Two hundred and thirty-seven consecutive, unselected primary total hip replacements were performed, in 213 patients, with use of a Harris-Galante porous-coated (HGP) acetabular component with screw fixation (Figs. 1-A, 1-B, and 1-C) between January 1984 and December 1987 by the senior one of us (W. H. H.). The femoral component varied depending on the configuration of the proximal part of the femur, the femoral bone stock, and the age of the patient. The femoral components that were used during the period of the study included the Precoat femoral component, the CDH Precoat femoral component, the straight-stem CDH component, and the Harris-Galante femoral stem without cement (all manufactured by Zimmer).
Twenty-four patients (twenty-seven hips) died before a minimum duration of follow-up of seven years, five patients (five hips) were too ill to return for a detailed follow-up examination, four patients (four hips) refused to return for clinical and radiographic follow-up (but one of these patients had adequate follow-up for one side of a bilateral total hip replacement), two patients (two hips) were lost to follow-up, and two patients (two hips) refused radiographic follow-up but had adequate clinical follow-up. One patient, who had had a bilateral total hip replacement, had a resection arthroplasty of one hip at another institution because of a late infection 115 months after the index operation. Of the five patients who were too ill to return for follow-up, four had no pain in the hip and one had moderate intermittent pain in the hip (as reported by the family or the caregiver). The two patients who had only clinical follow-up had Harris hip scores11 of 78 and 93 points at 120 and ninety months postoperatively. At the time of the latest follow-up visit, at an average of forty-six months (range, two to 115 months) postoperatively, all of the acetabular and femoral components in these forty-one hips were well fixed and none had been revised because of aseptic loosening. One of the thirty-eight patients in this group had a reoperation because of a trochanteric nonunion at twenty-four months. One of the forty-one hips had a retroacetabular osteolytic lesion, measuring ten by ten millimeters, in zone 2 of DeLee and Charnley6.
The remaining 196 hips (177 patients) were available for radiographic and clinical review after a minimum duration of follow-up of eighty-four months and are the primary focus of the present study. The average duration of follow-up was 122 months (range, eighty-four to 155 months). Ninety-eight patients were men and seventy-nine were women. The average age of the patients at the time of the operation was fifty-nine years (range, twenty-three to eighty-seven years). The average weight of the patients at the time of the operation was seventy-one kilograms (range, forty-three to 109 kilograms), and the average height was 170 centimeters (range, 149 to 194 centimeters). The diagnosis was osteoarthritis for 118 hips, congenital dysplasia of the hip for thirty-seven, osteonecrosis for twelve, posttraumatic osteoarthritis for nine, rheumatoid arthritis for six, multiple epiphyseal dysplasia for three, failure of an endoprosthesis for three, Paget disease for three, childhood infection for two, nonunion of the femoral neck for two, and poliomyelitis for one. Sixty-three patients were classified as Charnley class A (unilateral involvement); 101, as Charnley class B (bilateral involvement); and thirteen, as Charnley class C (involvement of other joints or systemic problems limiting activity)5.
A posterolateral approach was used in all hips. Forty hips had a trochanteric osteotomy in order to improve exposure (thirty-one hips) or to optimize the tension in the abductor muscles (nine hips). The HGP-1 component differs from the HGP-2 component by virtue of the thicker shell of the HGP-2 component (5.6 millimeters compared with 3.7 millimeters for the smaller [forty to forty-eight-millimeter-diameter] sizes of the HGP-1 component and 4.7 millimeters for the larger [more than forty-eight-millimeter-diameter] sizes of the HGP-1 component), the increased number of tines with the larger sizes of the HGP-2 component, and the greater diameter of the cancellous-bone screws used with the HGP-2 component (6.5 millimeters compared with 5.1 millimeters for the screws used with the HGP-1 component). The polyethylene liner was held in place with tines located around the mouth of the acetabular shell. In 189 hips, the acetabular bed was prepared with hemispherical reamers in the so-called line-to-line fashion; that is, the nominal outer diameter of the acetabular component was equal to the nominal outer diameter of the last reamer that was used. In the remaining seven hips, the acetabular component was press-fit after underreaming of the acetabulum by two millimeters; that is, the nominal outer diameter of the acetabular component was two millimeters larger than the nominal outer diameter of the last reamer that was used. All acetabular components were fixed with screws. An average of three screws (range, two to six screws) were used for fixation.
