
Uncemented Acetabular Components
9 Bauer et al.
169
Fig. 3. Specimen radiograph of the left acetabular component (case 1) before sectioning. Although orientation is not optimal, no radiolucencies are identified.
.q
B
of one of the screw holes (Plate 3B). Polarized light revealed rare, very small particles consistent with polyethylene in the screw hole, although no cellular reaction is present and there is no osteo]ysis. Comparing the histologic sections with a specimen radiograph (Fig 3) suggests that the extent of bone apposition might be overestimated by evaluating the clinical or specimen radiograph alone.
Case 2: HA-coated Threaded Acetabular Components Fig. 2. Higher magnification of the dome of the right cup from case l. The gap between the metal implant (M) and trabecular bone (B) contains loose, fibrous tissue (F). The ttA coating is not apparent, but granules of IIA and probable metal debris are present within scattered macrophages (arrow). No polyethylene debris have been identified. Other areas of the dome showed intact HA (hematoxylin and eosin with partially polarized light).
4.5 months. Trabecular bone area adjacent to the rim again appears to be greater than that over the dome (Plate 4), and bone apposition is quantitatively greater on the rim (Table 1). The pattem of bone apposition to the rim appears to show a nonrandom distribution, with bone apposition most prominent at the "peaks," and least prominent in the "valleys" of this surface (Plates 5, 6). The distribution of HA coating on the rim parallels that of bone. Both bone and HA have been resorbed from the "valleys" of the rim (Plates 5, 6). The dome of this implant is histologically similar to the contralateral cup. Although HA can be identified over greater than 90% of the surface, bone apposition is approximately 35%. The screw holes are incompletely covered with bone, and loose, fibrous tissue is present in the well
A 37-year-old woman with steroid-treated autoimmune chronic liver disease underwent total hip arthroplasty for avascular necrosis in April 1988 (left) and March 1989 (right). She did well orthopedically, but died of complications of liver disehse in May 1990. ttistologic sections of the femoral components showed that HA was present over 65-95% of the circumference, with 35-75% bone apposition. ~ Low-magnification photomicrographs (Plate 7) of the left acetabular component (25 months) illustrate more prominent bone apposition on the rim compared with the dome. Higher magnification again shows nonrandom apposition, with bone present primarily on the "peaks" of the threads (Plate 8). Bone and HA are essentially absent in the "valleys." Hydroxyapatite and bone are focally present against the dome of the implant, but less prominent histologically than might have been suggested by clinical or specimen radiographs (Table 1). The right acetabular component (13 months) also shows the most prominent bone apposition to the peaks of the threads around the rim, with bone and HA-16S~"from tile valleys between the threads (Plate 9). Overall, bone appositi'on is more prominent around the rim than the dome (Table 1).
170
The Journal of Arthroplasty Vol. 8 No. 2 April 1993
Case 3: HA-coated Threaded Acetabular Component With Uncemented Revision A 54-year-old man with a history of osteonecrosis developed aseptic loosening of his cemented femoral component. In June 1989, he undenvent revision arthroplasty, receiving uncemented, HA-coated femoral and screw-cup acetabular components. He was fully weight bearing at 4 months, but died 6 months after the arthroplasty from complications of a perforated intestinal diverticulum. Sections of the retrieved femoral component showed tlA over 80-98% of the surface; b o n e apposition was 35-40% proximally and greater than 60% near the distal aspect of the HA coating. ~ Histologic sections illustrate relatively little bone apposition (Plate 10). In a few areas of the rim, bone apposition shows the same pattern as described above. An area of bone loss is present adjacent to the threads, however, and this appears continuous with a gap over tile adjacent dome (Plate 10B). While this space is consistent with the concept of an "expanded joint space, ''m it might also reflect incomplete filling of the space Occupied by the previous cemented acetabular component. No particles of polyethylene were identified within the fibrous membrane, and n a histiocyte-mediated osteolysis was recognized. The space is occupied by a sparsely cellular fibrous membrane without giant cells, although a few particles of probable metal debris are present. Hydroxyapatite can be recognized over approximately 65% of the dome, but bone apposition is minimal (Table 1).
Fig. 4. Radiographs of a porous dual-geometry acetabular "component (case 4) show features suggestive of ingrowth especially along the superior aspect of the rim.
