One hundred consecutive, nonselected patients (115 hips) with an age of fifty years or less who underwent total hip arthroplasty from June 1994 to December 1999, with use of cementless femoral and cementless acetabular components, were prospectively followed for a minimum ten-year interval. The Prodigy femoral component (DePuy, Warsaw, Indiana) was utilized in all hips. The component is (1) chromium cobalt material, (2) fully porous coated, (3) has a medial cutout to decrease flexural rigidity, and (4) has a polished tip distally to avoid tip-cortex contact. The femoral component has 10° of anteversion incorporated into the neck. It was developed to potentially decrease thigh pain and stress-shielding. The Prodigy femoral components were mated with forty-three Harris-Galante-I acetabular components with polyethylene that had been sterilized with gamma radiation in air (Zimmer, Warsaw, Indiana) or with seventy-two Duraloc acetabular components (DePuy) (either sector types with two holes or a multihole component) with polyethylene that had been sterilized with gas plasma (sixty-nine hips) or moderately cross-linked polyethylene (three hips) (Marathon; DePuy).
The average age of the patients at the time of the index total hip arthroplasty was 40.1 years (range, seventeen to fifty years) (Table I). There were sixty-three men and thirty-seven women. Ten years after the index total hip arthroplasty, seventy-three patients (eighty-two hips) were alive, seventeen patients (twenty-three hips) had died, and ten patients (ten hips) were lost to follow-up. Of the remaining seventy-three patients (eighty-two hips), sixty-eight patients (seventy-seven hips) had a follow-up radiograph made at a minimum of ten years (average, twelve years; range, ten to seventeen years). Three of the remaining five living patients (three hips) had minimum five-year radiographs; the other two patients (two hips) had only an early postoperative radiograph.
Preoperative evaluation included clinical examination as well as the use of a standard terminology questionnaire for the reporting of clinical and radiographic results9. The most common preoperative diagnosis was primary osteoarthritis in thirty-nine hips and osteonecrosis in thirty-five hips (Table I). All surgical procedures were performed by the senior author (J.J.C.) through a posterolateral approach. The femoral canal was underreamed by 0.5 mm to the diameter of the coating of the implant utilized. Intraoperative radiographs were made with a broach in place to ensure proper fit and positioning. The canal was then machined for a larger size if there was any concern, on the basis of the intraoperative radiograph, that the component would be undersized or was malpositioned. The cementless acetabular component was secured to the acetabulum with two or three screws. The postoperative protocol consisted of having the patient weight-bearing with crutches for six weeks and progressing to full weight-bearing as tolerated.
Clinical Evaluation
Patients were asked to return for follow-up every two years or sooner if symptoms warranted. The most recent follow-up evaluation was at least ten years after the index operation. Patients returned to the clinic for follow-up or, if they were unable to return, sent current radiographs for evaluation. Patients were evaluated in person or interviewed by telephone with use of a standard terminology questionnaire9, and completed Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC)10, Harris hip11, University of California, Los Angeles (UCLA)12-14, and Tegner13,15 activity level questionnaires.
Radiographic Analysis
Early postoperative and interval follow-up radiographs included anteroposterior projections of the pelvis that showed the tip of the femoral prosthesis and lateral projections of the femur that included the hip. Observation measurements were based on the anteroposterior radiographs from the early postoperative period and those at the time of the latest follow-up. Minimum ten-year follow-up radiographs were available for sixty-eight of the seventy-three living patients (93% follow-up rate; seventy-seven of eighty-two hips [94%]). Two of us (J.J.C. and R.K.T.) reviewed the radiographs with interpretation reported by consensus. Correction for magnification was completed by standardizing all measurements to the known size of the femoral head.
