Salvage of Avascular Necrosis of the Talus by Combined Ankle and Hindfoot Arthrodesis Without Structural Bone Graft

Methods

We retrospectively reviewed 102 cases of ankle and hindfoot arthrodesis with compression retrograde intramedullary nail between September 2005 and January 2013. Inclusion criteria included patients with ankle and subtalar arthritis due to AVN of talus treated with a compressing retrograde intramedullary nail and a postoperative minimum follow-up of 12 months. Exclusion criteria included patients treated with IMTCA for other reasons than talar osteonecrosis.

Thirteen patients with 14 operated limbs met inclusion criteria and were evaluated. The mean follow-up was 26.1 months (range 12 to 47, SD ± 11.5). Average age of the patients was 54.3 (range 39 to 68, SD ± 9.9), and included 4 males and 9 females. Average BMI of the patients was 28.4 (range, 18.6 to 39.4, SD ± 5.5). The talar osteonecrosis was traumatic in 9 cases and nontraumatic in 5 cases. Medical records were reviewed for the extent of osteonecrosis, operative approach, concomitant procedures, bone graft used, and postoperative complications including nonunion, infection, nerve injury, wound healing issues, and the need for additional surgeries. Two patients had diabetes, and 2 patients were smokers.

Preoperative and postoperative outcome measures included the Visual Analogue Scale for pain, the AOFAS Ankle/Hindfoot Score, and the SF-36 questionnaire. It is of note that maximal postoperative AOFAS score was reduced from 100 to 86 due to the points assigned to range of motion.

All cases had radiographic evidence of severe talar AVN involving the body of the talus. Eight of the 13 patients had advanced imaging, composed of computed tomography in 7 and magnetic resonance imaging in 1. The remaining 5 patients had such severe and obvious radiographic findings that imaging was judged an expense that would not change the treatment decision making. All radiographs were done standing, except for those in the 2-month postoperative period of non-weight-bearing. Radiographs were evaluated for evidence of postoperative union at both levels of arthrodesis sites as seen in anteroposterior, oblique, and lateral projections at the time of last follow-up. Statistical analysis was done using the paired Student’s t test and Wilcoxon signed rank test. Significance was set at P < .05. All variables were tested for normal distribution using the Kolmogorov–Smirnoff goodness-of-fit test. For all variables the normal distribution was not rejected (P value between .7 and .9). Approval was obtained for the study from the institutional review board of the hospital.

Operative Technique

The lateral approach to the ankle and subtalar joints was utilized in 9 cases, and a combined anterior approach to the ankle joint and lateral approach to subtalar joints was utilized in 5 cases. Ten Achilles tendon lengthening were performed concomitantly, as well as 3 hardware removals and 2 peroneal tendon debridements. In 4 cases, the entire body of the talus required excision due to complete loss of bony integrity. In these cases a tibiocalcaneal arthrodesis was performed and the remaining talar head-neck portion was fused to the anterior aspect of the tibia to augment construct stability and extend fusion mass, while preserving transverse tarsal motion (Figures 1-2).

OPEN IN VIEWER

Figure 1.

(a) Preoperative radiograph of 65-year-old woman with talar osteonecrosis; 115 × 83 mm (300 × 300 DPI). (b) Preoperative radiograph of 65-year-old woman with talar osteonecrosis; 63 × 100 mm (300 × 300 DPI).

OPEN IN VIEWER

Figure 2.

(a) Postoperative radiograph demonstrating successful fusion, following excision of the necrotic talar body and integration of the remaining talar head to the fusion mass; 136 × 102 mm (300 × 300 DPI). (b) Postoperative radiograph demonstrating successful fusion, following excision of the necrotic talar body and integration of the remaining talar head to the fusion mass; 59 × 117 mm (300 × 300 DPI).

The operative technique for the compression retrograde nail has been previously reported. ⁴ None of the patients were treated with external or internal bone stimulators. Bone grafting was done in all cases (Table 1). Fusion of the talar head-neck portion to the tibialis anterior was done with cortical screws around the nail. Patients were kept non-weight-bearing for 8 weeks in a cast, followed by 1 month in a walking cast and then transitioning to a walking boot. Physical therapy for gait training and proximal joint strengthening was initiated 4 to 4.5 months postoperatively.

OPEN IN VIEWER

Table 1.

Bone Grafting Techniques.

