Total Hip A r t h r o p l a s t y A Comparison Between Cemented and Press-fit Noncemented Fixation
Anders Wykman, MD, PhD* Elisabeth Olsson RPT, PhD* Gabriella Axdorph, MD,-t- and Ian Goldie, MD, PhD*
Abstract: This study compares cemented (Chamley) with noncemented (ttonnart PateI-Garches) total hip arthroplasty. 150 patients with osteoarthritis, rheumatoid arthritis, and miscellaneous conditions were randomized into either group with 75 in each. Patients were evaluated over 5 postoperative years with examinations at 6 months, at 1 year, and then annually. Each examination involved clinical assessment, objective gait analysis, and radiographic examination. The Chamley group improved faster than the ttP-Garches group during the first 2 years. At the last evaluation the results were excellent or good in 79% in the Chamley group and 70% in the tlP-Garches group. In tile ttP-Garches group midthigh pain occurred in 64%. Five patients with Chamley prostheses (6.7%) and 14 patients with ttP-Garches prostheses {18.7% had revision surgery to correct mechanical loosening. Standard radiography revealed radiolucent zones in both groups, but this could not be related to the clinical outcome despite radiolucency being present in all noncemented replacements on the femoral side. Objective gait analysis confirmed inferior clinical results at 6 months in the ttP-Garches group. There was no significant difference between the groups at the most recent evaluation. Our findings are not consistent with earlier optimistic expectations on press-fit noncemented total hip arthroplasties. Key words: Noncemented procedures, hip arthroplasty, gait analysis, Charnley hip arthroplasty, ttP-Garches arthroplasty, radiography.
The first implants used in replacement surgery of the hip joint were inserted i~ direct contact"with bone. The success of early hip joint arthroplasties was c o m p r o m i s e d by the materials available at the time,
as well as the surgical techniques used, and the n o n availability of antibiotics. A b r e a k t h r o u g h was m a d e by. Sir J o h n Charnley, w h o believed that better results could be obtained using p o l y m e t h y l m e t h a c r y late, bone cement for stable fixation of the implant (7). By the early 1960s, total hip joint arthroplasty, including both femoral and acetabular c o m p o n e n t s , had become the standard procedure. Thirty years' experience has borne out Charnley's theory of a stable fixation of the implant in total hip arthroplasty. However, problems arise in the cases
Front the Departments of*Orthopaedics, and "~DiagnosticRadiology, Karolhuka Hospital Stockhohn, Sweden.
This investigation was supported by grants from the Karolinska Instltute's Research Funds, Loo and Hans Osterman Foundation, Ulla and Gustafaf Uggla Foundation, and the Swedish Association against Rheumatism. Reprint requests: Anders Wykman, MD, PhD, Department of Orthopaedics, Karolinska Hospital, S-10401 Stockholm, Sweden.
19
20
The Journal of Arthroplasty Vol. 6 No. 1 March 1991
which fail, mainly because of aseptic loosening. It is well k n o w n that revisions yield poorer results than primary total hip arthroplasties (33). In view of this, attempts have b e e n m a d e to reintroduce the old concept of total hip arthroplasty without the use of bone cement. N e w prostheses have been designed to adapt directly to bone. This study aims to c o m p a r e a standard cemented total hip arthroplasty with a n o n c e m e n t e d procedure in the following way: 1. To study prospectively 150 patients r a n d o m ized into two groups, one with a cemented total hip arthroplasty and one noncemented. 2. To base this comparison on clinical assessment, radiographic examination, and an objective gait analysis.
Patients and Methods Patients From the waiting list for total hip arthroplasty, 150 patients were consecutively r a n d o m i z e d into two groups, one to receive cemented and one noncemerited procedures. The indications for surgery included disabling pain and functional limitation of at least one hip. The investigation was approved by the ethics committee of the Karolinska Hospital, and all patients gave their informed consent. The operations were performed between September 1982 and December 1984. Postoperative examinations were conducted at 6 months, at 1 year, and then annually ( T a b l e 1).
Table 2. Age and Weight of Patients
Age, mean (range) years Weight, mean (range) kg
Charnley
HP-Garches
67.4 (48-82) 73.2 (45-110)
64.8 (29-82) 72.3 (49-98)
Table 3. Diagnoses
Charnley
HP-Garches
Total
Osteoarthritis Rheumatoid arthritis Miscellaneous
58 7 10
57 8 l0
l 15 15 20
Total
75
75
! 50
T h e H P - G a r c h e s G r o u p . There were 75 patienis, of w h o m 50 were female and 25 male, age 29 to 82 years, m e a n age 64.8 years, (Table 2), m e d i a n 65 years. In 60 patients only one hip was affected and there was no other condition that could interfere with walking. The remaining 15 patients suffered from bilateral hip disease, had some systemic disorder that interfered with locomotion, or both. Diagnoses are s h o w n in Table 3.
Implants Tile Charnley prosthesis consists of a stainless femoral stem with a 2 2 . 2 - m m spherical head and an ultrahigh molecular weight polyethylene cup. The components are anchored into b o n e with p o l y m e thylmethacrylate bone cement. The HP-Garches (Honnart Patel Garches) prosthesis consists of a conical femoral stem with a 3 2 - m m
T h e C h a r n l e y G r o u p . There were 75 patients, of w h o m 43 were female and 32 male, age 48 to 82 years, m e a n age 67.4 years (Table 2), median 68 years. In 60 patients only o n e hip was affected and there was no other condition that interfered with walking. The remaining 15 patients suffered from bilateral hip disease, had some systemic disorder that interfered with locomotion, or both. Diagnoses are s h o w n in Table 3. Table 1. Number of Patients Operated
Year
Chamley
HP-Garches
Total
1982 1983 1984
17 41 17
16 37 22
33 78 39
Total
75
75
150
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
j
Fig. 1. (Left) HP-Garches prosthesis. (Right) Chamley prosthesis.
Total Hip Arthroplasty
spherical head made of forged cobalt-chrome alloy and a threaded acetabular ring with an inner ultrahigh molecular weight polyethylene liner. The stem is fitted with a collar and has a rough surface with a pyramidal pattern (Fig. 1). All components are available in different sizes.
Surgical Procedures In the Chandey group, a lateral approach with trochanteric osteotomy was used and the procedure presented by Charnley in 1979 (10) was followed. Acrylic cement (polymethylmethacrylate) with barium sulphate (CMW) was used. No antibiotics were included in the cement. All operations were performed under epidural anesthesia. The operations lasted on average 123 minutes (70-180 min). Calculated total operative and postoperative blood loss was 1,875 nfl on average. In the HP-Garchesgroup, a posterolateral approach without trochanteric osteotomy was used. The acetabulum and femoral shaft were prepared with specially designed reamers. The acetabuliam was not tapped, and the threaded acetabular component was considered to be fully anchored in bone when the surgeon could no longer achieve any rotation on the T-handle. The operations lasted on average 104 minutes (60-180 rain) (p-value for comparison with the Charnley group p < 0.001). Calculated total operative and postoperative blood loss was 1,606 ml on average. The two surgical approaches used in the present study have previously been compared (30, 36, 41). No clinical difference was found in immediate postoperative course or later, and both approaches give satisfactory results (29). The operations in this study were performed by only experienced surgeons.
9 Wykman et al.
21
Table 4. Gradation of Harris Hip Score Excellent Good Fair Poor
90-100 80-89 70-79 - 30), and 10 patients had a pain score of 30
Assessment
Table 5. Results: Harris Hip Score (median)* H a r r i s H i p S c o r e ( T a b l e 5). Preoperatively, the Charnley group patients achieved a total score of 9 62 points, median 37.3 points. The average pain score (maximum 44 points; no pain) measured 13.5 points. There were no linear relationships between
Chamley Preoperative 6 months Most recent evaluation
37.3 89.4 95.3
* Reoperated patients included.
HP-Garches Sign. Diff. 38.1 74.3 88.7
n.s. p < 0.00l n.s.
