Abstract
Background:
Large bone defects present a difficult task for surgeons when performing single-stage, complex combined hindfoot and ankle reconstruction. There exist little data in a case series format to evaluate the use of frozen femoral head allograft during tibiotalocalcaneal arthrodesis in various populations in the literature.
Methods:
The authors evaluated 25 patients from 2003 to 2011 who required a femoral head allograft and an intramedullary nail. The average time of final follow-up visit was 83 ± 63.6 weeks (range, 10-265).
Results:
Twelve patients healed the fusion (48%). Twenty-one patients resulted in a braceable limb (84%). Four patients resulted in major amputation (16%).
Conclusion:
This series may allow surgeons to more accurately predict the success and clinical outcome of these challenging cases.
Level of Evidence:
Level IV, case series.
The lower limb is designed to adapt to terrain as the foot encounters ground reactive forces. To effectively accomplish this, the foot must be stable, plantigrade, and able to reach the ground. Ankle and subtalar joint (STJ) pain, arthritis, and deformity, both natural and iatrogenic, limit this ability. As one is challenged with severe hindfoot and ankle pathology, a salvage procedure may involve a tibiotalocalcaneal (TTC) fusion. TTC arthrodesis with an intramedullary nail was popularized by Russotti et al in 1988 to surgically correct these complicated scenarios. 26 Correction is driven at achieving a painless, stable, plantigrade foot, with or without bracing. Additional complexity may be encountered intraoperatively when a large bone void is present during a TTC fusion. 15
Loss of bone and/or loss of length in the lower extremity typically presents following avascular necrosis of the talus, failed total ankle arthroplasty, Charcot neuroarthropathy of the ankle, longstanding gross malalignment, and iatrogenic changes, both acute and in revision surgeries. Regardless of the etiology, surgical correction may address the loss of height to limit the need for external shoe modifications and bracing requirements. One method of regaining bony height is by adding a structural graft at the time of fusion. There may be considerable morbidity with taking large volume structural autograft.1,2,11,29 Large bone block frozen femoral heads are readily available and meet the volume and structural demands necessitated to restore bony defects at the level of the ankle.
Incorporating a large bone block allograft into a TTC arthrodesis in a single stage is a difficult task. Few reports exist to guide the surgeon regarding outcome expectations. The purpose of this study was to review patients who underwent a TTC arthrodesis utilizing femoral head allograft and a retrograde intramedullary rod as a treatment for the pathologic hindfoot associated with substantial loss of bone. Our Retrograde Arthrodesis Intramedullary Nail (RAIN) Database was used and has been shown to be useful in various conditions requiring TTC fusion.7-11
Materials
This review was taken from our RAIN Database, which is a comprehensive chart and radiographic review constructed of all patients who have undergone extended hindfoot and ankle arthrodesis with the use of a retrograde arthrodesis nail at our institution. This included 179 patients from 2003 to 2011, and its use has been demonstrated in several previous publications.7-11 All patients with femoral head allograft utilization performed at least 1 year prior to data collection were selected for evaluation. All conditions that were surgically treated with a retrograde intramedullary nail and femoral head allograft were evaluated. Exclusion criteria included patients treated primarily with forms of fixation other than intramedullary nail and restoration of bony voids without femoral head allograft. The arthrodesis was performed to relieve pain and disability, including restoration and/or maintenance of limb length.
Demographic data showed the average age of the patients was 57.3 ± 8.2 years old (range, 39-78 years) and included 12 males (48.0%) and 13 females (52.0%) (Table 1). The average BMI of the group was 32.4 ± 10.1 lbs/in2 (range, 21.3-46.2). Significant risk factors revealed 6 smokers, 11 patients with diabetes mellitus (DM), and 7 with coronary artery disease (CAD). The pathology etiology was as follows: 9 Charcot neuroarthropathy (36.0%), 6 failed total ankle arthroplasty (TAA) (24.0%), 4 avascular necrosis (AVN) (16.0%), 4 revisions of failed prior fusions (16.0%), and 2 posttraumatic degenerative joint disease (8.0%). The right side was affected in 9 patients (36.0%). All but 2 patients (92.0%) had some additional type of orthobiologic augmentation to the fusion site at the time of surgery, including 17 patients (68%) had adjunctive bone marrow aspirate application. Nineteen patients (76.0% of all patients) had bone stimulation at the time of the index procedure: 8 patients (42.1%) had implantable bone stimulation and 11 (57.9%) used external combined magnetic field bone stimulation. Adjunctive procedures included either a gastrocsoleus recession or a tendoachilles lengthening in 6 patients and complete talectomy in 12. No patient had a tendon transfer or tarsal fusion. The average time to the final follow-up visit was 83 ± 63.6 weeks (range, 10-265 weeks).
Cohort Demographics (N = 25).
