The use of a ceramic talar body prosthesis in patients with aseptic necrosis of the talus

Approximately 60% of the surface of the talus is covered in articular cartilage.^1,2 The talus has no muscular or tendinous attachments, and therefore has a relatively poor blood supply in comparison with other articular bones such as the femoral head.^1,2

Aseptic necrosis of the talus is usually secondary to a fracture of the talar neck or dome with interruption of its intraosseous blood supply.³ The incidence is predicted by the type of fracture, ranging from 0% to 10% in Hawkins³ type I fractures, to > 60% in type III fractures.⁴

Once collapse of the talus has developed, there are few options for treatment.⁵ Blair⁶ introduced an arthrodesis between the tibia and the talar neck and head for patients with fracture and collapse of the talar body, and subsequent modifications have been reported.^7-9 Although this procedure minimises the need for shortening, instability of the hindfoot remains and often leads to a pseudarthrosis.¹⁰ Tibiocalcaneal fusion appears to produce favourable functional results, but at the expense of shortening. Dennison et al¹⁰ and Kovoor et al¹¹ reported tibiotalar or tibiocalcaneal fusion for patients with bone loss of the talus combined with callotasis using the Ilizarov technique, which makes these operations technically demanding. Recently ankle or tibiocalcaneal fusion using allograft has been reported for patients with collapse of the talus and either osteonecrosis or severe bone loss.¹² Clements¹³ reported a case of traumatic osteonecrosis of the talus treated by tibiocalcaneal fusion using an allograft cellular bone matrix. However, these procedures affect the function of the ankle joint and allograft is restricted in some countries because of ethical concerns.

In 1999, in order to restore function of the ankle and maintain length, we designed an alumina-ceramic talar body prosthesis that replaces the defect in the talar body and preserves movement of the ankle.¹⁴ We examined two types of ceramic prosthesis, which may be provided with or without a peg for the neck of the talus (Fig. 1). The purpose of this study was to evaluate the clinical and radiological results of these new implants, and to assess the future treatment options for this condition.

Patients and Methods

Between July 1999 and April 2006, 22 patients (22 ankles) with necrosis of the talar dome underwent replacement with one of two types of ceramic prosthesis. They included four men and 18 women with a mean age of 65.5 years (37 to 80). Necrosis of the talus was caused by fracture of the talar neck in two patients and was idiopathic in 20. None was due to alcohol abuse or steroids. Our institution treats patients with chronic orthopaedic conditions from all over Japan, leading to a high proportion of patients with idiopathic disorders. The diagnosis was confirmed by MRI and we included patients with total involvement of the body of the talus.

In order to produce the prosthesis, measurements were made from radiographs and CT scans of the contralateral talus. A three-dimensional (3D) wire model and an implant were assembled and a stereolithographic model was cast, from which the alumina-ceramic prosthesis was produced (Fig. 1).

In the first-generation prosthesis a peg was placed on the implant to allow fixation to the neck of the surviving talus with bone cement. This method was used in eight patients via a lateral approach. There was one man and seven women, with a mean age of 64.9 years (37 to 80). Loosening appeared around the peg in all patients. A second version of the prosthesis was made, which was not fixed to the neck of the talus. This was introduced via an anterior approach in 14 patients, including three men and 11 women with a mean age of 65.8 years (45 to 74).

Operative technique

For introduction of the first-generation of the prosthesis, a lateral incision was made with a temporary osteotomy of the fibula at the level of the ankle joint allowing distal rotation of the hindfoot. The talar dome was removed and an anchor hole made in the talar neck to accept the cemented peg of the prosthesis.

The second-generation prosthesis was introduced using an anterior approach and the dome of the talus was prepared with a transverse osteotomy, to allow implantation of the prosthesis.

A below-knee non-weight-bearing cast was retained for the first two weeks. A half-weight-bearing cast was applied for a further two weeks after which movement was instigated. Full weight-bearing was allowed at five weeks post-operatively.

Patients were assessed clinically pre-operatively and at the time of the final examination using the American Orthopaedic Foot and Ankle Society (AOFAS) ankle/hindfoot scoring system.¹⁵ We classified the results as ‘excellent’ for scores ≥ 90, ‘good’ for 80 to 89, ‘fair’ for scores from 70 to 79, and ‘poor’ for scores < 70.

Radiological results were assessed by an author (AT), who had not been involved in the treatment, immediately post-operatively, at six months, one year, two years and three years after surgery or at the final follow-up. Radiological findings were evaluated against three criteria¹⁶: 1) presence of loosening; 2) presence of collapse of the remaining talar head; and 3) presence of osteophytes. All categories were scored from 0 to 3 points (Table I). Sclerotic change of the tibia and calcaneum, and biomechanical stability¹⁶ at the site of the neck of the talus were evaluated at final follow-up. When the radiograph was difficult to categorise, another author (KS) who had not been involved in the treatment also assessed the radiographs.