Total Hip Arthroplasty Table 6. Number of Patients at Most Recent Evaluation
Pre--op
Chamley
tlP-Garches
Total
3 years 4 years 5 years
6 20 42
4 27 39
10 47 81
Total
68
70
138
or less. One patient was reoperated at 2 m o n t h s with a Girdlestone hip because of infection. At the most recent evaluation 68 patients w e r e examined 3 to 5 years after surgery (Table 6). Of the missing 6 patients, 1 patient refused to attend, I patient had been reoperated with a Girdlestone hip because of infection, and 4 patients had died. 54 patients (79%) obtained a good or excellent result according to the Harris score (Fig. 3). Median total score was 95.3, m e a n value 88.65 patients (96%) reported slight pain or no pain at all. Of the HP-Garches group, the patients had a total preoperative score of 1 9 - 6 4 points, .median 38.1 points. The average pain score measured 14.7 points. There were no linear relationships between the patients' ages, their diagnoses, and the results of the preoperative assessment. At 6 m o n t h s 74 patients (1 patient died on the first postoperative day) had a median total score of 74.3 points. Thirty-one patients (42%) obtained a good or excellent result, ie, their Harris score exceeded 80 points. Forty-five of the patients (61%) had a pain score indicating slight pain or no pain at all (>30). The remaining 29 patients had a pain score of 30 or less. At the most recent evaluation 70 patients were examined 3 to 5 years after surgery (Table 6). One patient had m o v e d a w a y and 3 patients had died. Forty-eight patients (70%) obtained a good or excellent result according to the Harris score (Fig. 3). Median total score was 88.7, m e a n value 85.60 of the patients (86%) reported slight pain or no pain at all, 10 patients had a pain score of 30 or less. Preoperatively there were no significant differences b e t w e e n tile C h a m l e y and the HP-Garches groups. At 6 m o n t h s the patients of the Charnley
* Wykman et al. 6 months
23
Follow-up
Charnley
HP-Garches
,I
I 90"100
~
70-79
8O-89 B B < 70 Fig. 3. Result: ttarris Hip Score. group had a higher total score and pain score t h a n the patients of the HP-Garches group (P < 0.001). No significant differences b e t w e e n the groups w e r e found at the most recent evaluation. There were no linear relationships b e t w e e n the assessments preoperatively, at 6 months, or thereafter, and the patients' age, weight, or diagnosis. R e v i s i o n s . In all there were 22 revisions. In the Charnley group 2 patients had their prostheses rem o v e d because of deep infection; 6 patients underwent revision with exchange of either one or b o t h components. In 5 patients revision was due to implant loosening and in i to malpositioning of the prosthesis with recurrent dislocation. In the HPGarches group 14 patients had either a n e w noncemented femoral c o m p o n e n t , a cemented acetabular component, a cemented femoral c o m p o n e n t or b o t h c o m p o n e n t s changed for cemented alternatives owing to loosening (Table 7).
Other Factors Influencing Outcome At the most recent evaluation, 32 of the 68 patients in the Charnley group and 44 of the 70 patients in the tiP-Garches group suffered from c o n c o m i t a n t disorders. Bilateral hip disease was present in 15 pa-
Table 7. Revisions Chamley'
HP-Garches
Total
Extraction New noncemented stem New cemented stem New cemented cup New cemented stem and cup
2 3 2 1
I 8 1 4
2 l ll 3 5
Total
8
14
22
24
The Journal of Arthroplasty Vol. 6 No. 1 March 1991
dents in the Charnley group and 19 in the HPGarches group. Systemic factors of different kinds were also responsible for failure to achieve normal locomotion. In the Chamley group all patients with only unilateral hip disease (n = 36) had a good or excellent result. The results for these patients were significantly better than the results for the Charnley group as a whole (P < 0.05). The corresponding frequency for the patients in the HP-Garches group was 24 out of 26 (92%). The result for these patients did not differ significantly from the result for the whole HP-Garches group.
Midthigh Pain A number of patients with the noncemented HPGarches prosthesis experienced pain on weightbearing in the midthigh region of the operated leg. After six postoperative months midthigh pain occurred in 46 of 72 patients (64%). In 28 of these, pain spontaneously disappeared during the following 3 years, in 18 cases during the first year. However, in 18 patients, pain not only persisted but increased, necessitating a reoperation in 13. A further 5 appear to be candidates for reoperation. Of the 13 patients u n d e r g o i n g revision, 10 have become completely painfree whereas the remaining 3 still suffer tolerable pain (Fig. 4).
6 months
46/72
reop
Patients' Opinion At 6 months 93% of the patients in the Charnley group were satisfied with the operation; 7% were dissatisfied. The corresponding figures for the ltPGarches group were 70% and 30% respectively. The differences are significant (P < 0.05). At the most recent evaluation, 95% in the Charnley group and 90% of the patients in the HP Garches were satisfied.
Radiologic Evaluation The radiographic survey at the most recent evaluation comprised 49 complete Charnley series and 53 complete HP-Garches series. The remaining series for both groups were incomplete for various reasons and have therefore not been included in this report. T h e C h a r n l e y G r o u p . In the femoral component, the orientation of the stem in the AP view was neutral in 57%, varus in 23%, and valgus in 20%. In the lateral view the orientation was neutral in 92%, anreverted in 6%, and retroverted in 2%. The cement covered the tip of the implant in 92% of the cases. At follow-up evaluation a radiolucent zone exceeding 2 mm was found between bone and cement in 16%. No radiolucency was found between cement and implant. 16% of the stems had subsided more than 1 ram, 5% more than 4 ram. In 38% there was a resorption of the calcar femorale (>2 ram). Ectopic bone formation of slight degree was seen in 35% of cases, 0fmoderate degree in 7%, and of severe degree in 2%. The trochanteric osteotomy was healed in 84%. In the acetabular component, the position of the cup in the AP view was neutral (45 • 10~ to the horizontal) in 92%, horizontal (2 mm
40 52 54 62 25 52 65
54 27 6 25 63 31 13
continuous radiolucent zone between the femoral stem and the surrounding bone developed in all patients. The radiolucency in all cases was surrounded by a sclerotic rim. The localization of radiolucent zones according to Gruen and coworkers (12) is shown in Table 8. Ectopic bone formation was of slight degree in 33%, of moderate degree in 18% of cases, and of severe degree in 3%. Resorption of the calcar femorale (>2 ram) occurred in 58%. A zone of increased density (endosteal hypertrophy) appeared immediately distal to the tip of the femoral stem in 57%. In 66% of the cases there was localized cortical hypertrophy laterally, adjacent to the distal part of the stem. 66% of the stems had subsided more than 1 ram, 33% more than 4 ram. In the acetabular component, the orientation of the cup in the AP view was neutral (45 -+ 10 ~ to the horizontal) in 85%, vertical (55 ~) in 13%. In the lateral view the cup was in neutral position in 75%, anteverted in 19%, and retroverted in 6%. A noncontinuous radiolucent zone between the acetabular implant and bone deVeloped in 67%. The radiolucency exceeded 2 m m in 9%. The radiolucency appeared in most cases laterally around one or two flanges. In no hip did the zone totally surround the cup. However, 6% of the cups had changed their position radiographically. The change of position was due to a rotation.
*
Wykman et al.
25
ticipate, to change of address, or to causes already mentioned under clinical assessment. The preoperative walking tests in the two groups did not differ significantly. T h e C h a r n l e y G r o u p . The average maximal walkhlg speed (Table 9) before total hip arthroplasty was 93 cm/s. At 6 months the patients achieved 119 cm/s, and at the most recent examination the patients achieved 127 cm/s. Preoperatively, 41% of the patients used no assistive device indoors to reduce load on the affected leg. At 6 months, 77%, and most recently, 85% required no assistive devices. In the gait test, 60% of the patients walked without an assistive device preoperatively. At 6 months, 78% and most recently, 89% required no device. The averagefree walking speed (Table 9) was 57 cm/ s preoperatively, 84 cm/s at 6 months, and 100 cm/ s at the most recent evaluation. Single stancephase before operation was 27% of the gait cycle for the affected leg and 33% for the unaffected. Single stance phase in normal walking is 40%. At 6 months single stance phase was 33% for the affected leg and 35% for the unaffected. At the most recent examination single stance phase was 35% for the affected leg and 36% for the unaffected. Ttle preoperative weight acceptancephase was 21% of the gait cycle for the affected leg and 19% for the unaffected. Weight acceptance phase in normal walking is 11%. At 6 months the weight acceptance phase was 17% for the affected leg and 16% for the unaffected. Most recently the weight acceptance phase of each leg was 15%. The preoperative maximal vertical force (MVF) (Table 10) for the affected leg was 95% of b o d y weight, for the unaffected 102%. At 6 months MVF was 102% for the affected leg and 106% for the unaffected. Most recently MVF was 107% for the affected leg and 108% for the unaffected. T h e H P - G a r c h e s G r o u p . The average preoperative maximal walkhlg speed (Table 9) was 94 cm/s.
