- Email: [email protected]
Factors affecting femoral bone remodeling after cementless total hip arthroplasty
The Journal of Arthroplasty Vol. 14 No. 5 1999
Factors Affecting Femoral Bone R e m o d e l i n g After Cementless Total Hip Arthroplasty C. Anderson Engh, Jr., MD, Christi Sychterz, MS, and Charles Engh, Sr., MD
Abstract: We performed a postmortem comparison of femurs from two patients who had bilateral cementless total hip arthroplasties with femoral prostheses of different stiffness implanted in their right and left hips. Radiographs of transverse sections of the four femurs demonstrated that all the prostheses were bone ingrown with the most ingrowth occurring distally where the porous coating contacted diaphyseal bone. In both patients, dual-energy x-ray absorptiometry analysis revealed that the femur implanted with the stiffer prosthesis had a 65 % to 79% greater loss of proximal periprosthetic bone than the femur implanted with the more flexible prosthesis. One patient, however, had a dramatically greater total loss of bone from side to side than the other patient. In this patient, we believe that it was host factors more than the differences in stem stiffness that affected the bone-remodeling pattern. Although the two femurs with the stiffer prostheses had the greatest bone loss, the two femurs with the more flexible prostheses demonstrated radiographic signs of cantilever bending of the prosthetic stem and failure of proximal osseointegration. We are not aware of any other bilateral h u m a n postmortem analysis that so clearly illustrates the effect of stem stiffness on bone remodeling. Key words: cementless total hip arthroplasty, bone remodeling, bone-mineral content, femoral stems, femoral stem stiffness.
Femoral bone remodeling after total hip replacem e n t is inevitable. Stem design variables t h o u g h t to affect bone remodeling include the shape, the stiffness, and the a m o u n t of porous surface on the stem [1-4,8,9]. Host factors that influence bone remodeling include the patient's health status and activity level, as well as the physical properties of the f e m u r receiving the stem [10]. Femoral properties affecting remodeling include the geometry, bone density, and cortical thickness. In clinical studies of bone remodeling, it is difficult to differentiate the influence of the prosthesis from the equally important influence of host factors on bone remodeling. If one were to place totally different implants
in nearly identical femurs of the same patient, however, host differences in that patient would be eliminated, and variations in bone remodeling could t h e n be largely attributed to differences b e t w e e n the implants. In clinical practice, we use porous-coated femoral stems that differ in shape, stiffness, and extent of porous coating. We report here on an autopsy study of two clinical cases in which two different porous-coated implants were placed in the right and left femurs of two different patients. This analysis visually demonstrates some of the important differences in the resulting bone remodeling.
From Anderson Orthopaedic Research Institute, Alexandria, Virginia. Submitted March 9, 1998; accepted August 24, 1998. Reprint requests: Christi Sychterz, MS, 2501 Parker's Lane, Suite 200, Alexandria, VA 22306. Copyright © 1999 by Churchill Livingstone® 0883- 5403/99 t 1405-0021 $10.00/0
The identical set of implants used in each patient consisted of a 15-mm diameter anatomic medullary locking (AML) prosthesis implanted in each patient's right hip and a custom-designed prosthesis implanted in each patient's left hip. All stems were
Materials
637
638
The Journal of Arthroplasty Vol. 14 No. 5 August 1999
m a n u f a c t u r e d by DePuy (Warsaw, IN) and w e r e m a d e of a c o b a l t - c h r o m i u m alloy with a m o d u l u s of elasticity of 210 GPa for the AML and 233 GPa for the custom-designed prosthesis [ 11]. All stems w e r e n o n m o d u l a r with 3 2 - m m diameter heads. The differences b e t w e e n the two stem designs related to the extent of porous coating and to the crosssectional geometry. The AML stems had distal cylindrical diameters of 15 m m and were 165 m m in length, with porous coating covering 115 m m of their length. The custom-designed stems w e r e 180 m m in length, with porous coating on the entire stem. They had a tapered trapezoidal g e o m e t r y with a m u c h smaller cross-sectional area at all levels t h a n the AML, the greatest difference b e t w e e n areas being proximally. The stem type chosen for each f e m u r was based strictly on implant availability at the time of surgery; the custom stems were available before the AML stems. The s a m e surgeon i m p l a n t e d all four stems using the same surgical technique.
matoid arthritis. His preoperative height was 5'3"; his preoperative weight was 300 lbs. His preoperative pain and walking scores w e r e 3 a n d 2 on the left and 2 and 3 on the right. He was an indoor ambulator. Evaluation of his preoperative radiographs s h o w e d that he had Dorr b o n e quality type A with a canal-to-isthmus ratio of 4 7 % bilaterally. His right hip r e p l a c e m e n t was p e r f o r m e d 1 year after his left. He also received a right total k n e e r e p l a c e m e n t shortly after his right hip arthroplasty. Four years later, his right k n e e r e p l a c e m e n t failed because of sepsis and was successfully converted to a total k n e e fusion. Except for an a p p r o x i m a t e 6 - m o n t h interval in 1986 w h e n his right k n e e r e p l a c e m e n t was converted to the fusion, he a m b u l a t e d w i t h o u t assistive devices but r e m a i n e d an indoor ambulator. His pain and walking scores w e r e 6 and 4 at his last follow-up (10 and 11 years). He died of a myocardial infarction 13 years after his first hip replacement.
Case 1 The first patient was a 62-year-old male bus operator w h o developed bilateral idiopathic avascular necrosis of the hips. His preoperative height was 5'10", and his preoperative weight was 165 lbs. His preoperative D'Aubigne and Postel pain and walking scores were 3 and 4 on the left and 2 and 4 on the right. On this 6-point scoring system for pain and walking [12], a score of 6 and 6 indicates the absence of both pain and walking i m p a i r m e n t . Preoperative radiographic m e a s u r e m e n t s of both femurs revealed a calcar-to-canal isthmus ratio of 4 4 % with a Dorr b o n e quality type A. The Dorr classification of b o n e quality describes the appearance of the femoral cortex [13]. Type A bone, the best bone, is usually present in y o u n g patients. The f e m u r has a f u n n e l - s h a p e d a p p e a r a n c e of the proximal cortex on both anteroposterior and lateral radiographs. In case 1, the hip replacements w e r e p e r f o r m e d 1 m o n t h apart, the left one first. Three m o n t h s after the second hip replacement, the patient returned to w o r k at his same occupation and continued to w o r k until just a few m o n t h s before his death of a myocardial infarction, 9 years after his arthroplasties. During these years, the patient's hips functioned satisfactorily, and he was able to walk long distances outdoors w i t h o u t using a cane. At his last clinical evaluation, his pain and walking scores w e r e 6 and 6.
