Comparison of Porous Tantalum Acetabular Implants and Harrington Reconstruction for Metastatic Disease of the Acetabulum

Abstract

Background:

The periacetabular region is a common location for metastatic disease. Although large lytic acetabular defects are commonly treated with a hip arthroplasty with a cemented component according to a Harrington-style reconstruction, the use of highly porous uncemented tantalum acetabular components has been described. Currently, there are no direct comparisons of these reconstructive techniques. The purpose of this study was to compare the outcomes of the Harrington technique and tantalum acetabular component reconstruction for periacetabular metastases.

Methods:

From 2 tertiary sarcoma centers, we retrospectively reviewed 115 patients (70 female and 45 male) with an acetabular metastatic defect who had been treated between 2002 and 2015 with a total hip arthroplasty using either the cemented Harrington technique (78 patients) or a tantalum acetabular reconstruction (37 patients). The mean patient age was 61 years, and the most common Eastern Cooperative Oncology Group status was 3 (39 patients). The mean follow-up for surviving patients was 4 years.

Results:

An additional surgical procedure was performed in 24 patients (21%). Harrington-style reconstructions were more likely to require a reoperation compared with tantalum reconstructions (hazard ratio [HR], 4.59; p = 0.003). The acetabular component was revised in 13 patients (11%); 5 patients (4%) underwent revisions that were due to loosening of the acetabular component. The 10-year cumulative incidence of revision of the acetabular component for loosening was 9.6% in the Harrington group and 0% in the tantalum group (p = 0.09). The mean Harris hip score significantly improved following reconstruction (31 to 67 points; p < 0.001), with no significant difference (p = 0.29) between groups.

Conclusions:

In patients with periacetabular metastatic disease treated with total hip arthroplasty, an acetabular reconstruction strategy utilizing highly porous tantalum acetabular components and augments successfully provided patients with a more durable construct with fewer complications compared with the cemented Harrington-style technique.

Levels of Evidence:

Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.

The pelvis and acetabulum are common locations for the development of metastatic disease. Due to the mechanical forces transmitted through the hip, destructive periacetabular metastases lead to substantial pain and disability^1-3. Depending on the histology, treatment initially involves activity modifications, weight-bearing restriction, radiation therapy, chemotherapy, immunotherapy, bone-modifying agents, bisphosphonates, and analgesics; however, if these treatments fail, a patient has a fracture, or a patient is at high risk for fracture, surgical treatment is necessary³. A surgical procedure requires the removal of all gross tumor with intralesional curettage and reconstruction to restore mechanical stability of the hip, with the goal of allowing for immediate weight-bearing and pain and local tumor control³.

Historically, the Harrington reconstruction technique was the primary treatment option for patients with periacetabular metastatic disease and transmitted the weight-bearing load of the weakened acetabulum to the stronger, intact proximal aspect of the pelvis^1,4-8. Although these types of reconstruction have provided excellent short-term outcomes, there remains concern that this type of cemented implant may fail with longer follow-up. Highly porous tantalum acetabular components and augments have been used for primary and non-primary tumors, with promising initial results, even in the setting of radiation therapy^9-13. However, to our knowledge, the outcomes of these 2 techniques have not been directly compared. The aim of the current study was to compare the outcome of the Harrington technique with that of tantalum components for reconstruction of the acetabulum following intralesional resection of periacetabular metastatic disease.

Materials and Methods

Following multicenter institutional review board approval, we retrospectively reviewed the medical records of 119 patients who underwent a total hip arthroplasty for the treatment of metastatic disease in the periacetabular region between 2002 and 2015 at 2 tertiary sarcoma centers. Four surviving patients did not have at least 1 year of clinical follow-up and were excluded. The remaining 115 patients consisted of 70 female patients and 45 male patients with a mean age of 61 years (range, 26 to 90 years) at the time of the surgical procedure. There were 78 reconstructions (68%) using the Harrington technique and 37 reconstructions (32%) with tantalum components (Trabecular Metal Revision Shell and Augments; Zimmer Biomet).