Gait Analysis Preoperatively all patients in both groups we're examined by clinical assessment of gait as well as by objective means. At 6 months 67 patients in the Charnley group and 70 patients in the HP-Garches group were examined. Most recently, 60 patients in the Chamley group and 62 patients in the HPGarches group were examined. The patients missed in this part of the investigation could not be examined owing to technical failure of the gait analysis equipment, to the patients' o w n reluctance to par-
T a b l e 9.
Maximal Walking Speed and Free Walking Speed (cm/s).
iMax walking speed Preop 6 mo Follow-up Free w a l k i n g speed Preop 6 mo Follow-up
Charnley
HP-Garches
Sign. diff.
93 119 127
94 105 118
n.s. P < 0.001 n.s.
57 84 100
63 73 92
n.s. P < 0.001 n.s.
26
The Journal of Arthroplasty Vol. 6 No. 1 March 1991
At 6 months it was 105 cm/s, and most recently it was 118 cm/s. Preoperatively 41% of patients used no assistive device indoors to reduce load on the affected leg. At 6 months, 48% and most recently, 79% used no device. In the gait test 65% walked without an assistive device preoperatively, at 6 months 60%, and at the last examination 86%. The averagefree walking speed (Table 9) was 63 cm/ s preoperatively. It was 73 cm/s at 6 months and 92 cm/s at the most recent evaluation. Preoperatively the mean shzgle stance phase was 28% of the gait cycle for the affected leg and 33% for the unaffected. Single stance phase in normal walking is 40%. At 6 months the single stance phase was 31% for the affected leg and 33% for the unaffected. Most recently the single stance phase was 34% for the affected leg and 35% for the unaffected. Preoperatively the weight acceptance phase was 21% of the gait cycle for the affected leg and 18% for the unaffected. Weight acceptance phase in normal walking is 11%. At 6 months the mean weight acceptance phase for the affected leg was 18% and for the unaffected 17%. At the last examination the weight acceptance phase was 16% for the affected leg and 15% for the unaffected. The m e a n maximal vertical force (MVF) (Table 10) was 96% of body weight for the affected leg and 102% for the unaffected preoperatively. At 6 months mean MVF was 95% for the affected leg and 102% for the unaffected. Most recently it was 104% for the affected leg and 105% for the unaffected. C o m p a r i s o n o f D i f f e r e n t P a r a m e t e r s Bet w e e n G r o u p s . In comparing the different parameters between the two groups it emerged that at 6 months the differences were significant, whereas at the most recent evaluation no difference was significant. C o r r e l a t i o n B e t w e e n Clinical Results a n d G a i t A n a l y s i s . A significant correlation (p < 0.001, w h e n not specified) was found preoperatively as well as.at the most recent evaluation between
Harris hip score and free walking speed, single stance phase, and weight acceptance phase for the affected leg.
Maximal walking speed and free walking speed, single stance phase, weight acceptance phase for the affected leg, and Harris hip score. Patients' ophfion at the most recent evaluation and free walking speed, single stance phase, weight acceptance phase, and maximal walking speed (p < 0.01).
Discussion Before 1960, hip joint arthroplasties were performed with the implant material adapted directly to the bone w i t h o u t any intem~ediary layer. The early replacements met with only limited success even if many patients became painfree. Sir J o h n Charnley introduced the use of polymethylacrylate bone cement to hip a r t h r o p l a s t y - - o n e of the most important contributions to orthopaedic surgery in this century. Total hip arthroplasty rapidly became routine treatment for disabling hip disease. One of the problems encountered was infection. This, however, has in part been overcome (5, 9). Prosthetic c o m p o n e n t wear was another problem which has also in part been overcome (10). Noninfectious loosening has appeared as the most frequent cause of long-term failure of hip arthroplasties (20). The diagnostic criteria of loosening are controversial, which may account for considerable variation in reported rates of loosening. The reported rate of radiographic loosening ranges from less than 1% (32) to 38% (40) for the acetabular c o m p o n e n t and from 1% (14) to 47% (40) for the femoral component. The rate of revision for loosening 10 years after arthroplasty varies from 1.5% to 19% (2, 17). Bone cement acts by interlocking with the prosthetic irregularities and bony trabeculae (8). Failure of this mechanical interlocking may generate micromotion at the interfaces. The m o s t frequent longterm complication of cemented total hip arthroplasty is loosening at the b o n e - c e m e n t interface (1, 2, 12, 37, 39). This can be Particularly serious in younger patients (6, 11). This fact, combined with even poorer long-term results of revision total hip arthroplasties with cement (33) has stimulated consider-
Table 10. Maximal Vertical Force (% of Gait Cycle) Affected leg Preoperative 6 months Follow-up
95 102 107
Charnley Unaffectedleg 102 106 108
ttPrGarches Affected leg Unaffected leg 96 95 104
102 102 105
Sign. difL n.s. p < O.00l n.s.
Total Hip Arthroplasty * Wykman et al. able interest in a return to noncemented fixation methods. The presumption that bony tissue would grow adjacent to certain metallic surfaces and thus fix an endoprosthesis has made orthopedic surgeons cautiously enthusiastic about noncemented implants. Lord introduced a noncemented total hip prosthesis which consisted of a conical femoral stem with an irregular surface to facilitate bony ingrowth. The acetabular side was a threaded ellipsoid ring supporting a polyethylene interchangeable socket. In 1988 Lord and coworkers (21) reported 91.7% good or excellent results after 5 years in 625 hips. tIonnart and Patel in Garches designed a press-fit noncemented total hip prosthesis called the HP-Garches prosthesis. Good results were reported in 77.7% of the patients with a follow-up time of 3 to 5 years (31). When the present study was initiated (1981) the choice of noncemented prostheses was limited. The ttP-Garches prosthesis was chosen because it had become evident that the use of the Lord prosthesis was often complicated by intraoperative ,trochanteric fractures.
Midthigh Pain The patients in whom noncemented prostheses were placed developed an unexpected and at that time hardly known complication of midthigh pain. This symptom, which occurred on weightbeafing, was most frequently registered at 6 months following surgery (64% of the patients). The intensity of the pain often required the use of crotches or analgesics (42, 43). With time the symptoms diminished in most cases. In our series, 18 patients had persistent pain necessitating revision surgery. So far 13 patients have undergone revision owing to pain. Thigh pain has become more generally observed and reported in patients with noncemented total hip arthroplasties (15, 24, 34). It has been speculated that microfractures, micromotion or stress concentration oi a combination of these factors could cause these symptoms. Midthigh p~n may also be the forerunner of later changes characteristic of prosthetic loosening. In our study, midthigh pain most probably contributed to the intergroup differences in Harris hip score, maximal walking speed, and t h e patient's opinion 6 months postoperatively. Furthermore, these differences in clinical parameters were confirmed by all objective gait yariables chosen for this study.
27
Gait Analysis Murray and coworkers used gait analysis in comparative studies on the Chamley, Mfiller, and McKee-Farrar total hip procedures. No difference was found between the types of prostheses in the patients' functional performance after surgery (25, 26). Olsson, Goldie, and Wykman (28) used gait analysis to compare the early results after cemented and noncemented total hip arthroplasty in 119 of the patients in the present study. The patients treated with the cemented prosthesis demonstrated a faster free walking speed and a greater maximal vertical force. Hip function has also been measured by estimating the metabolic cost of walking, and this was found to be correlated to maximal walking speed as well as free walking speed (3, 22, 23). In the present study increased maximal walking speed and free walking speed indicated an improvement in hip function for both groups. These parameters appear to be a simpler way of estimating hip function than metabolic cost evaluation. Common to all methods aimed at the study of function after total hip anhroplasty is that normal values are seldom achieved despite good clinical results. In the present study we have analyzed the patient's walking ability by objective gait analysis and by clinical gait parameters. Maximal walking speed correlated strongly to the objective gait analysis. Maximal walking speed and free walking speed are easy to calculate with a stopwatch and can be recommended as simple measurements of changes in walking performance.
Radiologic Evaluation It has been speculated that localized bone resorption could be caused by bone cement (16), stress concentration/shielding (19, 38) surgical trauma (18, 35), or a combination of these factors. In our study, resorption of the calcar femorale was common in both the cemented and the noncemented group. "this indicates that bone cement can hardly be the only cause of localized bone resorption. The occurrence of endosteal as well as cortical hypertrophy distally in the femur characterized the noncemented ttP-Garches group. The relevance of these findings remains unclear. However, hypertrophy of bone often follows localized stress concentration. The radiographic findings in these cases may
28
The Journal of Arthroplasty Vol. 6 No. 1 March 1991
be the visible results of stress with or without micromotion at the tip of the femoral stem. In the HP-Garches group all femoral stems were surrounded by radiolucent zones of various sizes and locations. No relationship was found between the presence ofradiolucency and clinical outcome. However, a sinister sign was the high percentage of stems tilting over into varus postoperatively. There was no observable relationship between any radiographic signs and clinical outcome a m o n g the nonrevised hips.
