Of the 223 consecutive primary hip arthroplasties, 177 (79 per cent) were done because of osteoarthrosis; eighteen (8 per cent), because of rheumatoid arthritis; eighteen (8 per cent), because of osteonecrosis; eight (4 per cent), because of arthritis secondary to dysplasia of the hip; and two (1 per cent), because of post-traumatic osteoarthrosis. The mean age of the patients at the time of the operation was fifty-five years (range, sixteen to eighty-seven years). There were 108 male patients and 107 female patients. The mean preoperative weight was 74.5 kilograms (range, 34.1 to 127.3 kilograms).
The femoral components were all of the same cast cobalt-chromium design. The size of the stem was 9.0 millimeters in six hips (3 per cent), 10.5 millimeters in thirty-seven (17 per cent), 12.0 millimeters in twenty-nine (13 per cent), 13.5 millimeters in thirty-eight (17 per cent), 15.0 millimeters in seventy-one (32 per cent), 16.5 millimeters in thirty-one (14 per cent), and 18.0 millimeters in eleven (5 per cent).
The acetabular components were also made of cobalt-chromium and had the same type of porous surface as the stems. Each cup had three spikes for additional fixation. The thickness of the metal shell was 3.5 millimeters. The polyethylene articular surface was press-fit into the metal shell at the factory. Fifty-seven acetabular shells (26 per cent) had an outside diameter of less than fifty-two millimeters. Because of the 3.5-millimeter thickness of the shell, the polyethylene in these components was less than six millimeters thick. Eighty-three shells (37 per cent) had an outside diameter of fifty-two to fifty-four millimeters and contained polyethylene with a thickness of six to eight millimeters, and an additional eighty-three (37 per cent) had an outside diameter of more than fifty-four millimeters and contained polyethylene that was greater than eight millimeters thick.
The operations were done through a posterior approach. The acetabulum was prepared with so-called cheese-grater reamers and was fitted with an acetabular component that had the same outside diameter as the last reamer. Forceful impaction of the cup was needed to sink the three fixation spikes into the bone. These spikes provided the initial fixation of the cup.
For placement of the femoral component, a large pilot hole was made in the piriformis fossa to ensure correct alignment of the straight intramedullary drills used to prepare the femoral canal. Our goal was to use a femoral prosthesis that was large enough to contact both the medial and the lateral endosteal cortex of the femoral canal at the femoral isthmus. The canal was reamed to the same size as the femoral component. We graded the adequacy of the fill of the canal as seen on the anteroposterior radiographs of the pelvis that were made immediately postoperatively5,6. The radiographs showed the femoral stem to contact both endosteal cortices at the same level of the isthmus and, therefore, to be of adequate size in 178 hips (80 per cent). In the remaining forty-five hips (20 per cent), the stem appeared to be undersized.
The annual postoperative evaluations included administration of a questionnaire, a physical examination, and an analysis of four standard radiographs. One hundred and sixty-seven patients (174 hips) completed a questionnaire at least ten years postoperatively; of these patients, 131 (137 hips) had at least ten years of radiographic follow-up. The questionnaire included twenty-one items that fell into two categories: questions regarding over-all function and questions specific to the hip.
Four postoperative radiographs were used to determine the adequacy of the fixation of the prosthesis and the presence of osteolysis. The femoral component was considered to be either stably fixed by bone or fibrous tissue or not stably fixed8. Components that had not migrated but lacked definite signs of bone ingrowth were considered to be stably fixed by fibrous tissue. If the serial radiographs demonstrated a change in the position of the femoral component (that is, subsidence of two millimeters or more or varus or valgus tilting), the component was considered unstable.
The stability of the acetabular component was also assessed. Components that had tilted more than 5 degrees or had migrated two millimeters or more were considered unstable. A circumferential radiolucent line and shedding of beads from the prosthetic surface were also considered signs of instability.
Areas of localized loss of trabecular bone or localized cortical erosion were considered signs of osteolysis. Because the osteolytic lesions had discrete borders, their size could be measured by multiplying their length by their width on an anteroposterior radiograph. To correct for magnification, the known size of the femoral head was divided by the size measured on the radiograph. The resulting ratio was then multiplied by the length and width of the lesion before the size of the lesion was calculated. The osteolytic lesions were also characterized according to their location in the femur and the pelvis.
