Total hip arthroplasty has proved to be a reliable treatment for debilitating hip pain in patients who have osteoarthrosis. The procedure has provided good results when fixation with cement, fixation without cement, or a so-called hybrid procedure (an acetabular component fixed without cement and a femoral component fixed with cement) has been used4,8,14,17,19,22,30,32,34,35. Controversy centers on which implant and method of fixation are best for a given group of patients. Total hip arthroplasty performed with cement has been described as "the treatment of choice for severe degeneration of the hip in older and relatively inactive patients."15 In addition, the concept of so-called implant-matching.2,20 includes the premises that components fixed without cement should be reserved for young, active patients and components fixed with cement are more appropriate for older patients19,26.
We studied the long-term results of total hip arthroplasty performed without cement in patients who were sixty-five years of age or older. We believe that such information is currently lacking in the literature. Furthermore, this study provides a reference group for comparison with other series and offers the opportunity for developing objective, outcome-based guidelines for the selection of implants.
*One or more of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article. In addition, benefits have been or will be directed to a research fund, foundation, educational institution, or other nonprofit organization with which one or more of the authors is associated. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was DePuy (royalties to Dr. Engh).
†Anderson Orthopaedic Research Institute, P.O. Box 7088, Alexandria, Virginia 22307. E-mail address for Dr. McAuley and Dr. Engh: jim@aori.org.
Between January 1982 and July 1988, 212 primary total hip arthroplasties were performed in 203 unselected, consecutive patients who were sixty-five years of age or older (Table I). Of the 203 patients, ninety-five (47 percent) were sixty-five to sixty-nine years of age, sixty-three (31 percent) were seventy to seventy-four years of age, and forty-five (22 percent) were at least seventy-five years of age.
A posterior approach was used in each patient. All of the femoral components were of the anatomic medullary locking design (AML; DePuy, Warsaw, Indiana). The standard component was used in 161 hips (76 percent), the component with a modified medial aspect was used in forty-six hips (22 percent), and the lateralized component was used in five hips (2 percent). One hundred and ninety-eight (93 percent) of the components were five-eighths porous-coated, and the remaining fourteen (7 percent) were fully porous-coated. Diaphyseal fixation was obtained with sequential straight reaming until at least five to six centimeters of so-called scratch fit was obtained. A femoral component that was 0.5 millimeter larger in diameter than the final reamer was impacted to obtain good initial stability. If the component did not progress steadily during impaction, the diaphysis was reamed to the diameter of the stem. In very osteopenic bone with thin cortices, the force of reaming and impaction was modified but the same principles were followed.
A fully porous-coated acetabular component was used in all of the hips. Before 1985, the acetabular components were not modular, had three fixation spikes, and were inserted after reaming of the acetabulum to the same outer diameter as the cup. Since 1985, modular components have been used, with the acetabulum reamed to a diameter that is one to two millimeters smaller than the diameter of the cup (Table II).
All of the patients walked with protected weight-bearing and two crutches for six weeks after the operation. They then used one crutch for four weeks and, finally, used a cane for four weeks.
Clinical and radiographic data were obtained during standard follow-up visits. The patients were seen annually for the first two years and then every three years if they had good clinical and radiographic progress. Survivorship analyses were performed according to the life-table method31, with any reoperation, a femoral reoperation, and an acetabular reoperation as the end points.
Clinical evaluations, which included the use of standardized patient questionnaires and clinical information sheets, provided data on the location and severity of pain, walking and functional capabilities, and patient satisfaction. Collected data were stored in a computerized database. Statistical comparisons were made with chi-square and Fisher exact tests, with the significance level set at less than 0.05.
Standardized anteroposterior and lateral radiographs, made immediately postoperatively and at sequential follow-up visits, were analyzed for stability of the components; the components were classified, according to previously published criteria, as osseointegrated, fibrous stable, or unstable13. Components that showed so-called spot welds or stress-shielding, or both, at the level of the calcar were considered osseointegrated; those that lacked definite ingrowth but had no progressive lucency or change in position were designated as fibrous stable; and those with clear signs of loosening, including axial or angular migration, were classified as unstable. A change in the position of the acetabular component (a linear change of two millimeters or an angular change of 5 degrees), bead-shedding, and increasing circumferential lucency were considered to be signs of loosening.