Femoral components were evaluated for bone ingrowth, stable fibrous fixation, or unstable fibrous fixation according to the criteria of Engh et al.16. Alignment was classified as neutral, varus, or valgus. Hips in varus7 and valgus alignment5 were not >5° past neutral. Femoral component subsidence was determined with use of the relationship of the top of the lesser trochanter to the medial aspect of the stem collar, defined as a decrease of ≥5 mm (with magnification considered) between the initial postoperative radiograph and those from the last follow-up evaluation17,18. Osteolysis was defined as any nonlinear radiolucency at the bone-prosthesis interface that was ≥5 mm2 according to the seven femoral zones defined by Gruen et al.19. Radiolucencies were also recorded according to the seven femoral zones of Gruen et al.19. Femoral component stress-shielding was defined with use of a modification of the criteria defined by Engh and Bobyn20. Mild stress-shielding was limited to the upper third of the implant. Moderate stress-shielding extended to the middle one-third, and severe stress-shielding extended below the middle one-third. The acetabular components were evaluated for bone-prosthesis radiolucencies and acetabular component migration according to the criteria of Massin et al., as in our previous reports21-23. The definition of acetabular osteolysis was the same as that of femoral osteolysis.
As part of the routine clinical follow-up, gross evidence of wear on radiographs was noted. In cases of gross wear, patients were followed closely (yearly) and instructed to return sooner if they experienced any symptoms. If patients had symptoms or if their radiographs showed imminent wear-through of the acetabular liner, revision arthroplasty was indicated. If at the time of revision surgery the acetabular component was stable and well positioned, the preference was to replace the liner, rather than revise the entire acetabular construct.
Activity Evaluation
The activity level of fifty-nine of the seventy-three living patients (sixty-eight of eighty-two hips; 83%) was assessed with use of a StepWatch Activity Monitor (SAM; StepWatch, Orthocare Innovations, Mountlake Terrace, Washington)24,25. Output data included steps per day, minutes that the device was worn per day, and average steps for days with activity. The number of steps per year was then calculated by multiplying average steps per day by 365.
Additionally, sixty-four of the seventy-three patients (seventy-two of eighty-two hips; 88%) performed a six-minute walk26. Individual performance on the six-minute walk test was compared with the predicted distance for each patient using reference formulas incorporating the patient’s age, height, and weight. The reference formulas were generated on the basis of a study of selected, healthy, age-matched subjects reported by Enright and Sherrill27.
Wear
Patients included in the wear analysis had a radiograph made at a minimum of ten years along with sequential radiographs. Standard anteroposterior radiographs of the pelvis centered on the symphysis pubis were used. All initial postoperative radiographs were made at the University of Iowa Hospitals and Clinics. If patients were not able to return for their most recent radiograph, current radiographs were sent on compact discs. Radiographs were digitized, and comparison of the initial and final postoperative linear head penetration, as determined by the digital edge-detection technique described by Shaver et al., was used as a surrogate for wear28. Volumetric wear was calculated as described by Kabo et al.29. Seventy-two of eighty-two hips (sixty-four of seventy-three patients; 88%) had wear analysis performed with use of the minimum ten-year radiograph. Hips used for wear analysis had a minimum ten-year radiographic follow-up, unless a revision was performed, in which case the wear was based on the radiograph made just prior to the revision surgery. Average radiographic follow-up for all hips that had wear analysis was 11.4 years (range, eight to fourteen years). All wear measurements were performed by two authors (R.K.T. and N.A.B.).
Cohort Comparison
The results for our cohort were compared with those of a cohort of young patients who had Charnley total hip arthroplasty performed with cement by a single surgeon at another hospital7,8. This cohort had been followed for a minimum of twenty-five years and was evaluated with use of similar criteria and end points for Kaplan-Meier analysis30. The original cemented cohort consisted of sixty-nine patients with ninety-three Charnley flatback polished femoral components mated with an ultra-high molecular weight polyethylene acetabular component that had been gamma-irradiated in air. All patients were under fifty years of age at the time of the index surgery, which was performed between 1970 and 1976 by a single surgeon. The Charnley femoral stems with 22-mm-diameter heads were inserted with cement that was hand-packed. The ten to fifteen-year follow-up data from those studies were used for comparison. Radiographic analysis for both the cemented total hip arthroplasty cohort, as well as the present cohort, was performed with use of the same criteria, with one evaluator (J.J.C.) reviewing all radiographs in both cohorts. See Table I for comparisons between the cementless group and the cemented group.