Summary In the present study our interest was focused on differences in clinical outcome between the cemented Charnley prosthesis and the n o n c e m e n t e d press-fit HP-Garches prosthesis. Our investigation comprised 150 patients examined clinically, radiographically, and by gait analysis with a follow-up time of 3 to 5 years. In 59% of the:icases the followup time exceeded 4 years. Five patients (6.7%) with C h a m l e y prostheses and 14 patients (18.7%) with HP-Garches prostheses u n d e r w e n t revision owing to mechanical loosening. A further five patients in the n o n c e m e n t e d group appear to be candidates for revision. Thus, 25% of the patients in the HP-Garches group developed evident or highly suspected mechanical loosening. Accordingly, our findings do not comply with earlier optimistic expectations based on theoretical considerations (21, 3 I). It is well d o c u m e n t e d that cemented total hip arthroplasty is associated with a slowly increasing rate Of failure over time. W h e t h e r this also applies to n o n c e m e n t e d total hip arthroplasty is still u n k n o w n . Despite our findings, it m a y prove that the long-term results of HP-Garches prostheses are not worse than those obtained with cemented total hip arthroplasty.
References 1. Amstutz tIC, Lurie L,"Bullough P: Skeletal fixation with self-curing polymethyl-methacrylate. Clin Orthop 84:163, 1972 2. Amstutz ltC, Ma SM, Jirmah RH, Mai L: Revision of aseptic loose total hip arthroplasties. Clin Onhop 170:21, 1982 3. Arborelius M, Carlsson A, Nilsson B: Oxygen intake and walking speed before and after total hip replacement. Clin Orthop 121:113, 1976 4. Carlsson A, Gentz C:F: R6ntgenunders6kning vid total h6ftledsplastik. L~ikartidningen 74:760, 1977
5. Carlsson A, Josefsson G, Lindberg L: Revision with gentamicin-impregnated cement to deep infection in , total hip anhroplasties. J Bone Joint Surg 60A: 1059, 1978 6. Chandler H, Reineck F, Wixson R, McCarthy J: Total hip replacement in patients younger than thirty years old. J Bone Joint Surg 63A:1426, 1981 7. Chamley J: Anhroplasty of the hip: A new operation. Lancet 1:1129, 1961 8. Charnley J: The bonding of prostheses to bone by cement. J Bone Joint Surg 46B:518, 1964 9. Chamley J: The long-term results of low friction arthroplasty of the hip performed as a primary intervention. J Bone Joint Surg 54B:61, 1972 10. Chamley J: Low friction anhroplasty of the hip. Theory and practice. Springer Verlag, Berlin, 1979 11. Goldie I: Failed joint replacement. Current Orthop 1:3, 1986 12. Green T, McNeice G, Amstutz H: "Modes of failure" of cemented stem-type femoral components. A radiographic analysis of loosening. Clin Orthop 141:17, 1979 13. ttarris W: Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by mold arthroplasty. An end result study using a new method of result evaluation. J Bone Joint Surg 5 IA:737, 1969 14. Harris W, McCarthy J J, O'Neill D: Femoral component loosening using contemporary techniques of femoral cement fixation. J Bone Joint Surg 64A: 1063, 1982 15. Hedley A, Borden L, Hungerford D et al: Current status of non-cemented hip implants. Transactions of the 14th Annual Meeting of The Hip Society, New Orleans 1986, p. 49 16. ttienon C, Blomgren G, Lindgren U: Factors associated with calcar resorption in cemented total hip prostheses. Acta Onhop Scand 54:584, 1983 17. Kavanagh B, llstmp D, Fitzgerald R: Revision total hip anhroplasty. J Bone Joint Surg 67A:517, 1985 18. Krause W, Bradbury DKJ, Lunceford E: Temperature elevations in orthopaedic cutting operations. J Biomech 15:267, 1982 19. Lewis JL, Askew .MJ, Wixson RL et ah The inlluence of prosthetic stem stiffness and of calcar collar on stresses in the proximal end of the femur with a cemented femoral component. J Bone Joint Surg 66A:280, 1984 20. Lindberg II, Carlsson A: Mechanical loosening of the femoral component in total hip replacement, Branswick design. Acta Orthop Scand 54:557, 1983. 21. Lord G, Marotte J-it, Blanchard JP et ah Arthroplasties totales de hanche madr~poriques et polaris6es. Cementless madreporic polarised total hip prostheses. A ten-year review of 2688 cases. Revue Chir Orthop 74i3, 1988 22. Mattsson E, Brostr6m L-A, Linnarsson D. Walking efficiency after cemented and nonncemented total hip anhroplasty. Clin Orthop 254:170, 1990 23. McBeath A, Bahrke M, Balke B: Walking efficiency
Total Hip Arthroplasty before and after total hip replacement as determined by oxygen consumption. J Bone Joint Surg 62A:807, 1980 24. Mittelmeier H: Total hip replacement with the autophor cementfree ceramic prosthesis p. 225. In Morscher E (ed). The cementless fixation of hip endoprosthesis. Springer Verlag, Heidelberg, New York, 1984 25. Murray MP, Gore D, Brewer B e t al: Comparison of functional performance after McKee-Farrar, Chamley and Muller total hip replacement. Clin Orthop 121:33, 1976 26. Murray MP, Gore D, Brewer B e t al: Joint function after total hip arthroplasty: A four year follow-up of 72 cases with Chamley and Miiller replacements. Ciin Orthop 15"/:119, 1981 27. Olsson E, Oberg K, Ribbe T: A computerized method for clinical gait analysis of floor reaction forces and joint angular motion. Scand J Rehab Med 18:93, 1986 28. Olsson E, Goldie I, Wykman A: Total hip replacement. A comparison between Chamley and HP Garches fixation by clinical assessment and objective gait analysis. Scand J Rehab Med 18:107, 1986 29. Osbome G: The history of surgical access for hip replacement. Current Orthop 1:61, 1986" 30. Parker II, Wiesman H, Ewald F et al: Comparison of preoperative, intraoperative and early postoperative total hip replacements with and without trochanteric osteotomy. Clin Orthop 121:44, 1976 31.'Patel A, ttonnart F, Signoret F: HP-Garches Cementless THR in Young Patients. p. 260. In Sevastik J, Goldie I (eds). The young patient with degenerative hip disease. Almqvist & Wiksell Intemational, Stockholm, 1985 32. Paterson M, Fulford P, Dertham R: Loosening of the femoral component after total hip replacement. The thin black line and the sinking hip. J Bone Joint Surg 68B:392, 1986 33. Pellici P, Wilson PJ, Sledge C et al: Long term results
9 Wykman et al.
29
of revision total hip replacement. J Bone Joint Surg 67A:513, 1985 34. Reigstad A, Brandt M, Bye K: Results of a prospective comparison between a cemented and an uncemented hip prosthesis. Acta Orthop Scand (suppl) 227:31, 1988 35. Rhinelander F, Nelson C, Stewart R, Stewart C: Experimental reaming of the proximal femur and acrylic cement implantation. Clin Orthop 141:74, 1979 36. Robinson R, Robinson H, Salvati E: Comparison of the transtrochanterteric and posterior approaches for total hip replacement. Clin Orthop 147:143, 1980 37. Salvati EA, Wilson PJ, Jolley Met al: A ten-year follow-up study of our first one hundred consecutive Charnley total hip replacements. J Bone Joint Surg 63A:753, 1981 38. Sarmiento A, Turner TM, Latta LL et al: Factors contributing to lysis of the femoral neck in total hip arthroplasty. Clin Orthop 145:208, 1979 39. Stauffer R: Ten year follow-up study of total hip replacement. With particular reference to roentgenographic loosening of the components. J Bone Joint Surg 64A:983, 1982 40. Wejkner B, Wiege M: Correlation between radiologic and clinical findings in Chamley total hip replacement. A ten-year follow-up study. Acta Radiol Diagn 28:607, 1987 41. Wiesman H, Simon S, Ewald F et al: Total hip replacement with and without osteotomy of the greater trochanter. J Bone Joint Surg 60A:203, 1978 42. Wykman A, Goldie I: A prospective randomized study of cemented and non-cemented total hip prostheses-preliminary results after 3 years. Acta Orthop Seand 57:59, 1986 43. Wykman A, Goldie I: A prospective randomized study of cemented and non-cemented total hip prostheses-preliminary results after 4 years. Rev Chir Orthop 74:276, 1988
Anders Wykman, MD, PhD* Elisabeth Olsson RPT, PhD* Gabriella Axdorph, MD,-t- and Ian Goldie, MD, PhD*
Abstract: This study compares cemented (Chamley) with noncemented (ttonnart PateI-Garches) total hip arthroplasty. 150 patients with osteoarthritis, rheumatoid arthritis, and miscellaneous conditions were randomized into either group with 75 in each. Patients were evaluated over 5 postoperative years with examinations at 6 months, at 1 year, and then annually. Each examination involved clinical assessment, objective gait analysis, and radiographic examination. The Chamley group improved faster than the ttP-Garches group during the first 2 years. At the last evaluation the results were excellent or good in 79% in the Chamley group and 70% in the tlP-Garches group. In tile ttP-Garches group midthigh pain occurred in 64%. Five patients with Chamley prostheses (6.7%) and 14 patients with ttP-Garches prostheses {18.7% had revision surgery to correct mechanical loosening. Standard radiography revealed radiolucent zones in both groups, but this could not be related to the clinical outcome despite radiolucency being present in all noncemented replacements on the femoral side. Objective gait analysis confirmed inferior clinical results at 6 months in the ttP-Garches group. There was no significant difference between the groups at the most recent evaluation. Our findings are not consistent with earlier optimistic expectations on press-fit noncemented total hip arthroplasties. Key words: Noncemented procedures, hip arthroplasty, gait analysis, Charnley hip arthroplasty, ttP-Garches arthroplasty, radiography.