After each evaluation, all data were entered into a database. We then compared the data from the preoperative physical examination and questionnaire with those from the most recent postoperative physical examination and questionnaire. We also compared the serial radiographs that were made at each follow-up visit. These data were analyzed with the SPSS statistical package (SPSS, Chicago, Illinois). Continuous variables were tested for significance with a Student t test, and discrete variables were analyzed with a chi-square test. Survival of the components was calculated with use of the life-table method16.
Patients Who Were Lost to Follow-up
Twenty-one patients (twenty-two hips) were lost to follow-up less than ten years postoperatively. The mean age of these patients at the time of the operation was fifty-four years (range, twenty-three to seventy-eight years), and the mean duration of follow-up was 3.4 years (range, zero to nine years). Eighteen patients (seventeen hips) were followed for a minimum of one year; eleven patients (eleven hips), for a minimum of two years; and ten patients (ten hips), for a minimum of five years. None of these patients had a reoperation before they were lost to follow-up.
At the most recent evaluation, nine of the eleven patients (eleven hips) who were followed for at least two years reported that they had no pain. One patient had pain in the contralateral hip and in the back that limited activity; this patient used a walker. Another patient had pain in the buttock on the side of the involved hip but walked normally, without a limp; the cause of the pain was not determined. No patient had pain in the thigh. Ten of the eleven patients did not limp or use an assistive device for walking. The most recent radiographs revealed that all eleven femoral components had ingrowth of bone and all eleven acetabular components were stable. No hip had evidence of osteolysis.
Patients Who Died before the Ten-Year Evaluation
Twenty-seven patients (twenty-seven hips) died before the ten-year evaluation. None of the deaths were related to the hip arthroplasty, and none of the patients had had a reoperation. The mean age of the patients at the time of the operation was sixty-seven years (range, thirty-eight to eighty-seven years). The mean interval between the operation and death was 6.7 years (range, one to eleven years). The mean duration of follow-up was 4.5 years (range, zero to nine years). Two patients had less than one year of follow-up, and fifteen had at least five years. Twenty-two patients were followed for a minimum of two years. Twenty of the twenty-two patients reported that they had no pain at the last follow-up evaluation. One patient had pain in the contralateral hip (not included in the study) that was attributed to a loose acetabular component that had been inserted with cement. Another patient had pain, with no known cause, directly over the symphysis pubis. At the last evaluation, nineteen of the twenty-two patients used no assistive device for walking and had no limp. Three patients used crutches or a walker; two of the three were chronically ill, and both died of causes not related to the hip within a year after the last evaluation. The third patient was the patient who had pain in the contralateral hip. At the last radiographic evaluation, eighteen of the twenty-two stems appeared to have ingrowth of bone, and four appeared to be encapsulated by fibrous tissue but stable. All twenty-two cups appeared stable. One patient (one hip) had evidence of an osteolytic cyst in the greater trochanter on the seven-year postoperative radiograph. This patient died in the eighth postoperative year.
Patients Who Were Followed for a Minimum of Ten Years
Reoperations
One hundred and sixty-seven patients (174 hips) were evaluated at least ten years after the index arthroplasty. Twenty of these hips had an exchange of a component or part of a component. Three exchanges were done because of symptomatic loosening of the femoral component. The failure of bone ingrowth that eventually resulted in the loosening of the stem was first diagnosed one year postoperatively in two hips and two years postoperatively in one hip. These three hips did not become symptomatic enough to need a reoperation until 6.5 years postoperatively (one hip) and ten years postoperatively (two hips). The ages of these patients at the time of the initial operation were thirty-two, sixty-five, and thirty-eight years. The indication for the reoperation in two of the patients was pain in the thigh. One of the two patients had only a revision of the stem. The other patient had a revision of the stem and an exchange of the polyethylene liner within a well fixed acetabular shell. The third patient who had a reoperation because of loosening had an infection. Both components were removed. One year later, the femoral component was replaced with use of antibiotic-impregnated cement for fixation, and another porous-coated acetabular component was inserted. At the latest follow-up evaluation, radiographs of two of these three patients (the one who had an infection and the one who had a revision of the stem only) showed the new components to be stable two and four years after the revision, respectively. The third patient had the reoperation less than one year before the time of the review. The two patients who were evaluated more than two years after the reoperation were satisfied with the result; the hip functioned better than it had before the index arthroplasty. These three stems are the only ones that were revised in the entire series. All three were easily removed because they were loose. We did not have to remove a well fixed stem in this series, and no stem fractured.