Statistical Methods
Data were analyzed by univariate and multivariate linear regressions with use of Stata software (StataCorp). Pearson correlation coefficients were calculated to assess the relationship between clinical and radiographic wear variables. Wilcoxon rank sum tests were used to evaluate differences between categorical variables (such as sex and acetabular cup type) and clinical and radiographic variables. The t test was used to analyze differences between normally distributed activity data, such as six-minute-walk distances.
Kaplan-Meier survivorship curves were generated for five end points. The log-rank test was used to compare the survivorship curves of our cohort with the cemented cohort. The Bonferroni correction was used for making multiple pairwise comparisons.
Source of Funding
The study was partially funded by the Bierbaum Resident Fund.
Reoperations
Of the original 115 hips, seventeen in patients with a minimum follow-up of ten years had a reoperation (see Appendix). Twelve reoperations were for polyethylene wear (see Appendix), three were for periprosthetic fracture, and two were for dislocation. Six of the twelve hips that had reoperation for wear had a Harris-Galante-I acetabular component with sizes ranging from 48 mm to 52 mm, mated to either a 22-mm or 26-mm head. The other six had a Duraloc Sector acetabular component with sizes ranging from 48 mm to 54 mm, and mated to either a 22-mm or 26-mm head. The average time to reoperation for these twelve patients was 10.7 years (range, eight to 12.3 years). All twelve had bone-ingrown acetabular shells. Six underwent cementation of a new liner into the shell and a femoral head exchange One patient with Down syndrome required removal of the ingrown shell with reinsertion of a cementless shell and constrained liner. Five liners were placed with use of the preexisting capturing mechanism and moderately cross-linked polyethylene. Three hips had a reoperation for periprosthetic femoral fracture, one with retention of the femoral component and plating of the femur and two with removal of the ingrown component and placement of a longer stem. Two patients underwent reoperation for recurrent dislocation, and both were treated with a femoral head and liner exchange with no further dislocations. No hip was revised for infection or aseptic loosening.
Clinical Results
The average Harris hip score was 45 points (range, 15 to 81 points) preoperatively and 80 points (range, 20 to 100 points) at the time of the final follow-up. The average WOMAC scores for pain, stiffness, and function (with a high score indicating high function and 100 as the best score) were 38.2 points (range, 0 to 95 points), 37 points (range, 0 to 87.5 points), and 41.6 points (range, 0 to 95.6 points), respectively, at the preoperative examination and 85.4 points (range, 25 to 100 points), 68.9 points (range, 0 to 100 points), and 75.5 points (range, 11.3 to 100 points) at the time of final follow-up. The self-reported mean UCLA score for the study group was 5.7 (range, 2 to 10). The distribution of UCLA scores between male and female patients was not significantly different (p > 0.05). The self-reported mean Tegner score for the study group was 2.8 (range, 0 to 6). The distribution of Tegner scores between male and female patients was not significantly different (p > 0.05).
Thigh pain was at least episodically present in fourteen hips (17%) in patients specifically queried for thigh pain at the time of final follow-up. The pain was mild in eleven patients, moderate in one patient, and severe in two patients. No patient claimed that the pain limited his or her activity.