All patients had metastatic disease, most commonly from the breast (35 patients [30%]) and lung (18 patients [16%]). The preoperative Eastern Cooperative Oncology Group (ECOG)¹⁴ performance status was as follows: 27 patients (23%) with ECOG 1, 35 patients (30%) with ECOG 2, 39 patients (34%) with ECOG 3, and 14 patients (12%) with ECOG 4. Acetabular deficiencies were characterized according to the Harrington classification as follows⁴: Class I, 35 patients (30%); Class II, 19 patients (17%); and Class III, 61 patients (53%). At the time of the index total hip arthroplasty, 64 patients (56%) had a pathologic acetabular fracture, and 23 patients (20%) had pelvic discontinuities.

Ninety-six patients (83%) received radiation therapy to the hip. The radiation therapy was given preoperatively to 64 patients (56%), with a mean dose of 30 Gy (range, 8 to 80 Gy), or postoperatively to 32 patients (28%), with a mean dose of 21 Gy (range, 20 to 30 Gy). Fourteen patients died prior to receiving postoperative radiation therapy, 4 patients had a postoperative complication precluding radiation therapy, and 1 patient did not receive radiotherapy, but the reason for this could not be ascertained from the medical record. Furthermore, 22 patients (19%) underwent embolization of the tumor 24 hours prior to the surgical procedure, most commonly for renal cell carcinoma (17 patients [15%]).

All reconstructions were approached in a similar fashion. All gross tumor was resected, and the pelvic deficiency was assessed. The reconstruction was then performed on basis of the tumor defect and surgeon preference (Figs. 1-A through 1-I). At 1 center, the reconstruction was performed as a combined case with orthopaedic oncologists and adult reconstructive surgeons, and, at the other center, it was performed by orthopaedic oncologists alone. In the tantalum reconstructions, the senior author (D.G.L.) performed 70% (26 patients) of the acetabular reconstructions. During this time period, only tantalum reconstructions were performed at 1 center and only Harrington reconstructions were performed at the other center.

Figs. 1-A through 1-I Images of a patient with pathologic fracture with pelvic discontinuity through a metastatic adenoid cystic carcinoma.

Fig. 1-A.

Anteroposterior pelvic radiograph showing the pathologic fracture with pelvic discontinuity through a metastatic adenoid cystic carcinoma.

Fig. 1-B.

Photograph showing intralesional curettage. Because of pain and the pathologic fracture, the patient underwent a complex total hip arthroplasty. Following exposure of the acetabulum, a large acetabular defect was noted and the area underwent intralesional curettage to remove all gross tumor.

Fig. 1-C.

Photograph showing the defect filled with tantalum wedge augments and bone graft. After the intralesional curettage, the acetabulumwas reamed, and the defect was sized and was filled with tantalum wedge augments and bone graft.

Fig. 1-D.

Photograph showing the construct, which was unitized with a bead of cement placed between the augments and the acetabular shell, which was impacted into place and was fixed with multiple screws.

Fig. 1-E.

Photograph showing an over-the-top half-cage, which was also added to enhance fixation using screws through the cage and tantalum implant.

Fig. 1-F.

Radiograph showing the cemented polyethylene and the placement of the femoral component.

Fig. 1-G.

Positron emission tomography (PET) showing that the patient had diffuse disease progression and involvement around the hip.

Fig. 1-H.

Anteroposterior radiograph showing that the implant remained well fixed.

Fig. 1-I.

Lateral radiograph showing that the implant remained well fixed.