Three acetabular components were revised because of recurrent dislocations. The intervals between the arthroplasty and the reoperation for these patients were six months, four years, and five years. At the time of the index arthroplasty, these patients were forty-one, fifty-nine, and twenty-five years old. In two of the hips, a well fixed porous-coated shell was removed and was replaced with a slightly larger one in a different orientation designed to prevent dislocation. In the third hip, the original acetabular component could not be anteverted because of a post-traumatic defect of the posterior wall. At the time of the reoperation, a posterior allograft was needed to obtain adequate acetabular anteversion. None of the three hips redislocated. At the latest follow-up evaluation, two of the patients were satisfied with the result of the reoperation and believed that the function of the hip was better than it had been before the index arthroplasty. The third patient, who had the reoperation in the fourth postoperative year, was dissatisfied with the over-all treatment. He was seventy-two years old at the latest follow-up visit and did not limp, but he used a cane. The ten-year postoperative radiographs of all three patients showed that the new acetabular components were stable and the femoral components had ingrowth of bone.
Four patients had symptomatic loosening of the acetabular component, and the cup was revised eight, nine, ten, and twelve years after the index arthroplasty. Postoperative radiographs showed radiolucent lines at the bone-implant interface of all four cups, suggesting that the cup was at least partially encapsulated by fibrous tissue. Three cups appeared stable on the radiographs and caused no symptoms until the year of the reoperation. The fourth cup was determined to be unstable in the second year after the index arthroplasty but was not revised until the eighth year. The ages of these four patients at the time of the index arthroplasty were sixty, twenty-nine, forty-nine, and fifty-five years.
At the time of the revision of the cup, all four patients reported pain with weight-bearing. In three, the pain was in the groin and in one, in the buttocks. Three of these patients also had osteolytic lesions of the acetabulum. In one of the four patients, the acetabular component was revised after the patient fell and sustained an associated acetabular fracture. This patient needed structural bone-grafting of the acetabulum. In another, migration of the cup and osteolysis damaged the acetabular cavity and superior rim. A bilobed acetabular component (Johnson and Johnson, Stamford, Connecticut) was used for this reoperation. The other two patients had only slight damage to the acetabular bone, and the revision was done with use of an acetabular component similar to the one used for the index arthroplasty. All four patients had the revision more than one year before the latest follow-up evaluation. All four were satisfied with the result and reported that the function of the hip was still better than it had been before the index arthroplasty. Two of the patients were employed at the job that they had had before the revision. The other two patients were elderly and were not working. All four femoral components were stable.
Ten patients had a reoperation because we believed that the femoral head was about to penetrate the polyethylene liner, which would have damaged the femoral head and the acetabular shell. The mean age of these patients at the time of the index arthroplasty was 46.8 years (range, twenty-five to sixty-five years). Nine of the patients were asymptomatic at the time of the reoperation. None used a walking aid or had a limp before the revision. Eight of the patients had periarticular osteolysis in both the femur and the pelvis. One had an isolated periarticular femoral osteolytic lesion. In two of the nine patients who had femoral osteolytic lesions, there was also evidence of a healed trochanteric fracture through an osteolytic cyst. In eight patients, the revision consisted of only an exchange of the polyethylene liner within the cup; in all eight, the acetabular shell was well fixed and the cup was large enough to accommodate a new polyethylene liner with a thickness of at least six millimeters. In the other two patients, it was necessary to remove a well fixed porous-coated acetabular shell because it was too small to accommodate a new polyethylene liner of adequate thickness.