Case 2 The second patient was a 48-year-old m a n w h o s e bilateral hip replacements were p e r f o r m e d for rheu-
Methods Both femurs f r o m each patient were retrieved p o s t m o r t e m with the c o m p o n e n t s intact. All soft tissue was r e m o v e d , and the specimens were fixed in formalin. Anteroposterior and lateral specimen radiographs w e r e obtained. The periprosthetic b o n e - m i n e r a l content of each f e m u r was m e a s u r e d using d u a l - e n e r g y x-ray absorptiometry (DEXA) (DPX, Lunar Radiation, Madison, WI). B o n e - m i n e r a l content was m e a s u r e d in three periprosthetic levels that corresponded to the proximal, middle, and distal third of the AML implant (Fig. 1) [9]. A m a s k transfer algorithm included in the DEXA software was used to reproduce these femoral periprosthetic zones on the custom-designed stems because the c u s t o m stem was longer and t h i n n e r t h a n the AML design. With this algorithm, a mask of the AML stem is copied onto the f e m u r with the custom-designed stem. Femoral g e o m e t r y and l a n d m a r k s were used to orient the AML m a s k correctly (Fig. 1 ). Because the AML m a s k is wider t h a n the custom-designed stem, the m a s k excludes b o n e a r o u n d the c u s t o m designed stem that w o u l d not be m e a s u r e d on the f e m u r with the AML stem. Therefore, b o n e - m i n e r a l content f r o m equal areas a r o u n d the two stems was measured. After DEXA analysis was completed, the femurs w e r e e m b e d d e d u n d e r pressure in e t h y l m e t h a c r y late. The e m b e d d e d femoral pairs w e r e aligned according to femoral anatomy, t h e n sectioned transversely at five levels. The cuts were m a d e at the
Femoral Bone Remodeling After CementlessTHA
Fig. I. Bone-mineral content was measured in three periprosthetic levels that corresponded to the proximal, middle, and distal third of the AML implant. The mask or silhouette of the larger AML component was superimposed over the smaller custom-designed component. This positioning ensured that identical areas of bone were measured with dual-energy x-ray absorptiometry.
same femoral locations for the right and left femurs (Figs. 2 and 3). The actual locations of the five levels, however, corresponded to locations along the AML-stem and were made regardless of the seating level of the custom implant. Cross-sectional radiographs, made at each of these five levels, were used to study differences from side to side in the a m o u n t and location of bone ingrowth. Differences in cortical and trabecular bone remodeling also were evaluated. Because the cross-sections were made at the same femoral level, the difference in seating level of the custom implant b e t w e e n case 1 and case 2 did not affect comparisons. Because bone remodeling is, in part, related to stem stiffness, we calculated the difference in bending stiffness of the stem at each cross-sectional level. The stiffness of a femoral stern subjected to bending is a function of two factors: the elastic m o d u l u s of the metal and a geometric factor based on the cross-sectional shape of the object. The two stems used in this study were both made from c h r o m i u m cobalt alloys with similar elastic moduli (233 GPa and 210 GPa) [11]. Because the difference in m o d u lus b e t w e e n the two stems is small, the difference in
•
Engh et al.
639
Fig. 2. Case 1. The two femurs were sectioned transversely at the five levels shown. The AML stem (right) shows spot welds with proximal bone resorption, whereas the custom-designed stem (left) shows proximal radiolucencies.
Fig. 3. Case 2. The two femurs were sectioned transversely at the five levels shown. The AML stem (right) shows spot welds with proximal bone resorption, whereas the custom-designed stem (left) shows proximal radiolucencies.
640
The Journal of Arthroplasty Vol. 14 No. 5 August 1999
stem stiffness is almost entirely determined by the variations in the g e o m e t r y of the two stems. In anteroposterior and mediolateral bending, the geometric factor directly proportional to implant stiffness is the m o m e n t of inertia in these two directions. For the present study, these geometric factors were calculated by digitizing the cross-sectional radiographs of the implants and analyzing compon e n t shape with an AutoCAD p r o g r a m (Autodesk Incorporated, San Rapheal, CA). The differences in stem stiffness from side to side were then represented by the ratio of this geometric factor from one side to the other (Table 1). The differences in implant stiffness b e t w e e n sides are greater for case 1 than for case 2 at any given level, because the custom-designed stem in case 2 was placed more distally in the femur t h a n the AML stem (Fig. 3).
Fig. 4. Case 1. Proximal sections. The femur ~mplanted with the AML stem (right) shows decreased trabecular bone. A radiolucent gap with a pseudocortex nearly surrounds the custom-designed stem on the left.
Results The specimen radiographs illustrated in Figures 2 and 3 reveal qualitatively less periprosthetic bone a r o u n d the AML stems than a r o u n d the customdesigned tapered stems. This difference was evident on the anteroposterior view of both cases. In each case, the difference in remodeling was that the AML stems experienced cancelization of the cortex (proximal stress shielding). This pattern was most dramatic in case 2, in w h i c h there was also a loss of proximal cortex. S p o t w e l d s occurred near the termination of the porous coating of the AML stems. Below these spot welds, the cortical bone density and thickness were similar to those of the femurs containing the more flexible, fully porous-coated implants. The presence of proximal stress shielding and spot welds seen on the right hip of each patient is typical of a stable osseointegrated AML stem. The cross-sections of the four stems at the proximal and midproximal levels are s h o w n in Figures 4 and 5. A radiolucent gap with a pseudocortex nearly surrounds each custom-designed stem at these two levels. Similar gaps and pseudocortex are not present a r o u n d the AML stems (except at the antero-
medial side of case 1). The two hips with the AML stems appeared to have less cortical and cancellous bone in the greater trochanters. Cross-sections of the four stems at the middle and mid-distal levels are s h o w n in Figures 6 and 7. Bone ingrowth is visible on the medial and lateral sides of the rectangular, cross-sectioned, custom implants. These are the sides closest to the endosteal cortex. Cross-sections of the femur containing the larger, circular-shaped, canal-filling AML stems show a more circumferential pattern of bone ingrowth that is greater at the mid-distal stem level than at the
T a b l e 1. B e n d i n g Stiffness R a t i o s (Stiffness of the AML/Stiffness of the Custom-Designed C o m p o n e n t s )
Case 1 Antero-
posterior
Case 2
Mediolateral
Anteroposterior
Mediolateral
Mid-distal Middle Midproximal
2.2
4.6
i. 3
3.3
1.3 2.1
3.3 4.2
0.9 1.4
2.5 2.8
Mean
1.9
4.1
1.2
2.9
Fig. 5. Case 2. Proximal sections. A radiolucent gap with a pseudocortex surrounds the custom-designed stem at these two levels (left). A similar gap and pseudocortex are not present around the AML stem (right). The femur implanted with the AML stem appears to have less cortical and cancellous bone in the greater trochanter.
