Use of modular components in either primary or revision total knee arthroplasty allows the surgeon to achieve intraoperative customization, optimum kinematics, and improved fixation1-8,11,13,15. Modularity in total knee arthroplasty includes augmentation of the femoral component with distal or posterior metallic blocks, or both, and augmentation of the tibial component with metallic wedges or blocks. In addition, fixation of the intramedullary stems is possible, and this allows stress relief and load-sharing with the host bone1,6-8,11,12. Although such elaborate configurations are unnecessary for most primary total knee arthroplasties, many surgeons routinely use intraoperative customization in revision procedures and frequently this customization involves the stem1,2,4-8,12,14.
Concerns with respect to modularity in total knee arthroplasty have arisen because of the many junctions of modular components. Issues relating to metallic wear, fretting, and corrosion between the body of the prosthesis and the various modular components require the surgeon to understand the potential long-term problems2 and to proceed with caution. In addition, loosening of screws used to secure blocks, wedges, or the intramedullary stems can be a problem. Dissociation of modular components after implantation is quite rare; however, surgeons must be aware of the complication and be prepared to correct it2.
We present the cases of three patients who had disengagement of a locking screw from a modular stem that had been implanted in a revision total knee arthroplasty. In each instance, a screw had been used to reinforce the Morse taper junction between an intramedullary stem and the femoral component, as recommended by the manufacturer. Months after implantation, the screw disengaged from the femoral component and became an intra-articular loose body. We describe the diagnosis of such a failure, the assessment of the damage to the total knee replacement, and the management of each patient.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†The Hospital for Special Surgery, 535 East 70th Street, New York, N.Y. 10021. E-mail address for Dr. Westrich: westrichg@hss.edu.
CASE 1. An eighty-two-year-old man had a revision total knee arthroplasty in 1987 because of aseptic loosening of the femoral component. Substantial femoral osteolysis was encountered on removal of the femoral prosthesis. An Insall-Burstein-II constrained condylar knee prosthesis (Zimmer, Warsaw, Indiana) that included sixty-four-millimeter femoral and tibial components with intramedullary stems was implanted (Fig. 1-A). Locking screws were used to reinforce the Morse taper junctions between the intramedullary stems and the femoral and tibial components. The femoral stem measured fifteen millimeters in diameter and 200 millimeters in length. The femoral and tibial components were fixed to bone with acrylic polymethylmethacrylate cement, and the fluted stems were press-fit in the intramedullary canals without cement.
The patient did well until two years postoperatively, when he began to have persistent pain, a decreased range of motion, and crepitus in the right knee. On physical examination, the knee had a range of motion of 5 to 60 degrees, notable crepitus, parapatellar fibrosis, and atrophy of the quadriceps. Although radiographic examination of the knee revealed loosening and subsidence of the femoral component, the anteroposterior radiograph also showed a loose locking screw from the femoral stem in the intercondylar area. Interestingly, this locking screw was obscured on the lateral radiograph. A routine radiograph that had been made fifteen months postoperatively was carefully reviewed, and the locking screw, which had not been seen originally, was noted in the intercondylar area (Fig. 1-B). The patient had an arthrotomy, synovectomy, and débridement of the right knee with reinsertion of a new femoral locking screw between the femoral housing and the intramedullary stem. Intraoperatively, the loose locking screw was readily apparent in the anterior soft tissues of the quadriceps tendon with an adjacent severe fibrotic reaction and impingement of scar tissue between the femoral and tibial components. On retrieval, the locking screw was intact but it had evidence of wear. The postoperative course was complicated by additional arthrofibrosis, and manipulation was performed on the twelfth postoperative day. No appreciable wear of the tibial polyethylene was observed.
The patient continued to report pain and instability, and he fell several times. In August 1991, radiographs of the right knee revealed gross loosening of the femoral component with settling. The patient was managed with a second revision, at which time the locking screw was noted to be intact. Again, the postoperative course was complicated by arthrofibrosis, and manipulation was performed on the twenty-fourth postoperative day.