The first implants used in replacement surgery of the hip joint were inserted i~ direct contact"with bone. The success of early hip joint arthroplasties was c o m p r o m i s e d by the materials available at the time,
as well as the surgical techniques used, and the n o n availability of antibiotics. A b r e a k t h r o u g h was m a d e by. Sir J o h n Charnley, w h o believed that better results could be obtained using p o l y m e t h y l m e t h a c r y late, bone cement for stable fixation of the implant (7). By the early 1960s, total hip joint arthroplasty, including both femoral and acetabular c o m p o n e n t s , had become the standard procedure. Thirty years' experience has borne out Charnley's theory of a stable fixation of the implant in total hip arthroplasty. However, problems arise in the cases
Front the Departments of*Orthopaedics, and "~DiagnosticRadiology, Karolhuka Hospital Stockhohn, Sweden.
This investigation was supported by grants from the Karolinska Instltute's Research Funds, Loo and Hans Osterman Foundation, Ulla and Gustafaf Uggla Foundation, and the Swedish Association against Rheumatism. Reprint requests: Anders Wykman, MD, PhD, Department of Orthopaedics, Karolinska Hospital, S-10401 Stockholm, Sweden.
19
20
The Journal of Arthroplasty Vol. 6 No. 1 March 1991
which fail, mainly because of aseptic loosening. It is well k n o w n that revisions yield poorer results than primary total hip arthroplasties (33). In view of this, attempts have b e e n m a d e to reintroduce the old concept of total hip arthroplasty without the use of bone cement. N e w prostheses have been designed to adapt directly to bone. This study aims to c o m p a r e a standard cemented total hip arthroplasty with a n o n c e m e n t e d procedure in the following way: 1. To study prospectively 150 patients r a n d o m ized into two groups, one with a cemented total hip arthroplasty and one noncemented. 2. To base this comparison on clinical assessment, radiographic examination, and an objective gait analysis.
Patients and Methods Patients From the waiting list for total hip arthroplasty, 150 patients were consecutively r a n d o m i z e d into two groups, one to receive cemented and one noncemerited procedures. The indications for surgery included disabling pain and functional limitation of at least one hip. The investigation was approved by the ethics committee of the Karolinska Hospital, and all patients gave their informed consent. The operations were performed between September 1982 and December 1984. Postoperative examinations were conducted at 6 months, at 1 year, and then annually ( T a b l e 1).
Table 2. Age and Weight of Patients
Age, mean (range) years Weight, mean (range) kg
Charnley
HP-Garches
67.4 (48-82) 73.2 (45-110)
64.8 (29-82) 72.3 (49-98)
Table 3. Diagnoses
Charnley
HP-Garches
Total
Osteoarthritis Rheumatoid arthritis Miscellaneous
58 7 10
57 8 l0
l 15 15 20
Total
75
75
! 50
T h e H P - G a r c h e s G r o u p . There were 75 patienis, of w h o m 50 were female and 25 male, age 29 to 82 years, m e a n age 64.8 years, (Table 2), m e d i a n 65 years. In 60 patients only one hip was affected and there was no other condition that could interfere with walking. The remaining 15 patients suffered from bilateral hip disease, had some systemic disorder that interfered with locomotion, or both. Diagnoses are s h o w n in Table 3.
Implants Tile Charnley prosthesis consists of a stainless femoral stem with a 2 2 . 2 - m m spherical head and an ultrahigh molecular weight polyethylene cup. The components are anchored into b o n e with p o l y m e thylmethacrylate bone cement. The HP-Garches (Honnart Patel Garches) prosthesis consists of a conical femoral stem with a 3 2 - m m
T h e C h a r n l e y G r o u p . There were 75 patients, of w h o m 43 were female and 32 male, age 48 to 82 years, m e a n age 67.4 years (Table 2), median 68 years. In 60 patients only o n e hip was affected and there was no other condition that interfered with walking. The remaining 15 patients suffered from bilateral hip disease, had some systemic disorder that interfered with locomotion, or both. Diagnoses are s h o w n in Table 3. Table 1. Number of Patients Operated
Year
Chamley
HP-Garches
Total
1982 1983 1984
17 41 17
16 37 22
33 78 39
Total
75
75
150
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
j
Fig. 1. (Left) HP-Garches prosthesis. (Right) Chamley prosthesis.
Total Hip Arthroplasty
spherical head made of forged cobalt-chrome alloy and a threaded acetabular ring with an inner ultrahigh molecular weight polyethylene liner. The stem is fitted with a collar and has a rough surface with a pyramidal pattern (Fig. 1). All components are available in different sizes.
Surgical Procedures In the Chandey group, a lateral approach with trochanteric osteotomy was used and the procedure presented by Charnley in 1979 (10) was followed. Acrylic cement (polymethylmethacrylate) with barium sulphate (CMW) was used. No antibiotics were included in the cement. All operations were performed under epidural anesthesia. The operations lasted on average 123 minutes (70-180 min). Calculated total operative and postoperative blood loss was 1,875 nfl on average. In the HP-Garchesgroup, a posterolateral approach without trochanteric osteotomy was used. The acetabulum and femoral shaft were prepared with specially designed reamers. The acetabuliam was not tapped, and the threaded acetabular component was considered to be fully anchored in bone when the surgeon could no longer achieve any rotation on the T-handle. The operations lasted on average 104 minutes (60-180 rain) (p-value for comparison with the Charnley group p < 0.001). Calculated total operative and postoperative blood loss was 1,606 ml on average. The two surgical approaches used in the present study have previously been compared (30, 36, 41). No clinical difference was found in immediate postoperative course or later, and both approaches give satisfactory results (29). The operations in this study were performed by only experienced surgeons.
9 Wykman et al.
21
Table 4. Gradation of Harris Hip Score Excellent Good Fair Poor
90-100 80-89 70-79 - 30), and 10 patients had a pain score of 30
Assessment
Table 5. Results: Harris Hip Score (median)* H a r r i s H i p S c o r e ( T a b l e 5). Preoperatively, the Charnley group patients achieved a total score of 9 62 points, median 37.3 points. The average pain score (maximum 44 points; no pain) measured 13.5 points. There were no linear relationships between
Chamley Preoperative 6 months Most recent evaluation
37.3 89.4 95.3
* Reoperated patients included.
HP-Garches Sign. Diff. 38.1 74.3 88.7
n.s. p < 0.00l n.s.