Three of the eight patients who had an exchange of the polyethylene liner sustained a postoperative dislocation. The ages of these patients were twenty-five, forty-one, and fifty-nine years. Two of the three did not have recurrent dislocations; both were very satisfied with the over-all result of the treatment at the latest follow-up visit and stated that they had much better function than they had had before the index arthroplasty. The third patient, who had had bilateral total hip arthroplasty, continued to have dislocations of the hip and was therefore unhappy with the result of the reoperation. He was employed full time as a high-school teacher; before the arthroplasty, he had not been working.
Of the ten patients who had a reoperation because of wear, nine were satisfied with the result of the treatment and believed that the hip was functioning better than it had been before the index arthroplasty. All ten patients were pain-free and had stable femoral and acetabular components.
The mean age of the twenty patients who had a reoperation was less than that of those who did not have a reoperation (forty-six compared with fifty-four years; p < 0.01). Of the patients who had a reoperation, eleven (55 per cent) were less than fifty years old at the time of the index procedure, seven (35 per cent) were between fifty and sixty-five years old, and two (10 per cent) were more than sixty-five years old.
The preoperative level of activity was also associated with reoperation (p < 0.01). Ninety per cent (eighteen) of the twenty patients who had a reoperation had been able to walk more than three blocks without support before the index arthroplasty, compared with only 76 per cent (111) of the 147 who did not have a reoperation.
The number of reoperations increased with time. During the first eight postoperative years, the incidence of reoperation was 3 per cent (five hips); by the end of the tenth year, it was 6 per cent (eleven hips); and by the end of the twelfth year, it was 14 per cent (twenty hips).
The twelve-year probability of survival of the acetabular component (cup only) was 0.92 ± 0.03 (mean and standard error), and that of the femoral component was 0.97 ± 0.02. Because of the many exchanges of the polyethylene liner, the cumulative probability of survival of all parts of the hip prosthesis (stem, cup, and polyethylene liner) was 0.85 ± 0.03.
Our survivorship calculations did not include five patients for whom complications had not necessitated removal of the component. One patient was rehospitalized because of grade-4 heterotopic ossification1, and two were readmitted because of a postoperative dislocation, which was reduced under general anesthesia. Two other patients had complications at the time of the index arthroplasty that necessitated an extended hospital stay. One of them sustained a fracture of the femoral shaft during the index arthroplasty; the fracture was treated with application of a plate and insertion, without cement, of a femoral component that was five centimeters shorter than the one used in the index arthroplasty. The interval until the patient was allowed to bear full weight was also increased, but the end result was satisfactory. The other patient sustained a non-fatal pulmonary embolism during hospitalization. There were no acute infections in this group of patients. Grade-2 heterotopic ossification1 was present in twenty-three (17 per cent) of the 137 hips that were evaluated radiographically at least ten years postoperatively, and grade-3 ossification was seen in seven (5 per cent). The ossification did not cause functional limitations, and none of these hips needed a reoperation.
Ten-Year Clinical Results
One hundred and forty-seven patients (154 hips) who did not have an exchange of a component or part of a component completed a questionnaire at least ten years postoperatively. The mean age of these patients at the time of the index arthroplasty was fifty-four years (range, sixteen to seventy-nine years).
Seventy-six patients (52 per cent) routinely walked more than six blocks, thirty-one (21 per cent) could walk more than three to six blocks, and twenty-three (16 per cent) could walk two or three blocks. Thirteen patients (9 per cent) stated that they usually walked only indoors. Four patients (3 per cent) were confined to a wheelchair. One hundred and eight patients (73 per cent) were able to walk farther and better than they could before the operation.
One hundred and twenty-one patients (82 per cent) routinely walked without support, eighteen (12 per cent) always used a cane, six (4 per cent) always used crutches or a walker, and two (1 per cent) could not walk. One hundred and twelve patients (76 per cent) needed less support ten years postoperatively than they had needed preoperatively, twenty-three (16 per cent) needed the same amount of support, and twelve (8 per cent) needed more support to walk. Nineteen of the twenty-three patients who had no change in the amount of support that they used did not need any support preoperatively or postoperatively.
