Complex Revision Total Hip Arthroplasty with Modular Stems at a Mean of 14 Years

Abstract

We retrospectively reviewed 87 patients (92 hips) who had revision hip surgery to determine whether a proximally coated, modular femoral component would remain stably fixed at long-term followup. Thirteen patients died and 12 patients were lost to followup leaving 62 patients (67 hips) available for review with minimum followup of 8 years (mean, 14 years; range, 8-17 years). Preoperative radiographs were reviewed using Paprosky's classification. Postoperative radiographs were reviewed for osteolysis, endosteal hypertrophy, cortical hypertrophy, distal pedestals, component breakage, and loosening. Thirty-seven hips underwent femoral allografting, 10 of which were proximal femurs. With revision as the endpoint the Kaplan-Meier survivorship (including deaths and loss to followup) was 60% at 14 years. Forty-seven of the 57 (82%) noncemented stems were bone ingrown. All had relative proximal bone preservation and 33 of 57 (58%) had bone hypertrophy in the diaphysis around zones 2 and 6. There were five aseptic failures (9%). Each of those was Paprosky Class IIIB or IV preoperatively. There were no long-term failures in Paprosky Class II or IIIA. The aseptic failures have been re-revised. This modular stem resulted in reliable fixation with relative preservation of proximal bone stock at this intermediate interval in complex revisions in Paprosky Class II and IIIA. Paprosky Class IIIB and IV defects may need additional component fixation options.

Level of Evidence: Level IV, therapeutic study. See the Guidelines for Authors for a complete description of levels of evidence.

When faced with femoral component loosening accompanied by both cavitary and segmental bone loss, revision surgery is technically challenging. Extensile exposure, implant removal, and cement removal are difficult, and choosing the proper implant to allow preservation of existing bone and restoration of new bone is also challenging.

A modular stem allows independent metaphyseal/diaphyseal sizing, stem-to-neck length options, and adjustable offset and version. Modularity can also be advantageous to help offset leg length in situations with major discrepancies.^2,5 Although the porous ingrowth spout promotes preservation of proximal bone stock, the distal slot and longitudinal splines allow fixation distally, but not ingrowth, to minimize thigh pain.^4,21 Numerous potential concerns with all modular stems have been raised, including corrosion at the taper causing lysis, fear of breakage, and taper disengagement.^1,3

We asked whether a modular, proximally coated titanium femoral component would provide adequate fixation for complex revisions at mid- to long-term. Additionally, we asked whether modularity would adversely affect stable fixation due to breakage or taper disassociation and whether the construct would result in restoration of bone.

MATERIALS AND METHODS

We retrospectively reviewed 87 consecutive patients (92 hips) who had complex revision hip surgery between 1989 and 1993 using a proximally coated, modular femoral component (SROM; DePuy, Warsaw, IN). We defined complex surgery as femoral revisions with structural or cavitary defects or both. Thirteen patients died and 12 patients were lost to followup leaving 62 patients (67 hips) available for review; 57 of the hips were uncemented and 10 were cemented. The minimum followup was 8 years (mean, 14 years; range, 8-17 years). We obtained prior permission from our Institutional Review Board.

We (JCM, JL) classified the preoperative bone deficiencies as described by Paprosky et al.^18,19 Fifty-two (78%) of these hips were Paprosky Class III and IV, 19 of which were Class IIIB or IV (Table 1). Forty-two revisions were performed on previously cemented femoral components. Eight were second-stage conversions of previous resection arthroplasties. The remaining 17 cases were porous coated stems that failed.

TABLE 1: Classification of Preoperative Bone Deficiencies

Thirty-seven (71%) of the 52 hips with Paprosky Class III and IV had femoral allografting. Morselized graft was used in six hips for localized defects. Twenty-one were strut allografts used to provide additional strength for those femurs with large structural and cavitary defects resulting from osteolysis and prior surgery. In the hips with strut allografting we used bowed stems longer than 200 mm to bypass the deficiency. Ten hips (all Paprosky Class IV) had proximal femoral allografts because of massive proximal femoral bone loss. Because there was no host bone to accommodate ingrowth, the proximal cone was cemented into the allograft. Fifty-five hips underwent acetabular revision as well. Trochanteric osteotomy or slide was performed on 42 of the hips.

Postoperative radiographs were reviewed by the surgeon (JCM) for loosening, osteolysis, endosteal hypertrophy, cortical hypertrophy, distal pedestals, and breakage. Fixation was classified according to the criteria of Engh et al⁹ as well as by the bony response, the calcar for hypertrophy, endosteal resorption or hypertrophy, and of the cortical bone for hypertrophy.

We performed a Kaplan-Meier survivorship based on femoral component revision as the endpoint, including those lost to followup.¹²

RESULTS

With revision as the endpoint the Kaplan-Meier survivorship was 60% at 14 years (Fig 1). Bony ingrowth occurred in 47 (82%) of the 57 noncemented hips. Four stems had stable fibrous fixation and six were radiographically loose. These six radiographically loose hips were revised for aseptic loosening. There was one failure in Paprosky Class IIIA with an undersized stem that shifted into varus and failed early. There were two failures in the Paprosky Class IV category including one patient with osteogenesis imperfecta (OI). The other patient had the initial arthroplasty in his early 20s and it was subsequently revised three times (Fig 2). There were three failures in Paprosky Class IIIB. Three followed periprosthetic fractures but all resulted from progressive loosening and extensive bone loss from multiple prior failed fixations.

Fig 1: The Kaplan Meier survivorship shown for this modular femoral component is 60% at mean 14 years, including deaths and lost to followup, with femoral revision as the endpoint.