All reconstructions utilized multiple screws (mean, 6 screws [range, 2 to 15 screws]) or pins (mean, 7 pins [range, 3 to 10 pins]) depending on the type of reconstruction. In the Harrington group, an additional incision (37 patients [47%]) was made over the iliac crest to insert antegrade pins and 45 patients (58%) had a roof ring in addition to a cemented acetabular component. An acetabular antiprotrusio cage was utilized in 21 patients (27%) in the Harrington group and in 20 patients (54%) in the tantalum group. The mean acetabular component size was 54 mm (range, 44 to 68 mm). Three constrained acetabular components were utilized, all in the Harrington group. No dual-mobility components were used. In the Harrington group, all of the femoral components were cemented, and, in the tantalum group, 17 (46%) were uncemented and 20 (54%) were cemented. The mean operative time was 270 minutes (range, 113 to 570 minutes).

The mean follow-up for surviving patients was 4 years (range, 1 to 15 years). Of the 21 patients (18%) alive at the time of the final follow-up, 8 (38%) had survived at least 4 years, and an additional 10 (48%) had reached 2 years of clinical follow-up. Of all patients, 48 (42%) reached 2 years of follow-up.

Statistical Analysis

An unpaired Student t test was used to assess continuous variables, and categorical variables were compared with the Fisher exact test. Survival estimates were calculated using the cumulative incidence method, accounting for additive end points and associated censoring, including implant loosening and failure as well as patient death, as this has been demonstrated to be more appropriate than Kaplan-Meier analysis in competing-event situations. The Hanley rule was used to approximate the 95% confidence interval (CI) around non-observed failure events¹⁵. All tests were 2-sided. Significance was set at p < 0.05. A reoperation was defined as any additional surgical procedure on the hip. Acetabular revision was defined as any surgical procedure in which any component of the acetabular reconstruction was exchanged.

Results

Group Comparison

When comparing patients who underwent Harrington and tantalum reconstructions (Table I), there was no significant difference in the mean age (60 compared with 60 years; p = 0.89), female sex (55% compared with 73%; p = 0.10), ECOG status of ≥3 (50% compared with 38%; p = 0.23), Harrington Class-III acetabular defects (56% compared with 46%; p = 0.32), presence of a pathologic fracture (55% compared with 57%; p = 1.0), or pelvic discontinuity (19% compared with 22%; p = 0.80). Patients undergoing a tantalum reconstruction had a reduced mean operative time (and standard deviation) at 214 ± 73 minutes compared with the Harrington reconstruction group at 318 ± 81 minutes (p < 0.001).

TABLE I.

Demographic Characteristics of Patients Undergoing Periacetabular Reconstruction for Metastatic Disease

Demographic Characteristics	Harrington Reconstruction (N = 78)	Tantalum Reconstruction (N = 37)	P Value
Age* (yr)	60 ± 13	60 ± 14	0.89
Female sex†	43 (55%)	27 (73%)	0.10
ECOG ≥3†	39 (50%)	14 (38%)	0.23
Harrington Class-III acetabulum†	44 (56%)	17 (46%)	0.32
Pathologic fracture†	43 (55%)	21 (57%)	1.0
Pelvic discontinuity†	15 (19%)	8 (22%)	0.80
Operative time* (min)	318 ± 81	214 ± 73	<0.001

^*The values are given as the mean and the standard deviation.

^†The values are given as the number of patients, with the percentage in parentheses.

Overall Patient Survival

Ninety-four patients died of systemic disease progression at a mean of 16 months (range, 1 day to 9 years) postoperatively. The mean overall patient survival was 34% at 2 years, 16% at 5 years, and 15% at 10 years. There was no significant difference in the cumulative incidence of death at 10 years between patients in the Harrington group (85.4% [95% CI, 74.6% to 93.3%]) and the tantalum group (85.4% [95% CI, 70.9% to 95.1%]) (p = 0.30) (Fig. 2, Table II). Disease progression around the acetabulum occurred in 25 patients (22%), more commonly in patients with a tantalum reconstruction (13 patients [35%] compared with 12 patients [15%]; p = 0.02).

Fig. 2.

Cumulative incidence of death due to disease and acetabular component loosening for aseptic causes in the Harrington reconstruction and tantalum groups. Although the 10-year rate of death due to disease was similar, patients who underwent a Harrington reconstruction were more likely to undergo a revision of the acetabular component due to loosening (p = 0.09).