One hundred and thirty patients (88 per cent) stated that they had no difficulty or only slight difficulty in putting on their shoes and socks or stockings. Seventy-four patients (50 per cent) were able to ascend stairs normally (without using the handrail), and forty-seven (32 per cent) usually used the handrail. Eighteen patients (12 per cent) ascended and descended stairs by placing both feet on the same step. Eight patients (5 per cent) could no longer climb stairs.
One hundred and twenty-eight (87 per cent) of the 147 patients had no pain or mild pain that did not limit activity, fifteen (10 per cent) had pain that occasionally limited activity, and four (3 per cent) had pain that always limited activity. None of these patients rated the pain as being more severe than it had been preoperatively. Six patients (4 per cent) classified the pain as being of the same severity as it had been preoperatively, and 141 (96 per cent) rated the pain as being less severe than it had been preoperatively.
When asked specifically about limitations caused by the hip, seventy-eight patients (53 per cent) stated that they had no limitations; forty-six (31 per cent), that they had only slight limitations; eleven (7 per cent), that they had moderate limitations; and five (3 per cent), that they had severe limitations. Seven patients (5 per cent) did not respond to the question.
When asked about pain in the hip, fifteen patients (sixteen hips; 10 per cent) stated that the pain sometimes limited their activity. Often, the pain was in more than one area. Six patients had pain in the thigh that limited activity, and nine had pain over the greater trochanter. Four patients had pain in the buttocks, and three had pain in the groin. Eleven of the fifteen patients were satisfied with the result of the operation and believed that, although the pain sometimes limited their activity, it was of less magnitude than it had been before the operation.
Four patients who had pain in the hip that occasionally limited activity were dissatisfied with the over-all result of the treatment. All four were elderly. One, a seventy-five-year-old woman who had a stem that was encapsulated by fibrous tissue, had pain in the thigh that limited her activity. She did not become dissatisfied until the contralateral hip, in which the same type of prosthesis had been used, underwent osseointegration and did not cause similar pain. Another patient, an eighty-year-old woman, was dissatisfied because the hip replacement, done for a non-union of a fracture of the femoral neck, had not improved her walking ability. At the same time that she had sustained a fracture of the hip, she had also sustained a fracture of the ankle. We attributed her limited walking ability to persistent pain in the ankle. A third patient was dissatisfied because of a loose acetabular component. She was offered a reoperation, but she declined it. The fourth patient who was dissatisfied (discussed in the section on revision of the cup) had recurrent dislocations and eventually needed an exchange of the cup. Although the hip did not dislocate after this exchange, the patient remained apprehensive and dissatisfied.
Only thirteen patients (thirteen hips; 8 per cent) reported pain in the thigh. In seven patients, the pain was mild or did not limit their activity. The remaining six patients believed that the pain limited their activity, but only two were dissatisfied with the result of the operation. We were unable to establish any relationship between pain that limited activity and the size or type of fixation of the stem or the patient's diagnosis, age, or gender in this small group.
One hundred and thirty-eight (94 per cent) of the 147 patients responded that they were satisfied with the result of the arthroplasty. These same patients reported decreased pain and improved function of the hip. One hundred and thirty patients (88 per cent) believed that the operation had decreased their need for pain medication. The twenty patients who had had a revision procedure and the 147 who had not had a revision responded similarly to the questions about these matters.
Ten-Year Radiographic Evaluation
A total of seven acetabular components in this study were radiographically loose. Four were revised, and three were loose at the ten-year radiographic evaluation. The mean interval between the index operation and the diagnosis of loosening was 7.8 years (range, two to eleven years). Six of the seven loose cups were associated with osteolytic lesions.
Of the 117 femoral components that were not revised and for which a radiograph had been made at ten years, 105 (90 per cent) had ingrowth of bone, ten (9 per cent) had a stable fibrous interface, and two (2 per cent) were definitely loose. With the addition of the three revised femoral components, the total number of loose stems in the current study was five. All five stems were undersized. The diagnosis of loosening was made at a mean of 2.4 years (range, one to six years) postoperatively. No stem that had radiographic evidence of optimum fixation at two years subsequently became loose. One femoral component was rated as being stably fixed by fibrous tissue until the sixth postoperative year, when it migrated. The other four femoral components were diagnosed as loose by the time of the two-year evaluation.