The study had ethical approval.

Statistical analysis

A paired t-test was used to compare variables. A p-value of < 0.05 was considered to be significant.

Results

First-generation prosthesis

The clinical data and results are summarised in Table II. The mean follow-up was 96 months (21 to 174). Based on the AOFAS scale, the mean pain scores were 15 (0 to 20) pre-operatively and 32.5 (20 to 40) at final follow-up. The mean function improved from 24.8 (18 to 31) pre-operatively to 38.8 (25 to 47) and the mean alignment from 6.9 (0 to 10) to 8.9 (5 to 10), as assessed on the AOFAS alignment scale. The total mean score improved significantly from 46.6 (21 to 60) to 80.0 (50 to 97) (p < 0.001, paired t-test).

There was radiological evidence of loosening and necrosis of the talar head became evident in all patients one year post-operatively (Fig. 2).

The clinical results, however, were excellent in three patients, good in one, fair in three and poor in one. One patient with a poor result and one with a fair result developed severe loosening and a fracture in the remaining head and neck of the talus. Sclerotic changes in the tibia were seen in the patient with a poor result and sclerotic changes in both the tibia and calcaneum in a patient with a fair result. Consequently, a ceramic total talar implant was produced for each patient and introduced at a revision procedure after 16 and 52 months, respectively. Besides these two, most patients had no or slight pain and resumed normal daily activity (Fig. 3).

Two patients (numbers 3 and 7) died of unrelated causes. Their results at 52 months and 66 months post-operatively, respectively, were included.

Second-generation prosthesis

The clinical data and results are summarised in Table III. The mean follow up was 83 months (18 to 118). The mean pain scores were 20.0 (10 to 30) pre-operatively and 31.4 (20 to 40) at final follow-up. The mean function improved from 24.8 (9 to 37) pre-operatively to 40.0 (28 to 48) and the mean alignment from 5.4 (0 to 10) to 9.6 (5 to 10). The total mean score improved significantly from 50.4 (29 to 67) pre-operatively to 81.1 (64 to 98) (p < 0.001).

Radiological changes were seen six months post-operatively in all patients but subsequent progression was slight (Fig. 4). Sclerotic change in either the tibia or the calcaneum was seen in 12 patients.

The clinical results were excellent in three patients, good in five, fair in four and poor in two. Half of the patients had slight movement between the prosthesis and the talar neck. In two patients with a fair result and two with a poor result there was severe loosening with fracture in the talar neck and head at 19 and 43 months post-operatively respectively (Fig. 5a). They underwent revision using a total talar implant (Fig. 5b).

Discussion

A talar dome prosthesis was introduced by Harnroongroj and Vanadurongwan¹⁷ for the treatment of patients with necrosis of the talus. The prosthesis was made from stainless steel and templated from plain radiographs, and fixed to the talar neck using bone cement. In a series of 16 patients (12 with avascular necrosis and four after talar fracture) satisfactory results were obtained in 14 patients (88%) at a mean follow-up of ten years.¹⁷ However, there have been no further reports on these patients. The principal problems encountered with this prosthesis were loosening and sinkage.¹⁸ Ceramic components have been successfully used in hip surgery¹⁷ and also in total ankle replacement since 1980.^19,20 Yoshinaga²¹ reported a comparative study of 316 L stainless steel and alumina-ceramic prostheses in canine hip joints and concluded that alumina-ceramic was superior and tolerated by the articular cartilage. This material would thus be a suitable material for use in the talus where it would be surrounded by articular cartilage.

Four of eight patients (50%) had a favourable result (good or excellent) in the first-generation prosthesis, despite adverse radiological features. Two patients with fracture of the talar neck and head underwent revision using a total talar implant. We thus removed the peg in the design of the second generation implant to avoid concentration of the stress at the site of fixation. Eight of 14 patients (57%) had a favourable result and the mean post-operative score was > 80, and the radiological appearances were more satisfactory. Two patients in this group also underwent revision using a total talar implant

In view of these complications, we recommend the use of a total talar implant in patients with aseptic necrosis even if the talar neck and head are preserved.

The limitations of this study are the small number of patients and the retrospective design. However, considering the uncommon nature of this problem we believe that it has clinical value.

Talar dome prostheses cannot be recommended in patients with avascular necrosis; at least 50% of the patients will have a fair or a poor result. We now recommend the use of a total talar alumina ceramic implant in these patients.

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.