Case 4: Porous (Beaded) Dual-geometry Omnifit Acetabular Component An adult woman underwent total hip arthroplasty in 1987, but experienced dislocation several months after surgery. She subsequently experienced two to three dislocations a year, and the acetabular component was revised at 5 years. There was no evidence of infection and, except for frequent dislocations, the patient had been doing well (Fig. 4). At the time of revision, only the cup was removed with difficulty. Four representative histologie sections each demonstrate good bone apposition and ingrowth into the rim of the dual-geometry cup, with relatively little ingrowth over the dome (Fig. 5, Plate 11) (Table 1). In the rim there is extensive ingrowth into the first (superficial) layer of beads, but bone rarely penetrates the entire thickness of the bead layer to the substrate. Ingrowth is more prominent in the superior aspect of the rim than inferiorly. Overall bone. ingrowth in the rim is approximately 22%. There is no histologic evidence of polyethylene or metal wear debris. As in the previous cases, clinical and microra-
Fig. 5. Histologic section through the dome of the porous acetabular cup. No bone ingrowth is present in this microscopic field taken from the dome of the implant. There is no histologic evidence of infection or wear debris.
Uncemented Acetabular Components
diographs overestimated the degree of bone apposition, especiaIIy over the dome.
Discussion One purpose of evaluating retrieved implants is to determine whether the devices have functioned in vivo in the manner in which they were designed. For implants designed to attach directly to bone, the histology can provide direct evidence of the degree of attachment, and may allow some extrapolation of that attachment to mechanical performance. Relatively few studies have reported the histologic analysis of retrieved threaded acetabular components. Bobyn and co-workers 2 analyzed a single threaded type of acetabular component (MECRING) obtained at autopsy from a 49-year-old man who died of liver failure 27 months after revision arthroplasw. The clinical and radiographic performance of the implant up to the time of death was reported to be excellent, and specimen radiographs suggested good bone apposition to the grooves of the threads. Microradiographs of Bobyn's case suggested less bone apposV tion than expected, however, and histologic sections reveals that the interface was entirely fibrous with bone separated from the implant by a circumferential fibrous membrane of 0.5-1.0 mm minimum thickness. No ultra-high molecular weight polyethylene wear debris was detected around the dome of the implant. Our results with the three HA-coated threaded cups demonstrated bone apposition with persistent HA over the tips of the threads, but less bone in the grooves of the threads or over the dome. This suggests bone remodeling with trabecular hypertrophy in response to load, with possible stress shielding elsewhere. The difficulty of obtaining circumferential contact and uniform stress transfer with this design is apparent. The extent of ingrowth into clinically satisfactory porous acetabular components has also been incompletely documented. Cook 3 summarized the histologic findings of 42 uncemented porous acetabular components that had been retrieved for reasons other than clinical loosening. Bone ingrowth was absent in 33% of the implants studied, there was less than 2% ingrowth in 29%, and 2-5% ingrowth in 24%. Only 14% demonstrated bone ingrowth in greater than 5% of the available area. Specific patterns of ingrowth was most frequent around the pegs and screws, with substantially less bone seen at the dome. Tooke and co-workers ~2 compared threaded and porous titanium acetabular cup designs in dogs at 2 and 6 months. One of the three 6-month threaded
9
Bauer et al.