BMI, body mass index; CAD, coronary artery disease; DM, diabetes mellitus; ESRD, end stage renal disease.
Femoral head preparation required thawing in saline and partial decortication using a sagittal saw to allow placement into the void. Drill bit fenestration was utilized to allow a balance of strength and vascularization in-growth channels. In all cases the distal fibula was removed and morcellized into corticocancellous chips. This morcellized graft was added as adjunctive autograft to the femoral head allograft. Graft site preparation was performed by either curettage, saw cuts, or acetabular reamer (5) depending on the defect and host bone structure and viability. 5 A compressive Jones dressing with posterior splint was worn for the first postoperative week. A non-weight-bearing short leg fiberglass cast was worn for 4 weeks, a weight-bearing short leg cast for an additional 4 weeks, followed by a walker boot until regular shoe or brace was allowed.
A union was defined as both clinical and radiographic union. Stable pseudoarthrosis was reported in cases of clinical union and radiographic nonunion. Nonunion was reported as clinical signs of instability or pain and radiographic evidence of incomplete bony healing. We tried to minimize bias by having a surgeon not involved in the surgical care of the patient evaluate the charts and radiographs, but evaluation was still subjective and retrospective.
Results were broadly defined as success or failure. Success in this series was defined as maintenance of a functional limb or ability to ambulate in shoes or a brace regardless of radiographic findings. Failure was defined as a painful, unstable nonunion or loss of limb from major amputation.
Results
Overall, 4 patients were deemed clinical failures and underwent transtibial amputations. Of the remaining 21 patients, 12 were classified as clinical unions, 7 as pseudarthroses, and 2 as painful nonunions. These 21 cases progressed postoperatively to a functional brace and if possible to regular shoe wear. One of 6 (16.7%) smokers resulted in union.
During follow-up after the surgery, the average time to protected weight-bearing was 73.5 ± 39 days (range, 35-204 days), and the time to full weight-bearing was 117.7 ± 53.9 days (range, 43-209 days). Complications requiring a reoperation occurred in a total of 6 patients. This included postoperative incision and drainages that were soft-tissue-only in 1 patient and involved bone in 5 patients, hardware removal in 7, stress fractures in 3, and need for intravenous (IV) antibiotics in 8 (Table 2).
Surgical Variables (N = 25).
Discussion
In this series, the TTC fusion was a salvage attempt as an alternative to transtibial amputation. It is our experience that most patients, when given the option and a realistic description of the reconstruction process, choose an attempt at limb salvage. The goal of TTC arthrodesis is to create a painless, stable, plantigrade foot, with or without bracing. With these severe conditions and deformities, especially in the presence of bone loss, a solid fusion may not be achievable. Jeng et al also noted that the complete fusion rate was relatively low, but “success” defined as a functional limb can be higher. 15 The goal of this study was to quantify the outcomes of these severely morbid hindfoot reconstructions utilizing a retrograde intramedullary rod and large bone block femoral head allograft interpositioned at the ankle.
Our RAIN Database allowed a retrospective look at femoral head use in TTC arthrodesis at our institution, regardless of the presenting etiology.7-10 Establishing an indication for limb length restoration or maintenance across multiple etiologies when performing a TTC fusion, such is the case in this cohort, is difficult to summarize (Figures 1 and 2).
Fifty-year-old female who developed talar avascular necrosis after closed talar body fracture. (A) Lateral ankle view of the staged approach with a static external fixator to accommodate poor soft tissues and maintain limb length. (B) Initial postoperative anteroposterior (AP) and lateral ankle views demonstrating proper alignment of a retrograde intramedullary nail, femoral head allograft, and maintenance of the talonavicular joint. (C) and (D) 17 months after surgery weight-bearing AP and lateral views demonstrating appropriate hindfoot alignment and consolidation of the femoral head allograft.
Sixty-eight-year-old male with a history of a failed primary total ankle arthroplasty and an infected revision. (A) Preoperative anteroposterior (AP) radiograph reveals a failed ankle arthroplasty and peri-implant cysts. (B) Postoperative AP radiograph of a fixed bearing, stemmed total ankle replacement. (C) Postoperative AP radiograph after MRSA-positive cultures and explant and antibiotic-loaded cement. A tibiotalocalcaneal fusion was performed 8 weeks later. (D) and (E) AP and lateral radiographs, 5 months after TTC fusion with femoral head that healed.
The need for limb length maintenance was the impetus for interposition of large block allograft during hindfoot reconstructions. A shortened limb needs more proximal compensation including pelvic tilt and increased knee flexion of the long extremity and an equinus of the short, contralateral limb. 28 Length differential is also related to back pain, knee dysfunction and osteoarthrosis of the hip.12,17,18,20 These negative outcomes exist even in the apparent ‘successful’ timely fusion of the rearfoot, if the limb is unmatched in length. There is also a historically relatively low success rate of tibiocalcaneal type fusions. Detenbeck et al and Reckling found failure rates at 80%.6,25 Another study of tibiocalcaneal arthrodesis found only 4 of 11 (36 %) developed solid union and only 8 of these same 11 (73%) were braceable. 23 It is therefore useful to attempt to restore or maintain length when bony voids at the ankle are encountered.