Osteolytic lesions were seen in fifty-four hips (39 per cent) for which there were ten-year radiographs. These lesions were larger than 1.5 square centimeters in twenty-six hips (19 per cent). Osteolysis was associated with a younger age (mean age, forty-seven years for the patients who had osteolysis as compared with fifty-six years for those who did not), a smaller size of the stem, loosening of the cup, and reoperation (p < 0.01 for all four variables).
We have described the ten-year results of primary total hip arthroplasty performed with the anatomic medullary locking hip system, without use of cement, in a consecutive, non-selected series of patients. Although the implant systems were designed for insertion without cement in response to reported failures with use of cement, we were not initially sure that fixation without cement would be more successful.
Mulroy and Harris reported a rate of revision of the femoral component of less than 1 per cent at ten years with use of modern techniques of cementing in 195 patients (222 hips) who had a mean age of fifty-seven years. Sullivan et al. reported, after sixteen years of follow-up, a 3 per cent rate of revision of the Charnley stem in fifty-seven patients (seventy-four hips) who had a mean age of forty-two years at the time of the operation. We are not aware of any previous, comparable reports of the long-term rates of revision of femoral components inserted without cement. In the present series, in which only 80 per cent of the stems filled the canal (in 178 hips), the rate of revision of the femoral component was 1.5 per cent. There were only five loose stems, all of which were undersized. This resulted in a cumulative probability of survival of the femoral component of 0.97 ± 0.02 at twelve years. These rates compare favorably with those associated with modern techniques of cementing.
The rate of late revision due to aseptic loosening of acetabular components inserted with cement has ranged from 3 per cent (seven of 231 hips) to 6 per cent (eighteen of 322 hips)15,18,19. In the current series, there were seven loose acetabular components. Four were revised, and three were not because they were not causing symptoms. The rate of revision because of loosening of the acetabular component was 2 per cent (four of 201 hips for which the outcome was known). We found a relationship between loosening of the acetabular component and the development of osteolysis (p < 0.01). Although the over-all rate of revision of the acetabular component (the polyethylene liner or the shell) was 9 per cent (nineteen of 201 hips), only ten of the nineteen hips needed removal of the metal shell. The cumulative probability of survival of the metal shell was 0.92 ± 0.03 at twelve years, and the cumulative probability of survival of the acetabular component was 0.86 ± 0.03.
A major concern with regard to components that have been inserted without cement has been the difficulty of revision, especially when the femoral component is extensively porous-coated. We acknowledge that the removal of a well fixed, extensively coated component that was inserted without cement is difficult. However, we did not have to remove any well fixed femoral component in this consecutive series of patients. In the patient who had an infection, the femoral component had already been documented as loose, and none of the reoperations for recurrent dislocation necessitated revision of the femoral component. No patient had pain in the thigh that was severe enough to warrant revision of the femoral component. In the patients who had a revision of the femoral component, the diagnosis of loosening had been made in the first two years and the component was removed easily without substantial damage to the bone stock. At the time of the reoperation, in all but the patient who had an infection, a larger, completely porous-coated femoral component was successfully implanted.
Six well fixed acetabular components had to be removed. Circumferential exposure of the bone-prosthesis interface and the use of curved osteotomes allowed safe removal of all six components with minimum loss of bone stock. Five of the six hips had a successful revision with use of a larger porous-coated acetabular component. The four loose acetabular components were easier to remove than those that were well fixed. However, a loose acetabular component was sometimes associated with substantial loss of bone. Two of the four hips that had a reoperation because of loosening of the acetabular component had a large bone defect that necessitated use of a bulk allograft or insertion of an oblong-shaped acetabular component.
