Implant failure of the Compress prosthesis: a case report

Abstract

Background

The Compress is designed to achieve bone formation and stability by applying pressure at the bone–implant interface, minimizing the likelihood of aseptic loosening, which is a complication of stem implants. Herein, we report two cases of implant failure using the Compress.

Case presentation

Case 1 describes a 36 year-old Japanese man who underwent extraarticular tumor resection, Compress arthroplasty, and reconstruction with a gastrocnemius flap after preoperative chemotherapy for a secondary malignant giant cell tumor in the right distal femur. Postoperatively, partial weight-bearing was started at 6 weeks, and full weight-bearing was allowed at 10 weeks. One year after the surgery, a fall caused implant failure. No bone formation at the implant–bone interface was observed on radiographs immediately prior to the failure. Bone formation was achieved at the interface 1 year after revision arthroplasty, and the patient was able to walk unassisted with a brace.

Case 2 describes a 14 year-old Japanese boy who underwent wide surgical resection of osteosarcoma in the left tibia, Compress arthroplasty, and reconstruction with a gastrocnemius flap after preoperative chemotherapy. The postoperative weight-bearing schedule was the same as that of case 1. One year after the surgery, the patient experienced implant failure. A revision arthroplasty was performed. One year after revision surgery, the patient was able to walk unassisted.

Conclusion

Although the risk factors for Compress failure remain unknown, it is important to consider patient characteristics that may inhibit bone formation, implant selection, postoperative loading timing, and radiographs of bone formation at the implant interface when using the Compress.

Background

Primary reconstruction with a stemmed distal femoral megaprosthesis has become the most widely used reconstruction strategy after wide resection of malignant bone tumors around the knee [1]. However, with the growing number of young active survivors, long-term reconstructive failure presents in most patients. The most common reasons for failure include femoral stem loosening and infection with hardware failure [2,3,4]. Failures of stemmed megaprosthesis leave surgeons with few reconstructive solutions and present a challenge. Severe bone deficiency after the failure or loosening of a prior stem traditionally results in the revision of a cemented or large-diameter press-fit-stemmed implant, allograft–prosthetic composite reconstruction, or total femur replacement [5]. Short residual bone segments cannot be addressed by these options, and the constructs have high failure rates, often exceeding 30% at 5 years [4]. The current strategy needs to be improved to preserve the remaining bone stock as much as possible, and a durable construct must be created for revision surgery.

Compress (Biomet Inc., Warsaw, IN, USA) uses a different design to achieve bone ingrowth fixation and is intended to minimize stress shielding and osteolysis [6]. It uses stored energy from Belleville washers to provide compliant compression through a short traction bow device, which promotes bone ingrowth at the bone–prosthetic interface and induces new bone formation at the intervening cortex. The forces exerted on the involved extremities are transmitted directly from the implant to the host bone, thereby eliminating stress shielding. During osseointegration, the medullary canal is sealed and protected from wear particles that may induce osteolysis. Postoperative survival rates associated with the Compress have been estimated at > 80% over 10 years [6]; however, few reports of implant failure exist [7,8,9]. For recent stem-type tumor megaprosthesis, the all-cause 10-year implant survival rate has been reported at approximately 50% [5, 10]. The causes of megaprosthesis failure were classified according to Henderson et al. [11]: type 1 (soft tissue failure), type 2 (aseptic loosening), type 3 (structural failure), type 4 (infection), and type 5 (tumor progression). Stem breakage occurs in 5.4% of the cases, and stem size of ≤ 11 mm may be a risk factor for breakage [12]. Meanwhile, Compress fractures are considered to be type 1–3 injuries [6] and have been reported as occurring relatively early [13]; however, little is known about the risk factors associated with fractures [7].