Fig 2: This postoperative radiograph shows subsidence and loosening of the femoral component in a patient who had multiple surgeries and massive bone loss preoperatively.

We found no uncoupling or breakage of the components in this series. Eight patients subsequently underwent acetabular revisions for polyethylene wear and acetabular loosening but the modular femoral components remained solidly fixed. No fretting or corrosion was observed at the time of revision.

Endosteal hypertrophy (Fig 3), occurred in 33 (57%) of the 57 uncemented hips. We observed nonbridging pedestals (Fig 4) in 50% of cases from the distal stem abutting the anterior cortex. There was mild calcar resorption in 28%. Proximal lysis occurred in four hips (all went on to revision) and no distal lysis was observed.

Fig 3: This postoperative radiograph demonstrates a well-fixed stem with cancellous condensation around the spout.

Fig 4: This radiograph shows the distal femoral component had been deflected outward into the lateral cortex.

Six (10%) of the 67 hips in 62 patients had a postoperative infection, two of which remained infected despite subsequent interim staged resection arthroplasty. The two reinfected hips had cement fixation into an allograft. Of the 10 hips with cement into an allograft, the previously mentioned two became septic and three additional stems had progressive loosening.

DISCUSSION

Femoral component revision in the face of cavitary and segmental bone loss makes revision surgery challenging. A modular stem allows independent metaphyseal/diaphyseal sizing, stem-to-neck length options, and adjustable offset and version. We asked whether a modular, proximally coated titanium femoral component would provide adequate fixation for complex revisions at mid- to long-term and allow restoration of bone without the potential disadvantages of a modular component.

We note several limitations of this study, including incomplete clinical outcome data; therefore, we report only the radiographic outcomes and revisions. Varying duration of followup is a limitation, however, the rates of failure in Paprosky Class IIIB and IV were evident at the 8-year minimum followup in all hips revised for aseptic loosening. Additionally, we did not use a different component or different surgical technique to compare and contrast outcomes.

Our data suggest a proximally coated modular stem can achieve reliable fixation in femurs that have adequate proximal bone stock such as those in Paprosky Class II or IIIA. Ingrowth between host bone and porous coating requires maximum contact to prevent micromotion and subsidence. Design features that enhance bone ingrowth include a tight diaphyseal fit and maximum metaphyseal fit. In revision surgery there are often metaphyseal/diaphyseal mismatches due to cavitary or structural bone loss. Slotted stems and lightweight component material, such as titanium, decreases the stiffness of the component stem. This design transfers more of the weight-bearing load from the implant to the femur, which may minimize adverse remodeling of the proximal femur.²⁰ Additionally, the distal fluting provided increased rotational stability in our study and promoted excellent fixation.

Proximally coated monobloc stems in revision surgery have not reliably provided ingrowth fixation. Mulliken et al¹⁷ reviewed 52 stems at early followup (4-6 years) with a 10% revision rate but an additional 14% were radiographically unstable. That study also had a 40% intraoperative fracture rate. Mulkani et al¹⁴ reviewed 69 stems at early followup and had a similar 8.7% revision rate but a 29% mechanical failure rate. Woolson and Delaney²⁵ reviewed 28 hips with a 20% revision rate and a 45% mechanical failure rate.

Outcomes for modular stems with proximal coating are more favorable. Cameron⁶ reviewed 320 revision stems with only a 1.4% (3 of 211 long stems) loosening rate at midterm followup (2-12 years). Christie et al⁸ reviewed 163 revisions at early followup (4-7 years) with a 2.9% failure rate. However, they excluded Paprosky Type IV defects. Another study of 62 revisions with the same modular stem²³ reported only one failure, but also excluded patients with Paprosky Class III defects that had bone grafting. Similar to our study, Chandler et al⁷ reviewed 52 hips, 22 of which had femoral allografting; only five stems were mechanically loose at early followup. Our study reports longer followup.

We identified some limitations with this particular femoral component. Orientation of the neck dictates stem rotation and axial alignment of the spout dictates distal stem position. This can be an issue with an extensively bowed femur. If the spout is placed in anteversion, the stem turns with the component. This rotation can cause the stem to deflect anterolaterally and penetrate the cortex (Fig 4). The length of the stem dictates the neck length. There is no option for a short neck with a longer stem, which can be problematic in some situations such as reimplantation after resection arthroplasty. The proximally coated stem used in this study has limited ingrowth potential because only the sleeve is porous-coated. This limited fixation may be inadequate when there is extensive proximal bone loss as suggested by the aseptic failures in Paprosky Class IIIB and IV. In these situations an extensively coated stem may be a better option.

Extensively coated stems have varied success when there is extensive bone loss such as in Paprosky Class IIIB or IV at long-term followup.^{10,11,15,16,19} McCauley and Engh¹⁵ reported bone ingrowth occurred only 75% of the time at 10-year followup with Paprosky Class IIIA, although the mean time at revision was approximately 5 years. Sporer and Paprosky²² also reported high failure rates with Paprosky Class III and IV defects at mean 6-year followup in 51 hips. However, in a study with longer followup, the aseptic loosening rate in 170 hips was 3.5% with a 4.1% mechanical failure rate.²⁴ Engh et al¹⁰ reported a survival rate similar to ours (89%) with extensively coated stems that bypassed severe proximal bone loss at mean 13-year followup.

The early concerns with breakage and uncoupling of modular components were not substantiated by our data. Despite the limited fixation surface of the proximal sleeve, no stem that achieved early fixation (within the first year) developed lysis distal to the sleeve or became aseptically loose. The data suggest this stem design provides long-term fixation with Paprosky class II and IIIA bony deficiencies.