Another concern with regard to femoral components inserted without cement is pain in the thigh. We previously reported a prevalence of pain in the thigh of 14 per cent (forty-three of 307 patients) with use of the anatomic medullary locking stem10. Other authors have reported rates ranging from 16 per cent (twenty-four of 148 patients) to 51 per cent (thirty-six of seventy-one patients)2,10,13,14. Pain in the thigh was reported by thirteen patients (thirteen hips; 8 per cent) in the current study. This number was too small for us to determine if there was a relationship between the size of the stem and pain in the thigh. Pain in the thigh was the chief symptom for the five patients who had a loose stem. In two of these patients, the pain was not severe enough to warrant revision of the femoral component. The three patients in whom the stem was revised were pain-free at the most recent follow-up evaluation.
Ten of the twenty reoperations were done because we thought that the femoral head was about to wear through the polyethylene liner and we wanted to prevent fracture of the polyethylene and scratching of the femoral head. Guidelines for the timing of these exchanges of the polyethylene liner are difficult to describe. We follow, at yearly intervals, all patients who have excessive wear of the polyethylene. In some patients the wear appears to stabilize, whereas in others it continues. At the time of the reoperation, if the metal acetabular shell still has bone ingrowth after removal of the osteolytic granuloma, we simply replace the polyethylene liner. We have found this operation to be straightforward as long as the surgeon has ordered the appropriate polyethylene liner. Surgeons who plan this operation should warn their patients about the possibility of dislocation, which occurred after three of the ten reoperations done for this reason in the current study. We are still in the process of determining the indications and benefits of reoperation for wear of the polyethylene in asymptomatic patients.
We previously reported that the pattern of osteolysis differs according to whether the implant was inserted with or without cement21. Additionally, the location and the appearance of osteolytic lesions vary according to the design of the prosthesis. All of the lesions in the current study were periarticular—that is, they had an obvious connection to the articular surface—and they developed in fifty-four (39 per cent) of the 137 hips for which there were ten-year radiographs. Twenty-six hips (19 per cent) had periarticular lesions adjacent to the articular surface that were larger than 1.5 square centimeters. Schulte et al. found a 54 per cent prevalence of femoral osteolysis in zone 7 of Gruen et al. at twenty years in association with ninety-four Charnley prostheses inserted with cement. Mulroy and Harris reported a 7 per cent prevalence of femoral osteolysis at ten years in 222 hips, but they did not describe the location of the lesions. Kim and Kim, at a mean of five years, found a 38 per cent prevalence of femoral or acetabular osteolysis in association with eighty-two porous-coated anatomic systems (PCA; Howmedica, Rutherford, New Jersey) inserted without cement. In our patients, osteolysis was associated with late loosening of the acetabular component (p < 0.01). Eighteen hips (13 per cent) had femoral lesions that were larger than 1.5 square centimeters. All of the femoral lesions were in the greater trochanter or the calcar. None of these lesions caused loosening of the femoral component. However, two non-displaced trochanteric fractures occurred through osteolytic cysts in the greater trochanter. Both fractures healed with rest and non-operative treatment. We believe that these trochanteric osteolytic lesions were related to our operative technique. The large pilot hole that we created in the femur may have provided access for debris to reach the trabecular bone in the greater trochanter.
The young age of our patients was related to many aspects of the results. Many authors have reported that outcomes are worse in young patients than in older ones3,4,11,20. In our patients, a younger age was associated with osteolysis and reoperation (p < 0.01 for both). Of our thirty-seven patients who were sixty-five years old or more, only two (5 per cent) needed a reoperation. It should be kept in mind that 66 per cent of primary total hip arthroplasties done in the United States are in patients who are sixty-five years old or more12,17, whereas the mean age of our patients was fifty-five years.
Additionally, the consecutive, non-selected nature of our series must be considered. Today, surgeons commonly select implants on the basis of the patient's age, activity level, health status, and bone stock. As a consequence, the results for a particular implant or technique are potentially biased by this process of selection. Thus, we believe that the results of studies in which selection criteria are used should not be compared with the results of the current study.
Of the 147 patients who did not have a reoperation and who were followed for at least ten years, 138 (94 per cent) had increased function and less pain than before the arthroplasty. Because of the high degree of patient satisfaction and the relatively low rate of reoperation, we continue to use this prosthesis for most total hip replacements.