Femoral Bone Remodeling After CementlessTHA
Fig. 6. Case 1. Middle sections. Bone ingrowth is visible on the medial and lateral sides of the rectangular, crosssectioned custom-designed implant (left). Cross-sections of the femur containing the larger, canal-filling AML stem show a more circumferential pattern of bone ingrowth that is greater at the mid-distal level than at the middle stein level.
mid-stem level. Cortical porosity is greater a r o u n d the AML stems than a r o u n d the custom-designed stems, particularly in case 2. The cross-sections at the distal portion of the four stems are s h o w n in Figures 8 and 9. The rectangular, cross-sectioned, c u s t o m - d e s i g n e d implants both s h o w bone ing r o w t h to the porous surface on the lateral sides. The circular, cross-sectioned AML stems (without porous coating at the distal level) are s u r r o u n d e d by bone that closely approximates their grit-blasted surfaces. At this level, cortical bone density in all four cases is similar. The DEXA results of the four femurs are s h o w n in Figures 10 and 11. A r o u n d the proximal third of the
•
Engh et al.
641
Fig. 8. Case 1. Distal section. The custom-designed implant (left) shows bone ingrowth into the porous surface on the lateral and medial sides. The circular, crosssectioned AML stem (without porous coating at the distal level) is surrounded by bone that closely approximates the grit-blasted surface. At this level, cortical bone density is similar.
implant in case 1, the f e m u r implanted with the AML stem has 65% less bone mineral t h a n the femur implanted with the flexible stem. A r o u n d the proximal third of the implant in case 2, the difference was 78%. A r o u n d the middle third of the implants, the difference in bone mineral b e t w e e n the two sides was 17% in case 1 and 40% in case 2 (the side with the stiffer stem always having less bone). A r o u n d the distal third of the implants, there was little difference in the b o n e - m i n e r a l content a m o n g the four femurs. Case 2 had 1.7 times more side-to-side bone loss than case 1 (17.4% vs 29% overall decrease). This difference occurred even t h o u g h the two patients had stems that differed in stiffness the same a m o u n t and the four femurs had nearly identical preoperative bone quality.
Discussion In the analysis of autopsy-retrieved specimens, DEXA studies are particularly valuable because only t h r o u g h a u t o p s y e v a l u a t i o n of femoral crosssections, such as those s h o w n in these cases, can osseointegration of the implant be d o c u m e n t e d .
Fig. 7. Case 2. Middle sections. Bone ingrowth is visible at both levels with both stems. At both levels, cortical porosity is greater around the AML stem (right) than around the custom-designed stem (left).
Fig. 9. Case 2. Distal section. The custom-designed implant (left) shows bone ingrowth into the porous surface on the lateral side. The AML implant (without porous coating at the distal level) is surrounded by bone that closely approximates the grit-blasted surface. At this level, cortical bone density is similar.
642
The Journal of Arthroplasty Vol. 14 No. 5 August 1999 stiffness, directly affect bone remodeling (i.e., increasing stem size and stiffness increases the a m o u n t of bone loss) [10]. This effect of stem stiffness on bone remodeling has clearly been s h o w n in canine models. Bobyn et al. [1] showed dramatic bone loss differences in the hind legs of dogs implanted on one side with a stiff c o b a l t - c h r o m i u m stern and on the other side with a m o r e flexible, hollow titanium stem. Previous research on h u m a n autopsy cases has shown, however, that the strongest predictor of femoral bone remodeling is not stem stiffness, but rather the physical properties of the femur into which the stem is inserted [3,7,9]. A study of 1 l postmortem-retrieved cases demonstrated a strong linear correlation b e t w e e n bone loss because of bone remodeling and the preexisting bone-mineral content of the femur receiving the prosthesis [9].
Fig. I0. Case 1. Dual-energy x-ray absorptiometry analysis. The numbers indicate the difference in bone-mineral content of the femur implanted with the AML stem compared with the bone-mineral content of the femur implanted with the custom stem. A negative percentage indicates that the femur implanted with the AML stern has less bone mineral than the femur with the custom stein.
Without this histology, DEXA studies of bone remodeling cannot eliminate the possibility that differences in bone remodeling m a y be due to differences in the a m o u n t or location of bone ingrowth. This p o s t m o r t e m analysis is particularly important because it visually demonstrates three points: i) stem stiffness affects bone remodeling; ii) patient factors are also important; and iii) both excessive stem stiffness and excessive stem flexibility are potentially harmful. Although these concepts are not new, we present these two autopsy cases because we are u n a w a r e of any previously published h u m a n p o s t m o r t e m analysis that so visibly d e m o n strates these points. Finite element models of bone remodeling predict that the physical properties of a stem, particularly its
Fig. 11. Case 2. Dual-energy x-ray absorptiometry analysis. The numbers indicate the difference in bone-mineral content of the femur implanted with the AML stem compared with the bone-mineral content of the femur implanted with the custom stem. A negative percentage indicates that the femur implanted with the AML stem has less bone mineral than the femur with the custom stem.