Total Hip Arthroplasty Table 6. Number of Patients at Most Recent Evaluation
Pre--op
Chamley
tlP-Garches
Total
3 years 4 years 5 years
6 20 42
4 27 39
10 47 81
Total
68
70
138
or less. One patient was reoperated at 2 m o n t h s with a Girdlestone hip because of infection. At the most recent evaluation 68 patients w e r e examined 3 to 5 years after surgery (Table 6). Of the missing 6 patients, 1 patient refused to attend, I patient had been reoperated with a Girdlestone hip because of infection, and 4 patients had died. 54 patients (79%) obtained a good or excellent result according to the Harris score (Fig. 3). Median total score was 95.3, m e a n value 88.65 patients (96%) reported slight pain or no pain at all. Of the HP-Garches group, the patients had a total preoperative score of 1 9 - 6 4 points, .median 38.1 points. The average pain score measured 14.7 points. There were no linear relationships between the patients' ages, their diagnoses, and the results of the preoperative assessment. At 6 m o n t h s 74 patients (1 patient died on the first postoperative day) had a median total score of 74.3 points. Thirty-one patients (42%) obtained a good or excellent result, ie, their Harris score exceeded 80 points. Forty-five of the patients (61%) had a pain score indicating slight pain or no pain at all (>30). The remaining 29 patients had a pain score of 30 or less. At the most recent evaluation 70 patients were examined 3 to 5 years after surgery (Table 6). One patient had m o v e d a w a y and 3 patients had died. Forty-eight patients (70%) obtained a good or excellent result according to the Harris score (Fig. 3). Median total score was 88.7, m e a n value 85.60 of the patients (86%) reported slight pain or no pain at all, 10 patients had a pain score of 30 or less. Preoperatively there were no significant differences b e t w e e n tile C h a m l e y and the HP-Garches groups. At 6 m o n t h s the patients of the Charnley
* Wykman et al. 6 months
23
Follow-up
Charnley
HP-Garches
,I
I 90"100
~
70-79
8O-89 B B < 70 Fig. 3. Result: ttarris Hip Score. group had a higher total score and pain score t h a n the patients of the HP-Garches group (P < 0.001). No significant differences b e t w e e n the groups w e r e found at the most recent evaluation. There were no linear relationships b e t w e e n the assessments preoperatively, at 6 months, or thereafter, and the patients' age, weight, or diagnosis. R e v i s i o n s . In all there were 22 revisions. In the Charnley group 2 patients had their prostheses rem o v e d because of deep infection; 6 patients underwent revision with exchange of either one or b o t h components. In 5 patients revision was due to implant loosening and in i to malpositioning of the prosthesis with recurrent dislocation. In the HPGarches group 14 patients had either a n e w noncemented femoral c o m p o n e n t , a cemented acetabular component, a cemented femoral c o m p o n e n t or b o t h c o m p o n e n t s changed for cemented alternatives owing to loosening (Table 7).
Other Factors Influencing Outcome At the most recent evaluation, 32 of the 68 patients in the Charnley group and 44 of the 70 patients in the tiP-Garches group suffered from c o n c o m i t a n t disorders. Bilateral hip disease was present in 15 pa-
Table 7. Revisions Chamley'
HP-Garches
Total
Extraction New noncemented stem New cemented stem New cemented cup New cemented stem and cup
2 3 2 1
I 8 1 4
2 l ll 3 5
Total
8
14
22
24
The Journal of Arthroplasty Vol. 6 No. 1 March 1991
dents in the Charnley group and 19 in the HPGarches group. Systemic factors of different kinds were also responsible for failure to achieve normal locomotion. In the Chamley group all patients with only unilateral hip disease (n = 36) had a good or excellent result. The results for these patients were significantly better than the results for the Charnley group as a whole (P < 0.05). The corresponding frequency for the patients in the HP-Garches group was 24 out of 26 (92%). The result for these patients did not differ significantly from the result for the whole HP-Garches group.
Midthigh Pain A number of patients with the noncemented HPGarches prosthesis experienced pain on weightbearing in the midthigh region of the operated leg. After six postoperative months midthigh pain occurred in 46 of 72 patients (64%). In 28 of these, pain spontaneously disappeared during the following 3 years, in 18 cases during the first year. However, in 18 patients, pain not only persisted but increased, necessitating a reoperation in 13. A further 5 appear to be candidates for reoperation. Of the 13 patients u n d e r g o i n g revision, 10 have become completely painfree whereas the remaining 3 still suffer tolerable pain (Fig. 4).
6 months
46/72
reop
Patients' Opinion At 6 months 93% of the patients in the Charnley group were satisfied with the operation; 7% were dissatisfied. The corresponding figures for the ltPGarches group were 70% and 30% respectively. The differences are significant (P < 0.05). At the most recent evaluation, 95% in the Charnley group and 90% of the patients in the HP Garches were satisfied.
Radiologic Evaluation The radiographic survey at the most recent evaluation comprised 49 complete Charnley series and 53 complete HP-Garches series. The remaining series for both groups were incomplete for various reasons and have therefore not been included in this report. T h e C h a r n l e y G r o u p . In the femoral component, the orientation of the stem in the AP view was neutral in 57%, varus in 23%, and valgus in 20%. In the lateral view the orientation was neutral in 92%, anreverted in 6%, and retroverted in 2%. The cement covered the tip of the implant in 92% of the cases. At follow-up evaluation a radiolucent zone exceeding 2 mm was found between bone and cement in 16%. No radiolucency was found between cement and implant. 16% of the stems had subsided more than 1 ram, 5% more than 4 ram. In 38% there was a resorption of the calcar femorale (>2 ram). Ectopic bone formation of slight degree was seen in 35% of cases, 0fmoderate degree in 7%, and of severe degree in 2%. The trochanteric osteotomy was healed in 84%. In the acetabular component, the position of the cup in the AP view was neutral (45 • 10~ to the horizontal) in 92%, horizontal (2 mm
40 52 54 62 25 52 65
54 27 6 25 63 31 13
continuous radiolucent zone between the femoral stem and the surrounding bone developed in all patients. The radiolucency in all cases was surrounded by a sclerotic rim. The localization of radiolucent zones according to Gruen and coworkers (12) is shown in Table 8. Ectopic bone formation was of slight degree in 33%, of moderate degree in 18% of cases, and of severe degree in 3%. Resorption of the calcar femorale (>2 ram) occurred in 58%. A zone of increased density (endosteal hypertrophy) appeared immediately distal to the tip of the femoral stem in 57%. In 66% of the cases there was localized cortical hypertrophy laterally, adjacent to the distal part of the stem. 66% of the stems had subsided more than 1 ram, 33% more than 4 ram. In the acetabular component, the orientation of the cup in the AP view was neutral (45 -+ 10 ~ to the horizontal) in 85%, vertical (55 ~) in 13%. In the lateral view the cup was in neutral position in 75%, anteverted in 19%, and retroverted in 6%. A noncontinuous radiolucent zone between the acetabular implant and bone deVeloped in 67%. The radiolucency exceeded 2 m m in 9%. The radiolucency appeared in most cases laterally around one or two flanges. In no hip did the zone totally surround the cup. However, 6% of the cups had changed their position radiographically. The change of position was due to a rotation.
*
Wykman et al.
25
ticipate, to change of address, or to causes already mentioned under clinical assessment. The preoperative walking tests in the two groups did not differ significantly. T h e C h a r n l e y G r o u p . The average maximal walkhlg speed (Table 9) before total hip arthroplasty was 93 cm/s. At 6 months the patients achieved 119 cm/s, and at the most recent examination the patients achieved 127 cm/s. Preoperatively, 41% of the patients used no assistive device indoors to reduce load on the affected leg. At 6 months, 77%, and most recently, 85% required no assistive devices. In the gait test, 60% of the patients walked without an assistive device preoperatively. At 6 months, 78% and most recently, 89% required no device. The averagefree walking speed (Table 9) was 57 cm/ s preoperatively, 84 cm/s at 6 months, and 100 cm/ s at the most recent evaluation. Single stancephase before operation was 27% of the gait cycle for the affected leg and 33% for the unaffected. Single stance phase in normal walking is 40%. At 6 months single stance phase was 33% for the affected leg and 35% for the unaffected. At the most recent examination single stance phase was 35% for the affected leg and 36% for the unaffected. Ttle preoperative weight acceptancephase was 21% of the gait cycle for the affected leg and 19% for the unaffected. Weight acceptance phase in normal walking is 11%. At 6 months the weight acceptance phase was 17% for the affected leg and 16% for the unaffected. Most recently the weight acceptance phase of each leg was 15%. The preoperative maximal vertical force (MVF) (Table 10) for the affected leg was 95% of b o d y weight, for the unaffected 102%. At 6 months MVF was 102% for the affected leg and 106% for the unaffected. Most recently MVF was 107% for the affected leg and 108% for the unaffected. T h e H P - G a r c h e s G r o u p . The average preoperative maximal walkhlg speed (Table 9) was 94 cm/s.
Gait Analysis Preoperatively all patients in both groups we're examined by clinical assessment of gait as well as by objective means. At 6 months 67 patients in the Charnley group and 70 patients in the HP-Garches group were examined. Most recently, 60 patients in the Chamley group and 62 patients in the HPGarches group were examined. The patients missed in this part of the investigation could not be examined owing to technical failure of the gait analysis equipment, to the patients' o w n reluctance to par-
T a b l e 9.