Seventy-six hips had grafting at the sites of acetabular cystic defects with use of autogenous acetabular bone from reaming or morseled autogenous bone from the femoral head, and three hips had reconstruction of the acetabulum with use of a structural autogenous bone graft from the femoral head. All empty screw-holes in the acetabular shell were routinely packed with autogenous acetabular bone from reaming before placement of the acetabular liner. The diameter of the prosthetic femoral head was twenty-six millimeters in ninety-five hips, twenty-eight millimeters in sixty-one hips, twenty-two millimeters in thirty-three hips, and thirty-two millimeters in seven hips.
One hundred and one Precoat femoral stems, nineteen CDH Precoat femoral stems, and one straight-stem CDH component were inserted with use of contemporary cementing techniques21, which involved the use of a medullary plug, a cement gun, retrograde filling of the canal, lavage, and centrifugation and pressurization of the cement. Proximal and distal centralizers for the femoral stem were not used. Seventy-five Harris-Galante porous-coated femoral stems were inserted without cement. During the period of this study, the Harris-Galante stem generally was used for younger patients and patients who had good femoral bone stock. The average age of the patients in whom the stem was inserted with cement was sixty-three years (range, twenty-three to eighty-seven years), and the average age of the patients in whom the stem was inserted without cement was fifty-five years (range, thirty to seventy-four years).
The Harris hip score11 was determined preoperatively and at the time of the latest examination in order to evaluate the function of the hip. All patients were followed prospectively. At the time of the most recent follow-up, sixty-two patients (seventy-one hips) were evaluated with a personal interview, a physical examination, and a review of current radiographs. The remaining 115 patients (125 hips) were evaluated with use of a self-administered questionnaire and a review of current radiographs.
An anteroposterior radiograph of the hips and pelvis, a frog-leg lateral radiograph, and a true lateral radiograph of the treated hip were made for all patients preoperatively and at each follow-up examination. In addition, a Judet radiograph of the pelvis was made for 127 patients (143 hips) at the time of the latest examination. Preoperative, immediate postoperative, intermediate, and most recent radiographs were analyzed. The abduction angle of the acetabular component in relation to the interteardrop line was recorded. Acetabular coverage was estimated as the percentage of the shell that was covered with host bone as seen on the anteroposterior radiograph of the pelvis and hips. The position of the acetabular component was determined by measurement of the vertical and horizontal locations of the center of that component. The vertical position was defined as the perpendicular distance from the center of the acetabular component to the interteardrop line. The horizontal position was defined as the distance from the intersection of the interteardrop line and the perpendicular line through the femoral head to the inferior tip of the teardrop. The acetabular component was classified as migrated or loose if there was at least a four-millimeter change in the horizontal or vertical position of the center of the component compared with that seen on the immediate postoperative anteroposterior radiograph of the pelvis19. The acetabular interface on the anteroposterior radiograph was divided into three zones as described by DeLee and Charnley6. Specific attention was directed toward the interface between the acetabular component and the host bone (Figs. 1-B and 1-C). Regions in which the surface of the acetabular component was not in contact with bone on the immediate postoperative radiographs were classified as gaps, to distinguish them from radiolucent lines that appeared on subsequent radiographs in areas where no gaps had existed initially20. All gaps, radiolucent lines, and areas of osteolysis were measured and recorded. Acetabular components with a discontinuous radiolucent line in all three zones or a continuous radiolucent line around the component were not classified as radiographically loose unless the component had migrated. Polyethylene wear was determined by comparing the apparent minimum thickness of the polyethylene on the latest follow-up radiograph with that on the immediate postoperative radiograph, as described previously14. (The minimum thickness of the polyethylene was defined as the shortest distance between the outer edge of the femoral head and the outer edge of the acetabular component.) The difference between these two measurements, after correction for magnification with use of the known size of the femoral head, indicated the total amount of polyethylene wear. The average annual rate of wear then was determined by dividing the total amount of wear by the duration of follow-up (in years). These values represent a two-dimensional assessment of wear and probably are an underestimation of the true amount of wear. Heterotopic ossification was graded according to the system of Brooker et al.3.