Femoral Bone Remodeling After CementlessTHA
Femurs with relatively low b o n e - m i n e r a l content experienced greater b o n e resorption t h a n femurs with n o r m a l or high preexisting b o n e - m i n e r a l content, despite the size of the stem that had b e e n implanted. In a s u b s e q u e n t study by M a l o n e y et al. [7], this b o n e - m i n e r a l content variable a p p e a r e d to h a v e a greater influence on resorptive b o n e remodeling t h a n did the physical properties of the stem. The difference b e t w e e n these previous a u t o p s y studies and Bobyn's canine study was that the canine m o d e l was able to eliminate the effect of host variables (such as femoral b o n e - m i n e r a l content) on bone remodeling. Stems of different stiffness w e r e placed in the hind limbs of the same dog and, therefore, into bones with nearly identical physical properties. In this way, B o b y n et al. [1] w e r e able to isolate better the direct effect of s t e m stiffness on b o n e remodeling and illustrate greater resorptive bone remodeling in femurs i m p l a n t e d with stiffer implants. Similarly, in the current study, we were able to isolate the b o n e - r e m o d e l i n g response to femoral stems of different stiffness implanted into the hips of the same patient. Each of our two h u m a n autopsy cases d e m o n s t r a t e d greater resorptive b o n e remodeling in the f e m u r i m p l a n t e d with the stiffer implant. Although in b o t h autopsy cases the f e m u r implanted with the stiffer stem lost m o r e b o n e t h a n the f e m u r i m p l a n t e d with the m o r e flexible stem, the difference in b o n e loss f r o m side to side in case 2 was m u c h greater t h a n in case 1 (Figs. 10 and 1 i). This overall difference in the a m o u n t of b o n e loss b e t w e e n our two a u t o p s y cases d e m o n s t r a t e s the influence of host factors on periprosthetic b o n e remodeling. The same stiff and flexible stems w e r e used in b o t h patients, but the less active r h e u m a t o i d patient d e m o n s t r a t e d m u c h greater b o n e loss on the side with the stiffer AML prosthesis t h a n occurred in the m o r e active patient with avascular necrosis. Additionally, the r h e u m a t o i d patient had sepsis and a knee fusion on the same leg with the AML implant. This fusion affected his gait and potentially caused disuse osteoporosis, adding to the total b o n e loss on the side i m p l a n t e d with the stiffer stem. We believe it was primarily the difference in patient factors, not the c o m p o n e n t or b o n e properties, that resulted in an almost two-fold greater femoral b o n e loss from side to side in case 2 c o m p a r e d with case 1 (29% vs 17% overall difference). Surgeons have b e c o m e increasingly a w a r e of the susceptibility of osteoporotic patients to the stress shielding that can be p r o d u c e d by a stiff endoprosthesis. In this study, the reasonably stiff AML stem
•
Engh et al.
643
p r o d u c e d considerable periprosthetic b o n e loss in b o t h patients. As a result, m a n u f a c t u r e r s h a v e a t t e m p t e d to minimize or decrease stress shielding by decreasing implant stiffness. Although decreasing stem stiffness m a y reduce periprosthetic b o n e loss, it can introduce other problems. The two cases in o u r study illustrate some of these problems. The custom-designed stems in these two autopsy cases, although decreasing b o n e loss because of stress shielding, m a y actually be too flexible. In b o t h patients, these distally fixed, flexible, c u s t o m designed stems d e m o n s t r a t e d radiographic signs of cantilever bending. We believe that the resulting m i c r o m o t i o n of the u p p e r part of the stem prev e n t e d proximal b o n e ingrowth. In these two cases, the reactive n e o c o r t e x visible a r o u n d the proximal part of the custom-designed stems appeared to provide e n o u g h support to p r e v e n t stem breakage. This was not always the case with this particular implant design. In our clinical experience using these m o r e flexible stems, failures resulting f r o m p r o x i m a l stem fracture h a v e occurred [14]. Stem fractures and failed proximal b o n e ingrowth h a v e not b e e n problems with the stiffer AML prosthesis, but the dramatic proximal b o n e loss that sometimes occurs is equally undesirable and could potentially lead to complications if stem revisions were necessary. We believe that these two cases clearly illustrate some of the variables associated with bone remodeling after cementless arthroplasty, By eliminating host factors, we were able to e x a m i n e the m o r e subtle factors affecting b o n e remodeling and d e m o n s t r a t e visually the body's remodeling response to b o t h stiff and flexible b o n e - i n g r o w n , femoral implants.
References 1. Bobyn DJ, Mortimer ES, Glassman AH, et al: Producing and avoiding stress shielding: laboratory and clinical observations of noncemented total hip arthroplasty. Clin Orthop 274:80, 1992 2. Engh CA, Bobyn JD: The influence of stem size and extent of porous coating on femoral bone resorption after primary cementless hip arthroplasty. Clin Orthop 231:7, 1988 3. Engh CA, McGovern TF, Bobyn DJ, Harris WH: A quantitative evaluation of periprosthetic bone-remodeling after cementless total hip arthroplasty. J Bone Joint Surg Am 74:1009, 1992 4. Kilgus DJ, Shimaoka EE, Leanne S, et al: Femoral bone remodeling after total hip arthroplasty. Semin Arthroplasty 4:277, 1993 5. Kilgus DJ, Shimaoka EE, Tipton JS, et al: Dual-energy x-ray absorptiometry measurement of bone mineral
644
6.
7.
8.
9.
The Journal of Arthroplasty Vol. 14 No. 5 August 1999 density around porous-coated cementless femoral implants. J Bone Joint Surg Br 75:279, 1993 Kiratli BJ, Checovich MM, McBeath AA, et al: Measurement of bone mineral density by dual-energy absorptiometry in patients with the Wisconsin hip, an uncemented femoral stem. J Arthroplasty 11:184, 1996 Maloney WJ, Sychterz C, Bragdon C, et al: Skeletal response to well fixed femoral components inserted with and without cement. Clin Orthop 333:15, 1996 McCarthy CK, Steinberg GC, Agren M, et al: Quantifying bone loss from the proximal femur after total hip arthroplasty. J Bone Joint Surg Br 73:774, 1991 Sychterz CJ, Engh CA Sr: The influence of clinical
10.
11. 12.