Stress-shielding was graded on the two-year postoperative radiographs according to a previously published method12. Stress-shielding was classified as pronounced and of potential clinical concern if it extended to the level of the lesser trochanter or distal to it. The relationships between stress-shielding and prosthesis-related and patient-related variables and between stress-shielding and the outcome were analyzed.
We looked for osteolytic lesions and documented their size and location using a previously described technique35. Lesions that were more than 1.5 square centimeters were considered of potential clinical importance, and their location in the femur18 or the acetabulum10 was recorded.
The prevalence of heterotopic ossification was determined, and lesions were classified according to the criteria of Brooker et al.6.
Polyethylene wear was evaluated in two dimensions9 on anteroposterior radiographs, and any change of more than two millimeters in the position of the femoral head in the acetabular shell was noted.
For a more complete analysis of early complications, we extended this portion of the review to include the 192 hips that had been followed for at least one year. Because of their clinical importance, the failures that resulted in reoperations were studied in detail to determine the indications for the reoperation, the type of reoperation, and the outcome.
Of the entire series of 212 hips, thirty-seven (17 percent) were in patients who died and sixteen (8 percent) were in patients who were lost to follow-up less than five years after the operation; thus, 159 hips (75 percent) were followed for at least five years. The patients who were unable to return for a follow-up evaluation completed a telephone questionnaire, and they were asked to send recent follow-up radiographs. Because not all of these patients were able to send recent radiographs, only 142 hips were included in the radiographic analysis.
Clinical Findings
The cumulative probability of survival at twelve years was 0.92 (95 percent confidence interval, 0.85 to 0.99) with any reoperation as the end point, 0.97 (95 percent confidence interval, 0.91 to 1.00) with any femoral reoperation as the end point, and 0.92 (95 percent confidence interval, 0.85 to 0.99) with any acetabular reoperation as the end point (Fig. 1). Of the 152 patients who had been followed for at least five years (mean, 8.5 years; range, five to fourteen years) and had not had a revision, 120 (79 percent) reported having no hip-related pain. Nineteen patients (13 percent) had mild pain that did not limit activity: eight (5 percent) had tenderness and pain about the greater trochanter, and eleven (7 percent) had deep discomfort. Of the eleven patients who had deep discomfort, five (3 percent overall) had pain in the thigh and six (4 percent overall) had pain in both the thigh and the buttock. Thirteen (9 percent) of the 152 patients had pain that was severe enough to limit some or most of their activities: eight (5 percent overall) had trochanteric pain and tenderness, and the remaining five (3 percent overall) had deep pain in the thigh (four patients) or groin (one patient).
Even though the mean age was eighty years at the most recent follow-up evaluation, seventy-six patients (50 percent) could walk more than six blocks without an assistive device and another thirty-five (23 percent) could do so with a cane. Twenty-nine patients (19 percent) could walk only two or three blocks without support or used a support at all times when walking, and twelve patients (8 percent) were limited to walking only indoors. One hundred and thirty-eight patients (91 percent) had little difficulty putting on shoes and socks, and 111 patients (73 percent) could ascend and descend stairs.
One hundred and forty-four patients (95 percent) reported improved overall function, and 149 patients (98 percent) were satisfied with the functional outcome. Of the three patients who were not satisfied, one had fallen and had sustained a fracture of the pubic ramus. Although radiographs revealed a fracture through an osteolytic ischial lesion and showed a loose cup, the component was not revised for medical reasons. The dissatisfaction of the second patient was related to degenerative disease of the knee and a recurrent dislocation of the contralateral total hip prosthesis. The third patient had an osteoarthrotic ipsilateral knee that limited function. However, she declined an operation.