171
cups showed 9% bone apposition, while the other threaded implants were covered by only a fibrous membrane. The porous cups showed 0-70% ingrowth at 6 months. Ingrowth did not correspond consistently with anatomic location.X2 In another canine study, Harris and co-workers 9 evaluated uncemented hemispherical acetabular components with two to four layers of chromiumcobalt alloy beads. Microradiography from components retrieved at 6 and 8 months following implantation suggested that if a substantial gap existed between the dome of the component and bone at the time of insertion, bone was unlikely to fill the gap. Overall, these authors demonstrated that "the most extensive ingrowth often occurred at the periphery of the acetabulum, where dense trabeculae frequently were visible." These observations are consistent with our findings, in which periphei-al attachment was common. This was especially prominent in the dualgeometry cups (both HA-coated and porous) for which 0ptimal insertion includes peripheral interference fit. Oui" results also demonstrated focal loss of ttA, especially from the dome of the implants, but the mechanism of HA loss is unclear. Although osteoelastic resorption associated with bone remodeling is likely, especially in the shielded areas of the rim, micromotion of the implant or bone could have been associated with focal abrasion. Although small fragments of HA and metal were present within scattered histiocytes, the histologic appearance of the adjacent bone and fibrous tissue did not suggest a proliferative histiocytic response with secondary osteolysis. Clinical and retrieval results with HA-coated femoral components suggest that it is possible to obtain excellent initial bone apposition to HA-coated devices. We previously hypothesized that active bone remodeling occurs, and that the mechanical properties Of the device, coupled with the magnitude and direction of transmitted loads, ultimately determine the site and extent of bone attachment. 12 The results of this study appear to support that hypothesis. Bone apposition oriented relative to the textured surface at the corners of the dual-geometry component suggests preferential contact and loading in all three of these related designs, regardless of the specific surface modifications. The degree of bone attachment at the periphery of the ItA-coated devices suggests that, as with the femoral components, HA is highly conducive to bone apposition, but may be subject to active biologic and perhaps mechanical compromise. No ad~er'~6 consequences of the resorbed HA are apparent in these specimens. For both the femoraP and acetabular components,
172
The Journal of Arthroplasty Vol. 8 No. 2 April 1993
!:.
....... ~T.: : ~ " ' ( " : - ~ - ) . : .,&-~ _.,~ - - . _~ > ~ . _ ~ . ~ i > ~ j
.
~
-..--
i : ~ i . ~ Z ~ 9
~
J
'
5
-
", -....,,:, -
. .
,
. . . . _ , ._-. :. - ~ , ~ - : . ~ r
:. ;
....... - 7 - - - : ,. ,i..
:. : - ~- ~'." ,'?-,~i.~7,/., ,',',
. "
-'-.:,..,'a~,..:'~,J
,'.Jar
~'-]:.
.
9
'
"
~. ~.~.r
"..~..
.
",7-:,,~
i "1
Plate 1
Plate 2
.
Plate 3A
.
'
,
Plate 3B
P l a t e 1. Low-magnification photomicrograph of section through the fight acetabular cup of case 1. Hydroxyapatite cannot be recognized at this magnification, but there appears to be more trabecular bone near the rim than near the dome of the implant. This screw hole is only partially bridged by bone. P l a t e 2. A different area of the dome from the right acetabular component of case 1. Bone apposition to the metal implant (M) has occurred without a fibrous membrane. P l a t e 3. Photomicrographs of screw wells through the metal backing of case I: (A) right, (B) left. Pale stained fibrous tissue is present in each well, and although there is focal good bone apposition to HA (ttA), holes are only partially sealed with bone. Very rare, small particles of metal and polyethylene (not visible at this magnification) are present in the fibrous tissue.
Uncemented Acetabular Components
9
Bauer et al.
173
I
.
.4
Plate 4
Plate 5
Plate 6
P l a t e 4. Low magnification of the left acetabular component from case 1. Bone apposition is more prominent against the rim than the dome (absent bone from the right side is an artifact of slide preparation). P l a t e 5. Bone apposition is most evident on the "peaks" of the knurled~surface of the rim with relatively less bone in the "valley." P l a t e 6. At a higher magnification, both bone and HA (HA) are present over the "peaks," with a remodeling canal (RC) near the "'valleys." Remodeling appears to be associated with loss of both bone and IIA.
174
The Journal of Arthroplasty Vol. 8 NO. 2 April 1993
!
I
i 9 / f
~
Plate 7B
Plate 7A
Plate 8
P l a t e 7. Low-magnification photomicrographs from the left (25-months) acetabular component of case 2. Bone apposition is most prominent on the "peaks" of the threads, with less bone in the "valleys" of the threads or the dome. P l a t e 8. Higher magnification of the threaded portion of the left acetabular component. Both bone and ItA are present on the "peaks" of the threads with loss of both bone and ttA from the "valleys."
Uncemented Acetabular Components
9
Bauer et al.
175
L Plate 9B
Plate 9A
Plate 10A
i -
k
.
.
.
.
.-,~
..
.