13. 14.
factors on periprosthetic bone remodeling. Clin Orthop 322:285, 1996 Huiskes R, Weinans H, Rietbergen B: The relationship between stress shielding and bone resorption around total hip stems and the effects of flexible materials. Clin Orthop 274:124, 1992 Latrobe Steel Co, Technical Data Manual. Latrobe Steel Co, Latrobe, PA, 1980 D'Aubigne RM, Pastel M: Functional results of hip arthroplasty with acrylic prostheses. J Bone Joint Surg Am 36:45 l, 1954 Dorr LD: Total hip replacement using APR system. Tech Orthop 1:22, 1986 Engh CA, Bobyn JD: Biological fixation in total hip arthroplasty. Slack, Thorofare, NJ, 1988
Factors Affecting Femoral Bone R e m o d e l i n g After Cementless Total Hip Arthroplasty C. Anderson Engh, Jr., MD, Christi Sychterz, MS, and Charles Engh, Sr., MD
Abstract: We performed a postmortem comparison of femurs from two patients who had bilateral cementless total hip arthroplasties with femoral prostheses of different stiffness implanted in their right and left hips. Radiographs of transverse sections of the four femurs demonstrated that all the prostheses were bone ingrown with the most ingrowth occurring distally where the porous coating contacted diaphyseal bone. In both patients, dual-energy x-ray absorptiometry analysis revealed that the femur implanted with the stiffer prosthesis had a 65 % to 79% greater loss of proximal periprosthetic bone than the femur implanted with the more flexible prosthesis. One patient, however, had a dramatically greater total loss of bone from side to side than the other patient. In this patient, we believe that it was host factors more than the differences in stem stiffness that affected the bone-remodeling pattern. Although the two femurs with the stiffer prostheses had the greatest bone loss, the two femurs with the more flexible prostheses demonstrated radiographic signs of cantilever bending of the prosthetic stem and failure of proximal osseointegration. We are not aware of any other bilateral h u m a n postmortem analysis that so clearly illustrates the effect of stem stiffness on bone remodeling. Key words: cementless total hip arthroplasty, bone remodeling, bone-mineral content, femoral stems, femoral stem stiffness.
Femoral bone remodeling after total hip replacem e n t is inevitable. Stem design variables t h o u g h t to affect bone remodeling include the shape, the stiffness, and the a m o u n t of porous surface on the stem [1-4,8,9]. Host factors that influence bone remodeling include the patient's health status and activity level, as well as the physical properties of the f e m u r receiving the stem [10]. Femoral properties affecting remodeling include the geometry, bone density, and cortical thickness. In clinical studies of bone remodeling, it is difficult to differentiate the influence of the prosthesis from the equally important influence of host factors on bone remodeling. If one were to place totally different implants
in nearly identical femurs of the same patient, however, host differences in that patient would be eliminated, and variations in bone remodeling could t h e n be largely attributed to differences b e t w e e n the implants. In clinical practice, we use porous-coated femoral stems that differ in shape, stiffness, and extent of porous coating. We report here on an autopsy study of two clinical cases in which two different porous-coated implants were placed in the right and left femurs of two different patients. This analysis visually demonstrates some of the important differences in the resulting bone remodeling.
From Anderson Orthopaedic Research Institute, Alexandria, Virginia. Submitted March 9, 1998; accepted August 24, 1998. Reprint requests: Christi Sychterz, MS, 2501 Parker's Lane, Suite 200, Alexandria, VA 22306. Copyright © 1999 by Churchill Livingstone® 0883- 5403/99 t 1405-0021 $10.00/0
The identical set of implants used in each patient consisted of a 15-mm diameter anatomic medullary locking (AML) prosthesis implanted in each patient's right hip and a custom-designed prosthesis implanted in each patient's left hip. All stems were
Materials
637
638
The Journal of Arthroplasty Vol. 14 No. 5 August 1999
m a n u f a c t u r e d by DePuy (Warsaw, IN) and w e r e m a d e of a c o b a l t - c h r o m i u m alloy with a m o d u l u s of elasticity of 210 GPa for the AML and 233 GPa for the custom-designed prosthesis [ 11]. All stems w e r e n o n m o d u l a r with 3 2 - m m diameter heads. The differences b e t w e e n the two stem designs related to the extent of porous coating and to the crosssectional geometry. The AML stems had distal cylindrical diameters of 15 m m and were 165 m m in length, with porous coating covering 115 m m of their length. The custom-designed stems w e r e 180 m m in length, with porous coating on the entire stem. They had a tapered trapezoidal g e o m e t r y with a m u c h smaller cross-sectional area at all levels t h a n the AML, the greatest difference b e t w e e n areas being proximally. The stem type chosen for each f e m u r was based strictly on implant availability at the time of surgery; the custom stems were available before the AML stems. The s a m e surgeon i m p l a n t e d all four stems using the same surgical technique.
matoid arthritis. His preoperative height was 5'3"; his preoperative weight was 300 lbs. His preoperative pain and walking scores w e r e 3 a n d 2 on the left and 2 and 3 on the right. He was an indoor ambulator. Evaluation of his preoperative radiographs s h o w e d that he had Dorr b o n e quality type A with a canal-to-isthmus ratio of 4 7 % bilaterally. His right hip r e p l a c e m e n t was p e r f o r m e d 1 year after his left. He also received a right total k n e e r e p l a c e m e n t shortly after his right hip arthroplasty. Four years later, his right k n e e r e p l a c e m e n t failed because of sepsis and was successfully converted to a total k n e e fusion. Except for an a p p r o x i m a t e 6 - m o n t h interval in 1986 w h e n his right k n e e r e p l a c e m e n t was converted to the fusion, he a m b u l a t e d w i t h o u t assistive devices but r e m a i n e d an indoor ambulator. His pain and walking scores w e r e 6 and 4 at his last follow-up (10 and 11 years). He died of a myocardial infarction 13 years after his first hip replacement.
Case 1 The first patient was a 62-year-old male bus operator w h o developed bilateral idiopathic avascular necrosis of the hips. His preoperative height was 5'10", and his preoperative weight was 165 lbs. His preoperative D'Aubigne and Postel pain and walking scores were 3 and 4 on the left and 2 and 4 on the right. On this 6-point scoring system for pain and walking [12], a score of 6 and 6 indicates the absence of both pain and walking i m p a i r m e n t . Preoperative radiographic m e a s u r e m e n t s of both femurs revealed a calcar-to-canal isthmus ratio of 4 4 % with a Dorr b o n e quality type A. The Dorr classification of b o n e quality describes the appearance of the femoral cortex [13]. Type A bone, the best bone, is usually present in y o u n g patients. The f e m u r has a f u n n e l - s h a p e d a p p e a r a n c e of the proximal cortex on both anteroposterior and lateral radiographs. In case 1, the hip replacements w e r e p e r f o r m e d 1 m o n t h apart, the left one first. Three m o n t h s after the second hip replacement, the patient returned to w o r k at his same occupation and continued to w o r k until just a few m o n t h s before his death of a myocardial infarction, 9 years after his arthroplasties. During these years, the patient's hips functioned satisfactorily, and he was able to walk long distances outdoors w i t h o u t using a cane. At his last clinical evaluation, his pain and walking scores w e r e 6 and 6.