Radiographic Findings
One hundred and thirty-five hips that were not revised were followed clinically and radiographically for at least five years (mean, 8.2 years; range, five to fourteen years). Clear osseointegration was seen in 131 (97 percent) of the 135 femoral stems, and three other stems (2 percent) were classified as fibrous stable radiographically. The one radiographically loose stem, in a ninety-year-old man, was not symptomatic. With the two femoral revisions that will be described later, the rate of failure (loosening or revision) of the femoral component was three (2 percent) of 142.
Three (2 per cent) of the 135 hips had a radiographically loose acetabular component. Combined with the reoperations that will be described later (none of which were done because of a loose acetabular shell), the total rate of failure of the acetabular component was nine (6 per cent) of 142.
Stress-shielding was pronounced in forty-five (26 per cent) of the 174 hips that had been followed for at least two years. Despite this frequency, none of the hips with pronounced stress-shielding had loosening of the component, osteolysis, or a reoperation. Stress-shielding had no adverse effect on the clinical result. The prevalence of stress-shielding was significantly higher (p < 0.05, Fisher exact test) in the women and in the patients who had a stem that was fifteen millimeters or more in diameter.
Three hips that were not revised had osteolytic lesions of more than 1.5 square centimeters. In two hips, the lesions were located in zones 1 and 7 of Gruen et al.18, and in one hip the lesion was located in zone 1 (the ilium) of DeLee and Charnley10. Because the components were clinically successful and radiographically integrated, they were not revised because of the osteolysis. One ischial fracture due to osteolysis increased the rate of clinically important osteolysis to four (3 percent) of 142.
In addition to three fractured polyethylene liners that were revised, five (4 percent) of the hips had at least two millimeters of polyethylene wear on two-dimensional measurement. None had osteolysis, and none had a revision.
Eight hips (6 percent) had grade-3 heterotopic ossification6, and no hip had grade-4. None of these patients had functional impairment. The range of motion was mildly limited, and this mainly affected rotation, not flexion (for example, only one patient was unable to put on shoes and socks).
Complications
Nine (5 percent) of the hips in this series had a dislocation. All of these dislocations occurred more than six months postoperatively. Four (2 percent) of the hips had recurrent dislocations: three were treated with closed reduction and one, with a preoperation.
Intraoperatively, one patient (less than 1 percent) sustained an undisplaced fracture of the proximal aspect of the femur, which was secured by two cerclage wires. The period of protected weight-bearing was extended by three weeks, with no additional precautions, and the fracture healed without consequence.
One patient (less than 1 percent) had an infection that led to a revision, as described later.
Reoperations
There were seven reoperations, three of which were due to a fracture of the acetabular polyethylene liner (Table III). The fractures occurred in the sixth, seventh, and ninth postoperative years in patients who had a Triloc shell with an ACS (Acetabular Cup System) polyethylene liner (DePuy). These patients typically had audible squeaking, clicking, and symptoms of subluxation.
In one patient, the fractured polyethylene liner was replaced with a thicker liner, with downsizing of the modular thirty-two-millimeter head of the femoral component to a twenty-eight-millimeter head. In the other two patients who had a fracture of the polyethylene liner, this replacement was not possible because the thirty-two-millimeter head of the monoblock femoral component did not allow for adequate thickness of the polyethylene liner. In these two patients, the entire cup was revised. All three patients had a radiographically successful result and no pain when they were seen at six months (one patient) or more than two years (two patients) after the revision.
One patient needed a revision because of aseptic loosening of the femoral stem. As its typical with failures of osseointegration, this patient had pain in the thigh and radiographic signs of an unstable, undersized stem one year postoperatively. The femoral stem was revised to a larger, longer, fully coated anatomic medullary locking stem at 2.3 years. Seven years after the revision, the patient had clear radiographic stability and no pain.