Maximal Walking Speed and Free Walking Speed (cm/s).
iMax walking speed Preop 6 mo Follow-up Free w a l k i n g speed Preop 6 mo Follow-up
Charnley
HP-Garches
Sign. diff.
93 119 127
94 105 118
n.s. P < 0.001 n.s.
57 84 100
63 73 92
n.s. P < 0.001 n.s.
26
The Journal of Arthroplasty Vol. 6 No. 1 March 1991
At 6 months it was 105 cm/s, and most recently it was 118 cm/s. Preoperatively 41% of patients used no assistive device indoors to reduce load on the affected leg. At 6 months, 48% and most recently, 79% used no device. In the gait test 65% walked without an assistive device preoperatively, at 6 months 60%, and at the last examination 86%. The averagefree walking speed (Table 9) was 63 cm/ s preoperatively. It was 73 cm/s at 6 months and 92 cm/s at the most recent evaluation. Preoperatively the mean shzgle stance phase was 28% of the gait cycle for the affected leg and 33% for the unaffected. Single stance phase in normal walking is 40%. At 6 months the single stance phase was 31% for the affected leg and 33% for the unaffected. Most recently the single stance phase was 34% for the affected leg and 35% for the unaffected. Preoperatively the weight acceptance phase was 21% of the gait cycle for the affected leg and 18% for the unaffected. Weight acceptance phase in normal walking is 11%. At 6 months the mean weight acceptance phase for the affected leg was 18% and for the unaffected 17%. At the last examination the weight acceptance phase was 16% for the affected leg and 15% for the unaffected. The m e a n maximal vertical force (MVF) (Table 10) was 96% of body weight for the affected leg and 102% for the unaffected preoperatively. At 6 months mean MVF was 95% for the affected leg and 102% for the unaffected. Most recently it was 104% for the affected leg and 105% for the unaffected. C o m p a r i s o n o f D i f f e r e n t P a r a m e t e r s Bet w e e n G r o u p s . In comparing the different parameters between the two groups it emerged that at 6 months the differences were significant, whereas at the most recent evaluation no difference was significant. C o r r e l a t i o n B e t w e e n Clinical Results a n d G a i t A n a l y s i s . A significant correlation (p < 0.001, w h e n not specified) was found preoperatively as well as.at the most recent evaluation between
Harris hip score and free walking speed, single stance phase, and weight acceptance phase for the affected leg.
Maximal walking speed and free walking speed, single stance phase, weight acceptance phase for the affected leg, and Harris hip score. Patients' ophfion at the most recent evaluation and free walking speed, single stance phase, weight acceptance phase, and maximal walking speed (p < 0.01).
Discussion Before 1960, hip joint arthroplasties were performed with the implant material adapted directly to the bone w i t h o u t any intem~ediary layer. The early replacements met with only limited success even if many patients became painfree. Sir J o h n Charnley introduced the use of polymethylacrylate bone cement to hip a r t h r o p l a s t y - - o n e of the most important contributions to orthopaedic surgery in this century. Total hip arthroplasty rapidly became routine treatment for disabling hip disease. One of the problems encountered was infection. This, however, has in part been overcome (5, 9). Prosthetic c o m p o n e n t wear was another problem which has also in part been overcome (10). Noninfectious loosening has appeared as the most frequent cause of long-term failure of hip arthroplasties (20). The diagnostic criteria of loosening are controversial, which may account for considerable variation in reported rates of loosening. The reported rate of radiographic loosening ranges from less than 1% (32) to 38% (40) for the acetabular c o m p o n e n t and from 1% (14) to 47% (40) for the femoral component. The rate of revision for loosening 10 years after arthroplasty varies from 1.5% to 19% (2, 17). Bone cement acts by interlocking with the prosthetic irregularities and bony trabeculae (8). Failure of this mechanical interlocking may generate micromotion at the interfaces. The m o s t frequent longterm complication of cemented total hip arthroplasty is loosening at the b o n e - c e m e n t interface (1, 2, 12, 37, 39). This can be Particularly serious in younger patients (6, 11). This fact, combined with even poorer long-term results of revision total hip arthroplasties with cement (33) has stimulated consider-
Table 10. Maximal Vertical Force (% of Gait Cycle) Affected leg Preoperative 6 months Follow-up
95 102 107
Charnley Unaffectedleg 102 106 108
ttPrGarches Affected leg Unaffected leg 96 95 104
102 102 105
Sign. difL n.s. p < O.00l n.s.
Total Hip Arthroplasty * Wykman et al. able interest in a return to noncemented fixation methods. The presumption that bony tissue would grow adjacent to certain metallic surfaces and thus fix an endoprosthesis has made orthopedic surgeons cautiously enthusiastic about noncemented implants. Lord introduced a noncemented total hip prosthesis which consisted of a conical femoral stem with an irregular surface to facilitate bony ingrowth. The acetabular side was a threaded ellipsoid ring supporting a polyethylene interchangeable socket. In 1988 Lord and coworkers (21) reported 91.7% good or excellent results after 5 years in 625 hips. tIonnart and Patel in Garches designed a press-fit noncemented total hip prosthesis called the HP-Garches prosthesis. Good results were reported in 77.7% of the patients with a follow-up time of 3 to 5 years (31). When the present study was initiated (1981) the choice of noncemented prostheses was limited. The ttP-Garches prosthesis was chosen because it had become evident that the use of the Lord prosthesis was often complicated by intraoperative ,trochanteric fractures.
Midthigh Pain The patients in whom noncemented prostheses were placed developed an unexpected and at that time hardly known complication of midthigh pain. This symptom, which occurred on weightbeafing, was most frequently registered at 6 months following surgery (64% of the patients). The intensity of the pain often required the use of crotches or analgesics (42, 43). With time the symptoms diminished in most cases. In our series, 18 patients had persistent pain necessitating revision surgery. So far 13 patients have undergone revision owing to pain. Thigh pain has become more generally observed and reported in patients with noncemented total hip arthroplasties (15, 24, 34). It has been speculated that microfractures, micromotion or stress concentration oi a combination of these factors could cause these symptoms. Midthigh p~n may also be the forerunner of later changes characteristic of prosthetic loosening. In our study, midthigh pain most probably contributed to the intergroup differences in Harris hip score, maximal walking speed, and t h e patient's opinion 6 months postoperatively. Furthermore, these differences in clinical parameters were confirmed by all objective gait yariables chosen for this study.
27
Gait Analysis Murray and coworkers used gait analysis in comparative studies on the Chamley, Mfiller, and McKee-Farrar total hip procedures. No difference was found between the types of prostheses in the patients' functional performance after surgery (25, 26). Olsson, Goldie, and Wykman (28) used gait analysis to compare the early results after cemented and noncemented total hip arthroplasty in 119 of the patients in the present study. The patients treated with the cemented prosthesis demonstrated a faster free walking speed and a greater maximal vertical force. Hip function has also been measured by estimating the metabolic cost of walking, and this was found to be correlated to maximal walking speed as well as free walking speed (3, 22, 23). In the present study increased maximal walking speed and free walking speed indicated an improvement in hip function for both groups. These parameters appear to be a simpler way of estimating hip function than metabolic cost evaluation. Common to all methods aimed at the study of function after total hip anhroplasty is that normal values are seldom achieved despite good clinical results. In the present study we have analyzed the patient's walking ability by objective gait analysis and by clinical gait parameters. Maximal walking speed correlated strongly to the objective gait analysis. Maximal walking speed and free walking speed are easy to calculate with a stopwatch and can be recommended as simple measurements of changes in walking performance.
Radiologic Evaluation It has been speculated that localized bone resorption could be caused by bone cement (16), stress concentration/shielding (19, 38) surgical trauma (18, 35), or a combination of these factors. In our study, resorption of the calcar femorale was common in both the cemented and the noncemented group. "this indicates that bone cement can hardly be the only cause of localized bone resorption. The occurrence of endosteal as well as cortical hypertrophy distally in the femur characterized the noncemented ttP-Garches group. The relevance of these findings remains unclear. However, hypertrophy of bone often follows localized stress concentration. The radiographic findings in these cases may
28
The Journal of Arthroplasty Vol. 6 No. 1 March 1991
be the visible results of stress with or without micromotion at the tip of the femoral stem. In the HP-Garches group all femoral stems were surrounded by radiolucent zones of various sizes and locations. No relationship was found between the presence ofradiolucency and clinical outcome. However, a sinister sign was the high percentage of stems tilting over into varus postoperatively. There was no observable relationship between any radiographic signs and clinical outcome a m o n g the nonrevised hips.