Compared with traditional stem-type implants, the Compress has a significantly lower risk of aseptic loosening over the long term. Additionally, its shorter stem length is expected to preserve considerable bone stock in the event of a revision. However, as the Compress is designed to promote bone formation and achieve stability through the compression of the bone prosthetic interface (BPI), its use is limited in patients with bone fragility who may not withstand this compression. Such patients include those with osteoporosis, the elderly, smokers, and individuals with very thin cortical bone. At our institution, the Compress is the first choice for bony reconstruction after wide resection of malignant bone tumors in young patients and is expected to result in a healthy bone. We report two cases of relatively early implant failure after the resection of a malignant bone tumor and reconstruction using the Compress.

Case presentations

The study was approved by the Ethics Committee at our hospital (2015–03586850), and informed consent was provided by all patients. It adhered to any reporting/ethical guidelines.

Case 1 is about a 36-year-old Japanese man who underwent surgery for a giant cell tumor of the right distal femur using curettage and filling with cement for the defect (Fig. 1A, B). He had no remarkable medical history but was obese, with a body mass index (BMI) of 40 kg/m². After 3 years of postoperative follow-up, the patient refused to attend further follow-up appointments at his own discretion. Seven years after surgery, the patient experienced right femoral pain. A recurrence was suspected on the basis of radiography and magnetic resonance imaging (MRI) findings (Fig. 1C, D). An incisional biopsy was performed at the bone lesion, which was histologically diagnosed as a secondary malignant giant cell tumor. Preoperative chemotherapy with methotrexate, doxorubicin, and cisplatin (MAP) was administered before surgery. After extraarticular tumor-wide resection, the affected limb was reconstructed using the Compress (Fig. 2A). A tibial osteotomy was performed 4 mm from the original resection, and a 10-mm block was added to the 12-mm implant. The femur was osteotomized at 125 mm, and a 10-mm osteotomy was performed. The diameter of the femur at the osteotomy site was 32 mm, with the cortical thickness at the narrowest part measuring 4 mm. No intraoperative problems occurred with reaming or other manipulations. A 20-mm centering sleeve and a short anchor plug were selected to maintain the bone stock. The diameter of the osteotomy surface was measured; a 38-mm spindle was selected, the device force was set at 800 IBM, and an anti-rotation pin was inserted. Following implant insertion, plastic surgeons covered the defect with a gastrocnemius flap, and the surgery was completed. The patient was immobilized with a knee brace and unloaded postoperatively. After confirming wound healing, adjuvant chemotherapy with MAP was initiated 2 weeks after surgery. Partial weight bearing of 25% was permitted at 6 weeks postoperatively. It was increased gradually by 25% per week, with full loading permitted from 10 weeks postoperatively. Range-of-motion exercises of the right knee were also initiated 6 weeks postoperatively. However, patients sometimes struggled to adhere to the behavioral restrictions we set, and compliance was relatively poor. Adjuvant chemotherapy with MAP was completed 16 weeks postoperatively. Sixteen weeks after surgery, the patient experienced discomfort in the right knee, and radiography showed a slight varus deformity at the base of the anchor plug, but no obvious breakage was observed (Fig. 2B). The patient became aware of knee instability. After the completion of postoperative chemotherapy, the patient was able to work as the manager of a campground. However, 13 months after the surgery, the patient fell and was referred to the emergency department because of pain in the right knee. Radiography showed implant failure at the base of the anchor plug (Fig. 2C). Two weeks after the injury, revision arthroplasty was performed. Revision surgery was performed via an osteotomy proximal to the anchor plug, and a standard anchor plug was used to replace the prosthesis (Fig. 3A, B). After 2 weeks of immobilization with a knee brace, the patient was fit with a long-leg orthosis. Partial loading commenced after 6 weeks of unloading, with full loading permitted 10 weeks postoperatively. A total of 1 year and 2 months postoperatively, bone formation at the bone–implant interface was vigorous on radiography, and the patient could walk unassisted with the use of a brace and return to work on the campground (Fig. 3C).