Plate 10B
Plate 10C
P l a t e 9. Low-magnification photomicrographs of the right acetabular component of case 2. Although overall bone apposition is relatively low, remodeling appears non,random, with definite orientation relative to the peaks of the threads. P l a t e 10. Histologic sections and radiograph of a rough-cut block of uncemented revision threaded cup (case 3). Although focal bone apposition is present in the area of the threads, overall bone_apposition is relatively low. A fibrous membrane is present over a portion of the threads (F) and much of the dome. It is unclear whether this reflects motion or is simply a consequence of the previous cemented acetabular component. The apparent difference in metal thickness in the sections reflects increasingly tangenital cuts near the edge of a curved implant rather than true differences in thickness.
176
The Journal of Arthroplasty Vol. 8 No. 2 April 1993
.
-
Plate 11
Plate 11. Histologic section of the rim of the dual-geometry implant of case 4. There is good bone ingrowth through the outer layer of beads, although bone rarely reaches the metal substrate beneath the inner bead layer.
Uncemented Acetabular Components
bone apposition appears to be most prominent in areas of transmitted load, and bone apposition is least prominent where load is not transmitted, or in areas of relatively poor initial fit (eg, revision arthroplasty). Although remodeling of bone according to Wolfe's law was evident in both acetabular and femoral components, the magnitude of b o n e apposition was s o m e w h a t greater in the femoral stems than the corresponding acetabular cups. Because areas of load transmission usually correspond to contact between the implant and endosteal cortex, the differences in magnitude of b o n e apposition between these types of implants might reflect, in part, the greater extent of endosteal cortex available in the f e m u r c o m p a r e d to the primarily trabecular bed of the acetabulum. Finally, as noted by Bobyn and co-workers, 2 radiographic studies are likely to substantially overestim a t e the amoufit of bone apposition achieved by an acetabular c o m p o n e n t . This relates to attempting to use two-dimensional radiographs to detect apposition over a three-dimensional curved surface. Thus, one should anticipate that even in clinically satisfactory cases, a proportion of the bone-acetabular implant interface m a y be fibrous.
Acknowledgments The authors gratefully acknowledge the following surgeons for submitting specimens with radiographs for analysis: Richard Z i m m e r m a n , MD, Portland Orthopedic Clinic, Portland, Oregon, and J o h n Cardea, MD, Medical College of Virginia, Richmond, Virginia.
9
Bauer et al.
177
References 1. Bauer TW, Geesink RGT, Zimmerman R, McMahon JT: Hydroxyapatite coated femoral stems. J Bone Joint Surg 73A:1439, 1991 2. Bobyn JD, Engh CA, Glassman AII: Radiography and histology of a threaded acetabular implant. J Bone Joint Surg 70B:302, 1988 3. Cook SD: Clinical, radiographic and histologic evaluation of retrieved human non-cement porous coated implants. J Long Term Effects Med Implants 1:11, 1991 4. Cook SD, Thomas ttA, Kay JF, Jarcho M: ttydroxyapatite-coated porous titanium for use as an orthopedic biologic attachment system. Clin Orthop 230:303, 1988 5. D'Antonio JA, Capello WN, Jaffe WL: tIydroxyapatire-coated hip implants: multicenter three year clinical and radiographic results. Clin Orthop 285:102, 1992 6. Geesink RGT: Hydroxyapatite-coatcd total hip prostheses: two-year clinical and roentgenographic results of 100 cases. Clin Orthop 261:39, 1990 7. Geesink RGT, De Groot K, Klein CPAT: Chemical implant fixation using hydroxyapatite coatings. Clin Orthop 225:147, 1987 8. Geesink RGT, De Groot K, Klein CPAT: Bonding of bone to hydroxyapatite-coated implants. J Bone Joint Surg 70B:17, 1988 9. Harris WH, White RE, McCarthy JC et ah Bony ingrowth fixation of the acetabular component in canine hip joint arthroplasty. Clin Orlhop 176:7, 1983 10. Schmalzried TP, Kwong LM, Jasty M et ah The mecha9nism of loosening Of cemented acetabular components in total hip arthroplasty. Clin Orthop 274:60, 1992 1 I. Sumner DR, Jacobs J J, Turner TM et al: The amount and distribution of bone ingrowth in tibial components retrieved from human patients. Trans Orthop Res Soc 14:375, 1989 12. Tooke SM, Nugent PJ, Chotivichit Aet ah Comparison of in vivo cementless acetabular fixation. Clin Orthop 235:253, 1988