Case 2 The second patient was a 48-year-old m a n w h o s e bilateral hip replacements were p e r f o r m e d for rheu-
Methods Both femurs f r o m each patient were retrieved p o s t m o r t e m with the c o m p o n e n t s intact. All soft tissue was r e m o v e d , and the specimens were fixed in formalin. Anteroposterior and lateral specimen radiographs w e r e obtained. The periprosthetic b o n e - m i n e r a l content of each f e m u r was m e a s u r e d using d u a l - e n e r g y x-ray absorptiometry (DEXA) (DPX, Lunar Radiation, Madison, WI). B o n e - m i n e r a l content was m e a s u r e d in three periprosthetic levels that corresponded to the proximal, middle, and distal third of the AML implant (Fig. 1) [9]. A m a s k transfer algorithm included in the DEXA software was used to reproduce these femoral periprosthetic zones on the custom-designed stems because the c u s t o m stem was longer and t h i n n e r t h a n the AML design. With this algorithm, a mask of the AML stem is copied onto the f e m u r with the custom-designed stem. Femoral g e o m e t r y and l a n d m a r k s were used to orient the AML m a s k correctly (Fig. 1 ). Because the AML m a s k is wider t h a n the custom-designed stem, the m a s k excludes b o n e a r o u n d the c u s t o m designed stem that w o u l d not be m e a s u r e d on the f e m u r with the AML stem. Therefore, b o n e - m i n e r a l content f r o m equal areas a r o u n d the two stems was measured. After DEXA analysis was completed, the femurs w e r e e m b e d d e d u n d e r pressure in e t h y l m e t h a c r y late. The e m b e d d e d femoral pairs w e r e aligned according to femoral anatomy, t h e n sectioned transversely at five levels. The cuts were m a d e at the
Femoral Bone Remodeling After CementlessTHA
Fig. I. Bone-mineral content was measured in three periprosthetic levels that corresponded to the proximal, middle, and distal third of the AML implant. The mask or silhouette of the larger AML component was superimposed over the smaller custom-designed component. This positioning ensured that identical areas of bone were measured with dual-energy x-ray absorptiometry.
same femoral locations for the right and left femurs (Figs. 2 and 3). The actual locations of the five levels, however, corresponded to locations along the AML-stem and were made regardless of the seating level of the custom implant. Cross-sectional radiographs, made at each of these five levels, were used to study differences from side to side in the a m o u n t and location of bone ingrowth. Differences in cortical and trabecular bone remodeling also were evaluated. Because the cross-sections were made at the same femoral level, the difference in seating level of the custom implant b e t w e e n case 1 and case 2 did not affect comparisons. Because bone remodeling is, in part, related to stem stiffness, we calculated the difference in bending stiffness of the stem at each cross-sectional level. The stiffness of a femoral stern subjected to bending is a function of two factors: the elastic m o d u l u s of the metal and a geometric factor based on the cross-sectional shape of the object. The two stems used in this study were both made from c h r o m i u m cobalt alloys with similar elastic moduli (233 GPa and 210 GPa) [11]. Because the difference in m o d u lus b e t w e e n the two stems is small, the difference in
•
Engh et al.
639
Fig. 2. Case 1. The two femurs were sectioned transversely at the five levels shown. The AML stem (right) shows spot welds with proximal bone resorption, whereas the custom-designed stem (left) shows proximal radiolucencies.
Fig. 3. Case 2. The two femurs were sectioned transversely at the five levels shown. The AML stem (right) shows spot welds with proximal bone resorption, whereas the custom-designed stem (left) shows proximal radiolucencies.
640
The Journal of Arthroplasty Vol. 14 No. 5 August 1999
stem stiffness is almost entirely determined by the variations in the g e o m e t r y of the two stems. In anteroposterior and mediolateral bending, the geometric factor directly proportional to implant stiffness is the m o m e n t of inertia in these two directions. For the present study, these geometric factors were calculated by digitizing the cross-sectional radiographs of the implants and analyzing compon e n t shape with an AutoCAD p r o g r a m (Autodesk Incorporated, San Rapheal, CA). The differences in stem stiffness from side to side were then represented by the ratio of this geometric factor from one side to the other (Table 1). The differences in implant stiffness b e t w e e n sides are greater for case 1 than for case 2 at any given level, because the custom-designed stem in case 2 was placed more distally in the femur t h a n the AML stem (Fig. 3).
Fig. 4. Case 1. Proximal sections. The femur ~mplanted with the AML stem (right) shows decreased trabecular bone. A radiolucent gap with a pseudocortex nearly surrounds the custom-designed stem on the left.
Results The specimen radiographs illustrated in Figures 2 and 3 reveal qualitatively less periprosthetic bone a r o u n d the AML stems than a r o u n d the customdesigned tapered stems. This difference was evident on the anteroposterior view of both cases. In each case, the difference in remodeling was that the AML stems experienced cancelization of the cortex (proximal stress shielding). This pattern was most dramatic in case 2, in w h i c h there was also a loss of proximal cortex. S p o t w e l d s occurred near the termination of the porous coating of the AML stems. Below these spot welds, the cortical bone density and thickness were similar to those of the femurs containing the more flexible, fully porous-coated implants. The presence of proximal stress shielding and spot welds seen on the right hip of each patient is typical of a stable osseointegrated AML stem. The cross-sections of the four stems at the proximal and midproximal levels are s h o w n in Figures 4 and 5. A radiolucent gap with a pseudocortex nearly surrounds each custom-designed stem at these two levels. Similar gaps and pseudocortex are not present a r o u n d the AML stems (except at the antero-
medial side of case 1). The two hips with the AML stems appeared to have less cortical and cancellous bone in the greater trochanters. Cross-sections of the four stems at the middle and mid-distal levels are s h o w n in Figures 6 and 7. Bone ingrowth is visible on the medial and lateral sides of the rectangular, cross-sectioned, custom implants. These are the sides closest to the endosteal cortex. Cross-sections of the femur containing the larger, circular-shaped, canal-filling AML stems show a more circumferential pattern of bone ingrowth that is greater at the mid-distal stem level than at the
T a b l e 1. B e n d i n g Stiffness R a t i o s (Stiffness of the AML/Stiffness of the Custom-Designed C o m p o n e n t s )
Case 1 Antero-
posterior
Case 2
Mediolateral
Anteroposterior
Mediolateral
Mid-distal Middle Midproximal
2.2
4.6
i. 3
3.3
1.3 2.1
3.3 4.2
0.9 1.4
2.5 2.8
Mean
1.9
4.1
1.2
2.9
Fig. 5. Case 2. Proximal sections. A radiolucent gap with a pseudocortex surrounds the custom-designed stem at these two levels (left). A similar gap and pseudocortex are not present around the AML stem (right). The femur implanted with the AML stem appears to have less cortical and cancellous bone in the greater trochanter.