Fig. 1

Case 1: A 36 year-old man underwent surgery for a giant cell tumor in the right distal femur. A Preoperative coronal computed tomography (CT) image. B Coronal CT images after curettage and cement filling. C Coronal CT images showing recurrence (arrow). D Axial T2-weighted magnetic resonance imaging scan of the recurrence site (arrow)

Fig. 2

Case 1: A Radiograph taken immediately after the initial surgery with the Compress. B Radiographs taken 16 weeks after the initial surgery. A slight varus deformity is seen with evidence of suspected failure (arrow). C Radiographs obtained immediately after Compress failure

Fig. 3

Case 1: A Radiographs taken immediately after revision surgery. B Intraoperative findings at the time of revision. C Radiographs taken 14 months after surgery. Vigorous bone formation is observed at the bone–implant interface (arrows)

Case 2 is about a 14-year-old Japanese boy with osteosarcoma of the left tibia who underwent wide resection after neoadjuvant chemotherapy with MAP and reconstruction with a prosthesis using the Compress (Fig. 4A, B). He had no remarkable medical history but had a BMI of 25 kg/m². During surgery, the tumor was widely resected. The tibial osteotomy was performed 150 mm from the joint surface, with the diameter of the tibia at the osteotomy site measuring 25 mm and the cortical thickness at the narrowest part measuring 3.7 mm. Because tumor invasion into the posterior cruciate ligament (PCL) was suspected, a femoral osteotomy was performed 50 mm proximal to the articular surface, and the fibula was resected together. For bone reconstruction, the tibial and femoral components were measured. A short anchor plug was selected, and a spindle of 30 mm was chosen considering the small amount of surrounding soft tissue, although 38 mm was appropriate on the basis of the diameter of the osteotomy surface. The device force was 800 IBM, and an anti-rotation pin was inserted. After the implant insertion, the iliotibial ligament was harvested to reconstruct the separated patellar tendons. Finally, the defect was covered with a gastrocnemius flap and a split-thickness skin graft by a plastic surgeon, and the surgery was completed. Adjuvant chemotherapy with MAP was initiated 2 weeks after surgery. Partial weight bearing was allowed at 6 weeks, and full-weight bearing was allowed at 10 weeks postoperatively. At 8 months postoperatively, bony atrophy was noted on the medial side of the implant-bone interface (Fig. 4C). Nine months after the surgery, he experienced knee discomfort, fell, and visited the emergency department of our hospital. Radiography revealed an implant failure (Fig. 5A). The implant broke at the base of the anchor plug. A total of 2 weeks after the injury, revision arthroplasty was performed. During revision surgery, the tibia was fractured, and a cemented stem was used with supplemental locking plate fixation (Fig. 5B). Half-weight bearing was allowed at 7 weeks postoperatively, and full weight bearing was allowed at 9 weeks postoperatively. The patient was able to walk independently using a brace 7 months after revision arthroplasty.

Fig. 4

Case 2: A 14-year-old boy with osteosarcoma of the left tibia. A Radiographic and coronal T1-weighted fat-suppression contrast-enhanced magnetic resonance imaging scans obtained at the first visit. B Radiographic image obtained after initial surgery. C Radiographs taken 8 months postoperatively showing bone atrophy at the bone–implant interface (arrows)

Fig. 5

Case 2: A Radiograph obtained immediately after implant breakage. B Radiograph obtained after revision surgery

In both cases, the tissue surrounding the failure site was subjected to pathological and culture examinations. Both tests yielded negative results.

Discussion and conclusion

Herein, we report two cases of relatively early postoperative implant failure after the resection of malignant bone tumors that were reconstructed using the Compress megaprosthesis with a short-type anchor plug.

Healey et al. reported the clinical outcomes of 82 patients who had undergone compression knee arthroplasty for bone tumors of the distal femur [6]. The median follow-up period was 43 months, and 28 patients were followed up for > 5 years. The survivorship rates were 85% and 80% at 5 and 10 years, respectively. Eight failures of the interface occurred owing to aseptic loosening alone or aseptic loosening with periprosthetic failures that affected the interface. Five of the eight patients with aseptic loosening had bone failure, characterized by the absence of bone growth into the porous spindle, collapse of the bone–prosthesis interface, and associated failure between the anchor pins and the spindle. The authors classified the modes of bone failure as follows: type I, affecting the interface; type II, not requiring prosthetic revision; type IIA, failure proximal to the implant; and type IIB, bone failure that did not disrupt the interface or extend the anchor plug fixation pins. The patient in case 1 had discomfort in his knee at 16 weeks postoperatively and had been experiencing instability ever since. The radiographic findings suggested that bone formation at the interface was insufficient, resulting in implant failure, which was a phenomenon similar to the type I failure reported by Healey et al. The patient in case 2 also showed bone atrophy at the interface on the radiograph, and we assumed that the failure was caused by the same mechanism as the type I failure. In addition, the spindle size was small relative to the diameter of the tibial osteotomy surface, which may have also contributed to poor bone formation at the bone–implant interface.

