The current treatment of osteosarcoma includes neoadjuvant chemotherapy followed by wide resection of the tumor and postoperative chemotherapy. While amputation of the extremity was once the mainstay of treatment, medical and surgical developments over the last several decades have led to the success of limb-salvage surgery1,2. In the very young patient, this has led to the challenge of achieving limb-length equality. Recently, the introduction of an endoprosthesis, expandable by noninvasive means, has allowed for reconstruction of the affected limb while addressing limb-length discrepancy in skeletally immature patients3.
The latest generation of the expandable prosthesis is the Repiphysis expandable limb salvage system (Wright Medical Technology, Arlington, Tennessee), originally manufactured as the Phenix prosthesis. As the prosthesis has been followed for greater than ten years, failures are being described. Aseptic loosening and prosthetic breakage have been reported, but we know of no detailed report of the mechanism of failure of this prosthesis4.
We report the case of a patient who had failure of an expandable prosthesis attributable to internal prosthetic component failure that was salvaged with conversion to an adult modular prosthesis. The patient's family was informed that data concerning the case would be submitted for publication, and they consented.
A fourteen-year-old boy who had had resection of an osteosarcoma in the proximal end of the right tibia at the age of nine years and had reconstruction with use of a proximal tibial replacement, the Repiphysis expandable prosthesis, presented for a routine follow-up evaluation. The patient had undergone two prior limb-lengthenings without complication and had continued growth from the distal femoral physis documented by scanograms. At the time of the latest follow-up, the limb-length discrepancy was <1 cm. The patient complained of slowly progressive pain and a "spongy sensation" with walking. He denied any trauma and had no history of constitutional symptoms such as fever, night sweats, or weight loss.
Physical examination of the right lower extremity revealed no gross deformity. A knee effusion without erythema was present. He had full knee motion with slight crepitus and pain on palpation of the knee. Pistoning of the prosthesis was noted, in addition to varus and valgus instability distal to the knee. There were no sensory or motor deficits, and the remainder of the physical examination was unremarkable.
Laboratory tests, including a complete blood-cell count and serum chemistries, were normal. Radiographs demonstrated an effusion in the right knee as well as a collection of debris adjacent to the tibial prosthesis-bone interface, with the debris extending proximally to the posterior recess of the knee joint (Fig. 1). A computed tomography scan of the chest revealed a 4.5-mm pulmonary nodule, unchanged from prior imaging, in the lateral segment of the right middle lobe. A whole-body bone scan demonstrated increased uptake in the soft tissues about the right knee.
Because of the symptoms, the patient underwent surgical exploration of the limb. Once the skin incision was made through the previous scar, the muscle layer was exposed, revealing the rotational flap of the medial and lateral heads of the gastrocnemius, which had been performed for soft-tissue coverage during the initial reconstructive procedure. The rotational flap had no evidence of necrosis, and the medial part was taken down and retracted posteriorly to reveal the prosthesis.
The muscle around the prosthesis was covered with a dark gray film, extending from the distal part of the femur to the proximal portion of the tibia and to the bone-prosthesis interface (Fig. 2).
The body of the proximal tibial component and the entire distal femoral prosthetic component were removed. The tibial stem, which was well fixed on examination, was left in place and was attached to an adult modular prosthesis, while the distal end of the femur was resurfaced with a modular rotating-hinge femoral component (Fig. 3).
The gray soft tissue and debris surrounding the prosthesis and the removed implants were analyzed. Fibrous connective tissue with a foreign-body giant-cell reaction and pigment-laden macrophages was seen on histological examination, resulting in a final diagnosis of chronic inflammation. No tumor was seen. The casing of the prosthesis was opened, and the proximal portion of the cobalt-chromium spring was found to be broken (Fig. 4). In addition, a ceramic internal component of the implant (isolant ring) was broken into small pieces.
The patient began bearing weight as tolerated postoperatively. He was discharged on the third postoperative day. One month later, he began outpatient physical therapy for passive and active-assisted motion of the right knee. At eight months, the patient was pain-free and was able to walk without assistive devices. The knee motion was 0° to 90° with no extensor lag. Radiographs demonstrated stable, well-aligned femoral and tibial components.
The earliest expandable endoprosthesis widely used in the United States was the Lewis Expandable Adjustable Prosthesis (LEAP), introduced in the 1980s5. The expandable portion of the prosthesis was composed of a hollow, titanium-alloy tube assembled over a threaded shaft. The expansion of the prosthesis was accomplished with a chuck key that turned the screw mechanism, requiring a surgical incision and a short hospitalization. There were numerous reported complications with the LEAP, with one report of twenty-two patients that noted collapse of the prosthesis in four patients and mechanical failure in two patients5.