Femoral Bone Remodeling After CementlessTHA
Fig. 6. Case 1. Middle sections. Bone ingrowth is visible on the medial and lateral sides of the rectangular, crosssectioned custom-designed implant (left). Cross-sections of the femur containing the larger, canal-filling AML stem show a more circumferential pattern of bone ingrowth that is greater at the mid-distal level than at the middle stein level.
mid-stem level. Cortical porosity is greater a r o u n d the AML stems than a r o u n d the custom-designed stems, particularly in case 2. The cross-sections at the distal portion of the four stems are s h o w n in Figures 8 and 9. The rectangular, cross-sectioned, c u s t o m - d e s i g n e d implants both s h o w bone ing r o w t h to the porous surface on the lateral sides. The circular, cross-sectioned AML stems (without porous coating at the distal level) are s u r r o u n d e d by bone that closely approximates their grit-blasted surfaces. At this level, cortical bone density in all four cases is similar. The DEXA results of the four femurs are s h o w n in Figures 10 and 11. A r o u n d the proximal third of the
•
Engh et al.
641
Fig. 8. Case 1. Distal section. The custom-designed implant (left) shows bone ingrowth into the porous surface on the lateral and medial sides. The circular, crosssectioned AML stem (without porous coating at the distal level) is surrounded by bone that closely approximates the grit-blasted surface. At this level, cortical bone density is similar.
implant in case 1, the f e m u r implanted with the AML stem has 65% less bone mineral t h a n the femur implanted with the flexible stem. A r o u n d the proximal third of the implant in case 2, the difference was 78%. A r o u n d the middle third of the implants, the difference in bone mineral b e t w e e n the two sides was 17% in case 1 and 40% in case 2 (the side with the stiffer stem always having less bone). A r o u n d the distal third of the implants, there was little difference in the b o n e - m i n e r a l content a m o n g the four femurs. Case 2 had 1.7 times more side-to-side bone loss than case 1 (17.4% vs 29% overall decrease). This difference occurred even t h o u g h the two patients had stems that differed in stiffness the same a m o u n t and the four femurs had nearly identical preoperative bone quality.
Discussion In the analysis of autopsy-retrieved specimens, DEXA studies are particularly valuable because only t h r o u g h a u t o p s y e v a l u a t i o n of femoral crosssections, such as those s h o w n in these cases, can osseointegration of the implant be d o c u m e n t e d .
Fig. 7. Case 2. Middle sections. Bone ingrowth is visible at both levels with both stems. At both levels, cortical porosity is greater around the AML stem (right) than around the custom-designed stem (left).
Fig. 9. Case 2. Distal section. The custom-designed implant (left) shows bone ingrowth into the porous surface on the lateral side. The AML implant (without porous coating at the distal level) is surrounded by bone that closely approximates the grit-blasted surface. At this level, cortical bone density is similar.
642
The Journal of Arthroplasty Vol. 14 No. 5 August 1999 stiffness, directly affect bone remodeling (i.e., increasing stem size and stiffness increases the a m o u n t of bone loss) [10]. This effect of stem stiffness on bone remodeling has clearly been s h o w n in canine models. Bobyn et al. [1] showed dramatic bone loss differences in the hind legs of dogs implanted on one side with a stiff c o b a l t - c h r o m i u m stern and on the other side with a m o r e flexible, hollow titanium stem. Previous research on h u m a n autopsy cases has shown, however, that the strongest predictor of femoral bone remodeling is not stem stiffness, but rather the physical properties of the femur into which the stem is inserted [3,7,9]. A study of 1 l postmortem-retrieved cases demonstrated a strong linear correlation b e t w e e n bone loss because of bone remodeling and the preexisting bone-mineral content of the femur receiving the prosthesis [9].
Fig. I0. Case 1. Dual-energy x-ray absorptiometry analysis. The numbers indicate the difference in bone-mineral content of the femur implanted with the AML stem compared with the bone-mineral content of the femur implanted with the custom stem. A negative percentage indicates that the femur implanted with the AML stern has less bone mineral than the femur with the custom stein.
Without this histology, DEXA studies of bone remodeling cannot eliminate the possibility that differences in bone remodeling m a y be due to differences in the a m o u n t or location of bone ingrowth. This p o s t m o r t e m analysis is particularly important because it visually demonstrates three points: i) stem stiffness affects bone remodeling; ii) patient factors are also important; and iii) both excessive stem stiffness and excessive stem flexibility are potentially harmful. Although these concepts are not new, we present these two autopsy cases because we are u n a w a r e of any previously published h u m a n p o s t m o r t e m analysis that so visibly d e m o n strates these points. Finite element models of bone remodeling predict that the physical properties of a stem, particularly its
Fig. 11. Case 2. Dual-energy x-ray absorptiometry analysis. The numbers indicate the difference in bone-mineral content of the femur implanted with the AML stem compared with the bone-mineral content of the femur implanted with the custom stem. A negative percentage indicates that the femur implanted with the AML stem has less bone mineral than the femur with the custom stem.