TABLE II.

Cumulative Incidence Estimates of Death over Time^*

Reconstruction Type	1 Year	2 Years	5 Years	10 Years
Tantalum	40.5% (26.8% to 58.0%)	65.3% (50.0% to 80.0%)	81.8% (66.9% to 92.8%)	85.4% (70.9% to 95.1%)
Harrington technique	48.8% (37.7% to 59.8%)	66.6% (55.7% to 77.0%)	85.4% (74.6% to 93.3%)	85.4% (74.6% to 93.3%)

^*The values are given as the cumulative incidence, with the 95% CI in parentheses.

Reoperation, Acetabular Component Survival, and Complications

An additional procedure was performed in 24 patients (21%) at a mean time of 15 months (range, postoperative day 1 to 7 years) postoperatively, most commonly for dislocation (7 patients), acetabular loosening (5 patients), and infection (4 patients) (Table III). Of the patients who underwent a reoperation, there were 21 (27%) in the Harrington group and 3 (8%) in the tantalum group. Patients undergoing a Harrington-style reconstruction were more likely to require a reoperation (hazard ratio [HR], 4.59 [95% CI, 1.57 to 19.52]; p = 0.003). Patients undergoing a reoperation were noted to have a mean longer initial surgical procedure at 305 ± 96 minutes compared with those not undergoing a reoperation at 269 ± 92 minutes (p = 0.16).

TABLE III.

Postoperative Outcome Following Total Hip Arthroplasty for Metastatic Disease of the Acetabulum

Outcome	Harrington Reconstruction*† (N = 78)	Tantalum Reconstruction* (N = 37)	P Value
Local disease progression	12 (15%)	13 (35%)	0.02
Reoperation
All-cause	21 (27%)	3 (8%)	0.02
For dislocation	6 (8%)	1 (3%)	0.42
For deep infection	2 (3%)	2 (5%)	0.59
For pin migration	3 (4%)	0 (0%)	0.54
Acetabular revision
All-cause	10 (13%)	3 (8%)	0.54
For loosening	5 (6%)	0 (0%)	0.17
For instability	5 (6%)	1 (3%)	0.66
For infection	1 (1%)	2 (5%)	0.24

^*The values are given as the number of patients, with the percentage in parentheses.

^†One patient in the Harrington group underwent revision for combined loosening and instability.

Of these reoperations, a revision of the acetabular component was performed in 13 patients (11%) at a mean time of 14 months (range, 2 weeks to 7 years). Indications for revision included recurrent instability (4 patients), acetabular loosening (4 patients), infection (3 patients), and combined recurrent instability and acetabular loosening (1 patient). Instability was treated with conversion to a constrained liner (4 patients) and by revision of the cemented polyethylene to a more anteverted position and increasing the femoral-head size (1 patient). Patients with infection (3 patients) were treated with irrigation and debridement, exchange of the femoral head, and chronic antibiotic suppression. Acetabular revisions occurred in 10 patients (13%) in the Harrington group and in 3 patients (8%) in the tantalum group (Table III). There was no significant difference in overall acetabular revision rates (HR, 1.99 [95% CI, 0.60 to 8.93]; p = 0.27).

When specifically examining acetabular component loosening, 5 patients (4%) had loosening of the acetabular component necessitating a revision surgical procedure. In 2 of these patients, there was also disease progression in the acetabulum. These patients were treated with repeat curettage of the tumor, cementation, and acetabular component revision. In the other 3 patients, 2 had fracture of the acetabulum in the setting of irradiated bone leading to loosening of the roof ring and cement and the third underwent reconstruction in the setting of an irradiated pelvic discontinuity and had failure of the revision cage. All of these patients were in the Harrington reconstruction group. Following the revision procedure, there were no further revisions of these components. The 10-year cumulative incidence of revision (Fig. 2, Table IV) for acetabular component loosening was 9.6% (95% CI, 4.1% to 21.7%) in the Harrington group compared with 0% (95% CI, 0% to 8.1%) in the tantalum group (p = 0.09).

