Twenty-year Followup of an Uncemented Stem in Primary THA

Abstract

Background

Uncemented stems have been used in THA for well over two decades, but there are relatively few studies reporting on the results after 20 years.

Questions/purposes

The purpose of this study was to evaluate at a minimum followup of 20 years (1) hip scores; (2) radiographic findings, including stem fixation and osteolysis; (3) reoperations; and (4) survivorship free from aseptic loosening in a group of patients who underwent primary THA using a proximally porous-coated, plasma-sprayed, straight-stemmed, titanium-alloy femoral component.

Methods

Between 1987 and 1993, we performed 1517 primary THAs, of which 447 were cementless, and 157 used the implant under study here (representing 10% of the THAs during the period in question). General indications for the use of the study stem included (1) younger age with relatively stronger bone (average age 55 years in this study); (2) patients not involved in another study protocol (as were the 278 other cementless hips implanted during this time); and (3) patients who, at the surgeon’s discretion, would be compliant with protected weightbearing. Of those, 111 were available for followup at a minimum of 20 years (mean, 20 years), whereas 36 had died and nine (6%) were lost to followup or declined participation. The primary diagnosis was osteoarthritis in 77 hips (69%). The mean age at operation was 55 years. The clinical result was evaluated on the basis of the Harris hip score. Radiographic analysis was performed at each followup visit for distal cortical hypertrophy, spot welds, radiolucencies, stem subsidence, varus or valgus shift, osteolysis, and femoral component loosening. Kaplan-Meier analysis was performed to evaluate the survival of the femoral component.

Results

The mean Harris hip score improved from 46 points to 87 points (of a possible 100 points) at the most recent followup. All hips had evidence of proximal femoral remodeling consistent with osseous ingrowth. Femoral osteolyis was seen in 22 hips (20%). One stem was revised as a result of a periprosthetic fracture. No femoral component had evidence of loosening, and none was revised. Thirty-three acetabular components (30%) were revised.

Conclusions

This femoral component provided durable long-term fixation for over two decades after THA. The porous stem geometry is still in use today and will continue to be studied into the third decade of use.

Level of Evidence

Level IV, therapeutic study. See Instructions for Authors for a complete description of levels of evidence.

Introduction

Cementless fixation is the dominant approach for the femoral component in the United States [26]. Although registry data call this practice into question [25, 26], it has been estimated that 86% of all THAs in the United States are implanted using uncemented fixation [11].

However, there are relatively few reports on the minimum 20-year results of cementless THA [1–3, 7, 14, 16–19, 24, 28]. Only seven of these studies [2, 3, 14, 18, 19, 24, 28] report on stems that are still available for implantation today with stem survivorship between 86% [24] and 98% [1] at 20 years.

We therefore report at a minimum followup of 20 years (1) hip scores; (2) radiographic findings, including stem fixation and osteolysis; (3) reoperations; and (4) survivorship free from aseptic loosening in a group of patients who underwent primary THA at one institution using a proximally porous-coated, plasma-sprayed, straight-stemmed, titanium-alloy femoral component. The 10-year survivorship of this stem has been previously published [21].

Patients and Methods

Between May 1987 and December 1993, 157 THAs were performed in 142 patents with the use of a collarless proximally porous-coated, straight-stemmed, titanium-alloy femoral component. During this time period, 1517 primary THAs were performed at the same center. One thousand seventy THAs were performed with cemented stems and 447 THAs were performed with uncemented stems. These uncemented stems included 278 uncemented hydroxyapatite stems (BiMetric HAP; Biomet, Warsaw, IN, USA), seven Mallory-Head^® primary stems (Biomet), three BiMetric 220-mm long-stemmed femoral components (Biomet), one S-ROM (DePuy, Warsaw, IN, USA), one Omniflex stem (Stryker, Mahwah, NJ, USA), and 157 BiMetric stems (Biomet) stems that are the subject of this study. We included only the 157 BiMetrics and did not include the BiMetic HAP stems (1) to keep the study group relatively homogeneous; (2) because the BiMetric HAP was discontinued; and (3) because the BiMetric HAP had a different porous surface.

During this period, the general indications for the use of the study stem included (1) younger age with relatively stronger bone (average age 55 years in this study); (2) patients not involved in another study protocol (as were the 278 BiMetric HAP hips); and (3) patients who, at the surgeon’s discretion, would be compliant with protected weightbearing. All operations were performed by three surgeons (MAR, EMK, PMF) at one single center.