One hip was revised in the third postoperative year because of recurrent posterior dislocations. A lipped polyethylene liner was inserted to produce more anteversion, and a longer, modular femoral head was inserted to try to increase soft-tissue tension. After a five-year interval, the hip dislocated twice, with the most recent dislocation necessitating open reduction. The patient was managed with a brace and had not had a dislocation as of the most recent follow-up evaluation (one year postoperatively).
One patient who had a history of infection and an excision arthroplasty of the contralateral hip had symptoms and signs of septic failure in the eleventh postoperative year. She was managed at another institution with a successful two-stage revision with antibiotic-impregnated cement for fixation of the stem. Despite the presence of pus in the joint and the formation of an abscess, no organisms grew on culture.
The seventh reoperation was performed because an avulsion fracture had occurred through an osteolytic area in the lesser trochanter in the ninth postoperative year. Despite an asymptomatic osteolytic area in the superior aspect of the acetabulum that had not been noted on preoperative radiographs, the cup was not loose. At the reoperation, the bone-ingrown nonmodular acetabular component was revised because of polyethylene wear and the superior lytic area was packed with cancellous allograft. Four years after the revision, the patient had mild, occasional pain in the thigh and he was satisfied with the procedure; radiographically, he had stable components and no signs of recurrent osteolysis.
This study demonstrates that total hip arthroplasty without cement can be considered for patients who are sixty-five years of age or older because it predictably relieves pain, restores function, provides long-term stability of the implant, and is associated with few complications.
We found few published long-term studies with which to compare our results. We know of only one report that described the outcomes in a similar unselected, consecutive series of patients who were sixty-five years of age or older21. In that study, 547 Charnley low-friction prostheses fixed with so-called first-generation cementing techniques had an overall ten-year survival rate of 0.94 compared with 0.95 in the current series. Additionally, the authors reported that, in all 1041 hips that were studied, the overall rates of failure (revision or radiographic loosening) were 12 percent for the femoral components and 27 percent for the acetabular components. The rates of failure in our study were 2 percent (three of 142) for the femoral components and 6 percent (nine of 142) for the acetabular components.
Kobayashi et al.23 compared the results of the use of components fixed with so-called second-generation cementing techniques in patients who were fifty years of age or younger (fifty-three hips) with those in patients who were more than fifty years of age (273 hips). In the older group, the overall rates of failure of the femoral and acetabular components at sixteen years were 8 and 23 percent, respectively. Considering that this study involved different age criteria and longer follow-up, the rates of 2 and 6 percent in the current series compare favorably.
In recent reports of so-called hybrid fixation (fixation of the femoral component with cement and the acetabular component with cement) in predominantly older patients, rates of revision in series ranging from 125 to 153 hips were 1 to 7 percent4,8,17,30 at similar follow-up intervals. However, patients who were younger than sixty-five years of age were included in those series, so comparisons with the current series are less valid.
The accuracy of evaluating pain largely depends on the questions that the patient is asked and on the level of pain that the patient considers to be a problem. We found that defining the location of the pain was also important. When this was done, we found that a surprising number of patients had extra-articular or trochanteric pain.
By any estimate, discomfort that limits function is important, and the 3 percent prevalence of such pain in the thigh in this series is comparable with the 3 percent rate (three of eighty-eight) reported by Oishi et al.33, who used hybrid fixation with a so-called third-generation cementing technique. Similarly, in a recent study by Barrack and Paprosky3, no difference was found in the prevalence or the severity of pain in the thigh between matched groups of older patients (mean age, sixty-four years) in whom an extensively porous-coated stem or a stem implanted with a third-generation cementing technique had been used.
Other potential concerns with the use of femoral components fixed without cement in older patients include loss of stability of the component, failure of osseous ingrowth to occur, and stress-shielding when osseous ingrowth to occur, and stress-shielding when osseous ingrowth does occur. In the current series, the fixation failed only three times (2 percent): twice because of an undersized component and once because of infection. Femoral stability did not deteriorate on long-term follow-up. In fact, the concerns about the predictability and the durability of fixation of other, less extensively coated femoral components fixed without cement11,16,27 have not been supported by our overall experience with this component design14 or in the present consecutive series of older patients.
