Introduction

Use of modular components in revision total knee arthroplasty has allowed surgeons to achieve optimal soft-tissue balancing, improved fixation and therefore better knee function. However, junctions of such modular components can be at risk of failure [1]. Although such complications are rare, they can potentially have a serious effect on the outcome of the arthroplasty.

We report a case of revision total knee arthroplasty where the locking screw used for the modular femoral component disengaged, lodging in the joint and resulting in acute locking of the joint.

Case report

A 70-year-old, otherwise healthy man initially underwent left total knee replacement in late 1996. However, this was revised in two stages 6 months later for low-grade infection.

A cemented unconstrained revision arthroplasty using Insall-Burstein II prosthesis was carried out. The femoral component used was CCK-II modular, stemmed, 69 mm prosthesis. The stem diameter measured 22 mm and length 75 mm. A 5-mm distal augmentation block was also used. A locking screw was used to secure the Morse taper junction between the intramedullary stem and the femoral component. A standard 69 mm tibial tray without stem was used with a 12.5-mm posterior stabilized spacer. Both components were cemented using antibiotic-impregnated cement. The femoral medullary stem was press-fit. The patella was not resurfaced, but a lateral release was carried out. The post-operative course was uncomplicated. He had regained 95° flexion after the operation and was fully weight bearing.

After 2 years, the patient presented to the A&E department with sudden locking of the left knee joint, with pain. He was unable to put full weight on the knee and could not straighten the knee fully. On physical examination the patient was afebrile. The left knee showed a flexion deformity of about 20°. The operative wound was well-healed and there was no evidence of infection in the joint. The knee flexion was restricted to 40° with pain. Plain radiographs of the knee showed a well-cemented knee replacement. There was no evidence of loosening around either of the components. The radiographs also showed a screw lodged in the joint at the inferior pole of the patella, partly in the substance of the bone, just anterior to and above the tibial insert (Fig. 1).

Fig. 1
figure 1

a Anteroposterior radiograph of the knee with locking screw lodged in the joint. b Lateral radiograph of the knee showing the locking screw partially embedded in the bone at the inferior pole of the patella.

Full size image

An urgent arthroscopic procedure was carried out the following day. The screw was found partially embedded in the substance of the inferior pole of the patella. It was possible to disengage the screw with an arthroscopic grasper and remove it arthroscopically. Superficial scratches were noticed on the polyethylene spacer, near the midline. No obvious damage was seen on the femoral component. The joint was irrigated. As the radiographs demonstrated a well-fixed component, no further surgery was carried out. Post-operatively, the patient had immediate pain relief and was able to walk at the pre-operative level without assistance.

He was regularly followed in the outpatient clinic for 3 years. At 3-year follow-up consultation the patient was doing well, having retained pre-operative flexion of 95° in the joint.

Discussion

Modular components have been commonly used in revision knee replacements for the past few years. The Insall-Burstein II (CCK) design utilises a Morse taper at the junction of the stem and the femoral component. This is subsequently reinforced with a locking screw. This junction, however, remains an area of high stress concentration [2].

Westrich et al. [3] reported three cases where the locking screw disengaged from the modular femoral stem in revision knee arthroplasties. All three cases were Insall-Burstein constrained-knee arthroplasties. In two of these cases, the disengagement was associated with loosening of the femoral component. The suggested mechanism for this failure was either subsidence or micromotion of the stem—causing the Morse taper and the locking screw to become loose over a period of time—or improper assembly of the modular components before implantation, where the screw is not fully seated when tightened. The former is more likely, because the press-fit stem is well-fixed proximally and the femoral component is well-cemented distally. If indeed accompanied by faulty assembly at the time of implantation, then disengagement of the screw could occur.

The only other report of failure of the stem component junction, with gradual loosening of the femoral component, was by Lim et al. [4]. Again, all their five cases were highly-constrained knee arthroplasties, and this may have been a factor leading to distal loosening of the femoral component, requiring further revision. Late loosening of the locking screw in an unconstrained modular revision stem, with subsequent complication of the screw lodging in the joint with acute locking of the knee joint, has not been previously reported.

Once the screw has disengaged, it can not only cause mechanical locking in the joint, but can also cause premature wear of the insert and abrasions to the femoral metallic component. In our case, the disengaged screw was lodged in the inferior pole of the unresurfaced patella, probably as the patient attempted to straighten his knee. A prompt arthroscopy, as in this case where the diagnosis was made early, was successful in removing the screw from the joint.

Arthroscopy usually has a limited role to play in total knee arthroplasty [5], but in cases such as this, if undertaken promptly, it will avoid serious damage to the components and can also avoid more extensive procedures like arthrotomy, joint debridement or a second revision. We did find superficial abrasions on the polyethylene spacer, but in our case these were largely in the intercondylar area of the plastic insert owing to the position of the lodged screw in the inferior pole of the patella. This factor has perhaps prevented any long-term effects on the arthroplasty.