In the 1990s, designs focused on improving existing modular systems in which new modular segments were periodically inserted to match the growth of the contralateral limb. When lengthening was necessary, a midsection was replaced by a longer midsection, typically adding 2 cm of length; each lengthening required a surgical exposure. In one study of thirty-seven patients with a modular prosthesis, seventeen expansion procedures were performed6. One patient had long-term neurapraxia develop after expansion, and four patients had loosening of the prosthesis. The authors of that study stated that the failure rate of the modular endoprosthesis was similar to that of nonmodular systems.
The endoprosthesis that is expandable by noninvasive means, and was first used in the United States in the late 1990s4, has been proposed as a safer method to achieve limb-length equality following tumor resection. One such type of prosthesis, the Repiphysis (Wright Medical Technology), uses a mechanism that expands when the limb is exposed to an externally applied electromagnetic field, causing the outer polymer tube to soften and unlock from the inner titanium tube. A compressed spring then expands slightly, resulting in lengthening7,8. In a multicenter study of the Phenix prosthesis (now manufactured as the Repiphysis by Wright Medical Technology), eighteen prostheses were implanted in fifteen patients4. At the time of the latest follow-up, the Musculoskeletal Tumor Society (MSTS) functional scores averaged 90%. Eight revisions were performed because of loosening or fracture of the stem, the expandable body, or the nonexpandable component.
A similar endoprosthesis was developed in the United Kingdom and was studied in seven patients9. At a mean follow-up of 20.2 months, the patients had a mean lengthening of 25 mm and a mean MSTS score of 68%. One patient had a flexion contracture develop, and another died of "disseminated disease." There was no implant failure, loosening, or infection. While the mechanism of expansion of this prosthesis differs from that of the Repiphysis, it also is noninvasive.
Experience with the noninvasive expandable prosthesis has been increasing, but the body of literature surrounding these prostheses remains small. In particular, there is a lack of information about mechanisms of failure and salvage of such failures.
The Rephiphysis expansion mechanism is composed of a titanium tube embedded in a polyethylene housing cylinder. One end of the titanium tube has a flared "trumpet," which engages into the polyethylene, "locking" it into place (Fig. 5). Within the titanium tube is a compressed cobalt-chromium spring, which stores energy relative to its length, diameter, and thickness. The spring is insulated by a thin ceramic isolant ring, which functions to electrically isolate the trumpet flare from the spring.
When lengthening is desired, an external electromagnetic field is applied, heating an anular protuberance, which is located at the trumpet flare. This heat softens the outer polyethylene housing cylinder, allowing the trumpet flare to unlock from the polyethylene. The spring then expands slightly, allowing the inner titanium tube to slide within the outer polyethylene housing cylinder. Once the trumpet flare reaches a new and cooler portion of the polyethylene, it is locked back into place, limiting further expansion.
If the implant is subjected to a single catastrophic load or multiple cyclical loads that exceed the limits of the so-called locking mechanism, then repetitive movement may begin to occur between the trumpet flare and the polyethylene housing cylinder. Unless the locking mechanism between these two components is restored by a lengthening that places the trumpet flare into a new region of the polyethylene, the trumpet flare and titanium tube can continue to move vertically within the housing cylinder, further eroding the polyethylene and exacerbating the lack of resistance to a vertical load. Because the locking mechanism is lost, loading during normal gait is translated to the cobalt-chromium spring, which is not intended to be loaded in this fashion. Loading allows for repetitive compression and expansion of the spring, eventually distorting or breaking its coils. The distorted coils may then contact the thin ceramic isolant ring component surrounding the spring and, in turn, cause it to break. Metal and ceramic particles can then exit through sterilization vent holes in the polyethylene and collect around the implant and in the surrounding soft tissue, as was seen in our patient. Examination of the disassembled prosthesis from this patient showed a fracture of both the cobalt-chromium spring and the ceramic isolant.
To date, no report, as far as we know, has detailed the mechanism of intrinsic mechanical failure of the Repiphysis expandable limb salvage prosthesis. In order to best expand the available body of knowledge as well as ascertain the efficacy of these implants, it is important for surgeons to describe such failures and address how they are treated. In our patient, placement of the prosthesis when he was nine years old allowed for continued growth from the ipsilateral femoral physis and noninvasive lengthening of the tibia to achieve limb-length equality of within 1 cm. As the patient was at an appropriate age for conversion to an adult modular prosthesis at the time of failure of the Repiphysis, the complication was addressed in a single operation. In these patients, mechanical failure of an endoprosthesis should be suspected if radiographs demonstrate metallic or ceramic wear debris or if pistoning and varus and/or valgus instability are found on physical examination.
Note: The authors thank Bob Daily, manager and engineer, Custom Orthopaedics, Orthopaedic Reconstruction Division of Wright Medical Technology (Arlington, Tennessee) for his assistance and expertise.