Activity Results
At the time of the final follow-up, fifty-nine of the seventy-three patients (sixty-eight of eighty-two hips; 83%) had agreed to monitor their activity levels for up to three weeks using a StepWatch Activity Monitor and sixty-four of the seventy-three patients (seventy-two of eighty-two hips; 88%) had completed six-minute-walk tests. The study group walked a mean of 1.54 million steps per year (range, 77,000 to 3.26 million steps), with most (thirty-seven) of the fifty-nine patients walking between 1.0 million and 2.2 million steps per year. The mean six-minute-walk distance for the study group (sixty-four patients) was 323 m (range, eight to 503 m). When each patient was compared with his or her respective sex, age, and weight-adjusted predictive normative value, the observed distances were significantly lower (p = 0.0001) than the predicted values (mean, 519 m). Men (p = 0.0001) and women (p = 0.0001) did significantly poorer compared with the predicted, sex-matched, normative values, but there was no significant difference between the male and female groups (p > 0.05). Although it was a younger patient population, the group had individuals with substantial comorbidities. For example, in the sixty-four patients in the six-minute-walk group, twenty-five patients had no comorbidity, eleven patients had one comorbidity, and twenty-eight patients had two or more comorbidities. Regression analysis demonstrated that a greater number of comorbidities was predictive of a lower six-minute-walk distance (p = 0.0034).
Radiographic Results
A radiograph made at a minimum of ten years of follow-up was available for seventy-seven of the eighty-two hips in sixty-eight (93%) of seventy-three living patients. The average duration of radiographic follow-up for these seventy-seven hips was twelve years (range, ten to 16.8 years). The remaining five patients (five hips) were interviewed by telephone but did not send a follow-up radiograph. The average radiographic follow-up for these five patients was between 0.1 and nine years, and no patient had undergone revision arthroplasty. Radiographic evaluation of the seventy-seven hips with radiographs made at a minimum of ten years demonstrated osseous ingrowth in all hips. For the entire cohort of patients with at least one year of radiographic follow-up (108 hips), all demonstrated osseous ingrowth of the femoral stems and acetabular components. For the remainder of the radiographic analysis, we only included the seventy-seven hips with radiographs made at a minimum of ten years. Femoral stress-shielding was mild in forty-three hips (56%), moderate in twenty-six hips (34%), and severe in two hips (3%). The remaining six hips did not show radiographic evidence of stress-shielding. Proximal radiolucencies (zones I and/or VII) occurred in twenty hips (26%). Distal tip radiolucencies (zone IV) occurred in four hips (5%). In the intermediate zones (II, III, V, and VI), radiolucencies were present in thirteen hips (17%), with all in the area of the medial cutout in the stem. Proximal femoral osteolysis occurred in nineteen hips (25%). There were no cases of distal femoral osteolysis. Radiographic evaluation of the acetabular constructs demonstrated thirty-three hips (43%) with bone-prosthesis radiolucencies. None of these were >1 mm in width, and none were circumferential to include the screws (two or three screws were used in all cases). No acetabular components had migrated.
Wear Results
The mean linear polyethylene wear rate was 0.252 mm/y (range, 0.034 to 1.10 mm/y). The mean volumetric polyethylene wear rate was 79.95 mm3/y (range, 9.14 to 455.15 mm3/y). The mean gross linear and volumetric wear were 2.80 mm (range, 0.42 to 10.80 mm) and 887.63 mm3 (range, 115.17 to 4983.39 mm3), respectively. Three patients (three hips) had polyethylene that was moderately cross-linked (see Appendix). The mean linear and volumetric wears rates for two of these patients with ten-year radiographs was 0.06 mm/y (range, 0.03 to 0.09 mm/y) and 19.50 mm3/yr (range, 10.2 to 28.8 mm3/y), respectively. The mean linear polyethylene wear rates for Duraloc and Harris-Galante-I liners were 0.288 mm/y (range, 0.059 to 1.10 mm/y) and 0.208 mm/y (range, 0.034 to 0.491 mm/y), respectively. The mean volumetric polyethylene wear rates for Duraloc and Harris-Galante-I liners were 91.53 mm3/y (range, 11.69 to 455.15 mm3/y) and 65.39 mm3/y (range, 9.14 to 199.77 mm3/y), respectively.