Special attention was directed toward the identification of osteolytic lesions of the pelvis as seen on follow-up radiographs. The dimensions and location of osteolytic areas in the femur were assessed according to the zones described by Gruen et al.10. Any changes in the position of the femoral stem were noted. The stem was considered to be loose if it had subsided or changed position, if there was a fracture of the cement mantle, if there was a radiolucent line between the prosthesis and the cement (so-called debonding), or if the cement mantle itself had subsided. Any stem that had broken or bent was also considered loose.
Kaplan-Meier survivorship analysis13 was used to assess the life span of the acetabular component. All patients who had been lost to follow-up were included, and failure was defined as revision of the acetabular shell.
Eight (4 percent) of the 196 acetabular shells that were followed for at least eighty-four months (average, 122 months) were revised. Two well fixed acetabular components (1 percent) were revised, at fifty and seventy-two months after the primary total hip replacement, because of retroacetabular osteolysis. One of these components was in an extremely active, ninety-one-kilogram, sixty-seven-year-old man who exercised regularly (resulting in an estimated 4.5 million cycles of the hip per year), and the other was in a sixty-four-kilogram, thirty-seven-year-old man who bicycled eight kilometers per day. Three other well fixed acetabular components (2 percent) were revised, at forty-two, eighty-two, and 111 months after the primary total hip replacement, because of dissociation of the liner secondary to fractures of the tines around the periphery of the shell. Another three well fixed acetabular components (2 percent) were revised for various reasons at the time of revision of the femoral component. One of these components was revised, at seventy-three months, because of a fracture of a tine that occurred during exchange of the polyethylene liner at the time of revision of the femoral component. The second component was revised, at seventy-two months, because two of the three acetabular screws were grossly loose during the revision of the femoral component; although the acetabular component was stable at the time of the operation and did not have a continuous radiolucent zone surrounding the shell on the preoperative radiograph, the finding of the two loose screws led to concern that it had only fibrous ingrowth. In fact, once the component was removed, it became clear that the shell did have areas of osseous ingrowth. The third component was revised, at twenty months, because of metallosis and concern about the fact that one of the titanium acetabular screws was in contact with a stainless-steel screw that had been inserted during previous internal fixation of the acetabulum.
Eighteen femoral components were revised, at an average of sixty-eight months (range, twenty to 133 months) postoperatively, because of aseptic loosening. These eighteen components included fourteen (19 percent) of the seventy-five Harris-Galante stems, two of the nineteen CDH Precoat stems, one (1 percent) of the 101 Precoat stems, and the one straight CDH stem. In addition, one femoral component was revised at another institution because of a periprosthetic femoral fracture 131 months after the index operation, and one well fixed femoral component was revised at another institution because of recurrent dislocation forty-five months after the index operation. In eight hips, the acetabular liner was exchanged at the time of revision of the femoral component in order to allow the patient to start over with a fresh liner.
During the thirteen-year span of this study, no acetabular component migrated, was classified as radiographically loose, or was revised because of aseptic loosening. There were no complications associated with the insertion of the acetabular screws.
The average abduction angle of the 188 acetabular components that were in place at the time of the most recent follow-up was 34 degrees (range, 15 to 48 degrees). One hundred and eighty-three acetabular components were in anteversion, three were in a neutral position, and two were in retroversion. At least 90 percent of the acetabular component was covered with host bone in 166 hips, and at least 80 percent was covered in all 188 hips. There were no fractures of the acetabulum or the acetabular shell. One of 528 acetabular screws broke. There was no evidence of disruption or fragmentation of the titanium porous mesh of any cup. Four components had radiographic evidence of a fracture of a tine without dissociation of the polyethylene liner. The average rate of polyethylene wear was 0.10 millimeter per year (range, 0.00 to 0.41 millimeter per year). Heterotopic ossification was classified as grade 0 or 1 in 128 hips, grade 2 in twenty-four hips, grade 3 in twenty-eight hips, and grade 4 in eight hips.