Thirty-six patients (37 THAs) died before the 20-year followup from causes unrelated to the index arthroplasty between 1 and 19 years after THA. A total of 6% of the patients alive at a minimum of 20 years had either refused followup (four patients, four THAs; 3–15 years after surgery) or were lost to followup (five patients, five THAs; between 5 and 7 years after surgery). The prostheses in these 45 patients (46 THAs) were known to be functioning well at the time of the most recent followup. Thus, 111 hips in 97 patients were available for clinical and radiographic analysis (Table 1). The minimum followup was 20 years (mean, 22 years; range, 20–25 years). In the present study, the cohort of hips was expanded to 157 THAs to include all uncemented Bi-Metric^® stems implanted during this time period; 111 of the 157 THAs were included in the original study [21].

Table 1.

Demographic data

Variable	Value
Number of hips	111
Number of patients	97
Diagnosis (number of hips)
Osteoarthritis	77 (69%)
Osteonecrosis	20 (17%)
Developmental hip dysplasia	6 (5%)
Rheumatoid arthritis	6 (5%)
Paget’s disease	2 (2%)
Sex (number of patients)
Male	51 (53%)
Female	46 (47%)
Age at surgery (years)*	55 (27–72)

* Values are expressed as mean with range in parentheses.

Approval for a retrospective review of the THAs was obtained from the institution review board of St Francis Hospital (Indianapolis, IN, USA).

The plasma-sprayed porous coating covered the circumference of the proximal one-third of the stem and the stem has a 3° proximal-to-distal taper [21, 22] (Bi-Metric^®; Biomet) (Fig. 1).

Fig. 1.

The Bi-Metric^® femoral component is shown.

A posterior approach was used in all hips. All hips received an uncemented acetabular component (Nelson Cup; Biomet) and a 28-mm round titanium-alloy femoral head with a 12/14 taper. Deep vein thrombosis prophylaxis and the postoperative protocol have been previously published and were the same in all patients [21].

Patients were evaluated at 6 months, 1 year, 3 years, and every 2 to 3 years thereafter in the office. The clinical evaluation of function and pain was performed by the operative surgeon (MAR, EMK, PMF) with the use of the Harris hip score [9] and was prospectively recorded. Patients were specifically asked as to the degree and location of their hip pain. If present, activity-related thigh pain was specifically noted at each followup. Radiographs of the pelvis (AP) and the hip (Lauenstein lateral) were made at each visit. The technique has been described [21]. Distal cortical hypertrophy was recorded as either present or absent as were “spot welds” [6]. Radiolucencies and spot welds were recorded according to the zonal distribution of Gruen et al. [8]. Stem subsidence and varus or valgus shift were measured according to the criteria of Johnston et al. [12] and Loudon [15]. Osteolysis was considered to be present if a progressive radiolucent cavity was identified in the periprosthetic bone [21]. Osteolysis was not quantified but recorded according to the zonal distribution of Gruen et al. [8]. Femoral component loosening was evaluated according to the criteria of Engh et al. [5]. All radiographs were retrospectively reviewed by one author (JBM) who did not perform any of the THAs and was not blinded as to the surgeon.

Kaplan-Meier analysis [13] of the survival of the femoral component was performed for all 157 THAs from the original cohort. The best-case, standard-case, and worst-case scenarios [21, 23] were determined with the use of SAS^® software (SAS Institute, Inc, Cary, NC, USA). Stem failure was defined as aseptic loosening with or without revision.

Results

The mean Harris hip score improved from 46 points (range, 26–81 points; SD, 17.9) preoperatively to 87 points (range, 57–100 points; SD, 8.3); 100 points is the maximum score) at final followup. At each followup, the mean Harris hip scores were: 88 points (range, 53–100 points; SD, 17.2) at 6 months, 91 points (range, 58–100 points; SD, 16.4) at 1 year, 94 points (range, 66–100 points; SD, 7.5) at 3 years, 96 points (range, 76–100 points; SD, 4.9) at 5 years, 93 points (range, 50–100 points; SD, 9.3) at 7 years, 91 points (range, 54–100 points; SD, 11.4) at 10 years, 96 points (range, 63–100 points; SD, 6.7) at 12 years, 92 points (range, 68–100 points; SD, 12.6) at 15 years, 89 points (range, 27–100 points; SD, 15.9) at 17 years, and 87 points (range, 55–100 points; SD, 14.7) at 20 years. The mean pain score improved from 17 points (range, 0–30 points; SD, 7.8) preoperatively to 38 points (range, 30–44 points; SD, 6.2) (44 points is the maximum score, defined as no pain) at latest followup. At each followup, the mean pain scores were: 41 points (range, 10–44 points; SD, 5.3) at 6 months, 42 points (range, 20–44 points; SD, 3.8) at 1 year, 42 points (range, 20–44 points; SD, 4.7) at 3 years, 43 points (range, 30–44 points; SD, 2.8) at 5 years, 42 points (range, 20–44 points; SD, 5.1) at 7 years, 41 points (range, 10–44 points; SD, 4.3) at 10 years, 42 points (range, 10–44 points; SD, 6.6) at 12 years, 41 points (range, 20–44 points; SD, 8.5) at 15 years, 40 points (range, 10–44 points; SD, 8.3) at 17 years, and 38 points (range, 10–44 points; SD, 11.0) at 20 years. At the time of the most recent followup, no patient reported activity-related thigh pain.