Summary In the present study our interest was focused on differences in clinical outcome between the cemented Charnley prosthesis and the n o n c e m e n t e d press-fit HP-Garches prosthesis. Our investigation comprised 150 patients examined clinically, radiographically, and by gait analysis with a follow-up time of 3 to 5 years. In 59% of the:icases the followup time exceeded 4 years. Five patients (6.7%) with C h a m l e y prostheses and 14 patients (18.7%) with HP-Garches prostheses u n d e r w e n t revision owing to mechanical loosening. A further five patients in the n o n c e m e n t e d group appear to be candidates for revision. Thus, 25% of the patients in the HP-Garches group developed evident or highly suspected mechanical loosening. Accordingly, our findings do not comply with earlier optimistic expectations based on theoretical considerations (21, 3 I). It is well d o c u m e n t e d that cemented total hip arthroplasty is associated with a slowly increasing rate Of failure over time. W h e t h e r this also applies to n o n c e m e n t e d total hip arthroplasty is still u n k n o w n . Despite our findings, it m a y prove that the long-term results of HP-Garches prostheses are not worse than those obtained with cemented total hip arthroplasty.
References 1. Amstutz tIC, Lurie L,"Bullough P: Skeletal fixation with self-curing polymethyl-methacrylate. Clin Orthop 84:163, 1972 2. Amstutz ltC, Ma SM, Jirmah RH, Mai L: Revision of aseptic loose total hip arthroplasties. Clin Onhop 170:21, 1982 3. Arborelius M, Carlsson A, Nilsson B: Oxygen intake and walking speed before and after total hip replacement. Clin Orthop 121:113, 1976 4. Carlsson A, Gentz C:F: R6ntgenunders6kning vid total h6ftledsplastik. L~ikartidningen 74:760, 1977
5. Carlsson A, Josefsson G, Lindberg L: Revision with gentamicin-impregnated cement to deep infection in , total hip anhroplasties. J Bone Joint Surg 60A: 1059, 1978 6. Chandler H, Reineck F, Wixson R, McCarthy J: Total hip replacement in patients younger than thirty years old. J Bone Joint Surg 63A:1426, 1981 7. Chamley J: Anhroplasty of the hip: A new operation. Lancet 1:1129, 1961 8. Charnley J: The bonding of prostheses to bone by cement. J Bone Joint Surg 46B:518, 1964 9. Chamley J: The long-term results of low friction arthroplasty of the hip performed as a primary intervention. J Bone Joint Surg 54B:61, 1972 10. Chamley J: Low friction anhroplasty of the hip. Theory and practice. Springer Verlag, Berlin, 1979 11. Goldie I: Failed joint replacement. Current Orthop 1:3, 1986 12. Green T, McNeice G, Amstutz H: "Modes of failure" of cemented stem-type femoral components. A radiographic analysis of loosening. Clin Orthop 141:17, 1979 13. ttarris W: Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by mold arthroplasty. An end result study using a new method of result evaluation. J Bone Joint Surg 5 IA:737, 1969 14. Harris W, McCarthy J J, O'Neill D: Femoral component loosening using contemporary techniques of femoral cement fixation. J Bone Joint Surg 64A: 1063, 1982 15. Hedley A, Borden L, Hungerford D et al: Current status of non-cemented hip implants. Transactions of the 14th Annual Meeting of The Hip Society, New Orleans 1986, p. 49 16. ttienon C, Blomgren G, Lindgren U: Factors associated with calcar resorption in cemented total hip prostheses. Acta Onhop Scand 54:584, 1983 17. Kavanagh B, llstmp D, Fitzgerald R: Revision total hip anhroplasty. J Bone Joint Surg 67A:517, 1985 18. Krause W, Bradbury DKJ, Lunceford E: Temperature elevations in orthopaedic cutting operations. J Biomech 15:267, 1982 19. Lewis JL, Askew .MJ, Wixson RL et ah The inlluence of prosthetic stem stiffness and of calcar collar on stresses in the proximal end of the femur with a cemented femoral component. J Bone Joint Surg 66A:280, 1984 20. Lindberg II, Carlsson A: Mechanical loosening of the femoral component in total hip replacement, Branswick design. Acta Orthop Scand 54:557, 1983. 21. Lord G, Marotte J-it, Blanchard JP et ah Arthroplasties totales de hanche madr~poriques et polaris6es. Cementless madreporic polarised total hip prostheses. A ten-year review of 2688 cases. Revue Chir Orthop 74i3, 1988 22. Mattsson E, Brostr6m L-A, Linnarsson D. Walking efficiency after cemented and nonncemented total hip anhroplasty. Clin Orthop 254:170, 1990 23. McBeath A, Bahrke M, Balke B: Walking efficiency
Total Hip Arthroplasty before and after total hip replacement as determined by oxygen consumption. J Bone Joint Surg 62A:807, 1980 24. Mittelmeier H: Total hip replacement with the autophor cementfree ceramic prosthesis p. 225. In Morscher E (ed). The cementless fixation of hip endoprosthesis. Springer Verlag, Heidelberg, New York, 1984 25. Murray MP, Gore D, Brewer B e t al: Comparison of functional performance after McKee-Farrar, Chamley and Muller total hip replacement. Clin Orthop 121:33, 1976 26. Murray MP, Gore D, Brewer B e t al: Joint function after total hip arthroplasty: A four year follow-up of 72 cases with Chamley and Miiller replacements. Ciin Orthop 15"/:119, 1981 27. Olsson E, Oberg K, Ribbe T: A computerized method for clinical gait analysis of floor reaction forces and joint angular motion. Scand J Rehab Med 18:93, 1986 28. Olsson E, Goldie I, Wykman A: Total hip replacement. A comparison between Chamley and HP Garches fixation by clinical assessment and objective gait analysis. Scand J Rehab Med 18:107, 1986 29. Osbome G: The history of surgical access for hip replacement. Current Orthop 1:61, 1986" 30. Parker II, Wiesman H, Ewald F et al: Comparison of preoperative, intraoperative and early postoperative total hip replacements with and without trochanteric osteotomy. Clin Orthop 121:44, 1976 31.'Patel A, ttonnart F, Signoret F: HP-Garches Cementless THR in Young Patients. p. 260. In Sevastik J, Goldie I (eds). The young patient with degenerative hip disease. Almqvist & Wiksell Intemational, Stockholm, 1985 32. Paterson M, Fulford P, Dertham R: Loosening of the femoral component after total hip replacement. The thin black line and the sinking hip. J Bone Joint Surg 68B:392, 1986 33. Pellici P, Wilson PJ, Sledge C et al: Long term results
9 Wykman et al.
29
of revision total hip replacement. J Bone Joint Surg 67A:513, 1985 34. Reigstad A, Brandt M, Bye K: Results of a prospective comparison between a cemented and an uncemented hip prosthesis. Acta Orthop Scand (suppl) 227:31, 1988 35. Rhinelander F, Nelson C, Stewart R, Stewart C: Experimental reaming of the proximal femur and acrylic cement implantation. Clin Orthop 141:74, 1979 36. Robinson R, Robinson H, Salvati E: Comparison of the transtrochanterteric and posterior approaches for total hip replacement. Clin Orthop 147:143, 1980 37. Salvati EA, Wilson PJ, Jolley Met al: A ten-year follow-up study of our first one hundred consecutive Charnley total hip replacements. J Bone Joint Surg 63A:753, 1981 38. Sarmiento A, Turner TM, Latta LL et al: Factors contributing to lysis of the femoral neck in total hip arthroplasty. Clin Orthop 145:208, 1979 39. Stauffer R: Ten year follow-up study of total hip replacement. With particular reference to roentgenographic loosening of the components. J Bone Joint Surg 64A:983, 1982 40. Wejkner B, Wiege M: Correlation between radiologic and clinical findings in Chamley total hip replacement. A ten-year follow-up study. Acta Radiol Diagn 28:607, 1987 41. Wiesman H, Simon S, Ewald F et al: Total hip replacement with and without osteotomy of the greater trochanter. J Bone Joint Surg 60A:203, 1978 42. Wykman A, Goldie I: A prospective randomized study of cemented and non-cemented total hip prostheses-preliminary results after 3 years. Acta Orthop Seand 57:59, 1986 43. Wykman A, Goldie I: A prospective randomized study of cemented and non-cemented total hip prostheses-preliminary results after 4 years. Rev Chir Orthop 74:276, 1988