One hundred and two hips had no gaps between the bone and the acetabular component on the immediate postoperative anteroposterior radiograph (Fig. 1-B). Fifty-five of these hips had no radiolucent lines at the bone-implant interface on the latest follow-up anteroposterior radiograph, whereas the other forty-seven had a total of seventy-three radiolucent lines. One of these lines was 2.0 millimeters wide, and the rest were 1.0 millimeter wide or less. Twenty-eight lines were in zone 1, twelve were in zone 2, and thirty-three were in zone 3. Three of these forty-seven hips had a discontinuous radiolucent line in all three zones, and one had a continuous radiolucent line that was 0.5 millimeter wide in all three zones. Eighty-six hips had at least one gap between the bone and the acetabular component on the initial postoperative radiograph. There were a total of 108 gaps in this group: fifty-five were in zone 1, eight were in zone 2, and forty-five were in zone 3. One gap was 2.0 millimeters wide, and the rest were 1.0 millimeter wide or less. Sixty-five of these gaps persisted unchanged on the latest follow-up radiograph, and the other forty-three had resolved. In addition, a total of thirty-four new radiolucent lines that were 1.0 millimeter wide or less were identified on the latest follow-up radiographs of twenty-five hips. Five of these eighty-six hips had a discontinuous radiolucent line in all three zones, and one had a continuous radiolucent line that was 0.5 millimeter wide in all three zones. Overall, eight hips had a discontinuous radiolucent line that was 1.0 millimeter wide or less in all three zones at the time of the most recent follow-up. A total of four screws in three hips were associated with radiolucent lines.
Anteroposterior radiographs (available for all hips) and Judet radiographs (available for 143 hips) revealed osteolytic lesions, ranging in size from ten by five millimeters to seventeen by fifteen millimeters, adjacent to eight (4 percent) of the 188 cups that were in place at the time of the most recent examination. One additional osteolytic lesion was identified during operative exploration at the time of revision of a femoral component. Thus, a total of nine (5 percent) of the 188 hips had evidence of osteolysis: four hips had a lesion in zone 1 only, one hip had a lesion in zone 2 only, two hips had a continuous lesion in zones 1 and 2, one hip had discontinuous lesions in zones 1 and 2, and one hip had discontinuous lesions in zones 1, 2, and 3. Three of the four lesions that involved zone 1 only were identified radiographically; the fourth was identified during revision of the femoral component and was treated with open bone-grafting. Two of the four lesions that involved zone 1 only were associated with acetabular screws. The hip that had osteolytic lesions in all three zones was treated, at 136 months, with curettage and bone-grafting around the well fixed acetabular component, exchange of the acetabular liner, and exchange of the femoral head.
Seven of the eight osteolytic lesions that were seen radiographically were identified on the anteroposterior radiograph alone. Only one was identified on the Judet radiograph alone. Therefore, only one additional instance of pelvic osteolysis was identified on the basis of the 143 Judet radiographs that were made. The chance that the lack of a Judet radiograph led to an error in the identification of osteolysis in the forty-five hips for which only an anteroposterior radiograph was made appears to be small.
The average Harris hip score for the 188 hips (169 patients) that did not have revision of the acetabular shell improved from 47 points (range, 22 to 71 points) preoperatively to 89 points (range, 35 to 100 points) at the time of the latest examination. One hundred and thirty-four hips (71 percent) had an excellent result (90 to 100 points); twenty-six (14 percent), a good result (80 to 89 points); nineteen (10 percent), a fair result (70 to 79 points); and nine (5 percent), a poor result (less than 70 points). All nine patients (nine hips) who had a poor result had other factors that contributed to the low Harris hip score. Specifically, three patients had severe degenerative disease of the lumbar spine, one had had a resection arthroplasty of the contralateral hip because of a late infection, one had had a failed revision of the femoral component, one had severe pain and a limp in association with a revision of the femoral component that had been performed at another institution because of a periprosthetic fracture of the femur, one had difficulties with balance and gait because of a systemic mitochondrial disease, one had severe hemiplegia after a stroke as well as bilateral osteoarthritis of the knee, and one had severe neuropathy secondary to adrenoleukodystrophy.
