Compress® Knee Arthroplasty Has 80% 10-year Survivorship and Novel Forms of Bone Failure
- First Online:
- Cite this article as:
- Healey, J.H., Morris, C.D., Athanasian, E.A. et al. Clin Orthop Relat Res (2013) 471: 774. doi:10.1007/s11999-012-2635-6
- 1.5k Downloads
Compliant, self-adjusting compression technology is a novel approach for durable prosthetic fixation of the knee. However, the long-term survival of these constructs is unknown.
We therefore determined the survival of the Compress® prosthesis (Biomet Inc, Warsaw, IN, USA) at 5 and 10 actuarial years and identified the failure modes for this form of prosthetic fixation.
We retrospectively reviewed clinical and radiographic records for all 82 patients who underwent Compress® knee arthroplasty from 1998 to 2008, as well as one patient who received the device elsewhere but was followed at our institution. Prosthesis survivorship and modes of failure were determined. Followup was for a minimum of 12 months or until implant removal (median, 43 months; range, 6–131 months); 28 patients were followed for more than 5 years.
We found a survivorship of 85% at 5 years and 80% at 10 years. Eight patients required prosthetic revision after interface failure due to aseptic loosening alone (n = 3) or aseptic loosening with periprosthetic fracture (n = 5). Additionally, five periprosthetic bone failures occurred that did not require revision: three patients had periprosthetic bone failure without fixation compromise and two exhibited irregular prosthetic osteointegration patterns with concomitant fracture due to mechanical insufficiency.
Compress® prosthetic fixation after distal femoral tumor resection exhibits long-term survivorship. Implant failure was associated with patient nonadherence to the recommended weightbearing proscription or with bone necrosis and fracture. We conclude this is the most durable FDA-approved fixation method for distal femoral megaprostheses.
Level of Evidence
Level IV, therapeutic study. See Instructions for Authors for a complete description of levels of evidence.
Megaprostheses need improved bone fixation to reduce the rate of aseptic loosening associated with stemmed implants. Young patients cured of tumors have a long life expectancy and a compelling need for prosthetic fixation that is equally long-lasting. A recently developed strategy is compliant compressive fixation that uses compression, via a short traction bar, to stimulate osteointegration at the bone-prosthetic interface, promote hypertrophy of the loaded bone, and avoid stress bypass of the host bone around a stiff intramedullary stem . The Compress® Compliant Pre-Stress Implant (Biomet Inc, Warsaw, IN, USA), a rotating-hinge knee prosthesis, was approved by the FDA based on data from an unpublished short-term feasibility study, conducted by the manufacturer, that showed no difference in the acute complication rate and equivalent functional outcome scores compared with a cemented stem coupled to the same rotating-hinge articulation (Orthopaedic Salvage System [OSS™]; Biomet). Published studies of this device include an investigation in 26 patients, among whom only 10 had followup longer than 24 months , a study of 26 patients followed for a period of 0.3 to 9.2 years , and a study of 41 patients followed for 3 to 97 months . These studies suggest projected 10-year prosthetic survival is at least 80%, but the number of cases is small and the number followed for this duration is miniscule.
We therefore determined the survival of the Compress® prosthesis at 5 and 10 actuarial years and identified the failure modes for this form of prosthetic fixation. Finally, the results were compared with those reported in a comprehensive review of the literature to establish the superiority of this method of fixation compared to those previously reported.