At the latest follow-up examination, the patient was walking with the aid of crutches; he had only slight pain and no instability.
CASE 2. A seventy-one-year-old man had an infection with Staphylococcus aureus that had metastasized from the site of a spinal arthrodesis to that of a right total knee arthroplasty. A two-stage reimplantation procedure was performed. The knee replacement was removed, the tissue was debrided, and a cement spacer with antibiotics was implanted. After a six-week course of intravenous antibiotic therapy, an Insall-Burstein-II constrained condylar knee prosthesis that included sixty-four-millimeter femoral and tibial components with intramedullary stems was inserted. Locking screws were used to secure the Morse taper junctions between the intramedullary stems and the femoral and tibial components. The femoral stem measured eighteen millimeters in diameter and seventy-five millimeters in length. The femoral and tibial components were fixed to the underlying bone with Simplex-P cement (Howmedica, East Rutherford, New Jersey) impregnated with tobramycin (1.2 grams per forty-gram pack of cement). The intramedullary stems were press-fit.
The postoperative course was uncomplicated until three years after the reimplantation, when the patient had an acute onset of pain and locking of the right knee after hearing a loud snap while climbing stairs. On physical examination, the patient had pain on flexion of the knee and an audible creak. Radiographs made one day after the event revealed a screw in the intercondylar notch. An urgent arthroscopic procedure was performed, and the screw was found to be lodged in the posterior part of the intercondylar notch (Fig. 2). A small amount of metallic debris and several scratches on the polyethylene spacer were seen in both the medial and the lateral compartment. The locking screw was arthroscopically manipulated to the posterolateral corner and was retrieved with an arthroscopic grasper. On removal, the locking screw was intact but it had evidence of wear. Parapatellar fibrosis was also observed, and the area was debrided arthroscopically. As the radiographs demonstrated a well fixed femoral component with no radiolucency around the stem, a decision was made simply to remove the loose locking screw without replacing it.
Postoperatively, the patient had no pain and was able to walk, at the preoperative level, without an assistive device. At the latest follow-up examination, six months after the arthroscopic removal of the locking screw, he was still doing well.
CASE 3. A seventy-nine-year-old woman had a revision of a right total knee arthroplasty in 1989 because of infection. In April 1992, pain and effusion with subsequent drainage developed and radiographic examination revealed loosening of both the femoral and the tibial component consistent with infection. The patient was managed with a two-stage reimplantation protocol, which included removal of the prosthesis, débridement, a six-week course of intravenous antibiotic therapy, and subsequent reimplantation. An Insall-Burstein-II constrained condylar knee prosthesis that included fifty-nine-millimeter femoral and tibial components with intramedullary stems was inserted. Locking screws were used to secure the Morse taper junctions between the intramedullary stems and the femoral and tibial components. The femoral stem measured fourteen millimeters in diameter and seventy-five millimeters in length. The components were inserted with cement, and the stems were press-fit in the intramedullary canal.
At the three-year follow-up examination, the patient was doing well and was walking without an assistive device. The range of motion of the knee was 0 to 80 degrees without pain, crepitus, or locking. However, radiographs demonstrated loosening and subsidence of the femoral component and showed the femoral locking screw to be embedded within the polyethylene insert in the lateral compartment of the joint (Fig. 3). Although operative removal of the locking screw was recommended, the patient declined the procedure because of advanced age and a lack of symptoms (such as pain, swelling, locking, or instability). At the latest follow-up examination, she was being managed non-operatively with frequent clinical and radiographic examinations.
Although modular components in total knee arthroplasty allow for intraoperative customization of the prosthesis, optimization of kinematics, and improved fixation, concerns with respect to their use are not unwarranted2. The tolerance of machined components is variable, and the junctions of modular components are at risk for premature failure2. Furthermore, the effect of cyclical micromotion on prostheses with multiple sites of modularity is unknown. Metallic wear, fretting, and corrosion between the body of the prosthesis and the various modular components may be both theoretical and real problems2,5-8,10,11,15. A screw used for the fixation of a modular stem in a revision total knee arthroplasty failed in each of the three patients in the present report.