Kaplan-Meier survivorship analysis13, with failure defined as revision of the acetabular component, revealed a 96 percent chance of survival at 120 months (95 percent confidence interval, 0.92 to 1.0).
In the group of 196 hips that were followed for a minimum of eighty-four months, there were thirty-five reoperations and thirty-nine readmissions. The reoperations included twenty revisions of the femoral component (three of which were performed simultaneously with a revision of the acetabular component), five revisions of the acetabular component, five procedures performed for the excision of heterotopic bone, two bone-grafting procedures performed because of perforation of the femoral cortex, one bone-grafting procedure performed because of symptomatic localized thinning of the femoral cortex, one bone-grafting procedure performed because of retroacetabular osteolysis around a well fixed acetabular component, and one thrombectomy and repair of the femoral artery. Thirty-four of these thirty-five reoperations necessitated readmission to the hospital, and the other one was performed during the initial hospitalization. There were five additional readmissions: two patients were readmitted for reduction of a dislocation and application of a brace, one patient was readmitted twice for the treatment of recurrent cellulitis at the site of the operative wound, and one patient was readmitted for the nonoperative treatment of a nondisplaced periprosthetic fracture of the femur.
One of the 213 patients (237 hips) from the original cohort died within ninety days after the total hip replacement. This patient, an eighty-one-year-old man, died on the fifty-eighth postoperative day as a result of bronchopneumonia, which was unrelated to the operation on the hip. Although the patient had had moderately severe coronary artery disease and aortic stenosis preoperatively, the total hip replacement was performed for the treatment of severe pain secondary to a nonunion of the femoral neck.
The high prevalence of late aseptic loosening of cemented acetabular components has been well documented9,17,18,22-24,26, has not been reduced by contemporary cementing techniques15,17,18, and, according to many authors9,15,18,22,26, remains the major problem with total hip arthroplasty performed with cement. The frequency of this complication led to the development of acetabular components that were designed to be inserted without cement in an effort to improve long-term fixation. Nevertheless, the clinical and radiographic performance of such devices has varied greatly depending on the specific design features of the component that is used. For example, the threaded acetabular cup has been associated with unacceptably high rates of failure. Capello et al.4 reported that nine (17 percent) of fifty-four primary acetabular reconstructions performed with the Mecring device (Mecron, Berlin, Germany) failed at an average of twenty-six months postoperatively. Similarly, Fox et al.8 reported that sixteen (31 percent) of fifty-two primary acetabular reconstructions performed with a threaded titanium acetabular cup (T-TAP; Biomet, Warsaw, Indiana) failed at an average of six years postoperatively.
Heekin et al.12 reviewed the results of 100 primary total hip arthroplasties performed with the porous-coated anatomic prosthesis (PCA; Howmedica, Rutherford, New Jersey) and reported that 6 percent of the acetabular components had migrated by five years. A review of this same group of hips after nine to ten years of follow-up revealed that six acetabular components had been revised27. Specifically, two cups had been revised because of progressive osteolysis without loosening; two, because of progressive osteolysis with loosening; one, because of failure of the polyethylene liner; and one, because of malposition. Engh et al.7 reported on 223 consecutive primary total hip replacements performed with the anatomic medullary locking hip system (DePuy, Warsaw, Indiana) and a thirty-two-millimeter-diameter femoral head. In that series, the outcome was known for 201 hips and 174 hips had a minimum of ten years of follow-up. Those authors reported that seven (3 percent) of the 201 acetabular components were radiographically loose; four of these cups (2 percent overall) were revised, and the other three were not revised because they were not causing symptoms. An additional ten hips (5 percent) were revised because of excessive wear of the polyethylene liner. Zicat et al.28 reported that pelvic osteolysis was observed around thirteen (18 percent) of seventy-one Anatomic Medullary Locking (AML) cups (DePuy) after an average duration of follow-up of 8.8 years. Bono et al.2 described the short-term results associated with the Anatomic Medullary Locking Plus (AML+) cup used with the Acetabular Cup System (ACS) polyethylene liner (DePuy). At an average of 3.6 years postoperatively, eleven of the fifteen acetabular components that had failed and twenty-two (39 percent) of the fifty-seven acetabular components that had not failed were associated with pelvic osteolysis.