TABLE IV.

Cumulative Incidence Estimates of Acetabular Loosening over Time^*

Reconstruction Type	1 Year	2 Years	5 Years	10 Years
Tantalum	0.0% (0.0% to 8.1%†)	0.0% (0.0% to 8.1%†)	0.0% (0.0% to 8.1%†)	0.0% (0.0% to 8.1%†)
Harrington technique	0.0% (0.0% to 3.8%†)	2.8% (0.7% to 10.7%)	7.1% (2.7% to 18.4%)	9.6% (4.1% to 21.7%)

^*The values are given as the cumulative incidence, with the 95% CI in parentheses.

^†In areas in which there were no failures, the bottom limit of the 95% CI was 0 because there was no variance in the group (i.e., none of the implants were revised and all implants survived).

In addition to the operative complications listed above, nonoperative complications included deep venous thrombosis (3 patients), superficial infection (3 patients), and sciatic nerve palsy (2 patients). There was no difference in the incidence of complications between the 2 groups (p = 1.0).

Functional Outcome

The mean Harris hip score (HHS) was 31 points (range, 4 to 77 points) prior to a surgical procedure and improved following reconstruction (p = 0.0001) to 67 points (range, 19 to 93 points) at the last follow-up. There was no significant difference (p = 0.29) in the mean preoperative or postoperative HHS between patients undergoing the Harrington technique (68 ± 14 points) and those undergoing a tantalum reconstruction (65 ± 16 points) at the most recent follow-up. Patients with a preoperative ECOG status of 4 also had a reduced mean HHS (56 ± 5 points) compared with those with ECOG status levels of 1 (71 ± 3 points), 2 (66 ± 3 points), or 3 (67 ± 3 points) (p = 0.08).

Discussion

Since 1981, the Harrington technique has become the primary surgical method to reconstruct osseous defects associated with periacetabular metastatic disease. This type of reconstruction provides patients with an immediately stable and secure hip construct and pain relief and allows unrestricted weight-bearing. However, newer techniques using tantalum revision acetabular components have shown promise in providing an uncemented option for a variety of challenging reconstructions, including the settings of metastatic disease and irradiated bone. The results of the current series show that tantalum components provide similar functional outcomes compared with a Harrington-style reconstruction; however, the incidence of acetabular component loosening was reduced with tantalum reconstructions, and operative time was also reduced.

Previous implant survivorship studies examining the outcome of the Harrington technique have shown mechanical failure rates ranging from 0% to 9.1%^1,4,6-8. In the current series, we noted a similar rate of failure of the acetabular component. However, when considering only patients with aseptic acetabular loosening without tumor recurrence, the loosening rate was only 2.6% (3 patients), with all failures occurring after the mean time to death of patients in this series of 16 months. With newer treatments for metastatic disease and patients living longer with many types of metastatic cancer, there is concern that acetabular reconstructions in this setting need to last longer, and, as such, a reconstruction that provides a bone ingrowth component might be preferred. In the current series, we did not observe frank mechanical loosening of any of the acetabular implants in the tantalum group.

Local tumor control is an important goal in treating periacetabular metastatic disease. Previous series have shown local progression of 25% using Harrington-style reconstruction, which can lead to mechanical failure of the constructs¹. In the current series, there was a higher rate of progression in patients with a tantalum reconstruction; however, this did not lead to acetabular failure (Figs. 1-A through 1-I). The lack of failure could be due to spot-welding around the acetabular component; in addition, the judicious use of multiple screws, porous augments, and cup-cage techniques enhances fixation in the bone¹⁶.