Femoral Bone Remodeling After CementlessTHA
Femurs with relatively low b o n e - m i n e r a l content experienced greater b o n e resorption t h a n femurs with n o r m a l or high preexisting b o n e - m i n e r a l content, despite the size of the stem that had b e e n implanted. In a s u b s e q u e n t study by M a l o n e y et al. [7], this b o n e - m i n e r a l content variable a p p e a r e d to h a v e a greater influence on resorptive b o n e remodeling t h a n did the physical properties of the stem. The difference b e t w e e n these previous a u t o p s y studies and Bobyn's canine study was that the canine m o d e l was able to eliminate the effect of host variables (such as femoral b o n e - m i n e r a l content) on bone remodeling. Stems of different stiffness w e r e placed in the hind limbs of the same dog and, therefore, into bones with nearly identical physical properties. In this way, B o b y n et al. [1] w e r e able to isolate better the direct effect of s t e m stiffness on b o n e remodeling and illustrate greater resorptive bone remodeling in femurs i m p l a n t e d with stiffer implants. Similarly, in the current study, we were able to isolate the b o n e - r e m o d e l i n g response to femoral stems of different stiffness implanted into the hips of the same patient. Each of our two h u m a n autopsy cases d e m o n s t r a t e d greater resorptive b o n e remodeling in the f e m u r i m p l a n t e d with the stiffer implant. Although in b o t h autopsy cases the f e m u r implanted with the stiffer stem lost m o r e b o n e t h a n the f e m u r i m p l a n t e d with the m o r e flexible stem, the difference in b o n e loss f r o m side to side in case 2 was m u c h greater t h a n in case 1 (Figs. 10 and 1 i). This overall difference in the a m o u n t of b o n e loss b e t w e e n our two a u t o p s y cases d e m o n s t r a t e s the influence of host factors on periprosthetic b o n e remodeling. The same stiff and flexible stems w e r e used in b o t h patients, but the less active r h e u m a t o i d patient d e m o n s t r a t e d m u c h greater b o n e loss on the side with the stiffer AML prosthesis t h a n occurred in the m o r e active patient with avascular necrosis. Additionally, the r h e u m a t o i d patient had sepsis and a knee fusion on the same leg with the AML implant. This fusion affected his gait and potentially caused disuse osteoporosis, adding to the total b o n e loss on the side i m p l a n t e d with the stiffer stem. We believe it was primarily the difference in patient factors, not the c o m p o n e n t or b o n e properties, that resulted in an almost two-fold greater femoral b o n e loss from side to side in case 2 c o m p a r e d with case 1 (29% vs 17% overall difference). Surgeons have b e c o m e increasingly a w a r e of the susceptibility of osteoporotic patients to the stress shielding that can be p r o d u c e d by a stiff endoprosthesis. In this study, the reasonably stiff AML stem
•
Engh et al.
643
p r o d u c e d considerable periprosthetic b o n e loss in b o t h patients. As a result, m a n u f a c t u r e r s h a v e a t t e m p t e d to minimize or decrease stress shielding by decreasing implant stiffness. Although decreasing stem stiffness m a y reduce periprosthetic b o n e loss, it can introduce other problems. The two cases in o u r study illustrate some of these problems. The custom-designed stems in these two autopsy cases, although decreasing b o n e loss because of stress shielding, m a y actually be too flexible. In b o t h patients, these distally fixed, flexible, c u s t o m designed stems d e m o n s t r a t e d radiographic signs of cantilever bending. We believe that the resulting m i c r o m o t i o n of the u p p e r part of the stem prev e n t e d proximal b o n e ingrowth. In these two cases, the reactive n e o c o r t e x visible a r o u n d the proximal part of the custom-designed stems appeared to provide e n o u g h support to p r e v e n t stem breakage. This was not always the case with this particular implant design. In our clinical experience using these m o r e flexible stems, failures resulting f r o m p r o x i m a l stem fracture h a v e occurred [14]. Stem fractures and failed proximal b o n e ingrowth h a v e not b e e n problems with the stiffer AML prosthesis, but the dramatic proximal b o n e loss that sometimes occurs is equally undesirable and could potentially lead to complications if stem revisions were necessary. We believe that these two cases clearly illustrate some of the variables associated with bone remodeling after cementless arthroplasty, By eliminating host factors, we were able to e x a m i n e the m o r e subtle factors affecting b o n e remodeling and d e m o n s t r a t e visually the body's remodeling response to b o t h stiff and flexible b o n e - i n g r o w n , femoral implants.
References 1. Bobyn DJ, Mortimer ES, Glassman AH, et al: Producing and avoiding stress shielding: laboratory and clinical observations of noncemented total hip arthroplasty. Clin Orthop 274:80, 1992 2. Engh CA, Bobyn JD: The influence of stem size and extent of porous coating on femoral bone resorption after primary cementless hip arthroplasty. Clin Orthop 231:7, 1988 3. Engh CA, McGovern TF, Bobyn DJ, Harris WH: A quantitative evaluation of periprosthetic bone-remodeling after cementless total hip arthroplasty. J Bone Joint Surg Am 74:1009, 1992 4. Kilgus DJ, Shimaoka EE, Leanne S, et al: Femoral bone remodeling after total hip arthroplasty. Semin Arthroplasty 4:277, 1993 5. Kilgus DJ, Shimaoka EE, Tipton JS, et al: Dual-energy x-ray absorptiometry measurement of bone mineral
644
6.
7.
8.
9.
The Journal of Arthroplasty Vol. 14 No. 5 August 1999 density around porous-coated cementless femoral implants. J Bone Joint Surg Br 75:279, 1993 Kiratli BJ, Checovich MM, McBeath AA, et al: Measurement of bone mineral density by dual-energy absorptiometry in patients with the Wisconsin hip, an uncemented femoral stem. J Arthroplasty 11:184, 1996 Maloney WJ, Sychterz C, Bragdon C, et al: Skeletal response to well fixed femoral components inserted with and without cement. Clin Orthop 333:15, 1996 McCarthy CK, Steinberg GC, Agren M, et al: Quantifying bone loss from the proximal femur after total hip arthroplasty. J Bone Joint Surg Br 73:774, 1991 Sychterz CJ, Engh CA Sr: The influence of clinical
10.
11. 12.
13. 14.
factors on periprosthetic bone remodeling. Clin Orthop 322:285, 1996 Huiskes R, Weinans H, Rietbergen B: The relationship between stress shielding and bone resorption around total hip stems and the effects of flexible materials. Clin Orthop 274:124, 1992 Latrobe Steel Co, Technical Data Manual. Latrobe Steel Co, Latrobe, PA, 1980 D'Aubigne RM, Pastel M: Functional results of hip arthroplasty with acrylic prostheses. J Bone Joint Surg Am 36:45 l, 1954 Dorr LD: Total hip replacement using APR system. Tech Orthop 1:22, 1986 Engh CA, Bobyn JD: Biological fixation in total hip arthroplasty. Slack, Thorofare, NJ, 1988