Patients and Methods
Contraindications for use of the Compress® device for knee arthroplasty
Cortical thickness of less than 2.5 mm
Pre- or postoperative bone irradiation, precluding osteointegration
Extraarticular resection of knee (an articulated implant, such as the Burstein-Lane® implant, would be indicated)
Inadequate or unreconstructable soft tissue envelope (a very low-profile implant, such as the GUEPAR® implant, would be indicated)
Metastatic disease that mandates immediate weightbearing (precludes the requisite 3 months of protected weightbearing)
Inability to cooperate with the postoperative program of early, protected weightbearing
Patient demographic and clinical characteristics
Number of patients
Sex (number of patients)
Reconstruction surgery (number of patients)
Tumor diagnosis (number of patients)
High-grade osteogenic sarcoma
Giant cell tumor
Low-grade osteogenic sarcoma
No tumor (arthroplasty revision)
All patients underwent a similar rehabilitation regimen. Continuous passive motion was initiated on Postoperative Day 2 if there was no visible evidence of wound necrosis and continued for approximately 2 weeks for approximately 18 hours/day. Patients started walking on the first postoperative day using toe-touch weightbearing for 6 weeks, 50% weightbearing for an additional 6 weeks, and then progressive weightbearing as tolerated. All were fully weightbearing within a month. Patient adherence to these guidelines was presumed. However, at least one patient did not comply with recommended weightbearing proscription and sustained a periprosthetic fracture and implant failure after carrying a boat, necessitating revision. Chemotherapy was resumed 2 to 3 weeks postoperatively when appropriate for a patient’s diagnosis.
Patients were seen on a variable schedule based on the patient’s diagnosis and disease activity. For high-grade cancers, this was initially every 2 months and ultimately once per year after a 4-year disease-free interval. Patients with benign disease were seen every 3 to 6 months initially and then annually after 4 years. Standard AP and lateral radiographs were obtained at each visit. Radiographs were examined for any deformation of the implant that suggested bending or breaking of the device or fracture of the bone. Any new or increased pain was noted. Radiographic signs of loosening were not specifically quantified because, to our knowledge, there is no suitable methodology for evaluating loosening for this construct. Compress® fixation failure was defined as revision of the fixation mechanism (anchor plug, traction bar, spindle, sleeve, or fixation pin) for any reason. Revision of intraarticular components (eg, polyethylene tibial bearing) that did not affect the bony fixation was not counted as a fixation or implant failure. Symptoms and signs of periprosthetic bone failure were noted and typically manifested as patient complaints of thigh pain and tenderness at the spindle-bone junction or reports of thigh pain during examination when the hip was rotated in the 90–90 position (supine, hip flexed 90°, knee flexed 90°). In such instances, radiographs were examined for evidence of a lack of osteointegration (ie, lack of hypertrophy in the bony segment between the spindle and the anchor plug and the presence of radiolucent lines at the spindle-bone interface). The presence of implant, bone, or symptomatic worsening prompted surgery.
In three patients who underwent revision surgery, we analyzed the bone adjacent to the junction site, both by visual inspection of the interface during surgery and by microscopic analysis of standard hematoxylin and eosin staining of decalcified samples.
Prosthesis-independent complications occurred but seemingly at a rate similar to what we have observed for other joint megaprostheses. There were three local recurrences; two were managed by local excision that did not affect the prosthesis and one required an amputation that removed the intact fixation. There were five prosthetic infections, including four primary infections and one secondary infection. The primary infections were successfully treated with washout, a change of intercalary segments, and retention of the Compress® fixation; there were no recurrences of infection. The secondary infection necessitated amputation that included removal of the intact prosthesis. The intact implants removed by the two amputations were considered censored at the time of the removal.
We performed survival analysis of the device by the Kaplan-Meier log-rank technique using SPSS® Version 14.0 (SPSS Inc, Chicago, IL, USA). Survival was defined as the time from the date of surgical implantation to the date of prosthesis removal or latest followup.
A comprehensive review of the literature was performed to place our results in context. A total of 718 articles were systematically read and reviewed. These were selected by performing a PubMed search on July 14, 2012, using the following key words: “megaprosthesis”, “femoral prosthesis”, “knee replacement”, and “tumor”. Articles that reported on the results of at least 20 patients for a mean of 5 years’ followup were considered. The results were further stratified based on studies that specified the number of distal femoral resections/reconstruction, the diagnoses for the surgical indications, and the prosthetic survivorship.