Survivorship Analysis
Kaplan-Meier survivorship analysis demonstrated no differences between the cementless and cemented series in regard to reoperation for any reason at ten years to fifteen years (Fig. 1 and see Appendix). However, the fixation in the cementless series outperformed (moderately on femoral parameters) the cemented series (reoperation for acetabular loosening, reoperation for femoral loosening, and radiographic loosening of the femoral and acetabular components at ten to fifteen years; Table II). Log-rank analysis demonstrated that the cementless cohort performed better than the cemented cohort in terms of revision for aseptic loosening of the acetabular component (p = 0.007) and femoral component (p = 0.159), and radiographic loosening of the acetabular (p = 0.017) and femoral (p = 0.094) components (Table II).
This minimum ten-year follow-up study of a consecutive series of patients who underwent a total hip arthroplasty at the age of fifty years or younger is one of the few with longer follow-up and the only one using this second-generation, extensively porous-coated device (Table III). The limitations of the study include the number of patients lost to follow-up (10%), not all patients had minimum ten-year follow-up radiographs (93%), issues related to the intraobserver and interobserver variability in evaluating radiographs, and the pedometer and six-minute-walk data were obtained only at the time of final follow-up. In addition, the cemented historical control had more patients with dysplasia of the hip as a primary diagnosis. The small percentage of patients with rheumatoid arthritis (1% in the cementless group and 9% in the cemented group) is also much smaller than previous long-term follow-up studies in young patients31. The strengths of the study include the length of follow-up, the large number of hips with wear measurements, and activity assessments at the time of the final follow-up.
Our results with an extensively porous-coated cementless stem and cementless acetabular component corroborate the findings of Capello et al.1 and McLaughlin and Lee3 in groups of patients who were under fifty years old. Using proximally coated devices, they demonstrated the same durability of cementless fixation as was observed in the present study. Capello et al.1 revised only 0.9% of the hips for aseptic loosening, and McLaughlin and Lee3 had no revision for aseptic loosening. Kim et al. reported a 3% revision rate for aseptic loosening at this length of follow-up in the Asian population4. Duffy et al. reported a higher rate of revision using first-generation proximally coated devices2 (Table III). Fixation of the acetabular component was superior to our historical cemented control group; fixation of the femoral component was also better than our historical cemented control group, but the difference was not significant. Our observation of moderate to severe stress-shielding in 37% of the hips is similar to that reported by McLaughlin and Lee3 with the Taperloc femoral stem and is not surprising, given the work of Bugbee et al. who showed that 23% of a series of 208 extensively porous-coated hip replacements with a minimum two-year follow-up showed radiographic evidence of stress-shielding32.
We demonstrated an average linear wear rate of 0.252 mm/y, average steps per year of 1.54 million, and an average six-minute walk of 323 m. These findings are similar to those in previous reports; however, our relatively low six-minute-walk distances may be related to the comorbidities of some of the patients4,33-39. The high wear rate accounts for the 10% (twelve) of 115 hips that required liner exchange for polyethylene wear. We attribute the high wear rate to the younger population as well as to the fact that the majority of hips in the study cohort either had a Harris-Galante-I acetabular component (with potential concerns regarding the locking mechanism) or a Duraloc component with non-cross-linked polyethylene. This relatively high rate of revision for polyethylene wear is the main reason that the survivorship of the cementless cohort, in terms of reoperation for any reason, is not significantly different from our cemented cohort. While we are encouraged by the significant improvement of cementless total hip arthroplasty in terms of implant fixation at ten years, we are aware of the potential for more revisions for polyethylene wear with longer-term follow-up.
This study demonstrates the need to appropriately counsel young patients who are considering total hip replacement for disabling hip arthritis. Although we demonstrated durable fixation with uncemented porous-coated devices, a number of revisions were still required.