Radiographically, all stems were placed within 5° of neutral alignment. No measurable change in varus or valgus positioning was noted for any stem at any followup. Stem subsidence was noted in one hip (1%) measuring 8 mm at 6 months with no further change. At 20 years of followup, this patient was pain-free. At final followup, osteolysis was found around the stem in 22 hips (20%) and around the cup in 55 hips (50%). All 22 hips with osteolyis had lesions in Zone 1 and eight hips had lesions in Zone 7. There were no cases of femoral osteolysis noted in Zones 2, 3, 4, 5, or 6. Radiolucencies about the stem tip (Zone 4) were not uncommon and found in 11 hips (10%) between 1 and 7 years after THA. All of these were less than 2 mm and nonprogressive. In two other hips, a nonprogressive radiolucency was identified in Zone 3 at 1 year. There were no other radiolucencies noted. Distal cortical hypertrophy occurred in 92 hips (83%) at a mean of 2 years (range, 1–10 years). Spot weld formation was initially identified in 69 hips (62%) at a mean of 3 years of followup (range, 0.5–10 years) either in Zone 1 or 7 about the plasma-sprayed proximal portion of the stem. No spot welds were found distal to these zones. Thus, according to the criteria of Engh et al. [5, 6], stem osteointegration was noted in all THAs.

One stem had been revised during the study period for a periprosthetic femur fracture 3 years postoperatively (Fig. 2). There were no other loose or revised stems. However, 33 cups (30%) were revised secondary to loosening and/or wear. One other cup was revised for recurrent dislocation. There were no deep would infections in this cohort.

Fig. 2A–E.

(A) A radiograph shows a THA in the left hip of a patient 7 years postoperatively. (B) A periprosthetic fracture with a loose stem can be seen. (C–E) Radiographs show the hip after revision THA.

No stems were noted to be aseptically loose. Therefore, Kaplan-Meier analysis of the original 157 stems demonstrated best-case and standard-case scenarios of 100% survival at 20 years. Assuming that all of the patients who were not followed for a minimum of 20 years had their stems fail, then stem survivorship (with 95% confidence intervals) at 5, 10, 15, and 20 years were 94% (0.92–0.96), 94% (0.91–0.95), 92% (0.89–0.94), and 89% (0.83–0.92), respectively.

Discussion

Although the use of uncemented stem fixation in THA has increased worldwide [27], it has been suggested that this change is not backed by evidence that can be generalized to predict the outcome of THA [27]. Thus, it appears that more long-term clinical studies are needed to support the relative increase in the use of uncemented stems. In the present study, the clinical and radiographic results of 111 uncemented THAs in 97 patients were reviewed with a minimum followup of 20 years. The average Harris hip score improved from 46 points to an average of 87 points at the most recent followup. The average postoperative pain score was 38 points. No patient reported any activity-related thigh pain at final followup. All hips had evidence of proximal femoral remodeling consistent with osseous ingrowth. One stem was revised as a result of a periprosthetic fracture. No femoral component had evidence of loosening, although Kaplan-Meier survival analysis for aseptic loosening demonstrated a 89% survival rate at 20 years according to the worst-case scenario.

There are several limitations to this study. First, the surgeon evaluated the patients at the clinical visit and the recording of the Harris hip score and thigh pain could have been biased by this fact. All radiographs were independently reviewed, however, by someone who did not perform the operations. The authors believe that this minimized bias on the objective radiographic data. One author (EMK) was a consultant designer during the time of the THAs and was one of the operative surgeons but was not part of the radiographic review. Second, not all patients received this stem during the timeframe of this study. A majority of the THAs are our center were cemented during this time and implant selection, although geared to younger individuals, was primarily at the discretion of the surgeon. Third, some hips (6%) were unaccounted for. Although it is possible that they were revised elsewhere for stem loosening, it is unlikely that all patients who were not followed for a minimum of 20 years had failed stems as the worst-case scenario survival suggests. Fourth, the cup design used in this study has been associated with a high rate of wear (0.22 mm/year) [10]. Therefore, periarticular osteolysis was not uncommon in the present report. We have abandoned the use of this acetabular component. Similarly, we abandoned use of 28-mm titanium-alloy femoral heads in the early 1990s [21]. Fifth, osteolyis, although recorded, was not quantified and was recorded as either present or absent according to zonal distribution. It is likely that the actual degree of osteolysis is greater than would be suggested. Nevertheless, despite the cup design and the use of titanium-alloy heads, the uncemented stem is this study demonstrated no aseptic loosening. It is altogether reasonable that the clinical results of this stem would be similar with improved cup designs and better polyethylene and femoral heads. Further study of this stem under these circumstances is needed.