Autogenous bone, traditionally harvested from the iliac crest, offers the surgeon a biologically compatible graft, but it is not without donor site morbidity including fracture, hemorrhage, pain, nerve or arterial injury, and cosmetic disturbance.1,2,11,29 The volume of the femoral head as well as the corticocancellous nature is amenable to replacement of voids and restoration of structure at the level of the ankle. The fresh frozen femoral head offers a corticocancellous structure capable of withstanding reaming compatible with a retrograde rod. The cortical wall allows structural support, while the cancellous portion allows more rapid fusion site incorporation. 13 In a study of fresh frozen allografts in the foot and ankle, Myerson et al found allografts were suitable for use in the foot and ankle and avoided the potential complications associated with harvesting autografts, reduced operative time and cost, and effectively filled volumes too large for traditional autografts. 21
In the largest study of its kind to date, a multicenter study examined TTC arthrodesis in 55 patients (56 ankles) with various indications for surgery including posttraumatic injury (14 ankles), failed previous surgery (12 ankles), osteoarthritis (11 ankles), AVN (7 ankles), failed total ankle replacement (2 ankles), Charcot–Marie–Tooth disease (2 ankles), and Charcot foot (2 ankles). At an average follow-up of 26 months, fusion was achieved in 48 ankles and 48 patients reported being satisfied with the procedure, 86% and 87%, respectively. Autogenous iliac bone graft was utilized when they noted extensive bone loss. They also found the average leg length discrepancy postoperatively was 1.4 cm. 4
Moore et al found ambulatory success in 16 of 19 (84%) and union in 14 of 19 (74%) salvage TTC fusions utilizing an intramedullary rod in a cohort of rheumatoid arthritis, posttraumatic, and Charcot patients. 19 Papa et al looked at 13 patients having a TTC fusion for posttraumatic osteoarthrosis of the hindfoot and found a union rate of 86% with a mean 1.5 cm of shortening. 24 They concluded this type of procedure is a technical challenge but a reasonable alternative to amputation. Recently, Jeng et al reported 71% functional limb salvage when performing a TTC arthrodesis with femoral head allograft. They also found a significant relationship between DM and nonunion. 15 In comparison with these classic reviews, and given the use of a large bone block femoral allograft to further complicate the healing, our success rate of 21 out of 25 (84%) provides evidence that successful reconstruction using intramedullary rod and femoral head allograft can be expected.
In the case of failed TAA and a need for limb height maintenance and fusion, Thomason and Eyres utilized femoral head allograft and intramedullary fixation and achieved fusion in 3 of 3 cases at 3 months. 27 Both Hopgood et al and Carlsson et al reported on TTC fusion after failed TAA.3,14 They concluded that fusion with intramedullary nailing was preferable to fusion with external fixation. Kotnis et al produced an algorithm to guide surgeons on failed ankle replacement. 16 When infection is present, circular external fixation is favored with distraction osteogenesis to account for lost limb length. In aseptic cases, they prefer locked intramedullary compression nail, opting for an interposition of a fibular autograft.
This study does have weaknesses that are important to note. It is a retrospective study, and has all inherent limitations of such and is of a lower level of evidence. 22 The data reported are limited to those collected and compiled into our RAIN Database, which is intended to be comprehensive, but is still limited to predefined data collection. There were variables, such as limb length, specific decision making for complete talectomy, exact surgical approach, and previous surgeries, that were not standard measurements recorded at our institution and thus not available. In addition, the effect of specific surgical modifications such as retention of the fibula or biologics is beyond the scope of this article.
Certain other variables have been explored in previously published work using our RAIN Database.7-11 It involved only small numbers of patients. Both of these problems were the result of the relative infrequency of these conditions. TTC fusion utilizing large allogenic bone blocks is rare. Thus it may be ethically difficult or impossible to perform an adequate prospective and well controlled study on this condition. There was no evaluation of the position attained, either clinically or radiographically. Although inferences can be made based on shoe wear and braces needed afterward, no specific data on postoperative positioning were recorded in the RAIN Database. Finally, there was a lack of patient function and satisfaction scoring. This could not be rectified in this retrospective study design.
In summary, the authors have reported on the results of intramedullary nail arthrodesis of the hindfoot and ankle with the use of a large-volume, femoral head allograft. It was a powerful, 1-stage method of treating very complex pathology. The authors feel this is a useful technique as the success rate was reasonable for this difficult population, and the results can be useful to the surgeon faced with this difficult clinical presentation.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.