However, the risk factors for Compress failure are poorly understood. Although studies on the breakage of the Compress are limited, Kagan et al. reported that the risk factors significantly associated with mechanical failure were the location of the proximal femur and proximal tibia compared with the distal femur. The significant factors associated with overall failure were proximal tibial location and radiation therapy [7]. Goldman et al. reported 121 cases of distal femoral replacement with the Compress [9]. In the Cox proportional hazards model, age at surgery, chemotherapy, BMI, sex, and length of resection were not significantly associated with mechanical failure. In this context, the proximal tibia in case 2 fits the risk profile. In both cases, radiographs taken immediately prior to failure showed insufficient bone formation at the interface; as discussed by Healey et al., insufficient bone formation in this area could have been the cause of the failure. The relatively high BMI of the two patients (case 1: 40 kg/m², case 2: 25 kg/m²) may also have contributed to this failure. Perioperative chemotherapy was administered in both cases, which could have caused inadequate bone formation at the interface. In addition to perioperative chemotherapy, extraarticular resection in case 1 resulted in the removal of a large amount of soft tissue around the knee joint, which may have caused instability, and the patient’s relatively poor compliance with the rehabilitation programme may have contributed to poor bone formation. It is conceivable that, in case 2, the spindle size may have been small relative to the diameter of the osteotomy surface. The small diameter and cortical thickness of the tibia at the osteotomy site may have contributed to poor bone formation at the BPI and fractures during revision surgery. Although clear guidelines regarding the schedule for starting weight bearing are lacking, Tyler et al. reported that partial-weight bearing should be started at 6 weeks with a 25% increase per week, allowing full-weight bearing at 10 weeks [14]. Our patients followed this loading schedule, and no bone formation at the interface was observed at the start of loading. If bone formation is not observed at 6 weeks, it may be possible to delay the start of loading; however, the atrophy of the affected limb and a delay in rehabilitation due to delayed loading may also be a concern. The appropriate timing for the start of the loading should be fully considered for each case and examined in future studies.

Therefore, considering surgical techniques that do not disrupt bone formation is important. Preserving the periosteum and endosteum of the osteotomy surface as much as possible, selecting an appropriately sized central sleeve to avoid lateral instability, inserting an oblique pin according to the thickness of the cortex of the osteotomy surface, setting the appropriate pressure, and selecting an appropriately sized spindle are essential. The choice of anchor plug is also considered important. In cases with minimal residual bone after tumor resection, a short-type plug is the only option; otherwise, a choice must be made between the standard type and short type, which have different lengths and shapes. According to a Food and Drug Administration report from 2015 to October 2022, breakage occurred in 4 out of 855 cases of the standard-type and in 20 out of 972 cases of the short-type plug (0.47% versus 2.06%, respectively, p < 0.01) [15]. In two of our failure cases, short-type anchor plugs were used at the time of the initial surgery. Further research is required to elucidate the use and indications of these types of anchor plugs. However, breakage risk should be managed when using a short anchor plug.

Demonstrating the association between the Compress breakage and short-type anchor plugs in the two cases in this report has its limitations. Although the factors leading to the Compress failure remain unknown, careful follow-up, including setting the timing of the start of weight-bearing, is necessary in patients with reported risk factors and in patients in whom bone formation at the bone–implant interface is expected to be poor. It is also important to select appropriate components according to the cortical thickness and diameter of the osteotomy surface to avoid failure.