The intermediate-term clinical and radiographic results associated with Harris-Galante porous-coated acetabular components have been well documented in the literature. Latimer and Lachiewicz14, in a study of 136 primary total hip arthroplasties performed with the Harris-Galante-I porous-coated acetabular component, reported that no acetabular component had been revised for loosening, was radiographically loose, or had evidence of dissociation of the polyethylene liner after an average duration of follow-up of seven years. Two acetabular components (1 percent) were associated with asymptomatic osteolysis, which was treated with bone-grafting at the site of the lesion as well as exchange of the femoral head and the acetabular liner. Tompkins et al.25, in a recent review of the results of 173 primary total hip replacements performed with the Harris-Galante porous-coated acetabular component, reported that no cup had been revised because of loosening after an average duration of follow-up of 104 months. Two patients had had exchange of the acetabular liner because of excessive wear of the polyethylene, one patient had had bone-grafting because of massive retroacetabular osteolysis, and two patients had had revision of a stable acetabular component in association with revision of the femoral component. Radiographs, which were made for 165 hips at the time of the latest clinical examination, revealed that eight cups (5 percent) were possibly unstable. Six of these components had a radiolucent line that was 1.0 millimeter wide or less in at least four of five zones, and the other two had a complete radiolucent line in all zones. Importantly, although none of these components had migrated, all were classified as possibly unstable because the radiolucent lines probably represented fibrous ingrowth and a suboptimum interface.
In the present series of 196 Harris-Galante porous-coated acetabular components that were followed for an average of 122 months, no cup was revised because of aseptic loosening and no cup had radiographic evidence of migration. Eight cups (4 percent) were revised: three (2 percent) were revised because of dissociation of the liner in association with fractures of the tines, three (2 percent) were revised for various reasons at the time of revision of the femoral component, and two (1 percent) were revised because of retroacetabular pelvic osteolysis. Both of the patients who had a revision because of pelvic osteolysis were exercise enthusiasts. Of the 188 hips that did not have a revision of the cup (76 percent of which were evaluated with Judet radiographs as well as anteroposterior and true lateral radiographs), eight (4 percent) had minor osteolytic lesions of the pelvis and one (less than 1 percent) had an osteolytic lesion that necessitated bone-grafting. The present study and the two intermediate-term studies mentioned previously14,25 comprise 505 primary acetabular reconstructions that were performed with use of the Harris-Galante porous-coated acetabular component and screw fixation. The intermediate-term fixation of this component has been excellent, with none of the 505 acetabular shells having been revised because of aseptic loosening after five to thirteen years of follow-up. In addition, only one of the 505 components had radiographic evidence of migration25. That component, which was inserted into an acetabular bed that had been reconstructed with a bulk allograft from a femoral head, migrated more than two millimeters in the first two years and then stabilized without additional complications.
These data suggest that, after intermediate-term follow-up, the Harris-Galante porous-coated acetabular component continues to provide excellent fixation and excellent clinical results for most patients. Nevertheless, long-term follow-up is necessary in order to evaluate other potential mechanisms of failure of the acetabular component, including excessive polyethylene wear, dysfunction of the locking mechanism, dissociation of the liner, and pelvic osteolysis.
In response to the fractures of the tines that have been observed in association with the HGP-1 acetabular component, an improved version of this type of component, without tines, has been devised (Trilogy; Zimmer). This device has a split-ring locking mechanism. We continue to press-fit the acetabular component (that is, to insert an acetabular shell with a diameter that is two or three millimeters greater than that of the last reamer that was used) when the bone stock seems appropriate and to use some shells without screw-holes, but, for most primary procedures, we use a so-called cluster-type Trilogy shell with three holes and usually add two screws.