Fixation of the intramedullary stems was used in our three patients to provide stress relief and load-sharing in the femur and the tibia. In the Insall-Burstein-II constrained condylar knee prosthesis, the intramedullary stems are secured to the femoral and tibial components with a Morse taper that is subsequently reinforced with a locking screw. In each patient, the locking screw became dislodged and was free within the knee joint. A loose body, such as a screw, can cause not only mechanical locking in the knee joint but also premature wear of the polyethylene, a severe inflammatory response with subsequent fibrosis, and possibly premature osteolysis.
In anticipation of such a potential problem with the locking screw, design engineers for the Insall-Burstein-II constrained condylar knee prosthesis developed a capture mechanism for the screw in that it is preassembled into the femoral housing by the manufacturer before sterilization. In order for the screw to fail and to become disengaged from the femoral component, a series of events must occur. First, the screw must become loose within the intramedullary stem and counter-rotate until it has cleared the threads in the stem. Second, the screw must align itself with the threads that make up the capture mechanism of the femoral component. Third, the screw must counter-rotate again to disengage itself from the femoral component. Although the probability of any of these events occurring is low and the probability of all of them occurring is exceedingly low, this series of events must have occurred in our three patients.
Disengagement of the locking screw from the femoral component and the intramedullary stem may be caused by motion after implantation. Subsidence or micromotion of the stem may cause the Morse taper and the locking screw to become loose over time. Alternatively, the mechanism of failure may be improper preassembly of the modular components before implantation. If the stem is not impacted into the Morse taper of the femoral housing with enough force at the time of assembly, the screw is not fully seated when tightened. Any subsequent impaction of the stem may allow loosening of the screw. Thus, if during implantation the femoral component and stem assembly is impacted into the femur with enough force to seat the intramedullary stem further within the Morse taper, the locking screw may loosen prematurely. The construct then may be predisposed to loosen more quickly with the screw backing out as the result of cyclical loading and micromotion of the prosthesis. We believe that such disengagement of the screw may be prevented by a repeat verification that the screw is tightly seated after implantation of the femoral component and stem assembly. Finally, we recognize that this mechanism of failure may not be due to the screw fixation for modular stems per se but to failure of proper assembly of the modular components by the operative team.
Although arthroscopy has only limited application in the repair of a total knee prosthesis9, a loose locking screw within the knee joint, if diagnosed early, is amenable to arthroscopic removal and even replacement. However, if the loose body has been present for some time and if the total knee replacement has been damaged, then arthrotomy may be the more beneficial and definitive treatment because it allows for assessment of the components for loosening and for implantation of a new tibial polyethylene insert. A locking screw lodged between a metallic femoral component and a polyethylene liner inevitably will result in severe polyethylene wear and subsequent failure of the implant, either mechanically or secondary to osteolysis resulting from polyethylene debris. In one of our patients (Case 3), operative removal of the locking screw was recommended but was refused by the patient because of advanced age and a lack of symptoms. The screw was not replaced in another patient (Case 2) as well. However, the long-term effect of not replacing the locking screw is unknown.
Although the use of the Insall-Burstein-II constrained condylar knee prosthesis offers many advantages in the setting of revision total knee arthroplasty6, modularity per se is not without potential complications12. Modularity is useful in primary total knee arthroplasty when osseous defects and instability are encountered, and it is especially useful in revision total knee arthroplasty when osseous deficiencies cannot be fully appreciated preoperatively. In general, the high degree of modularity in total knee arthroplasty systems available today has obviated the need for custom components in most patients. However, the true success of modular systems depends on attention to detail, accurate preoperative planning, careful intraoperative assembly and implantation, and routine postoperative evaluation.