Although reconstruction of the acetabulum for metastatic disease is considered palliative and often requires extensive surgical procedures, it generally provides sufficient and reliable improvement in pain and function to be considered worthwhile for appropriately chosen patients^1,17. In the current series, acetabular reconstruction improved the functional status of patients at every functional level, regardless of the type of reconstruction undertaken. Patients with a poor performance status preoperatively, for example, based on an EGOC status of 4, also had a reduced HHS at the time of the final follow-up, which is similar to previous reports¹. Although these patients are more likely to have worse functional outcomes following a surgical procedure, reconstruction still provides them with improved pain relief, ambulatory status, and quality of life.

The Harrington technique was designed to transmit the forces imparted on the reconstructed acetabulum to the more proximal intact host bone using screws or pins⁴. In this series, contact with the host bone using a porous tantalum acetabular component or the cemented Harrington technique was accomplished with an oversized acetabular component relative to the local anatomy (mean diameter, 54 mm [range, 48 to 68 mm]) and multiple screws and augments in the tantalum group or multiple pins into the ilium in the cemented Harrington group. The results of this series highlight the potential advantage of tantalum reconstruction (compared with the cemented Harrington technique), which has the ability to provide durable long-term fixation from osseointegration compared with the cemented technique; however, it is possible that, with longer follow-up, there may be failure of these tantalum components. In addition, there was a 2:1 ratio of Harrington-style reconstructions compared with tantalum reconstructions; as such, with more patients treated with tantalum, failures of the acetabular component could occur.

Although the use of highly porous tantalum is an attractive option to treat these patients, there are contraindications to its use in certain patients, and, in such cases, a cemented Harrington reconstruction could potentially be utilized. A complete vertical fracture of the ilium, leading to discontinuity between the acetabulum and the posterior aspect of the pelvis, and/or sacropelvic discontinuities are contraindications to the use of a tantalum reconstruction. As such, it is important to carefully evaluate the preoperative radiographs and computed tomographic (CT) scan of the pelvis for these fractures and, if present, consider a different reconstruction technique.

Complications following a pelvic tumor surgical procedure are common, with previous series examining the Harrington style of reconstruction noting complications rates of 20%¹. Although complications are likely multifactorial and related to these large procedures and adjuvant treatments, the total operative time of the procedure is known to be a risk factor for complications such as infection¹⁸. Although the 2 groups of patients in this series were similar in terms of the complexity of the cases (presence of a pathologic fracture and Harrington defect classification), patients undergoing a tantalum reconstruction were noted to have a shorter operative time. Likely this is related to the need for additional incisions, pin insertion, and the time needed for the cement to cure on the femoral and acetabular sides in the Harrington group. These procedures require a skilled team of orthopaedic oncologists and reconstructive surgeons. As such, they should be performed at centers with appropriate personnel, resources, and experience to reduce operative time and maximize patient outcomes.

This study did have some specific limitations. The retrospective nature of the study limited the data that could be collected from the medical record; however, the prospective data collection of the 2 centers’ registries helps to reduce recall bias. During this time period, these were the only reconstructive arthroplasty techniques performed at the centers, which helps to reduce selection bias. In addition, multiple subspecialty-trained surgeons were involved with the reconstructions and, as such, there was no standardized method to the reconstruction; however, the reconstructions were approached in a similar fashion. In addition, we are unable to comment on the number and outcome of patients treated without a surgical procedure during the same time period and how they compared with the surgically treated patients in the current series. Finally, tumor size was not accurately documented prior to reconstruction and therefore a direct comparison of this factor between the 2 centers was not able to be undertaken. In our sample of 37 tantalum reconstructions, although no failures were observed, leading to an estimated 0% cumulative incidence of loosening, the upper bound of the 95% CI for failure would be expected to be up to 8.1%, highlighting the role for future investigations with larger sample sizes.

Overall, the use of porous tantalum acetabular components or the Harrington technique provided patients with a reliable means of reconstruction in the setting of metastatic periacetabular disease. Because both techniques provide a reliable treatment option, the treating surgeons should use the reconstruction technique with which they are most comfortable. However, for patients with extended life expectancy, the use of a tantalum reconstruction should be considered because of the durable implant fixation and lower complication rates.