Stress-shielding is common in patients who are sixty-five years of age or older, and it can be pronounced28, as it was in this series. By their nature, extensively coated cobalt-chromium femoral implants produce resorptive changes in the proximal aspect of the femur because diaphyseal stability is the goal of their use. However, these reactive changes did not adversely affect the clinical outcome or radiographic stability or predispose the distal aspect of the femur to osteolysis in our series. These results agree with those in our previous study7, which specifically addressed the long-term clinical consequences of stress-shielding in our overall series with a follow-up period of as long as eleven years. Therefore, the importance of stress-shielding is uncertain.
Although osteolysis remains an issue in arthroplasty performed without cement, it has been observed less frequently in patients who are sixty-five years of age or older. The problem of frequent osteolysis of the distal aspect of the femur, as described in some series of proximally, noncircumferentially coated designs11,16,24,27,29, did not occur in our series. The appearance of osteolysis in only the periarticular zones in our patients led to the conclusion that circumferential porous coating can protect against osteolysis of the distal aspect of the femur. Still, although osteolysis was infrequent (four [3 percent] of 142 hips) in our series, it was a concern because it produced a fracture of the lesser trochanter and it affected the pelvis.
There was only one iatrogenic fracture of the femur in our series. It was easily treated, and the hip had a successful union and good clinical and radiographic results. No other periprosthetic femoral fractures were encountered in our study.
Proponents of fixation of femoral components with cement in older patients have pointed out that durability is not an important issue in this group of patients and that the focus should be placed on the difference in cost20. Two recent studies have suggested that, when all costs are considered, the differences between the use of components fixed with cement and the use of components fixed without cement may be negligible1,25.
In our experience, failure of the acetabular component, not of the femoral component, was a problem; it resulted in four of the seven reoperations. In these four reoperations, design issues were key factors, as evidenced by the three fractures of the polyethylene liner that occurred when the Triloc cup and the ACS liner were used. The thickness of the polyethylene liner in this design was inadequate. For example, the polyethylene liner for a shell with a fifty to fifty-six-millimeter outer diameter, with a thirty-two-millimeter femoral head, had a thickness of 4.7 millimeters at the dome and even less at the periphery. The polyethylene liner for a fifty-eight to sixty-four-millimeter shell, also with a thirty-two-millimeter femoral head, had a thickness of 6.9 millimeters at the dome and 3.6 millimeters at the periphery. Our experience is mirrored by that of others who have shown a high rate of failure of this liner5. As a result, it is no longer used.
Even in our series of older and presumably lower-demand patients, five additional hips showed evidence of at least two millimeters of wear of the polyethylene liner on two-dimensional analysis, a method that obviously underestimates volumetric wear when compared with more recent, three-dimensional analysis. However, none of these patients had osteolysis or loosening of the component. With only three cases of osteolysis and three radiographically loose cups in the group of 152 hips that did not have a reoperation, it is difficult to relate lysis or loosening to variables such as the thickness of the liner. Similarly, because none of the patients who had a reoperation had a loose acetabular component despite catastrophic failure of the liner, relating wear or thickness of the liner to loosening is also difficult. The failure of the two femoral components was not related to the liner or to the generation of debris. One was an early failure of integration and the second, a late septic failure.
However, we recognize from our overall experience that the generation of debris and its resulting osteolysis remain unresolved issues. We currently favor twenty-eight millimeters as a standard size for the femoral head, provided that a polyethylene liner that is at least eight millimeters thick can be used. If not, a twenty-six-millimeter (or smaller) head should be employed. We consider using a thirty-two-millimeter head in older, lower-demand patients if the larger-diameter head improves stability of the hip and the thickness of the polyethylene liner is adequate.
The results of this series confirm that implants fixed without cement should be considered a viable and effective option for total hip arthroplasty in patients who are sixty-five years of age or older. The present consecutive series of unselected older patients provides a valid group for comparison with subsequent studies to clarify what role, if any, age should play in prosthetic selection.