To our knowledge, there are only seven 20-year followup studies reporting on uncemented stems that are still in use today [2, 3, 14, 18, 19, 24, 28]. Streit et al. [24] reported 22-year survivorship of 86% in 354 CLS^® stems (Zimmer, Warsaw, IN, USA). Twelve stems (3%) were revised for perprosthetic fracture. Clinical scores were not documented. Lombardi et al. [14] reported a 20-year survivorship of 99% in 196 Mallory-Head^® stems. The average Harris hip score in the original cohort of 2000 THAs at final followup was 83 (range, 18–100). The average pain score at final followup was 38 (range, 10–44). Six stems were revised for perprosthetic fracture. Corten et al. [3] also reported on the Mallory-Head^® stem and reported a 99% survival rate at 20 years in 126 THAs. Clinical scores and failure resulting from periprosthetic fractures were not documented. Vidalain [28] published a 23-year survivorship rate of 97% in 114 Corail^® stems (DePuy Synthes, Warsaw, IN, USA). Although the Harris scores were not used, “excellent results” were seen in 98% of cases and 83% of patients were pain-free. It was uncertain if any failures were the result of periprosthetic fractures. McLaughlin and Lee [18] published a 22-year survivorship rate of 99% in 145 Taperloc^® stems (Biomet). The average Harris hip score at final followup was 83 (range, 18–100). Pain scores were not presented and no stem was revised for periprosthetic fracture. In a followup report of the same cohort, they published a 26-year survivorship rate of 99% [19]. Finally, Belmont et al. [2] reported a 20-year survivorship rate of 98% in 223 AML^® uncemented stems. Clinical scores were not given. Three stems were revised from the original cohort. It is uncertain how many were revised because of periprosthetic fractures. Thus, if one includes the present report, it appears that, of the uncemented stems available for implantation today, only six are supported with 20-year survivorship data. Only one patient in this series underwent reoperation for a problem on the femoral side (peripropsthetic fracture); this compares favorably with other long-term cementless THA series that have documented the number of periprosthetic fractures [14, 18, 19, 23].

We reported a 100% 10-year survivorship rate of this stem using a standard-case scenario [21]. In that consecutive series of 105 THAs reviewed at minimum 10-year followup, no stem was revised. In the present study, the cohort of hips was expanded to 157 THAs using this stem implanted between May 1987 and December 1990; 111 of the 157 THAs were included in the original study. After another decade of followup, no stem in either study exhibited progressive migration or extensive radiolucencies. Thus, according to the criteria of Engh et al. [5], fixation of the stem by bone ingrowth occurred in all hips. Davies et al. [4] reported on 64 uncemented stems of the same design implanted in 54 patients with a minimum followup of 13 years (mean, 15.2 years) and noted no stem failures or stem revisions. A previous study from out institution reviewed 2321 uncemented THAs using this same stem for the purpose of comparing results in Dorr A, B, and C bone [20]. The mean followup was 5.9 years (range, 2–19.5 years). Only two stems in Class B bone were revised for aseptic loosening. No other loose stems were identified. These two patients were not included in the original cohort of the present study because they underwent THA in 1998 and 2002. The 5-, 10-, and 15-year survival rates were 100%, 100%, and 100%, respectively, in patients with Dorr A bone; 99.9%, 99.4%, and 99.4%, respectively, in patients with Dorr B bone; and 100%, 100%, and 100%, respectively, in patients with Dorr C bone. In 1997, we published a comparison of 103 collared stems and 100 collarless stems of this design that were part of a prospective randomized study [22]. No difference was noted in stem survival or periprosthetic bone remodeling at a mean followup of 45.7 months. Only one loose stem was identified in 203 THAs. Of note, although this stem has historically been manufactured in collarless, collared, and hydroxyapatite designs, only the collarless version of the stem was used in the present report.

In conclusion, this uncemented femoral component demonstrated durable fixation and excellent survivorship continuing into the third decade after implantation. Because the use of uncemented stems in THA continues to increase, further study is warranted to determine survivorship beyond 20 years.