Bone Grafting Technique	Number of Cases
Local autograft	11
Proixmal tibia autograft	5
Demineralized bone matrix	4
Bone morphogenic protein	1
Allograft with stem cells (Trinity® Evolution™)	1

Results

Evaluation of preoperative imaging showed that 12 of 14 (86%) limbs had AVN involving the entire body of talus, and 11 of 14 (78%) had bone collapse due to osteonecrosis. Union at both ankle and subtalar joints was achieved in all cases.

Preoperative VAS and AOFAS scores were available for 10 limbs, and available for all 14 limbs postoperatively. The mean preoperative VAS score was 6.9 (range 5 to 9, SD ± 1.5), decreasing to 1.7 (range 0 to 6, SD ± 2.2) postoperatively (P = .00008).

The mean preoperative AOFAS score was 32.7 (range 20 to 46, SD ± 8.7), increasing to 72.1 (range 46 to 86, SD ± 10.1) postoperatively (P = .00003).

Preoperative SF-36 scores were available for 8 limbs, and available for 13 limbs postoperatively. Statistically significant difference was analyzed using the Wilcoxon signed rank test for patients with pre- and postoperative data (n = 8). The mean preoperative SF-36 physical component score was 30.5 (range 21 to 42, SD ± 6.9), increasing to 42.8 (range 20 to 60, SD ± 11.4) postoperatively (P = .02). The mean preoperative SF-36 mental components score was 53.4 (range 28 to 66, SD ± 13.4), decreasing to 52.4 (range 27 to 64, SD ± 11.6) postoperatively (P = .64). Postoperatively, 1 patient needed an ankle foot orthosis (AFO) to ambulate. The remaining patients ambulated in regular shoe-wear without external walking aids.

One patient had a tibial stress fracture at the site of 1 of the proximal locking screws, which resolved following nail removal and cast immobilization. Three other patients required removal of symptomatic screws due to subcutaneous prominence, of which 2 were proximal tibial locking screws placed from the medial side and 1 was a screw from the posterior tibia to the talar neck-head fragment. One patient had a superficial anterior wound dehiscence in the early postoperative period, which was treated with intravenous antibiotics, and a VAC wound protocol that resolved without additional operative intervention.

Discussion

In the past, some authors advocated conservative treatment, especially in the early stages of the AVN. While a few authors reported treatment based on protected weight-bearing with immobilization,^5,37 others reported that prolonged periods of non-weight-bearing did not guarantee protection from talus collapse. ³⁷ Therefore, some authors suggested that once a talar fracture has healed, the weight-bearing regimen should be dictated by the AVN extent, patient lifestyle, and occupation (1). However, there are no data to support conservative treatment or 1 weight-bearing regimen over another.^20,29 Furthermore, a recent study showed that the likelihood of AVN is related to the extent of injury rather than the timing of operative treatment. ⁴³

Once structural changes and talus collapse ensues, the impaired healing potential of the bone and the associated deformity of the ankle and subtalar joints are obstacles to bony healing. With regard to non-intramedullary fixation for advanced talar AVN, the literature is scarce and primarily composed of case reports and small series of patients. A study by Dennison et al ¹² reported on 6 patients with talar AVN treated with talectomy and tibiocalcaneal arthrodesis fixed with an Ilizarov circular frame. Distraction osteogenesis was done in 4 of the patients to restore limb length. Union was achieved in all patients, and 5 patients had good or excellent clinical results. Kitaoka and Patzer ²⁷ reported on 19 cases of ankle joint arthrosis due to talar AVN, of whom 16 patients were treated with a tibiotalocalcaneal arthrodesis. Twelve patients were fixed with an external fixator, 3 with staples, and 1 patient with no fixation. Good to excellent clinical result were achieved in 11 patients, and the authors reported 3 (19%) cases of nonunion.

Urquhart et al ⁴² reported on 11 limbs with AVN treated with various hindfoot arthrodesis fixation techniques primarily using screws and plates. They had 2 cases of infected nonunions, of which 1 resulted in a below knee amputation. Using the Mazur grading system, 9 of the 11 arthrodeses (82%) achieved excellent scores.

Although retrograde intramedullary nail fixation for tibiotalocalcaneal arthrodesis for salvage of severe ankle and hindfoot arthritis and deformity is highly reported,^8,17,22,40 most series focus on the operative technique, rather than the underlying etiology, and most series include, at most, only a few cases of talar AVN. Thus, drawing conclusions pertaining to this patient population has been challenging.