Compliant, self-adjusting compression technology is a novel approach for durable prosthetic fixation of the knee. Early results have been encouraging, but longer followup reports are required from different centers. We therefore determined the survival of the Compress® prosthesis at 5 and 10 actuarial years and identified the fracture-associated failure modes for this novel form of prosthetic fixation.
Several factors limit the interpretation of this study. First, the absence of a control group makes it impossible to compare results directly with fixation from conventional cemented or uncemented intramedullary stems. This study does not prove this fixation method is better than historic options, even though the authors believe this is generally true. Previously, we had hoped to address this question through a prospective study protocol that would have compared Compress® fixation with press-fit and cemented stems. The protocol was proposed to a musculoskeletal oncology society, but the option of randomizing patients to the Compress® was rejected by the society’s surgeons; hence, definitive comparative conclusions may never be possible. However, in this study, we noted all 18 patients who were revised to a Compress® explicitly stated they were more comfortable with this implant, suggesting it was more stable and well fixed. Second, we only used the compliant fixation with a single design of rotating-hinge knee arthroplasty. Webber and Seidel  recently reported combining compliant fixation with a different body and articular design for pediatric limb salvage. Although the results are unlikely to have been different if a different knee design had been used, this question cannot be answered by this study. Patients in our cohort were treated over the course of a decade, and unrecognized differences in the population or surgical technique could have occurred. This problem plagues reviews of all low-incidence conditions that require many years to accumulate enough cases for analysis. Third, the minimum followup was set at only 1 year to allow inclusion of the two Type IIB cases of periprosthetic bone failure. The Kaplan-Meier method and survival curves allow the reader to see the time course of fixation failures and how the duration of followup may affect the prosthetic survival.
Summary of peer-reviewed literature reporting megaprosthesis survivorship
Number of patients
Implant survival (%)
Unwin et al. 
Langlais et al. 
GUEPAR® II or custom press-fit cemented
Myers et al. 
Zimel et al. 
Howmedica 39 OSS™ 8
Farfalli et al. 
Shehadeh et al. 
Bergin et al. 
Tan et al. 
Roberts et al. 
Horowitz et al. 
Kawai et al. 
Griffin et al. 
Bruns et al. 
Kinkel et al. 
Matsumine et al. 
Ritschl et al. 
Unwin et al. 
Unwin et al. 
Mascard et al. 
Mittermayer et al. 
Plötz et al. 
Bickels et al. †
Biau et al. 
Morgan et al. 
The unique fixation method of this prosthesis showed a unique spectrum of failure mechanisms. Aseptic loosening, commonly reported with other cemented and uncemented prostheses, also occurred with this implant. However, the aseptic loosening differed from that seen with other implants since bone ingrowth failed despite the continuously adjusting compression generated by the Belleville washers in the compression chamber. Retrieval specimens of these failures showed avascular necrosis of the underlying bone, in distinction to the viable bone found in well-fixed implants that were explanted for other reasons such as infection or tumor recurrence [7, 14]. A second, perhaps related unique finding was fracture or crumbling of the underlying bone between the anchor plug and the spindle. This was present in one patient who was included as part of a multicenter report on periprosthetic fractures around Compress® devices . The phenomenon has not been singled out for analysis. It could not be determined whether the osteonecrosis led to fatigue failure of the bone or the fracture caused osteonecrosis near the interface. The pathophysiology of these failures is unproven. Treatment of fractures related to prosthetic failure was not the focus of this study but is reportedly relatively easy and yields pain-free, functional reconstructions with few complications [1, 29].
Our analysis demonstrates a survivorship of 80% for Compress® knee arthroplasty; the only published report demonstrating better survivorship after 10 years is that of Langlais et al. , who utilized custom-made press-fit femoral revision stems in 20 of the 26 joint arthroplasties. Thus, this report is the most comprehensive to date on an FDA-approved device for this unique form of prosthetic fixation.
The authors thank the Major Fellowship in Musculoskeletal Oncology, the Pearlman Oncology Fund, and the Limb Preservation Fund for providing financial support.
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.
Supplementary material 1 (MPG 20372 kb)