Pelton et al ³⁶ reported on a series of 33 cases of IMTCA, out of which 5 patients had AVN of the talus as an underlying etiology. Although results were not reported separately for each subset of patients, 1 out of the 4 nonunion cases was a talar AVN case, in which a femoral head allograft was used. Kile et al ²⁶ reported on 30 cases of arthrodesis with intramedullary nail, out of which 3 cases were caused by talar AVN. Although specific results were not reported, the authors concluded that IM fixation is an “effective salvage for patients with posttraumatic degenerative arthrosis or avascular necrosis.” Moore et al ³³ reported on 19 cases of retrograde intramedullary nailing for ankle arthrodesis. Out of them 1 case was a salvage of a failed ankle arthrodesis following AVN of the talus. A recent study be Rammelt et al ³⁹ reported on 38 cases of tibiotalocalcaneal fusion using the hindfoot arthrodesis nail. Three (8%) of the patients were treated for severe arthrosis secondary to talar AVN, but no specific results were reported for this patient subgroup.

Jeng et al ²³ reported on 32 patients treated with tibiotalocalcaneal arthrodesis with bulk femoral head allograft. In their series 28% (9/32) patients had AVN of the talus. Out of theses 9 patients, union was achieved in only 6 patients. One out of 6 amputations in that study occurred in a patient with a history of infected AVN of the talus and a history of calcaneal osteomyelitis. As noted by the authors, as the cohort included multiple preoperative diagnoses, it is difficult to draw specific conclusions about outcomes. Chou et al ⁸ conducted a study on TTCA in 56 limbs. Out of those, an IM nail was used in 37 cases, of which 7 cases were for AVN of the talus. No specific conclusions can be drawn as no subgroup analysis was reported. To the best of our knowledge, only 1 previous study has reported on the use of IMTCA specifically in cases of arthritis caused by talar AVN. DeVries et al ¹⁴ reported on 14 patients treated with IMTCA specifically for AVN of the talus. This is the most comprehensive and comparable study to our report to date. Of the 14 patients, 12 (85%) achieved clinical and radiographic union, 57% (8/14) needed custom shoes or orthotics, and 42% (6/14) needed some kind of an ambulatory aid. Four patients had postoperative complications without the need for major revision surgery. The authors noted several shortcomings of their study, including some patients with follow-up as short as 5 months and the confounding effect of multiple therapeutic interventions including use of bone marrow aspirate, bone morphogenic protein, allograft bone graft, biological adjuncts, and the use of either internal and external bone stimulation in every patient. The authors reported rates of union but no patient-reported outcome data. One patient had a talectomy. Our current study showed satisfactory outcomes in a series of patients with AVN, some of whom had severe bone loss. One patient reported by DeVries et al had a talectomy, compared to 4 patients in our series with complete excision of the talar body, but with preservation of the head and neck of the talus. Our study showed satisfactory outcomes in those 4 patients using direct arthrodesis of the tibia to the calcaneus without bulk allograft, and with arthrodesis of the talar head and neck to the anterior tibia. To the best of our knowledge, this technique for preservation of transverse tarsal joint function with tibiocalcaneal arthrodesis has not been previously described. However, it could be seen as a variation of the Blair technique,^3,11,44 in which the talar head and neck are fused to the tibia, but in the absence of tibiocalcaneal arthrodesis.

This study is also distinguished from the DeVries et al series in the very minimal use of orthobiologic and adjunct bone healing modalities, despite which a satisfactory healing rate was achieved. We attribute this to both to the extensive debridement of nonviable avascular necrotic bone, as well as to the compression at the arthrodesis site. We were surprised that despite the diagnosis of diabetes in 2 patients, and 2 patients who were smokers, these factors did not appear to contribute to nonunion, although these numbers are very small. It is notable that the 2 diabetic patients did not have clinical evidence of neuropathic findings in the lower extremities, such as Charcot neuroarthropathy.

There are several limitations of this study. The study is composed of a small number of patients, the intermediate-term mean follow-up is 26 months, the study is retrospective in nature, and there were a number of patients whose clinical outcome data were incomplete. Last, the bone compression produced by this operative technique is sufficiently great as to make it difficult to assess the exact time at which osseous union was achieved.

Conclusions

This study showed that ankle and hindfoot arthrosis due to extensive talar AVN can be successfully treated with ankle and hindfoot arthrodesis using a compressing retrograde intramedullary nail. In addition it showed that even in some cases of complete loss of the talar body, tibiocalcaneal fusion can be successful without structural allograft. Despite the salvage nature of these cases, significant improvement, as measured by both radiographic and clinical outcomes, can usually be expected.